==> picture [215 x 22] intentionally omitted <==

Multiple Technology Appraisal

Therapeutics for people with COVID-19 [ID4038]

Committee Papers

==> picture [215 x 22] intentionally omitted <==

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

MULTIPLE TECHNOLOGY APPRAISAL

Therapeutics for people with COVID-19 [ID4038]

Contents:

The following documents are made available to stakeholders:

Final Scope , final Stakeholder list and Research Protocol

1. Assessment Report prepared by the School of Health and Related Research (ScHARR)

2. Consultee and commentator comments on the Assessment Report and model from:

• a. AstraZeneca

• b. Gilead

  • i. Assessment Report comments

  • ii. Assessment Group model comments

c. GlaxoSmithKline

• i. Assessment Report comments

• ii. Assessment Group model comments

• d. Merck, Sharp and Dohme

  • i. Assessment Report comments

  • ii. Assessment Group model comments

• e. Pfizer

  • i. Assessment Report comments

  • ii. Assessment Group model comments

• f. Roche g. Down’s Syndrome Association

• h. Kidney Care UK

• i. Long Covid SOS

• j. British Thoracic Society

• k. Faculty of Pharmaceutical Medicine l. Royal College of Physicians

• m. Renal Pharmacy Group n. UK Kidney Association o. Healthcare Improvement Scotland

3. Company targeted submissions from:

• a. AstraZeneca

• b. Eli Lilly and Company

• c. Gilead

• d. GlaxoSmithKline

• e. Merck, Sharp and Dohme

==> picture [215 x 22] intentionally omitted <==

f. Pfizer

4. Patient groups, professional group and NHS organisation submissions

• from:

• a. Blood Cancer UK-Anthony Nolan-Myeloma UK-Leukaemia CareLymphoma Action-Chronic Lymphocytic Leukaemia Support

• b. Down’s Syndrome Association

• c. Immunodeficiency UK

• d. Kidney Care UK

• e. Long Covid Kids

• f. Long Covid SOS

• g. LUPUS UK

• h. MS Society

• i. Faculty of Pharmaceutical Medicine

• j. UK Clinical Pharmacy Association

• k. UK Kidney Association

• l. UK Renal Pharmacy Group

• m. NHS England

• n. NHS England Deployment Costs Position Statement

5. Updated Assessment Report in response to stakeholder comments on the Assessment Report and model from School of Health and Related Research (ScHARR)

• a. Updated Assessment Group report

• b. Assessment Group Erratum

6. Pre-print : PANORAMIC trial molnupiravir preliminary analysis

Any information supplied to NICE which has been marked as confidential, has been redacted. All personal information has also been redacted.

==> picture [209 x 22] intentionally omitted <==

Therapeutics for people with COVID-19 [ID4038]

Assessment Report Commercial in Confidence stripped version for consultation

Produced by: School of Health and Related Research (ScHARR), The University of Sheffield

STRICTLY CONFIDENTIAL

==> picture [127 x 54] intentionally omitted <==

Baricitinib, casirivimab/imdevimab, lenzilumab, molnupiravir, nirmatrelvir/ritonavir, remdesivir, sotrovimab, and tocilizumab, for the treatment of COVID-19. An economic

evaluation

Produced by School of Health and Related Research (ScHARR), The University of Sheffield Authors Andrew Metry, Research Associate, ScHARR, University of Sheffield, Sheffield, UK

Abdullah Pandor, Senior Research Fellow, ScHARR, University of Sheffield, Sheffield, UK

Shijie Ren, Senior Research Fellow, ScHARR, University of Sheffield, Sheffield, UK

Andrea Shippam, Programme Administrator, ScHARR, University of Sheffield, Sheffield, UK

Mark Clowes, Information Specialist, ScHARR, University of Sheffield, Sheffield, UK

Paul Dark, Professor of Critical Care Medicine, University of Manchester, Honorary Consultant, Northern Care Alliance NHS Foundation Trust (Salford Care Organisation), UK

Ronan McMullan, Professor of Medical Microbiology, School of Medicine, Dentistry and Biomedical Sciences, Wellcome Wolfson Institute for Experimental Medicine, Queen’s University Belfast, UK Matt Stevenson, Professor of Health Technology Assessment, ScHARR, University of Sheffield, Sheffield, UK

Correspondence Author Andrew Metry, Research Associate, ScHARR, University of Sheffield, Sheffield, UK

Word Count 15,163

Date completed 30/06/2022

1

Source of funding : This report was commissioned by the NIHR Evidence Synthesis Programme as project number 135564.

Declared competing interests of the authors

Paul Dark is the National Deputy Medical Director of a National Institute for Health and Care Research Clinical Research Network. He is a Local Principal Investigator for both the RECOVERY and REMAPCAP platform trials, NIHR Urgent Public Health (UPH) pandemic research Advisory Group Lead Link for REMAP-CAP and specialist member of NIHR UPH Advisory Group. His NHS host hospital Research and Innovation Department has been contracted and paid to provide advice on the use of tocilizumab for Roche and sotrovimab for GlaxoSmithKline both in COVID-19. He supported the activity as a named NHS expert employed by the Northern Care Alliance NHS Foundation Trust (Salford Care Organisation) but received no personal payments. There are no other conflicts of interest within this project team.

Rider on responsibility for report

The views expressed in this report are those of the authors and not necessarily those of the NIHR Evidence Synthesis Programme. Any errors are the responsibility of the authors.

This report should be referenced as follows :

Metry A, Pandor A, Ren S, Shippam A, Clowes M, Dark P, McMullan R, Stevenson M. Therapeutics for people with COVID-19 [ID4038]. A Multiple Technology Appraisal . School of Health and Related Research (ScHARR), 2022.

Keywords: COVID-19; cost-benefit analysis economic evaluation; cost-effectiveness; cost-utility analysis

Copyright belongs to The University of Sheffield.

2

Abstract

Background: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the virus that causes coronavirus disease 2019 (COVID-19). Over six million deaths worldwide have been associated with COVID-19.

Objective. To assess the cost-effectiveness of eight treatments used for the treatment of COVID-19 in hospital or used in the community in patients with COVID-19 at high-risk of hospitalisation.

Perspective: Treatments provided in UK hospital and community settings.

Methods: Clinical effectiveness estimates were taken from the COVID-NMA initiative and the metaEvidence initiative. A mathematical model was constructed to explore how the estimated efficacy for interventions used in hospital and for those at high-risk in the community impacted on patient health, measured in quality-adjusted life years (QALYs) gained. The costs associated with treatment, including those of hospital care, were also estimated and used to form a cost per QALY gained value which was compared with thresholds published by the National Institute for Health and Care Excellence (NICE). Estimates of cost-effectiveness compared against current standard of care (SoC) were produced and a full incremental analysis performed.

Results: The treatments were estimated to be clinically effective although not all reached statistical significance. All treatments in the hospital setting were estimated to plausibly have a cost per QALY gained value below NICE’s threshold when compared with SoC. This conclusion held for interventions used in the community although cost per QALY values were higher than in the hospital setting. Full incremental analyses indicated that baricitinib may be the most cost-effective treatment in a hospital setting and that nirmatrelvir with ritonavir (at an estimated price) may be the most cost-effective treatment in the community setting. However, there is considerable uncertainty in the results of the full incremental analyses due to heterogeneity in the pivotal studies and imprecision in estimates due to the small number of observed events and some treatments may have cost per QALY values greater than NICE’s published thresholds.

Limitations: The decision problem has evolved in terms of improved SoC, vaccination status, history of being infected with SARS-CoV-2, and the prevalent SARS-CoV-2 variant. As such, studies do not reflect the current conditions. Therefore, many assumptions were required that limit the accuracy of the estimates of clinical- and cost-effectiveness. No head-to-head studies of interventions were identified for use in the model. Placeholder costs were used for some interventions and patient access schemes were not incorporated.

3

Conclusions: The results produced should be informative to decision makers, although conclusions regarding the most clinical—and cost-effective intervention in these settings should be tentative given the heterogeneity between studies, the evolving nature of the decision problem and the uncertainty in the costs of interventions.

Future work: Research assessing the relative clinical effectiveness of interventions within head-to-head studies would be beneficial. Contemporary information related to the probability of hospital admission and death for patients at high-risk in the community would improve the precision of the estimates generated as would ascertaining the average age of this population. Value of information analyses may efficiently direct future research.

Word Count 492

Funding: This report was commissioned by the National Institute for Health Research (NIHR) Evidence Synthesis programme as project number 135564. This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. XXX, No. XXX.

See the NIHR Journals Library website for further project information.

4

5

6

Figure 21: The NMB results for patients admitted to hospital who do not require supplemental oxygen when the duration of long COVID is halved and doubled ................................................ 66 Figure 22: The NMB results for patients in the community with COVID-19 who are at high-risk of hospitalisation when the duration of long COVID is halved and doubled ........................ 66 Figure 23: The NMB results for patients in the community with COVID-19 who are at high-risk of hospitalisation when the hospital admission percentage was changed ............................. 67 Figure 24: The NMB results for patients in the community with COVID-19 who are at high-risk of hospitalisation when the age was changed from 65 years to 60 years and 70 years ......... 68

7

Plain English Summary

COVID-19 is an infectious disease that can cause death and long-term ill-health. Treatments exist that can be provided in hospital to reduce the number of deaths from COVID-19. Treatments also exist which can be provided in the community for people at high-risk of needing to be admitted to hospital to reduce the number of admissions and to reduce the number of deaths from COVID-19. However, the value for money of these treatments have not been estimated. We took the clinical effectiveness of eight treatments from published literature sources and built a model that estimated the value for money of each treatment compared with care without these treatments. The results of the model showed that many treatments in a hospital setting had estimates of cost-effectiveness that would normally be seen to be good value for money using the thresholds published by the National Institute of Health and Care Excellence as did some treatments in a community setting. Comparing treatments directly was difficult as the studies which reported on the clinical effectiveness were different in many ways. These differences included 1) the treatments used in current care at the time the study was conducted, as better drugs are now used than when COVID-19 was first identified, 2) the proportion of people who have had vaccinations or who had previously had COVID-19 or the virus that causes COVID-19, and 3) the variant of the virus causing COVID-19. Because of these differences, and the unknown price of some interventions, we could not confidently say which treatment helped patients the most or which treatment represented the best value for money.

Scientific Summary

Background

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the virus that causes coronavirus disease 2019 (COVID-19). At the time of writing (June 2022) there had been over 540 million confirmed cases and over six million deaths worldwide associated with COVID-19. For the UK, these values are over 22 million cases and 175,000 deaths.

In addition to the widespread vaccination programme, treatments exist that can help people who have been hospitalised due to COVID-19 (casirivimab and imdevimab (henceforth casirivimab/imdevimab), tocilizumab, remdesivir, baricitinib, baricitinib and remdesivir, and lenzilumab) or be used in patients who have COVID-19 and are at high-risk of needing hospitalisation (casirivimab/imdevimab, molnupiravir, nirmatrelvir and ritonavir (henceforth nirmatrelvir/ritonavir), remdesivir, and sotrovimab). For reasons related to urgency, these treatments, unlike interventions in other disease areas, have not received positive guidance from the National Institute of Health and Care Excellence before being routinely used. As the pandemic subsides there is more need for a formal evaluation of the clinical and cost-effectiveness of these treatments.

Objectives

The objective of this study is to summarise the current knowledge related to the clinical efficacy of the interventions and to conduct an economic evaluation that estimates the cost-effectiveness of each intervention against standard of care (SoC), as of June 2022, and to perform a full incremental analysis, whilst noting the caveats in the comparison of all interventions simultaneously.

Methods

Given the timescale of the project, where there was less than three months between the publication of the final scope and the report deadline, a literature review following best practice was not possible. Instead, a pragmatic, alternative approach was undertaken where evidence was taken from two living systematic reviews (supported by the COVID-NMA initiative and the metaEvidence initiative). For interventions related to use in hospitals, data were extracted on time to death, clinical improvement, and time to discharge. For interventions which are used in the community for patients at high-risk of hospitalisation, data were extracted on the risks of hospitalisation or death, and the risks of death. These measures of efficacy were assumed transportable to June 2022 despite changes in background conditions which include the SoC, the percentage of people who have been vaccinated and a change in the dominant SARS-CoV-2 variant.

A mathematical model was constructed that used the data from the living systematic reviews to simulate the experiences of patients in hospital, and requirement for supplemental oxygen, until discharge or

death in hospital. Due to the (conditional) marketing authorisations of the interventions, the model was developed such that results could be produced for the supplemental oxygen group and the nonsupplemental oxygen group separately. The model structure utilised an eight-point ordinal scale that was used in clinical trials to categorise patients during their admissions. Outputs from this model included the costs associated with interventions and care, and the quality-adjusted life-years (QALYs) gained by the patient both within the hospital episode and after discharge, incorporating decrements in health-related quality of life associated with the lasting impact of COVID-19. For interventions used in the hospital, these values allowed a cost per QALY gained to be calculated for each treatment compared with SoC, and for a full incremental analysis to be conducted.

The costs of each intervention were taken from public sources where available. However, tocilizumab and baricitinib have confidential patient access schemes agreed, which discount the price of the intervention, and are not considered in this document, but were provided to the NICE Appraisal Committee in a separate confidential appendix. The price of some treatments (casirivimab/imdevimab, molnupiravir and nirmatrelvir/ritonavir) were unknown at the time of writing and placeholder prices were used in the report.

For patients at high-risk of hospitalisation treated in the community, a decision tree was put before the hospital model, to simulate the reduced need for hospitalisation associated with early treatment. The total costs and QALYs associated with treatment options were estimated to allow an evaluation of the cost per QALY of each treatment against SoC and for a full incremental analysis to be undertaken. The modelling did not assess the logistics of treatment in the community, but the External Assessment Group notes that this could be a large factor in deciding which treatments could be preferred, as oral treatments could be more acceptable to patients and healthcare systems than treatments that are given intravenously or subcutaneously.

Three scenarios were run changing the efficacy of interventions. The mean efficacy estimate used the mean of each distribution extracted from the living systematic reviews, the high efficacy estimate used the most favourable limits of the 95% CIs and the low efficacy estimate used the least favourable limits of the 95% CIs.

Three scenario analyses were run that explored: the impact of changing the assumed average duration of health impact associated with COVID-19 (henceforth denoted long COVID); the proportion that are admitted to hospital of people in the community with COVID-19 at high risk of hospitalisation; and the average age of people with COVID-19 at high risk of hospitalisation.

10

Results were presented in terms of incremental cost-effectiveness ratios (ICERs) measured in cost per QALYs gained.

Results

All treatments used for hospitalised patients, had a median hazard ratio (HR) for death below 1, indicating a benefit, although all confidence intervals (CIs) crossed unity apart from those for tocilizumab and baricitinib. The overlapping CIs, and heterogeneous studies meant that no firm conclusions could be made regarding the relative efficacy of these treatments. There was less data relating to the relative risks (RRs) of clinical improvement at 28 days and the HRs for the time to discharge, although these were generally close to unity and had CIs that crossed unity. No clear conclusions could be made on the relative efficacy of treatments for these two measures.

All treatments used in the community had favourable median RRs for hospitalisation and death at 28 days, although due to wide CIs no firm conclusions could be made regarding the relative efficacy of these treatments. The median RR associated with death at 28 days were favourable for all interventions, except for remdesivir where the median estimate was unity. The CIs were wide and spanned 1 for all treatments except for molnupiravir and nirmatrelvir/ritonavir. As such, no clear conclusions relating to the relative efficacy of the interventions could be made regarding avoiding death at 28 days.

For hospitalised patients requiring supplemental oxygen, all treatments except lenzilumab, had estimated ICERs compared with SoC below £10,000 in both the mean efficacy and high efficacy scenarios; the value for lenzilumab was below £20,000. However, in the low efficacy scenario only baricitinib and tocilizumab generated more QALYs than SoC and had estimated ICERs under £20,000.

For hospitalised patients not requiring supplemental oxygen, all treatments except lenzilumab had estimated ICERs compared with SoC below £10,000 in both the mean efficacy and high efficacy scenarios; the corresponding ICER for lenzilumab was below £25,000. However, in the low efficacy scenario only baricitinib generated more QALYs than SoC and the estimated ICER for baricitinib compared with SoC was under £5,000.

For interventions used in the community, the estimated ICERs compared with SoC were more varied. For all interventions except molnupiravir and nirmatrelvir/ritonavir, the ICERs compared with SoC were in excess of £65,000 in the mean efficacy scenario. In the high efficacy scenario, all interventions except molnupiravir and nirmatrelvir/ritonavir had ICERs compared with SoC above £20,000. In the low efficacy scenario, all interventions except molnupiravir and nirmatrelvir/ritonavir produced less QALYs than SoC. In the mean efficacy scenario and the high efficacy scenario both molnupiravir and nirmatrelvir/ritonavir had ICERs below £15,000. In the low efficacy scenario, the ICER for

11

nirmatrelvir/ritonavir compared with SoC was below £10,000, although the ICER for molnupiravir was greater than £65,000.

The efficiency frontiers based on the full incremental analyses differed based on setting and efficacy scenario. For patients in hospital requiring supplemental oxygen, baricitinib was the intervention that produced most QALYs and had an ICER below £10,000 compared with the previous intervention on the efficiency frontier in both the mean efficacy scenario and the low efficacy scenario. In the high efficacy scenario, baricitinib and remdesivir were the interventions on the efficiency frontier with most QALYs and had ICERs compared with the previous intervention on the efficiency frontier below £20,000.

For patients not requiring supplemental oxygen, baricitinib was the intervention that produced most QALYs and had a cost per QALY below £5000 compared with the previous intervention on the efficiency frontier in both the mean efficacy and low efficacy scenarios. In the high efficacy scenario, baricitinib and remdesivir were the interventions on the efficiency frontier with most QALYs and had ICERs compared with the previous intervention on the efficiency frontier below £15,000.

For patients at high-risk of hospitalisation treated in the community, nirmatrelvir/ritonavir was the intervention that produced most QALYs and had a cost per QALY below £10,000 compared with the previous intervention on the efficiency frontier in all of the efficacy scenarios explored.

However, the comparative results are highly uncertain due to the wide CIs associated with each intervention and the heterogeneity associated with the pivotal studies. An additional uncertainty was the unconfirmed prices of nirmatrelvir/ritonavir and molnupiravir at the time of writing and the use of list prices where patient access schemes are available.

In the scenario analyses, the proportion of people with COVID-19 in the community at high-risk of hospitalisation who are hospitalised when treated with SoC had a large impact on the ICERs with treatments becoming more cost-effective as the admission proportion increased. The average age of people in the community with COVID-19 at high-risk of hospitalisation also had a marked impact on the ICERs with younger people making the drugs more cost-effective. The assumed duration of long COVID had a lower impact on the ICERs than the previous scenarios, although shorter durations of long COVID were associated with the treatments becoming more cost-effective.

Conclusions

There is considerable uncertainty in the efficacy of treatments compared to SoC due to the small number of observed events in studies, which result in wide CIs for HRs and RRs. Additionally, the SoC, the

12

percentage of people who have had a vaccination, and the dominant SARS-CoV-2 variant could all vary between pivotal studies. Some treatments (tocilizumab and baricitinib in the hospitalised setting and molnupiravir and nirmatrelvir/ritonavir in the community setting) were estimated to have a statistically significant benefit related to death due to COVID-19, however, this may also have been shown for other treatments if the pivotal studies had had larger sample sizes.

Multiple treatments have been shown to be cost-effective against SoC for patients in hospital, and for patients at high-risk of hospitalisation in the community. Full incremental analyses have been conducted, which indicated in the mean efficacy analyses that baricitinib was the most cost-effective treatment in hospital if a cost per QALY of £10,000 was deemed acceptable, and that nirmatrelvir/ritonavir was the most cost-effective treatment in the community setting if a cost per QALY of £5000 was deemed acceptable. However, the results are uncertain due to the wide CIs, the heterogeneity between pivotal studies, and the unconfirmed prices of nirmatrelvir/ritonavir and molnupiravir. In some scenarios, baricitinib and remdesivir were the most cost-effective if a cost per QALY of £20,000 was deemed acceptable. Furthermore, some treatments have patients access schemes which have not been incorporated in the analyses and the prices of some interventions are currently unknown.

Word Count 1877

13

14

ABBREVIATIONS

15

1. BACKGROUND

1.1 Description of the underlying health problem

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the virus that causes coronavirus disease 2019 (COVID-19). At the time of writing (June 2022) there had been more than 540 million cases of COVID-19 worldwide and more than 6 million deaths; in the UK these values were more than 22 million cases and over 175,000 deaths.[1] In the UK, there have been waves of infections (peaking in late December 2021 and early January 2022), and waves of death (peaking in January 2021).[1]

The ratio of notified infections to death in the UK has changed markedly over time, being approximately 5 to 1 in April 2020, 45 to 1 in January 2021; and 700 to 1 in January 2022 (authors’ calculations based on worldometer data[1] ). Factors associated with the change in ratio include:

• better ascertainment of COVID-19 cases, which previously may have been left unobserved particularly early in the pandemic especially when mild or asymptomatic;

• increasing level of protection in the population, both acquired from previous SARSCoV-2 infection and vaccine-induced;

• improved levels of treatment, such as the use of dexamethasone;

• the likelihood of more fragile people dying in earlier waves; and

• the potential change in variants of SARS-CoV-2.

Should the risk of death following COVID-19 remain at low levels and SARS-CoV-2 becomes endemic in society, then treatments for patients with COVID-19 may no longer be treated differently to interventions for other conditions such as breast cancer or heart disease. If this were the case, then it could be considered logical and acceptable that pharmacological treatment for COVID-19 would be appraised by the National Institute for Health and Care Excellence (NICE) using its standard methods.[2]

1.2 The NICE scope

In April 2022, NICE issued a final scope[3] for the assessment of therapeutics for people with COVID-19; the NICE website also hosts the final protocol written by the External Assessment Group (EAG).[4] The remit of the final scope was to appraise the clinical and cost-effectiveness of eight interventions for treating (i) people with mild COVID-19 at high-risk of progressing to severe COVID-19 and (ii) people with severe COVID-19. The comparators included established clinical management in clinical practice with or without corticosteroids and appropriate respiratory support, and other interventions. The components of the decision

16

problem are discussed more fully in Section 1.4. The deadline for the EAG report was the 30th of June 2022, allowing less than three months for the estimates of the clinical effectiveness of each intervention to be made, for the mathematical models to be adapted and run, the results to be interpreted and the report to be written.

1.3 Description of current service provision

Patients with severe COVID-19 are typically hospitalised with the intensity of treatment dependent on the severity of the condition. Patients may be treated in intensive care units (ICUs), be provided with high-flow oxygen or low-flow oxygen, and be treated with interventions, including those in the NICE scope and with corticosteroids.

1.4 The Decision Problem

This section has been sub-divided into sections detailing the population, interventions, comparators, outcome measures, and subgroups.

1.4.1 Population

The population considered within the EAG report has been divided into two broad groups. The first group consists of people who have been hospitalised due to COVID-19 and the second group consists of people who are at high-risk of requiring hospital care due to COVID-19. Patients who were hospitalised for reasons other than COVID-19 and contracted COVID-19 in hospital and were at high-risk of requiring hospital care for COVID-19 in itself were categorised within the second group. For brevity, all patients not hospitalised due to COVID19 who are at high-risk of hospitalisation will be termed ‘non-hospitalised patients’ noting the aforementioned caveat regarding patients who contract COVID-19 in hospital, whereas patients who have been hospitalised directly because of COVID-19 are referred to as ‘hospitalised patients’.

Following discussions with NICE, the definition for patients at high-risk was aligned to that considered within the Platform Adaptive trial of NOvel antiviRals for eArly treatMent of COVID-19 In the Community (PANORAMIC) clinical study,[5] with the exception that being aged 50 years or over was not considered to be a high-risk factor.

The aim of treatment differs between each group. For patients hospitalised due to severe or critical COVID-19, the aim of treatment is to reduce the immunoinflammatory response of the body and prevent clinical deterioration. For non-hospitalised patients, the aim of treatment is to prevent viral replication and damp inflammation, thus reduce the probability of the development of severe symptoms that could lead to hospitalisation and death.

17

1.4.2 Interventions

The interventions listed within the NICE scope[3] , excluding anakinra which was withdrawn from the appraisal are shown in Table 1 to Table 3 based on marketing authorisation in the UK at the time of writing. Table 1 contains the interventions with marketing authorisation in the UK, Table 2 contains the interventions with conditional marketing authorisation in the UK, and Table 3 contains the interventions with no marketing authorisation in the UK. Each table contains the generic name of the intervention, its branded name and the company manufacturing it, the class of intervention, the mode of administration and recommended dose. Table 1 provides the indication for the drug, whilst Table 2 and Table 3 provide the population in key studies for the intervention.

Multiple interventions are indicated for the prevention of severe COVID-19. Severe disease in adults is defined as having clinical signs of pneumonia plus at least one of the following: respiratory rate >30 breaths/minute, severe respiratory distress, or saturation of peripheral oxygen <90% on room air and would require hospitalisation.[6]

1.4.3 Comparators

The comparators within the decision problem include all of the interventions contained in Table 1 to Table 3, when used in the same position as a particular intervention and additionally standard of care (SoC) which would be dependent on the severity of the patient’s illness. SoC is defined as any treatment widely accepted by the National Health Service (NHS) as SoC, which is routinely funded by the NHS with no strong rationale to appraise it, for example supplemental oxygen and dexamethasone. SoC has evolved throughout the COVID-19 pandemic, which means that randomised controlled trials (RCTs) conducted comparing interventions against SoC may not be directly comparable as SoC has improved over time.

18

Table 1: Interventions with marketing authorisation in the UK as of the 28[th] of June 2022

IV - intravenous, mAb – monoclonal antibody, SC – subcutaneous

19

Table 2: Interventions with conditional marketing authorisation in the UK as of the 28[th] of June 2022

IV - intravenous, mAb - monoclonal antibody, SC – subcutaneous, SmPC – summary of product characteristics

20

Table 3: Interventions with no marketing authorisation in the UK as of the 28[th] of June 2022

IV – intravenous, mAb – monoclonal antibody

21

1.4.4 Outcome Measures

The NICE scope[7] lists nine possible outcomes to explore: mortality; requirement for respiratory support; time to recovery; hospitalisation (requirement and duration); time to return to normal activities; virological outcomes (viral shedding and viral load); post-COVID-19 symptoms; adverse effects of treatments; and health-related quality of life (HRQoL). All model outcomes, except virological outcomes were assessed.

The cost-effectiveness of the eight treatments were expressed in terms of incremental cost-effectiveness ratios (ICERs) which were reported in terms of cost per quality-adjusted life year (QALY) gained. A patient lifetime horizon was used to take differential mortality between treatments into account.

1.4.5 Subgroups

Due to time constraints, the only subgrouping considered was related to whether oxygen was required upon admission to hospital entry. This was considered important as the licensed indication and the clinical outcomes for some of the appraised interventions depend on the level of oxygen support required. The EAG is aware that other possible criteria for selecting subgroups include, but are not limited to: age; immune system competence; comorbidities; seroprevalence; vaccination status; and the predominant SARS-CoV-2 variant but did not have the time to explore the impact of these characteristics.

22

2. CLINICAL-EFFECTIVENESS

2.1 Methods for the Rapid Evidence Review

Given the timelines of the project, the EAG could not follow best practice for systematically reviewing the clinical evidence relevant to the decision problem. Following discussions with NICE, a pragmatic, alternative approach was undertaken relying on the use of data extracted by third-parties which are referred to as ‘living systematic reviews’. The methods used, assumptions taken, and the summarised results are provided in this chapter.

2.1.1. Rationale for using living systematic reviews

COVID-19 clinical research has accelerated dramatically worldwide, with over 5000 registered trials investigating therapeutic interventions for COVID-19.[8] The need for rapid information on COVID-19 has resulted in a paradigm shift, especially in the communication of scientific results. Traditional systematic reviews can date quickly but ‘living’ systematic reviews search for evidence much more regularly than standard reviews and incorporate relevant new evidence as it becomes available. This is important in the context of COVID-19, in which the evidence-base is rapidly changing as new data emerge. The ability of a ‘living’ systematic review and network meta-analysis (NMA) to regularly update and incorporate relevant new evidence as it becomes available makes it the best type of evidence synthesis, in the opinion of the EAG, to inform this pragmatic rapid evaluation.

2.1.2. Selection criteria for the living systematic reviews

Several living systematic reviews that incorporate emerging trial data and allow for analysis of comparative effectiveness of multiple COVID-19 treatments, have been robustly developed and published.[8-11] Two sources were selected as they provided detailed relevant outcome data from individual studies and up-to-date evidence synthesis to inform the model.

The first source is the COVID-NMA initiative,[9, 12] supported by the World Health Organization (WHO) and Cochrane which is a living systematic review of registered randomised trials, in which all available evidence related to COVID-19 is regularly collected, critically appraised, and synthesised using pairwise comparisons and NMA methods. The analyses are updated every two weeks and results can be accessed via a web interface (https://covid-nma.com/).

The second source is the metaEvidence initiative,[10] supported by the University Hospital of Lyon and the University of Lyon which is also a living meta-analysis and evidence synthesis of therapies for COVID-19 and is an emerging online resource that provides direct access to the efficacy and safety results reported in the studies for potential drugs for the treatment of COVID-19. The risk of bias,

23

synthesised by meta-analysis, is also reported. The analyses are updated within a target time of less than 24 hours and results can be accessed through a web interface at http://www.metaevidence.org/COVID19.aspx.

Other sources of evidence, which primarily informed living guidelines,[8, 11] were deemed to lack full transparency in the extracted outcome data from individual studies. As such, they precluded further synthesis and evaluation and could even threaten the validity of the evidence synthesis.

2.1.3. Assumption of transportability of relative treatment effects

A consequence of the need to use data from the living systematic reviews was that the scope for the EAG to undertake nuanced analyses was reduced. An assumption was needed that all relative treatment effects were transportable to different settings. This meant that the same treatment effects, either hazard ratios (HRs) or relative risks (RRs), were assumed applicable regardless of study characteristics which include: the age, perceived severity, vaccination status, and history of SARS-CoV-2 infection of patients; the SoC at that time; the geographical location; and the dosage of the intervention used. It is acknowledged that this assumption may be incorrect, which adds additional uncertainty to the clinicaland cost-effectiveness results.

2.1.4 Inclusion criteria and data extraction

Data for the interventions contained in Table 1, Table 2 and Table 3 were extracted. Key model outcomes such as time to death, clinical improvement at day 28 or day 60 (defined as a hospital discharge or improvement on the scale used by trialists to evaluate clinical progression and recovery) and incidence of serious adverse events (SAEs) were initially extracted from the COVID-NMA living systematic review[9] . Where relevant outcome data were not available, these data were extracted from the metaEvidence living systematic review.[10] All data extractions (undertaken between the 16[th] of March to the 18[th] of May and updated between the 25[th] to the 31[st] of May 2022) were undertaken by one reviewer (AS) and checked by a second reviewer (AP), with any discrepancies resolved by a third reviewer (KR). All evidence synthesis analyses were extracted from data reported on the COVID-NMA and metaEvidence web interface; Double checks of the extracted data against the original RCT publications for accuracy could not be undertaken within the deadlines of the project.

2.1.5 Adjustments made for changing SoC, SARS-CoV-2 variant, vaccination status and prior infection

The conditions under which each study was evaluated were heterogeneous. Across time SoC has changed markedly, most particularly with reference to the widespread use of corticosteroids such as dexamethasone, and change in SARS-CoV-2 variants. The vaccine roll-out in England has provided

24

protection that was not available to patients recruited to early studies, similarly, there is likely to be an increased level of protection associated with prior infection. Ideally there would be attempts to establish the impact of different circumstances on the observed clinical effectiveness of interventions in studies, although this was not possible within the timescales of the project. As such, the EAG had to make a simplistic assumption that none of the changes were treatment effect modifiers, and that given this, the relative benefits observed in the studies were transportable and could be applied to the estimated outcomes for patients with COVID-19 in England in Summer 2022.

2.2 Results of the Rapid Evidence Review

This section reports key results from the analyses described in Section 2.1. A brief description of each included RCT, reproduced from the COVID-NMA Initiative,[9] is presented in Appendix 1. A summary of the extracted data for each intervention and relevant outcomes from the living systematic reviews is also presented in Appendix 1. The assumed clinical effectiveness for each intervention in hospitalised patients is detailed in Table 4, and in Table 5 for patients at high-risk of hospitalisation treated in the community. The interventions are listed in order of current marketing authorisation and alphabetical order. The values reported in Table 4 and in Table 5 are used in the economic evaluation. Where data were not available for clinical improvement or time to discharge a value of 1.0 was used as the model results were not sensitive to these values within the observed range associated with other interventions.

25

Table 4: Summarised clinical effectiveness data in patients hospitalised due to COVID-19

CI - confidence interval, HR - hazard ratio, RR - relative risk

26

Table 5: Summarised clinical effectiveness data for patients at high-risk of hospitalisation due to COVID-19

CI - confidence interval, HR - hazard ratio, RR - relative risk

• There were no deaths reported in either arm. This estimate is calculated assuming a continuity factor of 0.5 deaths and 1 extra observation was added to each arm.

To aid interpretation of the clinical efficacy data for interventions used to treat patients in hospital, plots of i) the HR for death at 28 days, ii) the RR for clinical improvement at 28 days, iii) the HR associated with time to discharge, iv) the probability that the intervention, based on the distribution extracted for clinical efficacy, is associated with more deaths at 28 days, and v) the ranked position of each intervention in 1000 joint samples of efficacy for all are shown in Figure 1, Figure 2, Figure 3, Figure 4 and Figure 5 respectively. Figure 1, Figure 2, and Figure 3 consist of two horizontal lines for each intervention which sit on a vertical line. The vertical line shows the lower and upper 95% confidence intervals (CI) whilst the lower horizontal line provides the median value, and the upper horizontal line provides the mean value from the distribution. When the mean and the median values are close these become indistinguishable in the figures.

As seen in Figure 1, all treatments have a beneficial mean estimate for the HR associated with death. The CIs of each treatment overlap showing that there is considerable uncertainty in the ranked order of clinical effectiveness. A similar conclusion related to the ranking of interventions for clinical improvement can be drawn from Figure 2, and for the ranking of treatments in relation to time to discharge from Figure 3, although only two interventions reported data on this measure.

27

==> picture [432 x 131] intentionally omitted <==

==> picture [432 x 130] intentionally omitted <==

Figure 1: The hazard ratio of avoiding death for interventions used to treat patients in hospital

==> picture [436 x 132] intentionally omitted <==

==> picture [436 x 133] intentionally omitted <==

Figure 2: The relative risk of clinical improvement at 28 days for interventions used to treat patients in hospital

28

==> picture [396 x 120] intentionally omitted <==

==> picture [396 x 120] intentionally omitted <==

Figure 3: The hazard ratio of discharge for interventions used to treat patients in hospital

Figure 4 indicates the probability that each intervention is associated with greater deaths than SoC at 28 days. For tocilizumab and baricitinib, this probability is very low. For casirivimab/imdevimab and lenzilumab, the probability is in excess of 0.1.

==> picture [396 x 122] intentionally omitted <==

==> picture [396 x 123] intentionally omitted <==

Figure 4: The probability that the intervention used in hospital is associated with increased mortality based on the lognormal distribution derived from the living systematic reviews

29

Figure 5 shows the large uncertainty in the ranking of each intervention in terms of efficacy, for example, baricitinib is the intervention with the greatest estimated probability of being ranked first, yet has similar probabilities of being ranked second, or of being third, fourth, fifth and sixth combined. To add additional uncertainty, the assumption that the efficacy estimate is transportable to different settings may be incorrect.

==> picture [428 x 93] intentionally omitted <==

==> picture [428 x 92] intentionally omitted <==

Figure 5: The estimated probability that each intervention is ranked first through to sixth for hazard ratio for mortality

To aid interpretation of the clinical efficacy data for interventions used to treat patients in the community, plots of i) the RR for avoiding hospitalisation or death at 28 days, ii) the RR for avoiding death at 28 days, iii) the probability that the intervention, based on the distribution extracted for clinical efficacy, is associated with more deaths at 28 days and iv) the ranked position of each intervention in 1000 joint samples of efficacy for all interventions are shown in Figure 6, Figure 7, Figure 8 and respectively. Figure 6 and Figure 7 consist of two horizontal lines for each intervention which sit on a vertical line. The vertical line shows the lower and upper 95% CIs whilst the lower horizontal line provides the median value, and the upper horizontal line provides the mean value from the distribution. When the mean and the median values are close these become indistinguishable in the figures.

From Figure 6, it can be seen that no CI crosses unity, although the width of the CIs differ, with that of nirmatrelvir/ritonavir having most precision, although the CI associated with this intervention overlaps with that of casirivimab/imdevimab, remdesivir, and sotrovimab indicating considerable uncertainty in the most clinically effective intervention.

30

==> picture [350 x 107] intentionally omitted <==

==> picture [350 x 107] intentionally omitted <==

Figure 6: The relative risk of avoiding hospitalisation or death at 28 days for interventions used to treat patients in the community

For the avoidance of death at 28 days, Figure 7 indicates wide CIs for all treatments excluding molnupiravir and nirmatrelvir/ritonavir, in which the upper confidence limits do not exceed 1.0. The wide CIs are primarily related to the sample size and the small number of observed events in each arm.

==> picture [346 x 106] intentionally omitted <==

==> picture [346 x 105] intentionally omitted <==

Figure 7: The relative risk of avoiding death at 28 days for interventions used to treat patients in the community

These wide CIs mean that there is a considerable probability (of more than 0.1) that all interventions except molnupiravir and nirmatrelvir/ritonavir could increase the risk of death, although this is a

31

frequentist interpretation of the distribution and does not consider any correlation between reduced hospitalisation rates and the reduced probability of death.

==> picture [330 x 111] intentionally omitted <==

==> picture [330 x 110] intentionally omitted <==

Figure 8: The probability that the intervention is associated with increased mortality based on the lognormal distribution derived from the living systematic reviews

Figure 9 shows large uncertainty in the ranking of each intervention in terms of efficacy, for example, whilst nirmatrelvir/ritonavir has a large estimated probability (greater than 60%) of being ranked first, it has a 19% chance of being ranked third or lower. To add additional uncertainty, the assumption that the efficacy estimate is transportable to different settings may be incorrect.

==> picture [439 x 95] intentionally omitted <==

==> picture [439 x 94] intentionally omitted <==

Figure 9: The estimated probability that each intervention is ranked first through to fifth for preventing mortality at 28 days

32

3. METHODS FOR THE COST-EFFECTIVENESS ANALYSIS

The model framework for assessing the cost-effectiveness of treatments for people hospitalised due to COVID-19 is an adaptation of the approach taken by Rafia et al .[13] This decision was made for two principal reasons. Firstly, that there is an overlap in the authors for both the Rafia et al . paper and this report, meaning that the model was available to the team reducing model construction time. Secondly, this model structure was used in a preliminary appraisal of remdesivir that was undertaken by a NICE panel meeting;[14] whilst no formal documents related to this meeting has been released an author of this report (MS) was on the panel and believes that no significant issues were raised relating to the model structure.

For non-hospitalised patients, the model structure was based on that outlined in an unpublished report by the NICE Decision Support Unit which provided an early economic evaluation of neutralising monoclonal antibodies and oral antivirals for treating COVID-19 prior to hospitalisation.[15] This consisted of a decision-tree approach where patients who ultimately required hospital admission were evaluated in the hospital-based structure, whereas those that didn’t, remained in the community.

This section initially describes the model structures briefly, with later sections providing detail on the population of the parameters values used to generate the results within this report.

3.1 Model Structures

3.1.1 General model structure for hospitalised patients

The economic model was developed in Microsoft Excel and uses a partitioned survival approach (often referred to as area under the curve (AUC) approach) with three mutually exclusive health states; (a) discharged from hospital and alive, (b) hospitalised with or without COVID-19 and (c) death from any cause (COVID-19 or due to other causes).

Movements between health states are not explicitly modelled. Instead, the partitioned model estimates health state occupancy at each time interval. A simplified schematic of the model structure is shown in Figure 10. A daily cycle length is used until the end of parametric extrapolation, at day 70, after which a weekly cycle length is used. An initial daily cycle length was chosen to allow changes in treatment and/or hospitalisation and oxygen requirements that happen early in a patient’s stay to be modelled at a granular level. A cohort partitioned survival approach was chosen due to the limited time and the absence of individual patient data (IPD). A limitation of this approach is that it is not possible to track individual patients in the model which may have allowed a better representation of the patient experience.

33

==> picture [438 x 118] intentionally omitted <==

==> picture [438 x 118] intentionally omitted <==

Figure 10: Simplified schematic of model structure (values are for illustration only)

Whilst in hospital, the 8-point ordinal scale of clinical status (an inverted version of the scale originally developed for severe influenza requiring hospitalisation as recommended by the World Health Organization (WHO)) used in the Adaptive COVID-19 Treatment Trial (ACTT-1) RCT,[16] and in the Remdesivir Effectiveness Evaluation Study (REES)[17] is used. This ordinal scale is described in Table 6 and is used in the model to (1) define the population at baseline in terms of oxygen requirements at the start of treatment, and (2) estimate changes in hospital/oxygen requirements during the hospital stay.

34

Table 6: Eight-points ordinal scale of clinical status used in ACTT-1[16]

When evaluating the interventions, patients enter the hospital model based on the marketing authorisation, where this has been granted, or in relation to the population in the key studies. A schematic of the positioning (or anticipated positioning when marketing authorisation has not been granted) of each intervention in Table 1 to Table 3 is provided in Table 7 with reference to the 8-point ordinal scale detailed in Table 6. Scale values of 1 or 2 describe patients with COVID-19 in the community whilst values 3 or higher describe patients in hospital. Only the latter group are relevant for the hospital model, although scale 3 does not require ongoing medical care.

35

Table 7: The positioning of treatments based on the 8-point ordinal scale

Cas and imd – casirivimab/imdevimab; Nirm and rit – nirmatrelvir/ritonavir; Bari and rem – baricitinib and remdesivir Δ – with one risk factor for developing severe illness,  - when receiving corticosteroids,  - in patients with pneumonia Interventions are permitted in cells shaded green and not permitted in cells shaded peach

Movements (improvement or worsening) between the different hospitalisation/oxygen requirements over time is modelled with each scale being associated with cost and HRQoL implications. During their hospital stay, patients are distributed according to their hospital/oxygen requirement derived from the placebo arm of the ACTT-1 study and additional assumptions where necessary. An illustration of movement between ordinal scales is shown in Figure 11 for patients who needed supplemental oxygen on hospital entry and when treated with SoC. The model assumes that all patients are discharged at 70 days. This may underestimate the costs and QALY losses associated with hospital care for the most efficacious drugs, although this is not expected to be a large limitation as the proportions of patients estimated to be in hospital at day 70 is relatively small.

36

==> picture [453 x 93] intentionally omitted <==

==> picture [453 x 93] intentionally omitted <==

==> picture [453 x 93] intentionally omitted <==

Figure 11: Illustration of ordinal scale occupancy during hospital stay of a cohort admitted to hospital requiring supplemental oxygen and receiving SoC treatment

Pivotal clinical trials/studies for treatments for COVID-19 used in this economic evaluation tend to follow patients and typically collect key clinical outcomes after 28 days of follow-up. It is, therefore, necessary to extrapolate beyond the duration of studies to capture the life expectancy and HRQoL following hospital discharge from COVID-19. Following discharge patients with COVID-19 are at an elevated risk of death,[18] emerging evidence suggest that some patients discharged with COVID-19 continue to experience symptoms and have a reduced quality of life,[19-28] may require re-admission due to COVID-19,[16, 29-33] and are at an elevated risk to experience multi-organ dysfunctions[18] (such as respiratory diseases, diabetes, cardiovascular, liver and kidney diseases) and may require long term management/monitoring.[34] Within the model, HRQoL reductions and additional costs associated with COVID-19 have been included; for brevity this has been termed ‘long COVID’. In addition, the possibility of patients having an increased risk of death following COVID-19 has been modelled using a standardised mortality rate (SMR) applied to the mortality rates for an age- and sex-matched population.

Consequently, a seven-step approach is employed:

• Step 1: use of a parametric function (hazard spline model with 3 knots) fitted to the relevant outcomes (time to death and time to discharge) for all patients on the SoC arm in RECOVERY study[35] for the first 28 days, as used in Rafia et al .[13]

37

• Step 2: This parametric function is adjusted to reflect the outcomes at day 28 as reported in the literature to reflect the benefit of using corticosteroids, which represent the current SoC for patients in need of supplemental oxygen.[36] The model was calibrated as detailed in Section 3.6.2,

• Step 3: Treatment effect in the form of hazard ratios (HRs) or RRs for the interventions were applied to the SoC curves. Data were missing for some interventions with respect to the HR for discharge and the HR for clinical improvement (see Section 2.2). The EAG noted that given the values for other interventions, neither were large drivers of the cost-effectiveness results, and that there was no clear relationship between these and other variables. Therefore, as no values for interventions with data were markedly different from unity when compared with SOC, the EAG decided to use the values for SoC where data were missing,

• Step 4: As shown in Figure 11, ordinal scale occupancy in hospital is assumed to last until the distribution for overall survival (OS) and the distribution for time to discharge intersect. It was assumed in the model that none of the hospitalised cohort would remain in hospital after 70 days,

• Step 5: parametric extrapolation is employed to estimate the rates of death between day 28 until day 70 in the base case,

• Step 6: use of mortality rates from the general population, adjusted by an SMR for the assumed mean duration of long COVID to reflect the elevated risk of death in patients with COVID-19 discharged from hospital,

• Step 7: use of unadjusted mortality rate from the general population after the assumed mean duration of long COVID.

3.1.2 General model structure for non-hospitalised patients

The model structure used for assessing interventions that can be provided to patients with COVID-19 and at high-risk of hospitalisation is depicted in Figure 12. This is comprised a decision tree which simulates whether hospitalisation is required or not, and for those patients who are hospitalised, whether supplemental oxygen is required on admission.

38

==> picture [440 x 238] intentionally omitted <==

Figure 12: Structure of the decision tree used for the non-hospitalised cohort

Hospitalisation rates for patients on SoC were taken from Nyberg et al .[37] where recent risks of hospitalisations associated with the Omicron variant were reported. Hospital admission up to 14 days after positive test was approximately 0.9% (over 9,000 patients admitted from a reported million cases). This value differed from the rates reported in the UK Coronavirus dashboard, in this source there were two million positive cases in the past three months and a hundred thousand admissions (implying a rate of 5%). Clinical advice given to the EAG is that although the dashboard has a much larger sample size, the data is less nuanced and does not allow attributions of COVID-19 to admissions, and it may be the case that half of patients with COVID-19 in hospital, were not hospitalised due to COVID-19. Hence, the EAG adopted the 0.9% rate in its base case and increased it in sensitivity analyses (see Section 3.4).

Since interventions for this group of patients are indicated for those with high risk of hospitalisation, the underlying risk had to be inflated from the average. The EAG reviewed data presented in HippisleyCox et al .[38] where the QCovid3 model was used to calculate cause specific hazard ratios for COVID19 hospital admissions after vaccination for subgroups with different comorbidities. Based on these data and clinical advice, the EAG applied a multiplier of 2 to the average hospitalisation rate for all patients to estimate the rate in people at high-risk of hospitalisation in the base case and increased it in sensitivity analyses (see Section 3.4). The proportion of hospitalised patients requiring supplemental oxygen was estimated from an ISARIC report[39] where the requiring oxygen of any level on admission was calculated at 81% (55% high flow oxygen, 16% non-invasive ventilation, and 10% invasive ventilation).

39

The model applies an RR to account for other medical attended visits (MAVs) (i.e., visits other than hospital admission) compared to admissions. This RR was estimated from data in Nyberg et al.[37] and was equal to 1.37 (1.23% MAV rate divided by 0.9% hospitalisation rate). Only costs were considered for MAVs and incorporated a visit to an accident and emergency department.

Two key clinical outcomes were extracted from the living systematic reviews: RRs for hospitalisation or death, and RRs for day 28 all-cause mortality, which are shown in Figure 6 and Figure 7 respectively. The RR for hospitalisation or death was assumed to apply for hospitalisations only due to the relatively low mortality rate compared to the admission rate. A separate RR was calculated for each intervention for deaths within hospital such that the overall RR for death at 28 days was consistent with the published estimate reported in Table 4 and Table 5. This methodology assumes that there were no deaths amongst non-hospitalised patients in the first 28 days of the model. The EAG believes that this limitation would have a negligible impact on the ICER.

For patients treated in the community it was assumed that there would be no further active treatment in hospital, and thus patients receive SoC only. This decision was based on the following factors: that the RRs for mortality for some of the interventions used in the community were substantially lower than the HRs for those treatments used in hospital where the midpoint efficacy was beneficial. For example, the RR for death for nirmatrelvir/ritonavir was 0.04 whilst the midpoint HR for death for baricitinib was 0.61, indicating that the residual effect of nirmatrelvir/ritonavir was larger than the impact of baricitinib, which was the most efficacious hospital intervention based on midpoint values. Furthermore, there is no evidence for the synergistic effects (or not) of using multiple interventions.

The modelling did not assess the logistics of treatment in the community, but the EAG notes that this could be a large factor in deciding which treatments could be preferred, as oral treatments could be more acceptable to patients and healthcare systems than treatments that are given intravenously or subcutaneously.

3.2 Clinical Parameters and Inputs Used in this Rapid Assessment

3.2.1 Baseline characteristics after discharge

Age and gender distribution are used in the economic model to estimate both the rate of mortality beyond the duration of clinical evidence and to estimate HRQoL beyond hospitalisation and for the non-hospitalised cohort. The baseline mean age for the modelled hospitalised cohort was calculated from weekly Office for National Statistics (ONS) data[40] reported in the middle of May 2022. For patients with COVID-19, these data included rates of hospital admissions per 100,000 people and number of deaths, by age bands. These values were multiplied by population data obtained from the

40

ONS[41] to estimate the absolute number of admissions and deaths by age band. The estimated number of discharged patients was calculated by subtracting the number of deaths from the number of admissions. Table 8 presents the estimated numbers and percentages calculated for admission, death and discharge conditional on age band.

Table 8: Hospital Admission and Death weekly numbers and percentages by age band compared to the whole population (mid May 2022)

If the midpoint of each age band represented the entire band, mean ages for admission, death and discharge are estimated at 70.6, 82.8 and 68.7 years respectively. For the non-hospitalised cohort, it was presumed that the average age would be lower than for the hospitalised group, as older age was believed to be associated with a greater risk of hospitalisation. Without data to accurately estimate the age for people with COVID-19 at high-risk of hospitalisation who do not get hospitalised, an arbitrary value of 65 years was assumed with sensitivity analyses using 60 and 70 years; patients who are hospitalised due to COVID-19 have the same characteristics as patients in the hospital model.

The distribution between sexes was taken from an Intensive Care National Audit & Research Centre report[42] which reported that 38.3% of patients admitted to hospital from May 2021, in a critically ill state due to confirmed COVID-19, were female.

3.2.2 Time to hospital death in patients initiating SoC (with or without corticosteroids)

The following steps were used to estimate the survival of patients admitted to hospital due to COVID19 and receiving SoC based on current conditions such as vaccination status, SARS-CoV-2 variant, seropositivity and the widespread use of corticosteroids.

41

The Kaplan-Meier (KM) estimate for OS was taken from the control arm of the RECOVERY study,[35] and was digitised which allowed pseudo-IPD to be reconstructed based on the algorithm developed by Guyot et al (2012).[43] A spline model (hazard scale) with 3 knots was subsequently fitted to the pseudoIPD using the R package flexsurv and employing a natural cubic spline function. This model was selected over standard parametric functions (such as the exponential, Weibull, Gompertz, Log-Normal, Log-Logistic, Gamma, Generalized Gamma) to increase the accuracy in the estimate and because parametric extrapolation beyond the observed period of the trial was limited to a maximum of 70 days. This distribution was then calibrated to the current data such that 73.5% of patients were alive for the population in need of oxygen and 86.0% of patients were alive for the population admitted with no need of supplemental oxygen at 28 days. These values were taken from a NICE rapid guideline[11] assuming that the outcomes for patients without corticosteroid use were generalisable to patients requiring supplemental oxygen and the outcomes for those patients corticosteroids were generalisable to patients not requiring supplemental oxygen. This decision was made as corticosteroids were only seen to be efficacious in patients not requiring supplemental oxygen. For illustration, Figure 13 shows the OS curves used in the model for SoC and remdesivir by oxygen requirement at hospital admission.

==> picture [482 x 118] intentionally omitted <==

==> picture [482 x 118] intentionally omitted <==

Figure 13: Illustration of OS curves used for the hospitalised cohort for SoC and remdesivir by oxygen requirement at entry

3.2.3 Time to discharge for patients initiating SoC

The KM estimate for time to discharge was taken from the control arm of the RECOVERY study,[35] and was digitised which allowed pseudo-IPD to be reconstructed based on the algorithm developed by Guyot et al (2012).[43] A spline model (hazard scale) with 3 knots was subsequently fitted to the pseudo-

42

IPD and was selected over standard parametric functions (such as the exponential, Weibull, Gompertz, Log-Normal, Log-Logistic, Gamma, Generalized Gamma) to increase the accuracy in the estimate and because parametric extrapolation beyond the observed period of the trial was limited to a maximum of 70 days. This distribution was then calibrated to the current data such that 64.0% of patients for the population in need of supplemental oxygen and 80.4% of patients with no need of supplemental oxygen were discharged at 28 days. These values were taken from a NICE rapid guideline[11] assuming that the outcomes for patients without corticosteroid use were generalisable to patients requiring supplemental oxygen and the outcomes for patients using corticosteroids were generalisable to patients not requiring supplemental oxygen. This decision was made as corticosteroids were only seen to be efficacious in patients not requiring supplemental oxygen. For illustration, Figure 14 shows the time to discharge curves used in the model for SoC and casirivimab/imdevimab by oxygen requirement at hospital admission.

==> picture [481 x 118] intentionally omitted <==

==> picture [481 x 118] intentionally omitted <==

Figure 14: Illustration of time to discharge curves used for the hospitalised cohort for SoC and casirivimab/imdevimab by oxygen requirement at entry

3.2.4 Redistribution of patients according to supplemental oxygen/hospitalisation requirements

In order to estimate costs and QALYs during an average hospital stay, it was necessary to model how patients move between the 8-point ordinal scale as each scale has different consequences in terms of the costs of treatment and the HRQoL of the patient. Hospitalised patients with COVID-19 may receive supplemental oxygen, defined as LFO, HFO, and mechanical ventilation (MV). However, during their

43

hospital stay, patients may require more or less intensive management. Hospitalised patients are divided into five states, which correspond to ordinal scales 3 to 7.

3.2.4.1 Assumed distribution of patients on the 8-point ordinal scale on hospital entry

By definition, all patients admitted to hospital due to COVID-19 without the need for supplemental oxygen are in ordinal stage 4. For patients requiring supplemental oxygen, data from ACTT-1[16] which reported the distribution of ordinal score by treatment for placebo on admission to hospital were used. These data however do not reflect the distribution of current admissions as the percentage requiring IMV or ECMO (ordinal stage 7) was 46%, however a recent value suggests that this was only 1%.[42] The distribution from ACTT-1 was adjusted such that only 1% of patients resided in ordinal stage 7 with those patients reallocated from ordinal stage 7 being redistributed between ordinal stages 5 and 6, according to their relative weight in the ACTT-1 study. Table 9 and Table 10 show the proportions of patients across the ordinal health stages at baseline for those requiring supplemental oxygen and those not requiring supplemental oxygen respectively.

3.2.4.2 Distribution of hospitalised patients between the ordinal stages on SoC at day 14

Beigel et al . report data from the ACTT-1 study[16] for the placebo arm which detailed the ordinal stage distribution at baseline and 14 days later. Because of small numbers, which would have meant that movement between some stages was impossible, a continuity correction was added for all possible transitions, splitting 1 new observation at day 14 equally over the five ordinal scales.

However, ACTT-1 was an early study and there have been many changes such as a vaccination programme, increased use of corticosteroids and changes in SARS-CoV-2 variants. These changes have meant that the results from this study are no longer generalisable to the UK, particularly in terms of the proportion of patients who reach ordinal scale 7 and require IMV or ECMO. In ACTT-1, the EAG calculated that the percentage of patients’ time spent in ordinal scale 7 was 48%, contrastingly, this value has been reported in May 2022 to be only 4.12%.[44] The ACTT-1 data was calibrated so that the percentage of time in ordinal stage 7 was equal to 4.12%, with the patients no longer allocated to ordinal scale 7 being allocated to ordinal stage 6 instead. The decision to allocate to ordinal stage 6 was to avoid a situation where the predicted outcomes for patients at stage 7 on hospital entry were better than those for patients admitted at ordinal stage 6. The estimated proportions of patients in hospital across the ordinal health stages at day 14 are shown in Table 9 and Table 10 for patients not requiring supplemental oxygen and those requiring it respectively.

44

Table 9: The distribution of hospitalised patients not requiring supplemental oxygen on entry to hospital and at day 14

Table 10: The distribution of hospitalised patients requiring supplemental oxygen on entry to hospital and at day 14

3.2.4.3 Movement between ordinal scales between day 0 and day 14

We assumed that the distribution of patients changes linearly from the distribution at baseline to the proportions assumed at day 14; for simplicity these proportions were assumed to remain constant after day 14. Figure 15 provides the assumed splits between ordinal scales over a 28-day period.

45

==> picture [470 x 122] intentionally omitted <==

==> picture [470 x 121] intentionally omitted <==

Figure 15: Linear assumptions for distribution across the five ordinal scales during hospital stay

3.2.5 Treatment effects for interventions compared with SoC

The treatment effects for interventions are summarised in Table 4 and Table 5. Where data were not available for clinical improvement or time to discharge a value of 1.0 was used as the model results were not sensitive to these values within the observed range associated with other interventions.

3.2.6 Duration of treatment/number of doses

The dosage information data were taken from the NICE COVID-19 rapid guideline.[11] Where either the dosage or the duration of treatment was not available, this information was taken from alternative sources. Table 11 summarises the dosage information used in the model.

46

Table 11: Dosing information of the interventions included in the model

*The dosing information was not used in the model as the overall course cost was derived from an Institute for Clinical & Economic Review report[45] as requested by NICE

3.2.7 Mortality rate assumed post-hospitalisation and for those people who did not require hospital admission

The unadjusted rate of mortality for the general population is taken from the England and Wales life table 2018-2020.[46] After discharge, patients hospitalised with COVID-19 were assumed to be at an elevated risk of death whilst they have long COVID. An SMR of 7.7 (7.2 – 8.3) was applied based on the RR reported by Ayoubkhani et al .[18] which was estimated from 47,780 patients treated for COVID19 in NHS hospitals and discharged alive, using matched-controls and which had a median follow-up of 140 days. This SMR was also applied to patients at high-risk in the community for the period in which they were simulated to have long COVID.

3.2.8 Serious Adverse Events

Whilst the living systematic reviews allowed the relative risks related to SAEs to be extracted, on inspection these were not events related to the unwanted impacts of the interventions but were conditions related to severe COVID-19. As such, many interventions were associated with less SAEs than SoC, which is generally atypical for efficacious pharmacological treatments. As the model was explicitly tracking the severity of patients through the use of the 8-point ordinal scale the EAG decided to omit SAEs from the model.

47

3.2.9 Long Covid

The prevalence of long COVID within the wider community has been taken from an ONS report dated the 6[th] May 2022,[47] which in supplementary tables reports adjusted model estimates for long COVID of any severity and at any point since the last vaccine of: 8.7% of double-vaccinated patients and 8.0% of triple-vaccinated patients, who had the Omicron BA 1 variant; and 15.9% of double-vaccinated patients and 8.6% of triple-vaccinated patients, who had the Delta variant. Having noted the relatively wide CIs for the ONS estimates, the difference depending on vaccination status (with no data reported for unvaccinated patients) and the method it proposes to use for estimating the duration of long COVID (described below), the EAG assumed that 10% of patients in the community who were at high-risk of severe COVID-19 but did not need hospitalisation would experience long COVID. The EAG was not aware of any evidence on the impact of community treatment on the incidence of long COVID and thus it was assumed that this was independent of treatment.

The duration of long COVID-19 was estimated from an ONS publication dated the 1[st] of June 2022.[48] This stated that of people with self-reported long COVID, defined as “ symptoms continuing for more than four weeks after the first suspected coronavirus (COVID-19) infection that were not explained by something else ” 72% of people had been first infected by COVID-19 (or suspected they had) at least 12 weeks earlier, 42% were infected at least one year previously, and 19% at least two years previously. This publication also reports that 22% of people had suspected they were infected by COVID-19 less than 12 weeks previously; it was not clear to the EAG why the addition of the proportion of patients less than 12 weeks, and 12 weeks or more, did not add up to 100%, but only 94%.

Simple parametric distributions were fitted to the three reported estimates of at least 12 weeks duration (72% with long COVID at 12 weeks, 42% at 1 year, and 22% at 2 years). A Gamma distribution (shape = 100.547, scale 0.644), a Weibull distribution (shape =0.749, scale 57.268) and a lognormal distribution (mean = 3.468, standard deviation 1.562 (on the log scale) were observed to fit the data well. The mean survival times from these distributions were 64.7 weeks (Gamma), 68.3 weeks (Weibull) and 108.6 weeks (lognormal). The plots using the Gamma and lognormal distributions, which had the lowest and highest values are shown in Figure 16.

48

==> picture [409 x 230] intentionally omitted <==

Figure 16: Assumed duration of long Covid

For its base case the EAG assumed the lognormal distribution was most appropriate, but undertook sensitivity analyses halving and doubling the mean duration, the range of which includes the mean from the Gamma distribution. The reason for this was that based on the previous ONS report, on which the EAG had conducted similar analyses, it was seen that the mean time with long COVID had increased, and the data is relatively immature and may be administratively censored. The EAG notes that its analyses are simplistic as formal survival analysis methods have not been used, and that it does not assume that all patients must have long COVID for at least 4 weeks, as used in some definitions but believes that the analyses undertaken are informative for decision making despite this limitation.

From Evans et al .[49] it is estimated that at approximately 6 months, 51.7% of patients with non-missing data (n=830) reported that they had not recovered from COVID-19; this value increases to 71.2% when patients stating they were not sure if they had recovered were included. The patients included in the study were hospitalised early in the pandemic (between March and November 2020) and it is unclear how generalisable this result is to patients hospitalised in 2022. The best-fitting gamma and log-normal distributions shown in Figure 16 estimate the proportions of patients not recovered from long COVID to be 57.8% and 55.3% at 26 weeks which is similar to the value reported in Evans et al .[49] Given the uncertainty in patients who stated they were not sure if they had recovered, a simplistic assumption was made that all patients hospitalised due to COVID-19 would suffer long COVID. The EAG was not aware of any evidence on the impact of hospital treatment on the incidence of long COVID and thus it was assumed that this independent of treatment.

49

3.3 Costs and Health-Related Quality of Life

3.3.1 Drug acquisition costs

Drug acquisition costs were supplied to the EAG by NICE. This included the list price for remdesivir, tocilizumab, baricitinib, lenzilumab, and sotrovimab. However, list prices were not available for molnupiravir, casirivimab/imdevimab, and nirmatrelvir/ritonavir. NICE requested that placeholder prices be used which were estimated from an Institute for Clinical & Economic Review report (dated March 2022) for molnupiravir and nirmatrelvir/ritonavir,[45] and that the price for sotrovimab was used for casirivimab/imdevimab. All analyses in this report are conducted at the list or placeholder prices, with analyses using the Patient Access Scheme (PAS) discounts for tocilizumab and baricitinib included in a confidential appendix. For corticosteroids, daily costs were assumed negligible compared to the inhospital day cost and were not included for simplicity. Table 12 summarises the list prices used in the model with assumptions done when necessary.

Table 12: List prices of interventions used in the model

3.3.2 Administration costs

It was assumed that the costs associated with treatment administration whilst in hospital would be incorporated in the unit costs associated with hospitalisation (see Section 3.3.3). Additional administration costs were assumed for intravenous treatment in the community, but for simplicity, not for oral or subcutaneous treatments. For each intravenous administration, a cost of £221 was incurred

50

which was that of NHS reference code SB12Z.[50] Within the analyses it has been assumed that there is likely to be a delay in patients receiving intravenous casirivimab/imdevimab and that a subcutaneous version would be used instead.

3.3.3 Unit costs associated with hospitalisation

The unit costs per hospital bed day are taken from the NHS National Schedule of NHS costs 20192020.[50] The NHS codes used are detailed in Table 13.

51

Table 13: The unit costs by ordinal scale used in the economic model and Utility values/decrement in HRQoL

HFO: high-flow oxygen; IVM: invasive mechanical ventilation; LFO: low-flow oxygen; NIV: non-invasive ventilation

52

3.3.4 Costs associated with COVID-19 for outpatients or following discharge

3.3.4.1 Monitoring costs

For simplicity, monitoring/follow-up was assumed to occur in the first year only. Following discharge, patients were assumed to undergo 2 chest X-rays and 6 GP e-consultations on average related to their COVID-19 as in Rafia et al .[13] A one-off cost of £384 was applied to all patients assuming the cost of a chest X-ray was £44 (taken from Stroke et al .[53] and inflated to 2019/2020 prices using NHSCII pay and prices indices[54] ) and the cost associated with a GP e-consultation was £49.[54]

3.3.4.2 Costs associated with long COVID

The EAG assumed that management costs for long COVID was similar to the management of chronic fatigue syndrome. For time constraints, the EAG pragmatically searched for literature and found an economic evaluation study evaluating multidisciplinary rehabilitation treatment versus cognitive behavioural therapy for patients with chronic fatigue syndrome in the Netherlands.[55] Healthcare resource use included GP care, mental healthcare specialist, paramedical care, medical specialist care, hospital care, medications, alternative healers, company physicians, and the evaluated interventions. The EAG substituted the company physician cost with GP care and noted the similarity in costs between arms when intervention costs were excluded. An average of the two costs was used, which resulted in an annual cost of €1195. After conversion using the average of the HMRC rates[56] published in January and December 2016, and inflation using NHS cost inflation index pay and prices indices,[54] an annual cost of £1013 was estimated for patients with long COVID.

3.3.5 Health-related quality of life

3.3.5.1 Unadjusted baseline utility value by age

Baseline utility values (prior to any decrements/adjustments) are taken from Ara and Brazier based on the age-sex utility values (EQ-5D) in the UK.[57]

3.3.5.2 HRQoL during the hospitalisation episode

Due to the nature of this rapid assessment, no formal systematic review of the literature was conducted to identify the most appropriate utility values. Hence, utility values (or decrements) were sourced from Rafia et al .[13] which estimated the cost-effectiveness of remdesivir.

3.3.5.3 HRQoL related to long COVID

A paper by Evans et al.[49] reported the impact on HRQoL following hospitalisation due to COVID-19. The EQ-5D 5 level (EQ-5D-5L) prior to hospitalisation was observed to be 0.84 but was 0.71 after hospitalisation, suggesting a utility impact of long COVID of 0.13. This value is not dissimilar to a reported utility loss in patients following severe sepsis.[58] It was assumed that this disutility would apply to all patients for their duration of long COVID.

53

3.4 Analyses undertaken

Probabilistic sensitivity analysis (PSA) is the most appropriate method for providing the most accurate estimation of the ICER, however this could not be undertaken within the deadlines of the project. This was because there was a need to calculate the proportion of patients treated in the community who are admitted to hospital, and die within this episode, as the model assumed that deaths due to COVID-19 only occurred in the hospital (see Section 3.1.2). This calculation added considerable computational time.

To circumvent this problem three ‘deterministic’ analyses were run, which were i) using the mean value for clinical effectiveness data, and the median for all other parameters, ii) using the most favourable limit of the 95% CI for clinical effectiveness data, and the median for all other parameters, and iii) using the least favourable limit of the 95% CI for clinical effectiveness data, and the median for all other parameters. For brevity, the analyses have been referred to as ‘mean efficacy’, ‘high efficacy’ and ‘low efficacy’ respectively. One exception was made in relation to the ‘mean efficacy’ which was for the use of remdesivir in a community setting. This was because there were no observed deaths in either arm, and using a mean HR of 7.36 was assumed to be overly punitive and a value of 1.00 was used instead. When operationalising these analyses, problems were encountered for the low efficacy values for three treatments for patients with COVID-19 at high-risk of hospitalisation in the community. This was because Excel generated a numerical error when the multiplier for RR of death for hospitalised patients treated with SoC was greater than 121 as, due to the number of decimal places used in Excel, the package was attempting to calculate the natural log of zero. As such, the EAG assumed that the upper limit of the 95% CIs for the RR of mortality at 28 days were 1.82 for casirivimab/imdevimab, 3.07 for remdesivir and 1.99 for sotrovimab, which were the values calculated when a multiplier of 121 was applied to the RR of death for hospitalised patients treated with SoC. The EAG notes that for all analyses no attempts of incorporating prior beliefs have been conducted and a frequentist approach using distributions derived from the raw data is used. The EAG comments that it may be clinically implausible that treatments which have a statistically significant beneficial HR relating to hospitalisation or death would be associated with increased RR of death at 28 days, but this limitation could not be addressed in the timescales of the project.

These analyses were supplemented by sensitivity analyses and are believed to provide the NICE appraisal committee with pertinent information relating to the true uncertainty in the decision problem, which will be much larger than any difference between the mean results from a PSA and from a deterministic analysis using the mean of the distribution. As the efficacy of treatments are assumed to be independent, then there is considerable uncertainty in the true treatment effect (see Figure 5 and

54

Figure 9) and it is plausible that one intervention had its ‘low efficacy’ value whilst another had its ‘high efficacy’ value.

Three sensitivity analyses were performed, which explored the impact of changing i) the duration of long COVID (ranging from half to double that of the base case), ii) changing the product of the rate of hospital admission in the community and the RR associated with people being at ‘high risk’ of hospitalisation from a value of 1.8% to 1.35% and 5.00% and iii) changing the average age of patients at high-risk of hospitalisation in the community from 65 years to 60 and 70 years.

The results presented provide the ICER, measured in terms of cost per QALY gained, for each intervention compared to SoC and also the efficiency frontier, which contains all interventions that are not dominated or extendedly dominated. For the efficiency frontier, the willingness to pay (WTP) at which the preferred treatment changes, presented in terms of cost per QALY thresholds, is provided.

For the sensitivity analysis, in order that a large number of results can be shown simultaneously, an incremental net monetary benefit, shortened to net monetary benefit (NMB) approach was taken comparing all interventions with SoC. Within this framework, the largest NMB is associated with the most cost-effective strategy at the stated cost-per-QALY threshold, and multiple strategies can be compared simultaneously, as the absolute difference in strategies in terms of cost, having monetarised health differences, can be easily determined. The formula for calculating NMB is the increase in QALYs associated with an intervention multiplied by a stated cost per QALY threshold minus the additional costs of associated with the intervention compared with the costs associated with SoC. If NMB is positive the intervention is cost-effective compared with SoC at the selected threshold; if the NMB is negative then the intervention is not cost-effective compared with SoC at the selected threshold. When multiple interventions are considered, the intervention with the greatest NMB would be interpreted as the most cost-effective intervention. For the analyses presented in this report, the cost per QALY threshold was set at £20,000 per QALY which is the lowest of NICE’s published thresholds. NMBs were also provided in the base case results.

One limitation associated with the omission of PSA is that value of information analyses could not be conducted to assess the monetary implications of recommending an intervention that was not the most cost-effective and to put a ceiling on the expenditure of research addressing knowledge gaps. This is an area for future research.

55

3.5 The use of severity modifiers

The guidance from NICE is that if there is an absolute discounted QALY shortfall of less than 12 and that the proportional shortfall in discounted QALYs is less than 85% then no severity modifier should be applied in the decision problem, and that the ICER remains unchanged.

For patients admitted to hospital, the mean age was assumed to be 70.6 years and with 38.3% being female. Using these characteristics, the EAG calculated that the discounted QALYs associated with the general population would be approximately 9.05. Based on the results presented in Section 4, SoC is associated with estimated discounted QALYs of 4.65 for patients who require supplemental oxygen on admission and 5.84 for patients who do not require supplemental oxygen on admission. For those requiring supplemental oxygen, the absolute shortfall was 4.40 discounted QALYs and the proportional shortfall was 49%; these numbers are lower for those who do not require supplemental oxygen. As such, no severity modifier is applied for patients who are hospitalised due to COVID-19.

For patients at high-risk of hospitalisation in the community, the mean age in the base case was assumed to be 65 years. The 38.3% proportion of females used for hospitalised patients was assumed to be generalisable to patients at high-risk in the community. Using these characteristics, the EAG calculated that the discounted QALYs associated with the general population would be approximately 10.05. Based on the results presented in Section 4, the absolute shortfall in discounted QALYs for patients at high-risk of hospitalisation was less than 1, and the proportionate shortfall in discounted QALYs was 7%. Given these values, no severity modifier is applied for patients who are at high-risk of hospitalisation due to COVID-19.

56

4 COST-EFFECTIVENESS RESULTS

The cost-effectiveness results have been divided into three subsections. The first provides the results for hospitalised patients who require supplemental oxygen on admission, the second provides the results for hospitalised patients who do not require supplemental oxygen on admission with the third providing the results for patients at high-risk of hospitalisation in the community. Each of the three subsections are further divided to provide the results from the mean efficacy, high efficacy, and low efficacy scenarios.

The EAG stresses that, following NICE’s recommendations, some prices are placeholders and that the PASs for tocilizumab and baricitinib are not included. This means that the results presented are not accurate representations of the true ICERs for some drugs. Results incorporating PASs, and NICEsuggested prices rather than the placeholders used in this report are contained in a confidential appendix.

4.1 Results for hospitalised patients who need supplemental oxygen on admission

4.1.1 Mean efficacy results for patients requiring supplemental oxygen on admission to hospital

The results of the mean efficacy analysis for patients requiring supplemental oxygen on admission to hospital are shown in Table 14. All interventions were estimated to have a cost per QALY gained compared to SoC below £20,000, with the majority less than £10,000. A full incremental analysis indicates an efficiency frontier of SOC for a WTP up to £3951, casirivimab/imdevimab for a WTP between £3951 and £6226, and baricitinib for a WTP over £6226.

Table 14: Mean efficacy results for people who require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

57

4.1.2 High efficacy results for patients requiring supplemental oxygen on admission to hospital

The results of the high efficacy analysis for patients requiring supplemental oxygen on admission to hospital are shown in Table 15. All interventions were estimated to have a cost per QALY gained compared to SoC below £20,000, with the majority below £10,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £1310, tocilizumab for a WTP between £1310 and £6456, casirivimab/imdevimab between £6456 and £17,781, and baricitinib and remdesivir for a WTP over £17,781. The costs associated with tocilizumab and casirivimab/imdevimab are lower than for other drugs due to the assumed higher rate of discharge of patients.

Table 15: High efficacy results for people who require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.1.3 Low efficacy results for patients requiring supplemental oxygen on admission to hospital

The results of the low efficacy analysis for patients requiring supplemental oxygen on admission to hospital are shown in Table 16. All interventions except for baricitinib and tocilizumab were dominated by SoC due to increased hazards of death associated with the upper limit of the 95% CI being above 1 (see Table 4). The ICERs for baricitinib and tocilizumab were both below £20,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £4608 and baricitinib for a WTP over £4608.

58

Table 16: Low efficacy results for people who require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.2 Results for hospitalised patients who do not need supplemental oxygen on admission

4.2.1 Mean efficacy results for patients requiring no supplemental oxygen on admission to hospital

The results of the mean efficacy analysis for patients not requiring supplemental oxygen on admission to hospital are shown in Table 17. With the exception of lenzilumab, all interventions were estimated to have a cost per QALY gained compared to SoC below £10,000, with the ICER for lenzilumab being greater than £20,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £3053 and baricitinib for a WTP over £3053.

Table 17: Mean efficacy results for people who do not require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

59

4.2.2 High efficacy results for patients requiring no supplemental oxygen on admission to hospital The results of the high efficacy analysis for patients not requiring supplemental oxygen on admission to hospital are shown in Table 18. All interventions were estimated to have a cost per QALY gained compared to SoC below £15,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £2863, casirivimab/imdevimab between £2863 and £7646, baricitinib for a WTP between £7646 and £13,243, and baricitinib and remdesivir for a WTP over £13,243. The costs associated with casirivimab/imdevimab are lower than for other drugs due to the assumed higher rate of discharge of patients.

Table 18: High efficacy results for people who do not require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.2.3 Low efficacy results for patients requiring no supplemental oxygen on admission to hospital The results of the low efficacy analysis for patients not requiring supplemental oxygen on admission to hospital are shown in Table 19. With the exception of baricitinib, all interventions were estimated to be dominated by SoC due to the 95% CI for these interventions being greater than 1 (see Table 4). A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £2820 and baricitinib for a WTP over £2820.

60

Table 19: Low efficacy results for people who do not require supplemental oxygen on admission to hospital

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.3 Results for patients at high-risk of hospitalisation treated in the community

4.3.1 Mean efficacy results for patients at high-risk of hospitalisation

The results of the mean efficacy analysis for patients at high-risk of hospitalisation are shown in Table 20. Nirmatrelvir/ritonavir and molnupiravir were estimated to have a cost per QALY compared to SOC of below £15,000 with all other interventions having an ICER in excess of £60,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £4439 and nirmatrelvir/ritonavir for a WTP over £4439.

Table 20: Mean efficacy results for people at high-risk of hospitalisation

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

61

4.3.2 High efficacy results for patients at high-risk of hospitalisation

The results of the high efficacy analysis for patients at high-risk of hospitalisation are shown in Table 21. Nirmatrelvir/ritonavir and molnupiravir were estimated to have a cost per QALY compared to SOC of below £15,000 with all other interventions having an ICER in excess of £20,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £3895 and nirmatrelvir/ritonavir for a WTP over £3895.

Table 21: High efficacy results for people at high-risk of hospitalisation

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.3.3 Low efficacy results for patients at high-risk of hospitalisation

The results of the low efficacy analysis for patients at high-risk of hospitalisation are shown in Table 22. All interventions, except for nirmatrelvir/ritonavir and molnupiravir were estimated to be dominated by SoC. The ICER for nirmatrelvir/ritonavir compared with SoC was below £10,000 whereas that for molnupiravir was greater than £65,000. A full incremental analysis indicates an efficiency frontier of SoC for a WTP up to £7989 and nirmatrelvir/ritonavir for a WTP over £7989.

62

Table 22: Low efficacy results for people at high-risk of hospitalisation

 Assuming a threshold of £20,000 per QALY gained QALY – quality-adjusted life years; SoC – standard of care

4.4 Sensitivity Analysis Results

Three sets of sensitivity analyses were run that related to:

• Amending the assumed duration of long COVID from 108.6 weeks to 54.3 weeks and to 217.2 weeks

• Changing the product of the percentage of hospital admission due to COVID-19 in the community and the RR associated with people being at ‘high risk’ of hospitalisation from a value of 1.8% to values of 1.35% and 5.00%

• Changing the average age of patients at high-risk of hospitalisation in the community from 65 years to 60 and 70 years.

For reference, the NMBs of each intervention are shown in Figure 17, Figure 18 and Figure 19 for patients who are hospitalised and require supplemental oxygen, patients who are hospitalised but do not require supplemental oxygen, and patients with COVID-19 in the community who are at high-risk of hospitalisation respectively.

63

==> picture [401 x 125] intentionally omitted <==

==> picture [401 x 124] intentionally omitted <==

Figure 17: Base case net monetary benefits for patients admitted to hospital who require supplemental oxygen

==> picture [400 x 131] intentionally omitted <==

==> picture [400 x 130] intentionally omitted <==

Figure 18: Base case net monetary benefits for patients admitted to hospital who do not require supplemental oxygen

64

==> picture [412 x 135] intentionally omitted <==

==> picture [412 x 135] intentionally omitted <==

Figure 19: Base case net monetary benefits for patients with COVID-19 in the community and high-risk of hospitalisation

4.4.1 Amending the duration of long COVID

The NMB results when the duration of long COVID is doubled (to 217.2 weeks) and halved (to 54.3 weeks) are shown in Figure 20, Figure 21 and Figure 22 for people admitted to hospital requiring supplemental oxygen, those admitted to hospital with no need for supplemental oxygen, and those treated in the community at high-risk of hospitalisation respectively.

For patients in all settings, the absolute difference in NMB between scenarios where the duration of long COVID was halved and scenarios where the duration was doubled was markedly smaller than the absolute differences in NMB when using the high efficacy scenario and the low efficacy scenario (Figure 20, Figure 21 and Figure 22). This indicates that the duration of long COVID was of lesser importance in driving the ICER than the actual efficacy of the interventions. The interventions were more cost-effective when the duration of long COVID was shorter, as the interventions typically increased survival and more QALYs would be gained from patients who had long COVID for a shorter period.

65

==> picture [453 x 82] intentionally omitted <==

==> picture [453 x 82] intentionally omitted <==

Figure 20: The NMB results for patients admitted to hospital who require supplemental oxygen when the duration of long COVID is halved and doubled

==> picture [453 x 82] intentionally omitted <==

==> picture [453 x 81] intentionally omitted <==

Figure 21: The NMB results for patients admitted to hospital who do not require supplemental oxygen when the duration of long COVID is halved and doubled

==> picture [453 x 82] intentionally omitted <==

==> picture [453 x 82] intentionally omitted <==

Figure 22: The NMB results for patients in the community with COVID-19 who are at highrisk of hospitalisation when the duration of long COVID is halved and doubled

66

4.4.2 Amending the hospital admission percentage for people with COVID-19 in the community at high-risk of hospitalisation treated with SoC

The NMB results when the hospitalisation admission percentage for people with COVID-19 in the community at high-risk of hospitalisation treated with SoC was changed from 1.8% to 1.35% and 5.00% are shown in Figure 23. It is seen that in the mean efficacy scenario and the high efficacy scenarios the proportion of patients with COVID-19 at high-risk of being hospitalised being admitted to hospital makes a large difference to the NMB. All interventions had a positive NMB when the proportion of patients hospitalised was increased to 5.00% and the high efficacy scenario was used. This shows that the proportion of patients with COVID-19 at high-risk of hospitalisation is an important driver of the ICER with the interventions becoming more cost-effective as the admission proportion increases.

==> picture [464 x 84] intentionally omitted <==

==> picture [464 x 83] intentionally omitted <==

Figure 23: The NMB results for patients in the community with COVID-19 who are at highrisk of hospitalisation when the hospital admission percentage was changed

• 4.4.3 Amending the age of people with COVID-19 in the community at high-risk of hospitalisation treated with SoC

The NMB results when the age assumed for people with COVID-19 in the community at high-risk of hospitalisation treated with SoC was changed from 65 years to 60 years and 70 years are shown in Figure 24. It is seen that the change in NMB between the ages of 60 and 70 years is not dissimilar from the changes in NMB when the different efficacy scenarios are used. As such, age is an important driver of the ICER for treatment of patients with COVID-19 at high-risk of hospitalisation in the community, with the drugs being more cost-effective as the age of patients decrease.

67

==> picture [453 x 80] intentionally omitted <==

==> picture [453 x 79] intentionally omitted <==

Figure 24: The NMB results for patients in the community with COVID-19 who are at highrisk of hospitalisation when the age was changed from 65 years to 60 years and 70 years

68

5 DISCUSSION AND CONCLUSIONS

5.1 Summary Of Clinical-Effectiveness data

For time reasons, the EAG used data from two living systematic reviews and had to assume that the reported efficacy of treatments was transportable to other settings. This assumption may not be correct due to: the evolving nature of SoC; the impact of vaccination; the impact of previous SARS-CoV-2 infection; and the predominant SARS-CoV-2 variant. In addition, patient age, ethnicity, sex and immune system competence may be treatment effect modifiers.

All treatments were associated with a midpoint beneficial effect on preventing mortality, with the exception of remdesivir for patients at high-risk in the community where there were no deaths in either arm. Noting the caveats associated with assuming transportability of treatment effects and the relatively wide CIs associated with preventing mortality, the EAG did not feel confident that it could robustly identify a treatment that was more efficacious than others.

5.2 Summary of Cost-Effectiveness analyses

For patients who have been hospitalised due to COVID-19, all treatments had scenarios where the ICER was below £20,000 compared with SoC, however, in the low efficacy scenario only baricitinib and tocilizumab had ICERs under £20,000 compared with SoC. For patients with COVID-19 in the community at high-risk of hospitalisation, nirmatrelvir/ritonavir and molnupiravir appeared to have the lowest ICERs, with the ICERs for the remaining drugs never falling below £20,000 compared with SoC in any efficacy scenario. In the mean efficacy scenario and the high efficacy scenario both molnupiravir and nirmatrelvir/ritonavir had ICERs below £15,000 compared with SoC.

In the mean efficacy analysis, baricitinib was the intervention that produced most QALYs and had an ICER below £10,000 compared with the previous intervention on the efficiency frontier for people admitted to hospital, independent of supplemental oxygen requirements. For people treated in the community with COVID-19 at high-risk of hospitalisation, nirmatrelvir/ritonavir produced most QALYs and had an ICER below £5,000 compared with the previous intervention on the efficiency frontier in the mean efficacy analysis. However, fully incremental analyses should be treated with caution due to the SoC, the percentage of people who have had a vaccination and the dominant SARSCoV-2 variant which could vary between studies. Furthermore, the PASs for baricitinib and tocilizumab have not been incorporated in this document and some prices are placeholders at the request of NICE.

69

The analyses in this report are more favourable to remdesivir treatment in hospital than previous estimates reported by Rafia et al .[13] The primary reasons for this are differing assumptions in the models. In Rafia et al .[13] remdesivir was associated with an odds ratio for clinical improvement that indicated that remdesivir was harmful to a patient who did not die, compared with SoC and the proportion of patients in ordinal scale 7 receiving SoC was large (22% at day 14). In our analyses, remdesivir is now associated with improved outcomes for patients who do not die but also the proportion of patients in ordinal scale 7 who receive SoC was significantly reduced (9% at day 14). These changes result in a considerable saving in hospital costs, which results in a lower ICER in our work.

The analyses did not look at the logistical aspects of providing treatment. For patients in hospital this is unlikely to be a significant issue, however it could be for patients in the community if an IV treatment was preferred. Local decision makers would need to ascertain whether IV treatment for patients with COVID-19 is possible.

• 5.2.1 Strengths of the economic analysis include:

• The use of contemporary effectiveness data from living systematic reviews

• An attempt by the EAG to align the results of SoC produced by the model with data observed in mid-2022

• Uncertainty in the model inputs and assumptions has been explored in wide ranging sensitivity analyses

• The modelling attempts to capture movement between the 8-point ordinal scale to consider the costs and consequences of patient improvement and patient decline

• The modelling explicitly attempts to take the impact of the longer-term implications of COVID-19 into consideration

5.2.2 Limitations of the analysis include:

• The characteristics of the decision problem may have changed considerably since the pivotal trials for each intervention was conducted. Such changes include the introduction of a vaccination programme, new SARS-CoV-2 variants, history of prior SARS-CoV-2 infection, the level of supplemental oxygen requirement, and the widespread use of corticosteroids in SoC. The EAG assumed that none of these were treatment effect modifiers and that the treatment effects were transportable which may be incorrect.

• No head-to-head studies of interventions were identified that could be used in the modelling and the uncertainty regarding the most efficacious treatment is large.

70

• Some prices for interventions are placeholders only and that results included PASs could not be provided in a publicly available document

• Uncertainty remains in the underlying rates of hospitalisation in patients with COVID-19 at highrisk of hospitalisation under SoC

• Uncertainty remains in the underlying rates of death in patients hospitalised due to COVID-19 who receive SoC

• SoC only was assumed to be provided to patients in hospital if they had been treated with an intervention in the community

• Treatments used in hospital were not assumed to affect the proportion of discharged people with long COVID and that treatments used in the community were not assumed to affect the proportion of people not admitted to hospital with long COVID

• All patients were assumed to be discharged from hospital at day 70, which could favour the more efficacious treatments in reducing hospital costs

• No prior beliefs were incorporated relating to the clinical efficacy of the interventions

• No value of information analysis was conducted. This would allow funders to estimate the relative benefits of investing in future research

• No analysis was conducted on whether it is logistically possible to treat patients in the community with COVID-19 and a high-risk of hospitalisation with IV drugs

5.3 The use of patient and public involvement

There was no patient and public involvement in producing this report. This was not considered possible within the timescales of the project. However, the EAG is aware that at the NICE Technology Appraisal Committee that will discuss this topic, there will be patient and public involvement and representation and this may result in the EAG changing model parameters and generating revised results.

5.4 Equality, Diversity and Inclusion

As this report is secondary research, no patient participation was involved and the EAG did not need to consider the Equality, Diversity and Inclusion of participants. The primary research team was part of the ScHARR Technology Assessment Group contracted by the Department of Health, and this team is a diverse group representing a wide range of protected characteristics, consisting of a wide range of seniority, ages, ethnicity and religious beliefs and including both males and females. The clinical team represent experts within their field who have successfully worked with the ScHARR Technology Assessment Group on previous projects. The lead author is not the most senior member of the team.

71

Acknowledgements

None.

Contributions of authors

Andrew Metry (https://orcid.org/0000-0001-7412-6093) is a Research Associate at ScHARR, University of Sheffield. He was involved in all aspects of the project.

Abdullah Pandor (https://orcid.org/0000-0003-2552-5260) is a Senior Research Fellow at ScHARR, University of Sheffield. He extracted, reviewed and summarised the studies related to the clinical evidence used in the model.

Shijie Ren (https://orcid.org/0000-0003-3568-7124) is a Senior Research Fellow at ScHARR, University of Sheffield. She extracted and reviewed the evidence to populate the model.

Andrea Shippam (https://orcid.org/0000-0003-3866-0939) is a Programme Administrator at ScHARR, University of Sheffield. She extracted and reviewed the evidence to populate the model.

Mark Clowes (https://orcid.org/0000-0002-5582-9946) is an information specialist at ScHARR, University of Sheffield. He provided guidance on the preliminary searches undertaken by the team.

Paul Dark (https://orcid.org/0000-0003-3309-0164) is a Professor of Critical Care Medicine at the University of Manchester. He provided clinical advice throughout the project.

Ronan McMullan (https://orcid.org/0000-0001-6760-6072) is a Professor of Medical Microbiology at Queen’s University, Belfast. He provided clinical advice throughout the project.

Matt Stevenson (https://orcid.org/0000-0002-3099-9877) is a Professor of Health Technology Assessment at ScHARR, University of Sheffield. He led the project and advised on all aspects of the research.

All authors commented on the final monograph

72

6 REFERENCES

1. WorldoMeter. COVID Live - Coronavirus Statistics. Journal 2020. https://www.worldometers.info/coronavirus/ (Accessed June 2022)

2. National Institute for Health and Care Excellence. Introduction to health technology evaluation. NICE health technology evaluations: the manual. 2022.

3. National Institute for Health and Care Excellence. Health Technology Evaluation. Therapeutics for people with COVID-19. Pre-invite scope. 2022.

4. National Institute for Health and Care Excellence. MTA of Therapeutics for people with COVID-19 – Protocol. 2022.

5. PANORAMIC. Participant Information — PANORAMIC. 2022.

6. Practice BB. Treatment algorithm - Coronavirus disease 2019. 2022. https://bestpractice.bmj.com/topics/en-gb/3000201/treatment-algorithm#patientGroup-0-0 (Accessed 16 March 2022).

7. National Institute for Health and Care Excellence. Health Technology Evaluation of therapeutics for people with COVID-19. Pre-invite scope.; 2022.

8. World Health Organization. Therapeutics and COVID-19. Living Guideline 3 March 2022. 2022.

9. The COVID-NMA Initiative. Pharmacological treatments for Covid-19 patients. Journal 2022. https://covid-nma.com/ (Accessed 26 May 2022)

10. metaEvidence Initiative. Meta|Evidence - COVID-19. Living meta-analysis and evidence synthesis of therapies for COVID19. 2022.

11. National Institute for Health and Care Excellence. COVID-19 rapid guideline: managing COVID-19. NICE Guideline, No. 191. 2022.

12. Boutron I, Chaimani A, Meerpohl JJ, Hróbjartsson A, Devane D, Rada G , et al. The COVIDNMA Project: Building an Evidence Ecosystem for the COVID-19 Pandemic. Ann Intern Med 2020;173:1015-7.

13. Rafia R, Martyn-St James, M., Harnan, S., Metry, A., Hamilton, J., Wailoo, A. A CostEffectiveness Analysis of Remdesivir for the Treatment of Hospitalized Patients With COVID19 in England and Wales. Value Health 2022;25:761-9.

14. National Institute for Health and Care Excellence. Remdesivir for treating COVID-19 [ID3808]. 2022.

15. Rafia R, Metry A, Wailoo A. Early economic evaluation of neutralising mononoclonal antibodies/oral antivirals for the treatment of covid-19 pre-hospitalisation. 2021.

16. Beigel JH, Tomashek KM, Dodd LE, Mehta AK, Zingman BS, Kalil AC , et al. Remdesivir for the Treatment of Covid-19 - Final Report. N Engl J Med 2020;383:1813-26.

17. University of Liverpool. ISARIC 4C CCP-UK Remdesivir Effectiveness Evaluation Study: Statistical Report v1.0; 24 February 2021.

18. Ayoubkhani D, Khunti K, Nafilyan V, Maddox T, Humberstone B, Diamond SI , et al. Epidemiology of post-COVID syndrome following hospitalisation with coronavirus: a retrospective cohort study. medRxiv 2021; 10.1101/2021.01.15.21249885:2021.01.15.21249885.

19. Halpin SJ, McIvor C, Whyatt G, Adams A, Harvey O, McLean L , et al. Postdischarge symptoms and rehabilitation needs in survivors of COVID-19 infection: A cross-sectional evaluation. Journal of Medical Virology 2021;93:1013-22.

20. McCue C, Cowan R, Quasim T, Puxty K, McPeake J. Long term outcomes of critically ill COVID-19 pneumonia patients: early learning. Intensive Care Medicine 2021;47:240-1.

21. Garrigues E, Janvier P, Kherabi Y, Le Bot A, Hamon A, Gouze H , et al. Post-discharge persistent symptoms and health-related quality of life after hospitalization for COVID-19. Journal of Infection 2020;81:e4-e6.

22. Taboada M, Moreno E, Cariñena A, Rey T, Pita-Romero R, Leal S , et al. Quality of life, functional status, and persistent symptoms after intensive care of COVID-19 patients. British Journal of Anaesthesia 2021;126:e110-e3.

23. Huang C, Huang L, Wang Y, Li X, Ren L, Gu X , et al. 6-month consequences of COVID-19 in patients discharged from hospital: a cohort study. The Lancet 2021;397:220-32.

73

24. Wong AW, Shah AS, Johnston JC, Carlsten C, Ryerson CJ. Patient-reported outcome measures after COVID-19: a prospective cohort study. The European respiratory journal 2020;56:2003276.

25. Lerum Tøri V AT, Brønstad E, et al. Dyspnoea, lung function and CT findings three months after hospital admission for COVID-19. Eur Respir J; in press (https://doiorg/101183/1399300303448-2020) 2020.

26. Carfì A, Bernabei R, Landi F, Gemelli Against C-P-ACSG. Persistent Symptoms in Patients After Acute COVID-19. JAMA 2020;324:603-5.

27. Iqbal A IK, Arshad Ali S, et al. The COVID-19 Sequelae: A Cross-Sectional Evaluation of Post-recovery Symptoms and the Need for Rehabilitation of COVID-19 Survivors. Cureus 2021;13(2):.

28. Arab-Zozani M, Hashemi F, Safari H, Yousefi M, Ameri H. Health-Related Quality of Life and its Associated Factors in COVID-19 Patients. Osong Public Health Res Perspect 2020;11:296302.

29. Rokadiya S, Gil E, Stubbs C, Bell D, Herbert R. COVID-19: Outcomes of patients with confirmed COVID-19 re-admitted to hospital. The Journal of infection 2020;81:e18-e9.

30. Yeo I, Baek S, Kim J, Elshakh H, Voronina A, Lou MS , et al. Assessment of thirty-day readmission rate, timing, causes and predictors after hospitalization with COVID-19. Journal of Internal Medicine 2021;n/a.

31. UYAROĞLU OA, BAŞARAN NÇ, ÖZIŞIK L, DİZMAN GT, EROĞLU İ, ŞAHİN TK , et al. Thirty-day readmission rate of COVID-19 patients discharged from a tertiary care university hospital in Turkey: an observational, single-center study. International Journal for Quality in Health Care 2020;33.

32. Ginez T, Garaigorta U, Ramírez D, Castro V, Nozal V, ... Host-directed FDA-approved drugs with antiviral activity against SARS-CoV-2 identified by hierarchical in silico/in vitro screening methods. bioRxiv 2020.

33. Donnelly JP, Wang XQ, Iwashyna TJ, Prescott HC. Readmission and Death After Initial Hospital Discharge Among Patients With COVID-19 in a Large Multihospital System. JAMA 2021;325:304-6.

34. COVID-19 rapid guideline: managing the long-term effects of COVID-19 National Institute for Health and Care Excellence; 2020.

35. Group RC, Horby PW, Mafham M, Peto L, Campbell M, Pessoa-Amorim G , et al. Casirivimab/imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. medRxiv 2021; 10.1101/2021.06.15.21258542:2021.06.15.21258542.

36. National Institute for Health and Care Excellence. COVID-19 rapid guideline: Managing COVID-19 2022.

37. Nyberg T, Ferguson NM, Nash SG, Webster HH, Flaxman S, Andrews N , et al. Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. The Lancet 2022;399:1303-12.

38. Hippisley-Cox J, Coupland CA, Mehta N, Keogh RH, Diaz-Ordaz K, Khunti K , et al. Risk prediction of covid-19 related death and hospital admission in adults after covid-19 vaccination: national prospective cohort study. BMJ 2021;374:n2244.

39. Docherty AB, Harrison EM, Green CA, Hardwick HE, Pius R, Norman L , et al. Features of 20 133 UK patients in hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study. BMJ 2020;369:m1985.

40. Office for National Statistics. Hospital admissions by age. 2022.

41. Office for National Statistics. Early indicators of the age and sex structure of the UK population in 2020. 2022.

42. ICNARC. Intenseive Care National Audit & Research Centre. ICNARC report on COVID-19 in critical care: England, Wales and Northern Ireland 8 April 2022. 2022.

43. Guyot P, Ades AE, Ouwens MJNM, Welton NJ. Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves. BMC Medical Research Methodology 2012;12:9.

74

44. NHS Digital. Average length of stay in hospital for patients with COVID-19 or suspected COVID. 2022.

45. ICER. Institute for Clinical and Economic Review. Special Assessment of Outpatient Treatments for COVID-19. Evidence Report. 2022.

46. Office for National Statistics. National Life Tables - life expectancy in the UK: 2018 to 2020. 2021.

47. Office for National Statistics. Self-reported long COVID after infection with the Omicron variant in the UK: 6 May 2022. 2022.

48. Office for National Statistics. Prevalence of ongoing symptoms following coronavirus (COVID-19) infection in the UK: 1 June 2022. 2022.

49. Evans RA, McAuley H, Harrison EM, Shikotra A, Singapuri A, Sereno M , et al. Physical, cognitive, and mental health impacts of COVID-19 after hospitalisation (PHOSP-COVID): a UK multicentre, prospective cohort study. The Lancet Respiratory Medicine 2021;9:1275-87.

50. NHS. National Cost Collection for the NHS. 2022.

51. Wilcox MH, Ahir H, Coia JE, Dodgson A, Hopkins S, Llewelyn MJ , et al. Impact of recurrent Clostridium difficile infection: hospitalization and patient quality of life. Journal of Antimicrobial Chemotherapy 2017;72:2647-56.

52. Hollmann M, Garin O, Galante M, Ferrer M, Dominguez A, Alonso J. Impact of influenza on health-related quality of life among confirmed (H1N1)2009 patients. PloS one 2013;8:e60477e.

53. Stokes EA, Wordsworth S, Bargo D, Pike K, Rogers CA, Brierley RCM , et al. Are lower levels of red blood cell transfusion more cost-effective than liberal levels after cardiac surgery? Findings from the TITRe2 randomised controlled trial. BMJ Open 2016;6:e011311.

54. Jones K, Burns A. Unit Costs of Health and Social Care 2021. University of Kent, Canterbury.: Personal Social Services Research Unit.; 2021.

55. Vos-Vromans D, Evers S, Huijnen I, Köke A, Hitters M, Rijnders N , et al. Economic evaluation of multidisciplinary rehabilitation treatment versus cognitive behavioural therapy for patients with chronic fatigue syndrome: A randomized controlled trial. PLoS ONE, 2017;12:e0177260.

56. GOV.UK. HMRC exchange rates for 2016: monthly. 2016.

57. Ara R, Brazier JE. Populating an economic model with health state utility values: moving toward better practice. Value Health 2010;13:509-18.

58. Cuthbertson BH, Elders A, Hall S, Taylor J, MacLennan G, Mackirdy F , et al. Mortality and quality of life in the five years after severe sepsis. Critical Care 2013;17:R70.

59. Marconi VC, Ramanan AV, de Bono S, Kartman CE, Krishnan V, Liao R , et al. Efficacy and safety of baricitinib for the treatment of hospitalised adults with COVID-19 (COV-BARRIER): a randomised, double-blind, parallel-group, placebo-controlled phase 3 trial. The Lancet Respiratory Medicine 2021;9:1407-18.

60. Group RC, Horby PW, Emberson JR, Mafham M, Campbell M, Peto L , et al. Baricitinib in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, openlabel, platform trial and updated meta-analysis. medRxiv 2022; 10.1101/2022.03.02.22271623:2022.03.02.22271623.

61. Ely EW, Ramanan AV, Kartman CE, de Bono S, Liao R, Piruzeli MLB , et al. Efficacy and safety of baricitinib plus standard of care for the treatment of critically ill hospitalised adults with COVID-19 on invasive mechanical ventilation or extracorporeal membrane oxygenation: an exploratory, randomised, placebo-controlled trial. The Lancet Respiratory Medicine 2022;10:327-36.

62. Kalil AC, Patterson TF, Mehta AK, Tomashek KM, Wolfe CR, Ghazaryan V , et al. Baricitinib plus Remdesivir for Hospitalized Adults with Covid-19. N Engl J Med 2021;384:795-807.

63. Recovery Collaborative Group, Horby PW, Mafham M, Peto L, Campbell M, Pessoa-Amorim G , et al. Casirivimab/imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. The Lancet 2022;399:66576.

64. Somersan-Karakaya S, Mylonakis E, Menon VP, Wells JC, Ali S, Sivapalasingam S , et al. Casirivimab/imdevimab for the treatment of hospitalized patients with COVID-19. medRxiv 2022; 10.1101/2021.11.05.21265656:2021.11.05.21265656.

75

65. O’Brien MP, Forleo-Neto E, Sarkar N, Isa F, Hou P, Chan K-C , et al. Effect of Subcutaneous Casirivimab/imdevimab Antibody Combination vs Placebo on Development of Symptomatic COVID-19 in Early Asymptomatic SARS-CoV-2 Infection: A Randomized Clinical Trial. JAMA 2022;327:432-41.

66. Weinreich DM, Sivapalasingam S, Norton T, Ali S, Gao H, Bhore R , et al. REGEN-COV Antibody Combination and Outcomes in Outpatients with Covid-19. New England Journal of Medicine 2021;385:e81.

67. Weinreich DM, Sivapalasingam S, Norton T, Ali S, Gao H, Bhore R , et al. REGN-COV2, a Neutralizing Antibody Cocktail, in Outpatients with Covid-19. New England Journal of Medicine 2020;384:238-51.

68. Temesgen Z, Burger CD, Baker J, Polk C, Libertin CR, Kelley CF , et al. Lenzilumab in hospitalised patients with COVID-19 pneumonia (LIVE-AIR): a phase 3, randomised, placebocontrolled trial. Lancet Respir Med 2022;10:237-46.

69. Arribas José R, Bhagani S, Lobo Suzana M, Khaertynova I, Mateu L, Fishchuk R , et al. Randomized Trial of Molnupiravir or Placebo in Patients Hospitalized with Covid-19. NEJM Evidence 2022;1:EVIDoa2100044.

70. Caraco Y, Crofoot Gordon E, Moncada Pablo A, Galustyan Anna N, Musungaie Dany B, Payne B , et al. Phase 2/3 Trial of Molnupiravir for Treatment of Covid-19 in Nonhospitalized Adults. NEJM Evidence 2022;1:EVIDoa2100043.

71. Fischer WA, 2nd, Eron JJ, Jr., Holman W, Cohen MS, Fang L, Szewczyk LJ , et al. A phase 2a clinical trial of molnupiravir in patients with COVID-19 shows accelerated SARS-CoV-2 RNA clearance and elimination of infectious virus. Sci Transl Med 2022;14:eabl7430.

72. Jayk Bernal A, Gomes da Silva MM, Musungaie DB, Kovalchuk E, Gonzalez A, Delos Reyes V , et al. Molnupiravir for Oral Treatment of Covid-19 in Nonhospitalized Patients. New England Journal of Medicine 2021;386:509-20.

73. Tippabhotla SK, Lahiri DS, D RR, Kandi C, V NP. Efficacy and Safety of Molnupiravir for the Treatment of Non-Hospitalized Adults With Mild COVID-19: A Randomized, Open-Label, Parallel-Group Phase 3 Trial. 2022; http://dx.doi.org/10.2139/ssrn.4042673.

74. Hammond J, Leister-Tebbe H, Gardner A, Abreu P, Bao W, Wisemandle W , et al. Oral Nirmatrelvir for High-Risk, Nonhospitalized Adults with Covid-19. New England Journal of Medicine 2022;386:1397-408.

75. Ader F, Bouscambert-Duchamp M, Hites M, Peiffer-Smadja N, Poissy J, Belhadi D , et al. Remdesivir plus standard of care versus standard of care alone for the treatment of patients admitted to hospital with COVID-19 (DisCoVeRy): a phase 3, randomised, controlled, openlabel trial. The Lancet Infectious Diseases 2022;22:209-21.

76. Beigel JH, Tomashek KM, Dodd LE, Mehta AK, Zingman BS, Kalil AC , et al. Remdesivir for the Treatment of Covid-19 — Final Report. New England Journal of Medicine 2020;383:181326.

77. Mahajan L, Singh AP, Gifty. Clinical outcomes of using remdesivir in patients with moderate to severe COVID-19: A prospective randomised study. Indian J Anaesth 2021;65:S41-s6.

78. Wang Y, Zhang D, Du G, Du R, Zhao J, Jin Y , et al. Remdesivir in adults with severe COVID19: a randomised, double-blind, placebo-controlled, multicentre trial. The Lancet 2020;395:1569-78.

79. Spinner CD, Gottlieb RL, Criner GJ, Arribas López JR, Cattelan AM, Soriano Viladomiu A , et al. Effect of Remdesivir vs Standard Care on Clinical Status at 11 Days in Patients With Moderate COVID-19: A Randomized Clinical Trial. JAMA 2020;324:1048-57.

80. Goldman JD, Lye DCB, Hui DS, Marks KM, Bruno R, Montejano R , et al. Remdesivir for 5 or 10 Days in Patients with Severe Covid-19. New England Journal of Medicine 2020;383:1827-37.

81. Gottlieb RL, Vaca CE, Paredes R, Mera J, Webb BJ, Perez G , et al. Early Remdesivir to Prevent Progression to Severe Covid-19 in Outpatients. New England Journal of Medicine 2021;386:305-15.

82. ACTIV-3/Therapeutics for Inpatients with COVID-19 (TICO) Study Group. Efficacy and safety of two neutralising monoclonal antibody therapies, sotrovimab and BRII-196 plus BRII-

76

198, for adults hospitalised with COVID-19 (TICO): a randomised controlled trial. Lancet Infect Dis 2022;22:622-35.

83. Gupta A, Gonzalez-Rojas Y, Juarez E, Crespo Casal M, Moya J, Rodrigues Falci D , et al. Effect of Sotrovimab on Hospitalization or Death Among High-risk Patients With Mild to Moderate COVID-19: A Randomized Clinical Trial. JAMA 2022;327:1236-46.

84. Broman N, Feuth T, Vuorinen T, Valtonen M, Hohenthal U, Löyttyniemi E , et al. Early administration of tocilizumab in hospitalized COVID-19 patients with elevated inflammatory markers; COVIDSTORM-a prospective, randomized, single-centre, open-label study. Clin Microbiol Infect 2022;28:844-51.

85. The Remap- C. A. P. Investigators, Derde LPG. Effectiveness of Tocilizumab, Sarilumab, and Anakinra for critically ill patients with COVID-19 The REMAP-CAP COVID-19 Immune Modulation Therapy Domain Randomized Clinical Trial. medRxiv 2021; 10.1101/2021.06.18.21259133:2021.06.18.21259133.

86. REMAP-CAP Investigators, Gordon A, Mouncey P, al. e. Interleukin-6 Receptor Antagonists in Critically Ill Patients with Covid-19. New England Journal of Medicine 2021;384:1491-502.

87. Hermine O, Mariette X, Tharaux P-L, Resche-Rigon M, Porcher R, Ravaud P , et al. Effect of Tocilizumab vs Usual Care in Adults Hospitalized With COVID-19 and Moderate or Severe Pneumonia: A Randomized Clinical Trial. JAMA Internal Medicine 2021;181:32-40.

88. Hermine O, Mariette X, Porcher R, Resche-Rigon M, Tharaux P-L, Ravaud P. Effect of Interleukin-6 Receptor Antagonists in Critically Ill Adult Patients with COVID-19 Pneumonia: two Randomised Controlled Trials of the CORIMUNO-19 Collaborative Group. European Respiratory Journal 2022; 10.1183/13993003.02523-2021:2102523.

89. RECOVERY Collaborative Group. Tocilizumab in patients admitted to hospital with COVID19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet 2021;397:163745.

90. Rosas IO, Bräu N, Waters M, Go RC, Hunter BD, Bhagani S , et al. Tocilizumab in Hospitalized Patients with Severe Covid-19 Pneumonia. New England Journal of Medicine 2021;384:150316.

91. Rosas IO, Diaz G, Gottlieb RL, Lobo SM, Robinson P, Hunter BD , et al. Tocilizumab and remdesivir in hospitalized patients with severe COVID-19 pneumonia: a randomized clinical trial. Intensive Care Med 2021;47:1258-70.

92. Rutgers A, Westerweel P, van der Holt B, al. e. Timely Administration of Tocilizumab Improves Survival of Hospitalized COVID-19 Patients. . 2021; http://dx.doi.org/10.2139/ssrn.3834311.

93. Salama C, Han J, Yau L, Reiss WG, Kramer B, Neidhart JD , et al. Tocilizumab in Patients Hospitalized with Covid-19 Pneumonia. New England Journal of Medicine 2020;384:20-30.

94. Salvarani C, Dolci G, Massari M, Merlo DF, Cavuto S, Savoldi L , et al. Effect of Tocilizumab vs Standard Care on Clinical Worsening in Patients Hospitalized With COVID-19 Pneumonia: A Randomized Clinical Trial. JAMA Internal Medicine 2021;181:24-31.

95. Soin AS, Kumar K, Choudhary NS, Sharma P, Mehta Y, Kataria S , et al. Tocilizumab plus standard care versus standard care in patients in India with moderate to severe COVID-19associated cytokine release syndrome (COVINTOC): an open-label, multicentre, randomised, controlled, phase 3 trial. The Lancet Respiratory Medicine 2021;9:511-21.

96. Stone JH, Frigault MJ, Serling-Boyd NJ, Fernandes AD, Harvey L, Foulkes AS , et al. Efficacy of Tocilizumab in Patients Hospitalized with Covid-19. New England Journal of Medicine 2020;383:2333-44.

97. Talaschian M, Akhtari M, Mahmoudi M, al. e. Tocilizumab Failed to Reduce Mortality in Severe COVID-19 Patients: Results From a Randomized Controlled Clinical Trial, 06 May 2021, PREPRINT (Version 1). 2021; https://doi.org/10.21203/rs.3.rs-463921/v1.

98. Veiga VC, Prats JAGG, Farias DLC, Rosa RG, Dourado LK, Zampieri FG , et al. Effect of tocilizumab on clinical outcomes at 15 days in patients with severe or critical coronavirus disease 2019: randomised controlled trial. BMJ 2021;372:n84.

99. Wang D, Fu B, Peng Z, Yang D, Han M, Li M , et al. Tocilizumab in patients with moderate or severe COVID-19: a randomized, controlled, open-label, multicenter trial. Frontiers of Medicine 2021;15:486-94.

77

7 APPENDICES

Appendix 1: Summary of clinical studies used to inform the economic model

Table 23: Summary of study and patient characteristics of included studies with relevant outcomes to inform the economic model (all data extracted from https://covid-nma.com/,[9] unless specified otherwise)

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

NR, not reported; RCT, randomised controlled trial; RoB, risk of bias

a Different remdesivir loading dose

b Data from http://www.metaevidence.org/covid19.aspx10

c For this outcome (hospital discharge), data reported for seronegative patients only: REGN-COV2, n=1633; standard care, n=1520

d For this outcome (hospital discharge), data reported for combined doses only: 2.4 g REGEN-COV (1.2 g casirivimab and1.2 g imdevimab), 8.0g REGEN-COV (4.0 g casirivimab/4.0 g imdevimab), n=804

95

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Assessment Report

Therapeutics for people with COVID-19 [ID4038]

Please read the checklist for submitting comments at the end of this form. We cannot accept forms that are not filled in correctly. Deadline for comments 5pm on Friday 29 July . Please submit via NICE Docs .

XXXXXXXXXXXXX AstraZeneca Disclosure Please disclose any past or current, direct or indirect links to, or funding from, the tobacco industry.

Please return to: NICE DOCS

Please comment on the assessment report

We thank ScHARR for the detailed Assessment Report and we understand the compromises that had to be made due to stringent timelines. We have separated the comments by topic and detailed additional considerations, analyses and potential inputs sources that we believe would strengthen the report and provide a more comprehensive evaluation of therapeutics for people with COVID-19. Our comments mostly focus on the outpatient treatment analysis (i.e. people with mild COVID-19 at high-risk of progressing to severe COVID-19 – page 16 ), if not otherwise stated.

Comment 1. Population of interest for the analysis

The current definition of patients at high-risk is aligned with the PANORAMIC trial[1] , which is a relatively wide definition that includes diseases such as chronic obstructive pulmonary disease, coronary heart disease, atrial fibrillation and congestive cardiac failure, and in whom hazard ratios for COVID-19 hospital admission are marginally higher than 1 (falling in the range 1.13-1.37, based on findings by Hippisley-Cox and colleagues[2] ).

While the PANORAMIC trial definition is broadly aligned with outpatient randomised controlled trial inclusion/exclusion criteria[3-6] , it is not aligned with the definition of the “ highest-risk clinical subgroups ” cohort from the independent advisory group[7] commissioned by the Department of Health and Social Care (DHSC), and the advisory group[8] commissioned by the Deputy Chief Medical Officer and supported by the NHS England RAPID-C19 team. The purpose of the set of conditions outlined by these advisory groups was to prioritise treatment with monoclonal antibodies and antivirals, and in whom one can reasonably expect the risk of COVID-19 hospitalisation on SoC to be significantly higher than the 1.8% used in the Assessment Report. These conditions are summarised below:

1. Down’s syndrome

2. Solid cancer

3. Haematological disease and stem cell transplant

4. Renal disease

5. Liver disease

6. Immune-mediated inflammatory disorders

7. Immune deficiencies

8. HIV/AIDS

9. Solid organ transplants

10. Rare neurological conditions

For instance, focusing on COVID-19 hospitalisation, Hippisley-Cox et al[2] reports a HR = 1.73 – 7.37 for chronic kidney disease stage 4/5, HR = 6.81 – 12.82 for bone marrow or solid organ transplant, HR=2.55 for Down’s syndrome and HR=2.3 for rare neurological conditions. A similar pattern is observed for COVID-19 death, with marked risk increase associated with chemotherapy recipients, CKD stage 4/5, Down’s syndrome, solid organ transplants and rare neurological conditions. This is further supported by disease specific evidence, with Pinato et al .[9] reporting 87.39% risk of hospitalisation in the UK cohort of the study (adult patients with cancer and consequent COVID-19 diagnosis) and Crolley 2020[10] reporting 88% risk of hospitalisation in adults with COVID-19 with cancer treated with SACT. Furthermore, Bierle, et al.[11] (US study in fully vaccinated individuals eligible for outpatient treatment with monoclonal antibodies) showed that the risk of hospitalisation markedly increases from 2.1% for patients with a MASS score of 0 (BMI = 25-35, hypertension and

Please return to: NICE DOCS

age <55, chronic lung disease and age <55) up to 29.9% for patients with MASS score of 4+ (3 MASS points given for patient with chronic kidney disease or immunocompromised).

In conclusion, we believe that NICE decision making would benefit from the inclusion of subgroup analyses for (1) the Highest-risk clinical subgroups cohort; and (2) Disease-specific clinical sub-groups (e.g. immunocompromised, chronic kidney disease, solid organ transplants). This is because outpatient treatments (monoclonal and antiviral therapeutics) are likely to provide substantial benefit for these sub-groups, given the markedly higher hospitalisation risk associated with these conditions (vs. the 1.8% used in the Assessment Report) and that the risk of hospitalisation on SoC is one of the key drivers of costeffectiveness ( page 12 ). Failure to adopt such an approach may run the risk of denying access to patients who serve to benefit most from treatment.

Comment 2. Populations contra-indicated to nirmatrelvir/ritonavir

• (1) Antiviral treatments are associated with a number of contra-indications as described in the UK Interim Clinical Commissioning Policy. Notably, nirmatrelvir/ritonavir is contra-indicated in severe liver and kidney disease and solid organ and islet transplants, and there are multiple significant drug-drug interactions to consider (as detailed on the COVID-19 drug interactions tool developed by the University of Liverpool[12] ), some of which pertaining to commonly prescribed drugs[13] such as atorvastatin (tool outcome: potential interaction) and apixaban (tool outcome: do not co-administer). This highlights the need to consider separately patient groups who may only be eligible for treatment with monoclonal antibodies, or monoclonal antibodies and molnupiravir/remdesivir.

Therefore, we recommend including a sub-group analysis focusing on the specific high/highest-risk population not eligible to receive nirmatrelvir/ritonavir. Note this suggestion overlaps with Comment 1, since part of the highest-risk cohort (e.g. solid organ transplant, chronic liver and kidney disease, broader immunocompromised population) is likely to be the key sub-groups in whom nirmatrelvir/ritonavir is contraindicated.

• (2) It is also worth noting there is a growing understanding that some patients do not respond to a normal 5-day course of nirmatrelvir/ritonavir.[14,15] Such patients may be treated with an extended course of nirmatrelvir/ritonavir monotherapy, however this does not always suppress viral replication and therefore alternative strategies are needed i.e. combination therapies with / without monoclonal antibodies, or monoclonal antibodies alone. We appreciate there may not be available data to inform treatment sequencing in the model, however this does highlight the need for other therapeutic strategies beyond antivirals.

Comment 3. Risk of hospitalisation on SoC

We note the UK Coronavirus dashboard estimate of 5% risk of hospitalisation was discarded from the base case ( page 39 ) in favour of a more conservative estimate of 0.9% from Nyberg, et al .[16] . Given the smaller sample size of Nyberg et al vs. the UK coronavirus dashboard, we suggest that an intermediate value between the two estimates (e.g. ~3%) would be more appropriate. Based on the latest available data from the UK Coronavirus Dashboard for England[17,18] from 26 January to 8 July, the risk of hospitalisation can be estimated at 4.24% ([798,595-591,828]/[19,507,435-14,630,705]), while Menni 2022[19] reported a hospitalisation rate of 1.9% (94/4,990). While these sources have their own

Please return to: NICE DOCS

limitations (likely underestimation of the denominator in the UK Coronavirus Dashboard; limited sample size in Menni 2022), it would appear that 0.9% represents a lower bound estimate rather than an average value representative of the risk of COVID-19 hospitalisations in the UK general population.

Furthermore, applying a HR of 2 ( page 39 ) leads to a risk of hospitalisation on SoC for the high-risk cohort of 1.8%, which is lower than the general population estimates from the UK Coronavirus dashboard and Menni 2022, and therefore would appear to lack face validity. In addition, recent UK clinical expert advice (July 2022) highlighted that a ~2-3% risk of hospitalisation on SoC for the high-risk population seems to be an underestimation, with a more reasonable estimate being in the 5%-10% range. Furthermore, the expert advice suggested that for the highest risk population, a risk of hospitalisation >10% would be expected. Finally, the advice highlighted that the risk of hospitalisation is bound to increase given that continuing to vaccinate the high-risk cohort 3-4 times per year, as done so far, is not feasible.

In conclusion, we propose that the risk of hospitalisation on SoC for the high-risk cohort should be revised upwards to a more plausible evidence-based estimate. In addition, as highlighted in Comment 1, separate sub-group analyses should be undertaken for the highest risk cohort and for specific diseases with markedly higher risk of hospitalisation.

Comment 4. Impact of monoclonal antibody treatment on the reduction of oxygen needs (low-flow oxygen, high-flow oxygen or non-invasive ventilation, invasive mechanical ventilation or extracorporeal membrane oxygenation).

In the current Assessment Report, the outpatient treatment effect is captured with respect to hospitalisation or death and all-cause mortality ( Table 5 – page 27 ). However, the monoclonal antibody trials of the treatments considered in the analysis, sotrovimab and casirivimab/imdevimab and, in the future, tixagevimab-cilgavimab have demonstrated a reduction of oxygen needs among hospitalised patients versus placebo (e.g. sotrovimab progression to severe COVID-19: 0.26 [0.12-0.59][3] ; tixagevimab-cilgavimab prevention of respiratory failure 71.9% [0.3% - 92.1%][20] ).

Data on the impact of antivirals in reducing oxygen needs are not available in the main trial publications for nirmatrelvir/ritonavir[21] , molnupiravir[6] and remdesivir[5] . Nevertheless, UK clinical expert advice highlighted that antivirals are not expected to have an impact on reducing oxygen needs due to the different mechanism of action vs. monoclonal antibodies.

In conclusion, we believe there are significant uncaptured QALYs and would therefore urge the EAG to include the reduction in oxygen needs associated with monoclonal antibodies in the model and analysis, given the large utility decrements and costs associated with low-flow oxygen, high-flow oxygen and invasive mechanical ventilation vs. no supplemental oxygen.

Comment 5. Mean age of the cohort

It is unclear why an arbitrary base case value of 65 years of age was assumed for people with COVID-19 at high risk of hospitalisation ( page 41 ). Given the lack of UK specific data, the outpatient clinical trial estimates may be a relevant source to consider, as summarised below.

• Sotrovimab - median age 53 years[3]

• Nirmatrelvir/ritonavir - median age 45-46.5 years[4]

• Molnupiravir – median age of 42-44 years[6]

• Casirivimab/imdevimab - median age of 48-50 years[22]

Please return to: NICE DOCS

• Remdesivir – age of 50-51 years[5]

Comparing the trial-based estimates above, it is clear that the value of 65 years used in the Assessment Report overestimates the outpatient cohort age. As highlighted in the report ( page 67 ), the outpatient cohort starting age is an important driver and strongly impacts the ICER with drugs becoming more cost-effective with decreasing age.

We therefore suggest amending the outpatient cohort age using the outpatient trial patient characteristics (as summarised above), until UK specific data to accurately estimate the outpatient cohort age becomes available.

Comment 6. Re-infections are not included in the model

Growing evidence suggests that COVID-19 natural immunity wanes over time and that COVID-19 vaccine-induced immunity is more protective than infection-induced immunity.[23,24] Reynolds et al.[25] observed that the likelihood of reinfection following Omicron is higher compared to other variants and therefore natural infection with Omicron is a poor booster of COVID-19 immunity.

In contrast to antivirals, monoclonal antibodies have the potential to confer longer lasting protection against re-infection than natural immunity from the virus itself. Data to differentiate protection derived from the virus as opposed to monoclonal antibodies is currently emerging; we therefore suggest that (1) this aspect should be discussed by the NICE Technology Appraisal Committee, where patient and public involvement will be leveraged; and (2) the model structure should be amended so that this value may be appropriately captured and explored.

Comment 7. Logistics of treatment

As highlighted in the report ( page 40 ), the logistics of treatment in the community have not been assessed. However, the report does recognise that this could be an important factor in deciding which treatments may be preferred.

Whereas IV treatments such as sotrovimab, remdesivir, casirivimab/imdevimab can only be administered in a specialist, controlled environment (i.e. hospital setting) and have a requirement for observation over time, UK clinical expert advice suggests that oral and intramuscular treatments may be administered in a primary care setting without the need to access the COVID Medicines Delivery Unit, community hospital or out-patient facilities. This would bring benefits for patients in terms of easier access and, additionally, benefits to the NHS in terms of resource use.

This is an important aspect that should be discussed by the NICE Technology Appraisal Committee, where patient and public involvement will be leveraged. It may also be appropriate to undertake a full incremental analysis for the IV and oral/subcutaneous/ intramuscular treatments separately.

Please return to: NICE DOCS

Checklist for submitting comments

• Consultation responses must not be longer than 10 pages , excluding the references.

• Use this comment form and submit it as a Word document (not a PDF).

• • Complete the disclosure about links with, or funding from, the tobacco industry. • Combine all comments from your organisation into 1 response. We cannot accept more than 1 set of comments from each organisation.

• Do not paste other tables into this table – type directly into the table.

• Please underline all confidential information, and separately highlight information that is submitted under ‘commercial in confidence’ in turquoise, all information submitted under ‘academic in confidence’ in yellow. If confidential information is submitted, please also send a second version of your comments with that information replaced with the following text: ‘academic/commercial in confidence information removed’. See the Process and methods manual (section 5.4) for more information.

• Do not include medical information about yourself or another person from which you or the person could be identified.

• • Do not use abbreviations

• Do not include attachments such as research articles, letters or leaflets. For copyright reasons, we will have to return comments forms that have attachments without reading them. You can resubmit your comments form without attachments, it must send it by the deadline.

• If you have received agreement from NICE to submit additional evidence with your comments on the pro-forma response document, please submit these separately.

Note: We reserve the right to summarise and edit comments received during consultations, or not to publish them at all, if we consider the comments are too long, or publication would be unlawful or otherwise inappropriate.

Comments received during our consultations are published in the interests of openness and transparency, and to promote understanding of how recommendations are developed. The comments are published as a record of the comments we received, and are not endorsed by NICE, its officers or advisory committees.

Please return to: NICE DOCS

References

1. Participant Information — PANORAMIC (panoramictrial.org). https://www.panoramictrial.org/participant-information?885eb090-0cdc-11ed-882a0a5fbb384b76. Accessed 26 July 2022.

2. Hippisley-Cox J, Coupland CA, Mehta N, et al. Risk prediction of covid-19 related death and hospital admission in adults after covid-19 vaccination: national prospective cohort study. Bmj. 2021;374:n2244.

3. Gupta A, Gonzalez-Rojas Y, Juarez E, et al. Effect of Sotrovimab on Hospitalization or Death Among High-risk Patients With Mild to Moderate COVID-19: A Randomized Clinical Trial. JAMA. 2022;327(13):1236-1246.

4. Hammond J, Leister-Tebbe H, Gardner A, et al. Oral Nirmatrelvir for High-Risk, Nonhospitalized Adults with Covid-19. N Engl J Med. 2022;386(15):1397-1408.

5. Gottlieb RL, Vaca CE, Paredes R, et al. Early Remdesivir to Prevent Progression to Severe Covid-19 in Outpatients. N Engl J Med. 2022;386(4):305-315.

6. Jayk Bernal A, Gomes da Silva MM, Musungaie DB, et al. Molnupiravir for Oral Treatment of Covid-19 in Nonhospitalized Patients. N Engl J Med. 2022;386(6):509-520.

7. Coronavirus » Interim Clinical Commissioning Policy: Therapies for patients with symptomatic hospital-onset COVID-19 – clinical guide (england.nhs.uk). https://www.england.nhs.uk/coronavirus/documents/c1572-interim-clinical-commissioning-policytherapies-for-patients-with-symptomatic-hospital-onset-covid-19-clinical-guide/. Accessed 26 July 2022.

8. Defining the highest-risk clinical subgroups upon community infection with SARS-CoV-2 when considering the use of neutralising monoclonal antibodies (nMABs) and antiviral drugs: independent advisory group report - GOV.UK (www.gov.uk). https://www.gov.uk/government/publications/higher-risk-patients-eligible-for-covid-19-treatmentsindependent-advisory-group-report/defining-the-highest-risk-clinical-subgroups-upon-communityinfection-with-sars-cov-2-when-considering-the-use-of-neutralising-monoclonal-antibodies. Accessed 26 July 2022.

9. Pinato DJ, Scotti L, Gennari A, et al. Determinants of enhanced vulnerability to coronavirus disease 2019 in UK patients with cancer: a European study. Eur J Cancer. 2021;150:190-202.

10. Crolley VE, Hanna D, Joharatnam-Hogan N, et al. COVID-19 in cancer patients on systemic anticancer therapies: outcomes from the CAPITOL (COVID-19 Cancer PatIenT Outcomes in North London) cohort study. Ther Adv Med Oncol. 2020;12:1758835920971147.

11. Bierle DM, Ganesh R, Tulledge-Scheitel S, et al. Monoclonal Antibody Treatment of Breakthrough COVID-19 in Fully Vaccinated Individuals with High-Risk Comorbidities. J Infect Dis. 2022;225(4):598-602.

12. University of Liverpool - COVID-19 Drug Interactions. https://www.covid19druginteractions.org/checker

13. What are the most commonly prescribed medicines? Top 10 prescribed medicines in NHS England primary care for 2019. https://www.bennett.ox.ac.uk/blog/2020/03/what-are-the-mostcommonly-prescribed-medicines-top-10-prescribed-medicines-in-nhs-england-primary-care-for2019/

14. Alshanqeeti S, Bhargava A. COVID-19 Rebound After Paxlovid Treatment: A Case Series and Review of Literature. Cureus. 2022;14(6):e26239.

15. Rubin R. From Positive to Negative to Positive Again-The Mystery of Why COVID-19 Rebounds in Some Patients Who Take Paxlovid. Jama. 2022;327(24):2380-2382.

16. Nyberg T, Ferguson NM, Nash SG, et al. Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. Lancet. 2022;399(10332):1303-1312.

17. Cases in England | Coronavirus in the UK (data.gov.uk). https://coronavirus.data.gov.uk/details/cases?areaType=nation&areaName=England. Accessed 26 July 2022.

Please return to: NICE DOCS

18. Healthcare in England | Coronavirus in the UK (data.gov.uk). https://coronavirus.data.gov.uk/details/healthcare?areaType=nation&areaName=England. Accessed 26 July 2022.

19. Menni C, Valdes AM, Polidori L, et al. Symptom prevalence, duration, and risk of hospital admission in individuals infected with SARS-CoV-2 during periods of omicron and delta variant dominance: a prospective observational study from the ZOE COVID Study. Lancet. 2022;399(10335):1618-1624.

20. Montgomery H, Hobbs FDR, Padilla F, et al. Efficacy and safety of intramuscular administration of tixagevimab-cilgavimab for early outpatient treatment of COVID-19 (TACKLE): a phase 3, randomised, double-blind, placebo-controlled trial. Lancet Respir Med. 2022.

21. Hammond J, Leister-Tebbe H, Gardner A, et al. Oral Nirmatrelvir for High-Risk, Nonhospitalized Adults with Covid-19. New England Journal of Medicine. 2022;386(15):1397-1408.

22. Weinreich DM, Sivapalasingam S, Norton T, et al. REGEN-COV Antibody Combination and Outcomes in Outpatients with Covid-19. N Engl J Med. 2021;385(23):e81.

23. Cavanaugh AM, Spicer KB, Thoroughman D, Glick C, Winter K. Reduced Risk of Reinfection with SARS-CoV-2 After COVID-19 Vaccination - Kentucky, May-June 2021. MMWR Morb Mortal Wkly Rep. 2021;70(32):1081-1083.

24. Bozio CH, Grannis SJ, Naleway AL, et al. Laboratory-Confirmed COVID-19 Among Adults Hospitalized with COVID-19-Like Illness with Infection-Induced or mRNA Vaccine-Induced SARS-CoV-2 Immunity - Nine States, January-September 2021. MMWR Morb Mortal Wkly Rep. 2021;70(44):1539-1544.

25. Reynolds CJ, Pade C, Gibbons JM, et al. Immune boosting by B.1.1.529 (Omicron) depends on previous SARS-CoV-2 exposure. Science. 2022;377(6603):eabq1841.

Please return to: NICE DOCS

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Assessment Report

Therapeutics for people with COVID-19 [ID4038]

Please read the checklist for submitting comments at the end of this form. We cannot accept forms that are not filled in correctly. Deadline for comments 5pm on Friday 29 July . Please submit via NICE Docs .

1

Summary

Gilead acknowledges the unique and inherent challenges of carrying out an assessment of the cost-effectiveness of treatments for COVID-19. We support the EAG’s conclusion that remdesivir has been found to be plausibly cost-effective alongside other treatments for COVID-19, when compared to standard of care (SoC). Considering the challenges relating to the emergence of new variants, evidence shows that anti-virals, such as remdesivir, show reliability of effectiveness in COVID-19 regardless of the variant (Takashita et al., 2022)[1] .

Whilst we acknowledge the unique challenges of rapidly appraising medicines in a pandemic or post-pandemic setting, Gilead has significant concerns about the conduct of this technology appraisal, primarily with regard to robustness, fairness, and a lack of methodological transparency.

NICE has substantially deviated from usual MTA process in terms of the process and the methods of technology appraisal. The Evidence Assessment Group (EAG) were commissioned and the Evidence Assessment Report (EAR) was published without formally starting the technology appraisal process. No other stakeholder submission has been considered as this stage in the technology appraisal process. Gilead is of the opinion that NICE have potentially failed to act fairly by breaching its own published process and methods for the development of guidance. We believe that NICE has potentially acted outside of its remit by formally starting the technology appraisal despite claims that process has not formally begun and has potentially exceeded its powers by ‘resequencing the steps of the MTA’. Companies have potentially been treated unfairly during the first phase of this assessment. It is unclear if there will be any mechanism to appeal, rectify or otherwise meaningfully contribute to the development of robust guidance on the use of COVID-19 therapeutics.

In this new and undefined HTA process, it is not clear the extent to which companies will be permitted to submit evidence, in particular clinical study reports (CSRs). HTAs which adhere to the published NICE process and methods guidelines permit the submission and consideration of CSRs. CSRs represent an important source of clinical and safety data on interventions and comparators. Companies have not been invited to submit an evidence submission before the development of the EAR; it is unclear if and how this evidence will be considered by the EAG and the committee. It is also unclear if, how and when commercial in confidence patient access schemes (PAS) net price discounts will be considered in the technology appraisal process, which in turn unfairly constrains companies’ participation in the technology appraisal process and may limit patient access to effective treatments.

These issues, and others, are fundamental to the way the technology appraisal is being conducted. While challenges such as changing vaccination status and changing COVID-19 variants are unavoidable, we would respectfully request amendments regarding the following specific elements of the EAG’s analysis.

1. The incremental cost-effectiveness analysis included in Section 4.1. is unhelpful - and potentially obfuscating for decision making

2

The incremental analysis undertaken by the EAG, which includes all comparators within each of the populations, is inappropriate and methodologically flawed. Gilead is of the opinion that any recommendation based upon the incremental analysis may be unreasonable given the methodological flaws in the analysis and due to the substantial deviation from normal NICE process and methods. The EAG notes that “the EAG did not feel confident that it could robustly identify a treatment that was more efficacious than others” (pp.69) therefore it is important to carefully consider how the incremental analysis is positioned within the report and any subsequent NICE guidance.

Typically, incremental analyses are used to assess comparators that are mutually exclusive and could in practice displace each other; however, this is not the case for the treatments compared within this appraisal. For example, in clinical practice remdesivir, baricitinib, tocilizumab or lenzilumab can be used in the same patient at different stages of their disease progression. This is due to their very different but complementary mode of actions making these therapies role in the treatment of COVID-19 not mutually exclusive: remdesivir targets viral replication while tocilizumab, baricitinib and lenzilumab act against the inflammatory phase induced by the patient’s own immune system. This is clearly evidenced by the fact that remdesivir is used as a backbone or component of SoC in a number of trials for other medicines included in the MTA. These include baricitinib and lenlizumab, and in fact data on remdesivir’s combination with the former is referred to by the EAG in the overview of evidence for this intervention (ACTT-2). There is a growing body of literature supporting the positioning of remdesivir as a backbone for therapy, for example, Ngo. D. et al (2022), which explored the combination of dexamethasone, remdesivir and baricitinib in severe COVID-19. As we outlined in our response to the scoping exercise, there is further evidence of remdesivir being considered standard of care in the clinical trials of other treatments being studied for use in hospitals, for example:

• The US National Institute of Allergy & Infectious Diseases ACTT and ACTIV platform trials:

  • ACTT-2: evaluated the combination of baricitinib and remdesivir compared to remdesivir alone

  • ACTT-3: evaluated the combination of interferon beta-1a and remdesivir compared to remdesivir alone

  • ACTT-4: evaluated the combination of baricitinib and remdesivir compared to dexamethasone and remdesivir

  • ACTIV-3: evaluating the combination of ACTIV-3 investigational treatments and remdesivir compared to remdesivir plus placebo

  • ACTIV-5 (BET-A): evaluated the combination of remdesivir and Risankizumab compared to remdesivir plus placebo

1 Takashita E, Yamayoshi S, Simon V, van Bakel H, Sordillo EM, Pekosz A, Fukushi S, Suzuki T, Maeda K, Halfmann P, Sakai-Tagawa Y, Ito M, Watanabe S, Imai M, Hasegawa H, Kawaoka Y. Efficacy of Antibodies and Antiviral Drugs against Omicron BA.2.12.1, BA.4, and BA.5 Subvariants. N Engl J Med. 2022 Jul 20. doi: 10.1056/NEJMc2207519. Epub ahead of print. PMID: 35857646.

3

• ACTIV-5 (BET-B): recruiting to evaluate the combination of lenzilumab and remdesivir compared to remdesivir plus placebo

• ACTIV-5 (BET-C): recruiting to evaluate the combination of danicopan and remdesivir compared to remdesivir plus placebo in patients younger than or older and equal to 70 years old

• ITAC trial (NIAID, NIH and INSIGHT): evaluating the combination of hyperimmune immunoglobin to SARS-CoV-2 (hIVIG) and remdesivir compared to remdesivir plus placebo

• Casirivimab and Imdevimab for Treatment of Hospitalized Patients With COVID-19 Receiving Low Flow or No Supplemental Oxygen

In this example, the pairwise analysis showed that all of these treatments were cost-effective against SoC, but the incremental analysis suggested that all were dominated or extendedly dominated. The real-world comparators of remdesivir in combination with tocilizumab, baricitinib or lenzilumab are not considered due to lack of evidence, although their inclusion may have theoretically altered the results of the incremental analysis.

While current NICE guidelines recommend that ideally a full incremental analysis is carried out, in this instance, inclusion of the incremental analysis in the main body of the report encourages decision-making based on flawed evidence. We therefore suggest the incremental analysis is removed or reported within the appendices for information purposes only, and further cautionary notes are included regarding the interpretation of the EAG’s incremental analysis.

2. The analysis does not appropriately segment patient in the ‘in hospital’ setting by oxygen use, and should do so

Related to the clinically irrelevant comparisons between some interventions, is the inappropriate patient segmentation within the hospital setting. The EAG fails to segment the patient population according to oxygen use, which is important to do, not only to reflect clinical practice and treatment decisions and sequencing, but also to reflect the correct wording of the regulatory labels of the various interventions. For example, the labels for nirmatrelvir/ritonavir, molnupiravir and sotrovimab specifically state there should be no requirement for supplemental oxygen. The label for tocilizumab refers to supplemental oxygen and mechanical ventilation, while the data included in the EAG report for lenlizumab (LIVE AIR) only considered survival without ventilation as an endpoint. Although baricitinib is yet to receive a label, the NHSE commissioning highlights in order to treat patients with baricitinib they need to be receiving supplemental oxygen or respiratory support. Meanwhile, the label for remdesivir, which has recently received positive CHMP opinion for a full marketing authorisation from a conditional approval, includes its use in patients not requiring supplemental oxygen up until patients requiring, low flow, high flow and non-invasive mechanical ventilation.

4

Furthermore, the use of these therapies at different stages of their disease progression is important to understand. For example, the use of therapies with an immunomodulatory mode of action too early (such as in a patient not yet requiring supplementary oxygen support) can be detrimental to a patient’s outcomes as outlined in the RECOVERY trial for dexamethasone. The EAG seem to discount this clinically important note when assessing clinical and cost effectiveness of the therapies in the different disease settings.

Therefore, we request that the EAG structures the model to appropriately reflect the various patient groups within the hospitalised setting which are people with COVID-19 who are in hospital and require low flow supplemental oxygen for the management of their COVID-19 disease, people with COVID-19 who are in hospital and require high flow oxygen, and people with COVID-19 who are in hospital and require mechanical ventilation / ECMO. This split would therefore enable the EAG to conduct clinically relevant comparisons, rather than seeking to assess medicines in a patient population without recognising key stages of disease progression. Due to this reasoning, this makes many of the figures (such as figure 5) included in the report uninterpretable and potentially misleading.

3. The model structure excludes the impact of more severe disease on longer-term outcomes

The existing EAG model structure assumes that a proportion of patients experience ‘long COVID’, with an associated reduction in quality of life (QoL) and higher mortality risk. This approach currently does not differentiate between hospitalised patients with lower-intensity oxygen requirements and those with more severe disease requiring admission to ICU and/or need for mechanical ventilation or ECMO. The Sheinson et al. (2021), economic evaluation of treatment for patients hospitalised with COVID-19 applies a disutility for patients requiring mechanical ventilation for a period of 5-years post-discharge based on evidence that even relatively young patients who survived acute respiratory distress syndrome (ARDS) had persistent exercise limitations and a reduced physical quality of life 5 years after their critical illness.

The same model applied a hazard ratio for post-discharge mortality for ventilated patients vs. general population for 5 years based on evidence that there is an increased risk of death (33%) and hospital readmission rate (22%) in patients surviving an episode of intensive care compared with hospital control subjects in the 5 years after discharge from hospital, after adjusting for important confounders.

These structural additions would be expected to increase QALY gain and therefore reduce the ICER for comparators demonstrated to reduce the need for mechanical ventilation or ECMO.

5

4. Rationale for selection of certain sources of evidence and studies are unclear and should be more fully justified by the EAG

It is not clear which sources of evidence the EAG has chosen to include and exclude. Equally, where some sources of evidence were excluded, these are not described in detail, and the EAG states “were deemed to lack full transparency in the extracted outcome data”.

Regarding extraction of particular studies, the EAG states that all data extractions were undertaken by the end of May 2022 but certain key studies which one would reasonably expect to be included, for example the SOLIDARITY study, were not. The EAG provide no explanation of the reasons for excluding such studies. With regard to SOLIDARITY in particular, this is the full data set for which DISCOVERY is a sub study and was included (see table 23 of the EAG report), so it is not clear why the EAG has not used the full data set, which would enable a more comprehensive appraisal of the available evidence. In addition, NICE has recently updated the living guidelines for the management of COVID-19 using the SOLIDARITY data set which confirms the relevance of this source of evidence. The guideline also splits patient groups in hospital by oxygen usage, so we would hope the EAG would reflect this in its report and ensure alignment between various NICE guidelines associated with the treatment and management of COVID-19.

Furthermore, on the studies selected to estimate the efficacy of each treatment, the EAG notes how many studies were used but does not indicate which studies these were. The approach to identifying and selecting key sources of clinical evidence and lack of transparency regarding data selection is unsystematic and contrary to the normal NICE methods.

It appears that the selection of interventional studies, and therefore the appraisal of the benefit is inequitable and unbalanced. For example, the EAG’s assessment of nirmatrelvir/ritonavir is based on only one study (EPIC-HR) despite the fact that there are multiple additional sources of evidence to consider. Equally, the evidence for baricitinib is drawn from COV BARRIER, when RECOVERY offers a meta-analysis of all trials for the intervention, including COV BARRIER. Therefore, we do not believe the most relevant or applicable data has been selected for many of the interventions, and this flaw in the analysis highlights the importance of a systematic literature review underpinning any HTA.

Additionally, in regard to the studies that were specifically selected for remdesivir, the EAG gave no clear rationale for why certain outcomes were chosen from these studies. An example is the inclusion of the pivotal study ACTT-1 (Biegel et al, 2020) – this study had a clear primary endpoint of time to recovery, however the EAG chose to select this study to look at the outcome of time to death. If the EAG had included SOLIDARITY in their assessment, this outcome would make more sense to be retrieved from this study given the primary endpoint of investigating mortality in a much larger population. The EAG discount the outcome of time to discharge for remdesivir, which is an outcome that could easily be retrieved from ACTT-1. Furthermore, the EAG have chosen to include Wang et al, a study that was halted early due to the lockdown in China and led to a study that could not be completed efficiently and gave an underpowered trial, which taken alone, gives inconclusive findings. Given that this study was selected to assess the outcomes time to death and clinical improvement, both ACTT-1 and SOLIDARITY amongst others are far more robust

6

data sets from which to retrieve these outcomes for assessment. Gilead is of the opinion that the EAG’s arbitrary selection of studies and outcomes may lead to unreasonable recommendations in light of the evidence that is available.

Gilead strongly recommends that the EAG conduct a fully systematic review of all literature (clinical, humanistic and economic), and fully describes the rationale for inclusions and exclusion of sources of evidence and the studies themselves as well as rationale for selecting certain outcomes from each study selected. The information should be presented in a PRISMA diagram and the appraisal should adhere to the NICE Reference Case.

7

Checklist for submitting comments

• Consultation responses must not be longer than 10 pages , excluding the references.

• Use this comment form and submit it as a Word document (not a PDF).

• • Complete the disclosure about links with, or funding from, the tobacco industry. • Combine all comments from your organisation into 1 response. We cannot accept more than 1 set of comments from each organisation.

• • Do not paste other tables into this table – type directly into the table. • Please underline all confidential information, and separately highlight information that is submitted under ‘commercial in confidence’ in turquoise, all information submitted under ‘academic in confidence’ in yellow. If confidential information is submitted, please also send a second version of your comments with that information replaced with the following text: ‘academic/commercial in confidence information removed’. See the Process and methods manual (section 5.4) for more information.

• • Do not include medical information about yourself or another person from which you or the person could be identified.

• • Do not use abbreviations • Do not include attachments such as research articles, letters or leaflets. For copyright reasons, we will have to return comments forms that have attachments without reading them. You can resubmit your comments form without attachments, it must send it by the deadline.

• If you have received agreement from NICE to submit additional evidence with your comments on the pro-forma response document, please submit these separately.

Note: We reserve the right to summarise and edit comments received during consultations, or not to publish them at all, if we consider the comments are too long, or publication would be unlawful or otherwise inappropriate.

Comments received during our consultations are published in the interests of openness and transparency, and to promote understanding of how recommendations are developed. The comments are published as a record of the comments we received, and are not endorsed by NICE, its officers or advisory committees.

8

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Gilead Scienced Ltd – Pro-forma Response

Executable Model

Therapeutics for people with COVID-19 [ID4038]

The economic model enclosed and its contents are confidential and are protected by intellectual property rights, which are owned by ScHARR . It has been sent to you for information only. It cannot be used for any other purpose than to inform your understanding of the appraisal. Accordingly, neither the model nor its contents should be divulged to anyone other than those individuals within your organisation who need to see to them to enable you to prepare your response. Those to whom you do show the documents must be advised they are bound by the terms of the Confidentiality Agreement Form that has already been signed and returned to the Institute by your organisation.

You may not make copies of the file and you must delete the file from your records when the appraisal process, and any possible appeal, are complete. If asked, you must confirm to us in writing that you have done so. You may not publish it in whole or part, or use it to inform the development of other economic models.

The model must not be re-run for purposes other that the testing of its reliability .

Please set out your comments on reliability in writing providing separate justification, with supporting information, for each specific comment made. Where you have made an alteration to the model details of how this alteration was implemented in the model (e.g. in terms of programme code) must be given in sufficient detail to enable your changes to be replicated from the information provided. Please use the attached pro-forma to present your response.

Please prepare your response carefully. Responses which contain errors or are internally inconsistent (for example where we are unable to replicate the results claimed by implementing the changes said to have been made to the model) will be rejected without further consideration.

Results from amended versions of the model will only be accepted if their purpose is to test robustness and reliability of the economic model. Results calculated purely for the purpose of using alternative inputs will not be accepted.

No electronic versions of the economic model will be accepted with your response.

1

Responses should be provided in tabular format as suggested below (please add further tables if necessary).

July 2022

2

Issue 1 Model size

Issue 2 Long covid discounting

3

Issue 3 Lack of attribution of impact of ICU admission on long-term outcomes

Issue 4 Utility decrements

4

decrements to be incorporated this might be a more material issue.

Issue 5 Inadequate uncertainty analyses

5

6

Issue 7 Implementation of clinical improvement within hospital

Issue 8 Use of remdesivir in patients at ordinal state seven

7

8

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Assessment Report

Therapeutics for people with COVID-19 [ID4038]

Please read the checklist for submitting comments at the end of this form. We cannot accept forms that are not fille d in correctly. Deadline for comments 5pm on Friday 29 July . Please submit via NICE Docs .

Please return to: NICE DOCS

Please comment on the assessment report

In the last few months, millions of UK COVID-19 infections have been recorded, resulting in a corresponding upturn in both hospitalisations and mortality (Alastair McLellan 2022). Considering the potential virulence of some reported variants of concern/interest, it is vital that all efforts are exhausted to ensure the most accurate assessment of available therapeutics, so as not to exclude interventions that provide significant therapeutic benefit and value in the continuing management of the COVID-19 pandemic.

• After reviewing the External Assessment Group (EAG) report, a number of significant issues were identified relating to the process, methodology, and inputs used by the EAG in this analysis. The issues identified include: • Baseline hospitalisation rate • Exclusion of age from high-risk population definition • Effect Modifiers • Comparative Efficacy • ICER threshold and outbreak status • Administration cost • Absence of Safety Profile • Clinical inputs

Issues specifically related to the implementation of the economic analysis are provided in the separate proforma, as requested.

1. Baseline hospitalisation rate ( section 3.1.2, pg.38)

Description of problem

“Hospitalisation rates for patients on SoC (Standard of Care) were taken from Nyberg et al. where recent risks of hospitalisations associated with the Omicron variant were reported. Hospital admission up to 14 days after positive test was approximately 0.9% (over 9,000 patients admitted from a reported million cases).”

“Based on these data and clinical advice, the EAG applied a multiplier of 2 to the average hospitalisation rate for all patients to estimate the rate in people at high-risk of hospitalisation in the base case and increased it in sensitivity analyses”

Proposed amendment

Given that this parameter is a key driver of the cost-effectiveness analysis, we believe that the non-systematic approach to identifying a source for this value is a significant limitation of the analysis.

The authors highlight limitations in the study that could have impacted the estimated hospitalisation rate. (Nyberg et al. 2022) The authors noted the study excluded hospitalisations in the following cases:

“Cases were excluded if the NHS number recorded was missing or invalid (since such cases could not be linked to hospitalisation or vaccination records); information was missing for any adjustment variables; there were more than 14 days between the date of the first positive test and the date of the test which led to the variant being identified (via sequencing, genotyping, or S-gene positivity); or the specimen date was after an individual had died. We also excluded a small number of cases in individuals who had received (1) a vaccine other than Oxford–AstraZeneca, Pfizer–BioNTech, or Moderna or more than three doses of vaccine; (2) a third dose of vaccine that was not Pfizer or Moderna; or (3) a third dose of vaccine less than 80 days after the second dose.”

In addition, viral evolution, as reported by UK Health Security Agency(UKHSA 2022), impacts the severity of COVID-19, with the risk of hospitalisation varying with variants of the SARS-CoV-2 virus. Since hospitalisation rate is the major driver of cost effectiveness for non-hospitalised interventions, we believe more work is required to identify and validate a more realistic hospitalisation rate.

Please return to: NICE DOCS

The EAG attempt to adjust the Nyberg et al(Nyberg et al. 2022) hospitalisation rate by applying a multiplier of 2 to account for people in high-risk populations, a value derived from “…these data and clinical advice.” Few details are given about how this value was derived, including the uncertainty around such a critical model parameter. Given the significant uncertainty and lack of clarity about how these values were identified and validated, we request that the EAG conduct a more formal targeted review of the literature to inform this parameter.

GSK notes the EAG does not believe that the estimated 5% hospitalisation rate derived from the UK Coronavirus dashboard is valid, given “…it may be the case that half of patients with COVID-19 in hospital, were not hospitalised due to COVID-19.” (Page 39), However, the EAG report states “The population considered within the EAG report has been divided into two broad groups. The first group consists of people who have been hospitalised due to COVID-19 and the second group consists of people who are at high-risk of requiring hospital care due to COVID-19. Patients who were hospitalised for reasons other than COVID-19 and contracted COVID19 in hospital and were at high-risk of requiring hospital care for COVID-19 in itself were categorised within the second group” , so patients who are treated for COVID-19 infection acquired within the hospital are still eligible for these treatments and are therefore considered as part of the population for community treatment.

Finally, the baseline hospitalisation rate assumes that this value will hold for the future for which any NICE guidance may apply. However, the hospitalisation rate could plausibly increase if future variants are more virulent, or if protection offered by vaccines wanes. If treatments are reserved for specific ‘ultra-high risk’ populations, then the baseline hospitalisation rate should be adjusted to account for the increased baseline risk of hospitalisation in any appropriate subgroup analyses.

Expected Impact on result

A higher baseline hospitalisation rate will lower the estimated ICER versus standard of care for effective interventions given in the community to patients at high-risk of hospitalisation for COVID-19.

2. Exclusion of Age from High-Risk Population Definition (section 1.4.1 Paragraph 2, pg.17)

Description of problem

“Following discussions with NICE, the definition for patients at high-risk was aligned to that considered within the Platform Adaptive trial of NOvel antiviRals for eArly treatMent of COVID-19 In the Community (PANORAMIC) clinical study(PANORAMIC 2022), with the exception that being aged 50 years or over was not considered to be a high-risk factor.”

Evidence from the UK suggests that the risk of poorer outcomes from COVID-19 infection increases substantially with age among healthy adults and adults with underlying health conditions. People older than 65 are by far the most at risk, and the risk increases with age (Hippisley-Cox et al. 2021).

Proposed amendment GSK disagrees with the exclusion of age as a risk factor and asks for this assumption to be reconsidered for the following reason(s):

Office of National Statistics (ONS 2022) data shows there is a clear correlation between age and hospitalisation rate. Data for the week ending 3rd July 2022, showed overall hospital admissions were highest for those aged 85 years and over and lowest for those aged between 5 and 14 years. Real world evidence has shown that

Please return to: NICE DOCS

“Although overall hospital admission rates have consistently been highest in the oldest age group, the highest ICU (Intensive Care Unit) and HDU admission rates have varied across groups aged 55 years and over.” (ONS 2022) The ONS data and above statement confirmed a consistent trend throughout the pandemic.

Based on consultation with UK clinical experts, we request the inclusion of people aged over 50 years and also, the inclusion of clinical frailty score as risk factors for the high-risk group.

Expected Impact on result

The inclusion of age as a risk factor could / is likely to increase the hospitalization rate of the high-risk group. This key clinically relevant parameter will lower the estimated ICER for effective interventions given in the community to patients at high-risk of hospitalisation for COVID-19.

In addition, including age 50 years and over could have an increasing impact on the proportional shortfall calculation, which could influence the application of a severity modifier and a higher cost effectiveness threshold.

3. Analytical perspective: Effect Modifiers (section 2.1.5 pg.24)

Description of problem

The EAG stated “The conditions under which each study was evaluated were heterogeneous. Across time SoC has changed markedly, most particularly with reference to the widespread use of corticosteroids such as dexamethasone and change in SARS-CoV-2 variants. The vaccine roll-out in England has provided[25] protection that was not available to patients recruited to early studies, similarly, there is likely to be an increased level of protection associated with prior infection. Ideally there would be attempts to establish the impact of different circumstances on the observed clinical effectiveness of interventions in studies, although this was not possible within the timescales of the project. As such, the EAG had to make a simplistic assumption that none of the changes were treatment effect modifiers, and that given this, the relative benefits observed in the studies were transportable and could be applied to the estimated outcomes for patients with COVID-19 in England in Summer 2022.”

Each study assessed by the EAG in this analysis was conducted at different periods of the pandemic. Taking into consideration the continuous evolution of effect modifiers (SoC, SARS-CoV-2 variant, vaccination status, case mix, and prior infection) throughout the pandemic, we believe the analysis should have considered adjustment for these effect modifying factors.

Because these effect modifiers have not been accounted for, we believe that naïve comparisons being performed to assess the potential relative efficacy and subsequent incremental cost-effectiveness between interventions are misleading.

Contributing to the challenge of comparing treatments against each other for the purpose of conducting a fully incremental analysis, is the fact that placeholder values have been used to inform the treatment cost of key comparators. This invalidates any comparative ICERs and adds to our view that these ICERs should be removed.

Proposed amendment

Consider the adjustment of treatment efficacy by accounting for the impact of effect modifying factors present at the time of the relevant trials, before attempting any form of comparative efficacy or fully incremental costeffectiveness analysis.

Expected Impact on result

It is not possible to approximate the impact of this amendment on the results. If it is not possible to formally account for these treatment effect modifiers, then we suggest that conclusions regarding comparative efficacy and fully incremental cost-effectiveness analyses should be removed from the EAG report, and instead each treatment should be evaluated independently versus standard of care in the authorised patient population.

Please return to: NICE DOCS

4. Analytical perspective: Comparative Efficacy (section 2.1.1 pg.23)

Description of problem

“The need for rapid information on COVID-19 has resulted in a paradigm shift, especially in the communication of scientific results. Traditional systematic reviews can date quickly but ‘living’ systematic reviews search for evidence much more regularly than standard reviews and incorporate relevant new evidence as it becomes available. This is important in the context of COVID-19, in which the evidence-base is rapidly changing as new data emerge. The ability of a ‘living’ systematic review and network meta-analysis (NMA) to regularly update and incorporate relevant new evidence as it becomes available makes it the best type of evidence synthesis, in the opinion of the EAG, to inform this pragmatic rapid evaluation.”

Proposed amendment

We do not consider the study populations in the trials for the assessed treatments to be similar. A large category of patients identified as high-risk, as highlighted by the PANORAMIC study (PANORAMIC 2022) are not captured in the study population of oral antiviral trials.

A considerable number of high-risk factors were excluded from EPIC-HR(Medicine 2022), due to the adverse prognosis of patient outcomes. With this considered, we think that owing to the level of differences in the study populations and the heterogeneity of study methodology, it is inaccurate to imply which treatment is the most/least cost-effective, especially without an Indirect Treatment Comparison (ITC)/Network Meta Analysis (NMA) to inform comparative efficacy.

In the EAG analysis as it is, sotrovimab and other monoclonal antibody therapies are put at an unfair disadvantage due to the fact they had broader inclusion criteria, meaning the trial results are a truer reflection of their effectiveness in a high-risk group.

Expected Impact on result

An ITC/NMA could provide a more robust estimate of comparative efficacy, and therefore we believe a feasibility assessment should have been considered.

5. ICER threshold and ‘post-pandemic/endemic’ (section 1.1 pg.16)

Description of problem

“Should the risk of death following COVID-19 remain at low levels and SARS-CoV-2 becomes endemic in society, then treatments for patients with COVID-19 may no longer be treated differently to interventions for other conditions such as breast cancer or heart disease.”

Proposed amendment

Given cases, hospitalisations, and deaths are currently rising in the UK during the time of this consultation (15 July 2022), we would challenge the assumption that we are in a post-pandemic or endemic situation. There are currently minimal public health interventions aiming to control the spread of the virus, relative to earlier in the pandemic, and approximately three million people in England remain unvaccinated, and are at greater risk of becoming hospitalised or dying because of COVID-19 than if they were vaccinated (Commons and Accounts 4 July 2022). Therefore, treatments that prevent progression of COVID-19 have an important role to play in avoiding significant morbidity and mortality at a population level. We challenge the view that these treatments can be evaluated using the standard NICE decision-making framework and the assumptions relating to low rates of hospitalisation.

By using the lower end of the £20,000 to £30,000 per QALY (Quality Adjusted Life Year) threshold range to judge potential cost-effectiveness, the value of treatments in reducing transmission of the virus has not been accounted for within the EAG’s analysis. This is contrary to treatments for other infectious diseases that have been appraised by NICE, such as HCV ((NICE) 2015). As well as the potential benefit of reducing onward transmission, the availability of treatments with the capacity to avoid hospitalisations have additional value if the numbers of cases are high and bed/ICU capacity is being exceeded. In this situation, there are significant uncaptured benefits and non-health factors which are valid considerations to enable the Committee to apply a higher ICER threshold as part of their decision-making process.

Expected Impact on result

Please return to: NICE DOCS

While we do not expect the ICERs to change unless the model structure is revised to formally account for the potential benefit of reduced onward transmission, we believe that the Committee should consider applying a higher ICER threshold for determining if the treatments are cost-effective given the fact that this is an infectious disease (that can lead to significant morbidity and increased mortality, particularly in the immunocompromised populations where effective vaccination is not always possible), and that there are significant uncaptured benefits and non-health factors due to the nature of the COVID-19 pandemic.

6. Administration Cost (section 3.3.2 pg.50-51)

Description of problem

“It was assumed that the costs associated with treatment administration whilst in hospital would be incorporated in the unit costs associated with hospitalisation (see Section 3.3.3). Additional administration costs were assumed for intravenous treatment in the community, but for simplicity, not for oral or subcutaneous treatments. For each intravenous administration, a cost of £221 was incurred which was that of NHS reference code SB12Z. Within the analyses it has been assumed that there is likely to be a delay in patients receiving intravenous casirivimab/imdevimab and that subcutaneous version would be used instead.”

The EAG use the NHS reference code SB12Z for the cost of administration of intravenous treatments. This code is specifically for parenteral infusion of chemotherapy. As per the recommending dosing schedule as indicated in Table 2 on page 20 of the EAG report, and consistent with the sotrovimab SmPC ((MHRA) 2022), treatment with sotrovimab is administered intravenously over 30 minutes. We believe using the chemotherapy unit cost is likely to be an overestimate, given chemotherapy can often be associated with a longer infusion duration. Also, given that in-hospital treatment with sotrovimab is a possibility for a proportion of patients who acquire COVID-19 within hospital (approximately 10% of sotrovimab administration is in-hospital, based on the weekly COVID-19 therapeutics summary), these individuals may receive the infusion with sotrovimab as part of their bed stay cost (as is assumed for the in-hospital treatment population), and therefore the sotrovimab administration cost is likely an overestimate for this population.

Similarly, assuming all patients eligible for intravenous casirivimab/imdevimab would instead be completely switched to a subcutaneous version is unlikely to be reasonable, and subcutaneous administration is likely to be associated with some level of resource. The subcutaneous route for casirivimab/imdevimab involves higher needle burden and a monitoring period of at least one hour post administration impacting the resource use which should be captured in the administration cost.

Finally, the model does not include any administration or logistical resource use for shipping oral-antiviral therapies direct to patients. This therefore likely underestimates the true cost to the NHS for provision of these treatments.

Proposed amendment

We request that the EAG re-consider the assumptions and inputs related to administration costs for all therapies included within the analysis. In particular, considering the following amends:

• Using a more appropriate infusion administration NHS unit cost. For example, £173.01 (2019 prices) from TA676 ((NICE) 2021).

• Including an adjustment for infusion administration costs where treatment is provided to patients already hospitalised. For example, approximately 10% of sotrovimab is administered in-hospital (NHS-England checked 22/07/2022).

• Including a proportion of patients treated with casirivimab/imdevimab via intravenous administration

• Include an estimate of the cost to the NHS of shipping/ dispatching oral antivirals direct to patients’ homes.

Expected Impact on result

We anticipate relatively small improvements in the ICERs for sotrovimab versus SoC.

Please return to: NICE DOCS

7. Absence of safety profile (section 2.1.4 pg.24)

Description of problem

“As such, many interventions were associated with less SAEs (Serious Adverse Events) than SoC, which is generally atypical for efficacious pharmacological treatments. As the model was explicitly tracking the severity of patients through the use of the 8-point ordinal scale the EAG decided to omit SAEs from the model.”

Proposed amendment

According to their Summary of Product Characteristics, some of the interventions in this analysis, especially the oral antivirals like nirmatrelvir/ ritonavir (Care 09/02/2022 update), have the potential to cause clinically serious adverse events(Hammond et al. 2022), the incidence of these adverse events could be higher when taken by some comorbid patients in the high-risk group.

For example, the COVID-19 rapid guideline ([NG191] 14 July 2022 Update), states the following: “Ritonavir is a potent CYP3A inhibitor and has interactions with many other medicines, some of which may lead to severe, lifethreatening or fatal events”. A full medication review (including over-the-counter and herbal medicines) is needed before prescribing nirmatrelvir and ritonavir (Care 09/02/2022 update).

Excluding the consideration of adverse events and the risk of drug-drug interactions in the clinical and cost effectiveness analysis will result in an inaccurate estimation of the clinical safety and potential cost-effectiveness of these treatments. We request that this assumption is revisited, and the safety implications of these therapies are included either formally or in a qualitative way, to ensure that any resulting guidance is appropriate for the high-risk and comorbid patient population.

Expected Impact on result

We expect that the inclusion of adverse events would be a more accurate reflection of the overall impact of the therapeutics.

8. Incorrect relative risks for sotrovimab

Description of problem

We would like to call the EAG’s attention to the disparity in the COMET-ICE trial results and those presented by the EAG in Figures 6-7 and Table 5 and subsequently used in the analysis and the impact of this error on the analysis. The EAG’s report stated:

“All treatments used in the community had favourable median RRs for hospitalisation and death at 28 days, although due to wide CIs no firm conclusions could be made regarding the relative efficacy of these treatments. The median RR associated with death at 28 days were favourable for all interventions, except for remdesivir where the median estimate was unity. The CIs were wide and spanned 1 for all treatments except for molnupiravir and nirmatrelvir/ ritonavir. As such, no clear conclusions relating to the relative efficacy of the interventions could be made regarding avoiding death at 28 days.”

In COMET-ICE and as reported in (Gupta et al. 2021), the adjusted relative risk for all-cause hospitalisation lasting >24 hours for acute illness management or death due to any cause through 29 days was 0.21. A corresponding relative risk of 0.20 for all-cause mortality can be derived (0/528 from sotrovimab group, 2/529 from placebo group). The EAG model wrongly uses a median relative risk of 0.65 derived from the confidence intervals, instead of the mean relative risk in its baseline analysis, contrary to best practice. The resulting impact is an inflated estimate of mortality in the sotrovimab treated population, with 75% of patients dying if hospitalised on sotrovimab, compared with 25% of patients dying if standard of care was received. This lacks clinical and face validity and is inconsistent with the outcomes of the clinical study.

Proposed amendment

We recommend that the base case analysis be re-run using the correct clinical data.

Expected Impact on result

An improvement in the ICER for sotrovimab with event outcomes including mortality that are consistent with the clinical study.

Please return to: NICE DOCS

Checklist for submitting comments

• Consultation responses must not be longer than 10 pages , excluding the references.

• Use this comment form and submit it as a Word document (not a PDF).

• • Complete the disclosure about links with, or funding from, the tobacco industry. • Combine all comments from your organisation into 1 response. We cannot accept more than 1 set of comments from each organisation.

• Do not paste other tables into this table – type directly into the table.

• Please underline all confidential information, and separately highlight information that is submitted under ‘commercial in confidence’ in turquoise, all information submitted under ‘academic in confidence’ in yellow. If confidential information is submitted, please also send a second version of your comments with that information replaced with the following text: ‘academic/commercial in confidence information removed’. See the Process and methods manual (Section 5.4) for more information.

• Do not include medical information about yourself or another person from which you or the person could be identified.

• • Do not use abbreviations

• Do not include attachments such as research articles, letters or leaflets. For copyright reasons, we will have to return comments forms that have attachments without reading them. You can resubmit your comments form without attachments, it must send it by the deadline.

• If you have received agreement from NICE to submit additional evidence with your comments on the pro-forma response document, please submit these separately.

Note: We reserve the right to summarise and edit comments received during consultations, or not to publish them at all, if we consider the comments are too long, or publication would be unlawful or otherwise inappropriate.

Comments received during our consultations are published in the interests of openness and transparency, and to promote understanding of how recommendations are developed. The comments are published as a record of the comments we received, and are not endorsed by NICE, its officers or advisory committees.

Please return to: NICE DOCS

References

• (MHRA), Medicines & Healthcare products Regulatory Agency. 2022. 'Summary of Product Characteristics for Xevudy'.

• (NICE), National Institute for Health and Care Excellence. 2015. 'Technology appraisal guidance [TA363]. Ledipasvir–sofosbuvir for treating chronic hepatitis C'.

• ———. 2021. 'Technology appraisal guidance [TA676]: Filgotinib for treating moderate to severe rheumatoid arthritis'.

• [NG191], NICE guideline. 14 July 2022 Update. 'COVID-19 rapid guideline: managing COVID-19. '. Alastair McLellan, Kamran Abbasi. 2022. 'The NHS is not living with covid, it’s dying from it'.

• Care, Department of Health and Social. 09/02/2022 update. 'Summary of Product Characteristics for Paxlovid'.

• Commons, House of, and Committee of Public Accounts. 4 July 2022. 'The rollout of the COVID-19 vaccine programme in England'.

• DHSC, Department of Health & Social Care. 2022. 'Independent report: Defining the highest-risk clinical subgroups upon community infection with SARS-CoV-2 when considering the use of neutralising monoclonal antibodies (nMABs) and antiviral drugs: independent advisory group report', MHRA, Accessed 26/07/2022.

• Gupta, A., Y. Gonzalez-Rojas, E. Juarez, M. Crespo Casal, J. Moya, D. R. Falci, E. Sarkis, J. Solis, H. Zheng, N. Scott, A. L. Cathcart, C. M. Hebner, J. Sager, E. Mogalian, C. Tipple, A. Peppercorn, E. Alexander, P. S. Pang, A. Free, C. Brinson, M. Aldinger, and A. E. Shapiro. 2021. 'Early Treatment for Covid-19 with SARS-CoV-2 Neutralizing Antibody Sotrovimab', N Engl J Med , 385: 1941-50.

• Hammond, Jennifer, Heidi Leister-Tebbe, Annie Gardner, Paula Abreu, Weihang Bao, Wayne Wisemandle, MaryLynn Baniecki, Victoria M. Hendrick, Bharat Damle, Abraham Simón-Campos, Rienk Pypstra, and James M. Rusnak. 2022. 'Oral Nirmatrelvir for High-Risk, Nonhospitalized Adults with Covid-19', 386: 1397-408.

• Hippisley-Cox, Julia, Carol AC Coupland, Nisha Mehta, Ruth H Keogh, Karla Diaz-Ordaz, Kamlesh Khunti, Ronan A Lyons, Frank Kee, Aziz Sheikh, Shamim Rahman, Jonathan Valabhji, Ewen M Harrison, Peter Sellen, Nazmus Haq, Malcolm G Semple, Peter W M Johnson, Andrew Hayward, and Jonathan S Nguyen-Van-Tam. 2021. 'Risk prediction of covid-19 related death and hospital admission in adults after covid-19 vaccination: national prospective cohort study', 374: n2244.

• Medicine, U.S. National Library of. 2022. 'EPIC-HR: Study of Oral PF-07321332/Ritonavir Compared With Placebo in Nonhospitalized High Risk Adults With COVID-19'.

• NHS-England. checked 22/07/2022. 'COVID-19 Therapeutics (antivirals, neutralising monoclonal antibodies and interleukin 6 inhibitors)'.

• Nyberg, T., N. M. Ferguson, S. G. Nash, H. H. Webster, S. Flaxman, N. Andrews, W. Hinsley, J. L. Bernal, M. Kall, S. Bhatt, P. Blomquist, A. Zaidi, E. Volz, N. A. Aziz, K. Harman, S. Funk, S. Abbott, R. Hope, A. Charlett, M. Chand, A. C. Ghani, S. R. Seaman, G. Dabrera, D. De Angelis, A. M. Presanis, and S. Thelwall. 2022. 'Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study', Lancet , 399: 1303-12.

ONS, Office for National Statistics. 2022. ' Hospital admissions by age'.

PANORAMIC. 2022. 'Participant Information — PANORAMIC.'.

• Shields, A. M., S. O. Burns, S. Savic, and A. G. Richter. 2021. 'COVID-19 in patients with primary and secondary immunodeficiency: The United Kingdom experience', J Allergy Clin Immunol , 147: 870-75.e1.

UKHSA, UK Health Security Agency. 2022. 'COVID-19 (SARS-CoV-2) variants'.

Please return to: NICE DOCS

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Executable Model

Therapeutics for people with COVID-19 [ID4038]

The economic model enclosed and its contents are confidential and are protected by intellectual property rights, which are owned by ScHARR . It has been sent to you for information only. It cannot be used for any other purpose than to inform your understanding of the appraisal. Accordingly, neither the model nor its contents should be divulged to anyone other than those individuals within your organisation who need to see to them to enable you to prepare your response. Those to whom you do show the documents must be advised they are bound by the terms of the Confidentiality Agreement Form that has already been signed and returned to the Institute by your organisation.

You may not make copies of the file and you must delete the file from your records when the appraisal process, and any possible appeal, are complete. If asked, you must confirm to us in writing that you have done so. You may not publish it in whole or part, or use it to inform the development of other economic models.

The model must not be re-run for purposes other that the testing of its reliability .

Please set out your comments on reliability in writing providing separate justification, with supporting information, for each specific comment made. Where you have made an alteration to the model details of how this alteration was implemented in the model (e.g. in terms of programme code) must be given in sufficient detail to enable your changes to be replicated from the information provided. Please use the attached pro-forma to present your response.

Please prepare your response carefully. Responses which contain errors or are internally inconsistent (for example where we are unable to replicate the results claimed by implementing the changes said to have been made to the model) will be rejected without further consideration.

Results from amended versions of the model will only be accepted if their purpose is to test robustness and reliability of the economic model. Results calculated purely for the purpose of using alternative inputs will not be accepted.

No electronic versions of the economic model will be accepted with your response.

Responses should be provided in tabular format as suggested below (please add further tables if necessary).

Issue 1 Incorrect RR for mortality for sotrovimab

==> picture [345 x 227] intentionally omitted <==

Issue 2 Error in Formula used to capture the impact of Long Covid on QALY

Issue 3 Lack of Model Validation

Confidential

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Assessment Report

Therapeutics for people with COVID-19 [ID4038]

Please read the checklist for submitting comments at the end of this form. We cannot accept forms that are not filled in correctly. Deadline for comments 5pm on Friday 29 July . Please submit via NICE Docs .

1

Confidential

Please comment on the assessment report

The Company would like to thank NICE and the EAG for the opportunity to review and respond to the assessment report. Given the challenges in the data, we considered the report a fair representation of what is currently known about COVID-19 therapeutics. However, we consider it important to highlight that treatment of COVID-19 is an area of intense scientific development, as evidenced by the rapidly evolving clinical evidence base, which could impact on the relevance of the conclusions drawn from the multiple technology appraisal (MTA). Having reviewed the Evidence Assessment Group (EAG) report, we consider it important to expand on points raised by the EAG in the following areas:

• Inappropriateness of the ranking of treatments by cost effectiveness;

• Lack of exploration of potential impact of drug–drug interactions on cost effectiveness;

• Lack of evaluation of the logistics of use of community-based treatments on cost effectiveness;

• Additional considerations for health-related quality-of-life;

• Lack of consideration of the direct cost to the NHS of absences in the NHS and social care workforce due to COVID-19;

• Additional considerations for a societal perspective;

• Relevance of the findings to the current landscape of treatment for COVID-19.

The proposed ranking of COVID-19 interventions based on the estimated costeffectiveness is flawed given the high level of uncertainty in the relative efficacy of treatments.

On page 9, the EAG report states:

“Given the timescale of the project, where there was less than three months between the publication of the final scope and the report deadline, a literature review following best practice was not possible. Instead, a pragmatic, alternative approach was undertaken where evidence was taken from two living systematic reviews (supported by the COVIDNMA initiative and the metaEvidence initiative).”

The indirect treatment comparison assumes transportability of relative treatment effects. The EAG note that this means that “the same treatment effects, either hazard ratios (HRs) or relative risks (RRs), were assumed applicable regardless of study characteristics which include: the age, perceived severity, vaccination status, and history of SARS-CoV-2 infection of patients; the SoC at that time; the geographical location; and the dosage of the intervention used. ”

This assumption cannot be justified due to the differences in study design and settings, which limits the generalisability of outcomes. A specific example is in comparison of the outcomes observed in studies of nirmatrelvir/ritonavir and molnupiravir where the trial populations were quite different. Specifically, regarding co-morbidities and their concomitant medication, patients taking certain concomitant medications were not allowed in the nirmatrelvir/ritonavir

2

Confidential

study; therefore, the patient population treated were on average less medically complicated than the MOVe-OUT study. Indeed, the EAG acknowledge the weakness of the assumption of transportability by stating “it is acknowledged that this assumption may be incorrect” .

Given the differences across studies, it would be expected that adjustments would be made for time-related factors, such as vaccination rates and variant type. Additionally, adjustments would be expected to account for variation across studies in baseline antibody levels (reflecting prior exposure to SARS-CoV-2 and thus likely some degree of innate protection against COVID-19) and other aspects of population heterogeneity. These factors would have contributed to the observed differences in the event rates of study outcomes in placebo arms. For example, a UK study reported that race, social background, gender, age, and presence of severe asthma are all key risk factors associated with death from COVID-19.[1] Since these differences between studies were not adjusted for in the network meta-analysis (NMA), the company considers it inappropriate to assume transportability of the results on comparative clinical effectiveness derived from the NMA. This is demonstrated in Figure 9 of the EAG report as there is considerable uncertainty in the ranking of efficacy amongst interventions.

Having reviewed the studies informing the NMA, we have concluded that it is not possible to mitigate the study differences to derive robust estimates of relative treatment effects between interventions. For example, whilst a matched adjusted indirect comparison would likely capture unobserved prognostic factors and effect modifiers distributed differently between trials, it would remain impossible to adjust for time-related factors such as variant type. Therefore, we posit that it is inappropriate to formally rank interventions for the purposes of the MTA, and instead the focus should be on head-to-head comparisons versus standard of care (SoC). We are also aware that data from use of antivirals in the community setting are currently being collected. We consider that even these may not resolve some of the clinical uncertainties due to the evolution of the virus over time.

Therefore, the company consider that robust conclusions cannot be drawn on relative clinical effectiveness between treatments (recognised by the EAG in page 3) and ask that only head-to-head comparisons versus SoC are considered for decision making purposes.

Indeed, despite the EAG concluding that “nirmatrelvir with ritonavir (at an estimated price) may be the most cost-effective treatment in the community setting” (on page 3 of the report), this statement was followed by another statement highlighting the considerable levels of uncertainty in the incremental analysis. Uncertainty relating to relative treatment effects that inform pairwise comparisons of cost-effectiveness was mentioned on pages 12, 13, 24, 27, 30, 32, and 70 of the report. As such, it is not appropriate to conclude one therapy as “the most cost-effective treatment” based on the network meta-analysis. The limitations associated with such analyses should be elaborated under each table presenting such results to provide context.

3

Confidential

Molnupiravir has fewer drug-drug interactions than its comparators and the full implications of this are not explored in the EAG assessment.

Molnupiravir has reduced complications compared to treatments with proven drugdrug interactions

COVID-19 treatment is targeted at a high-risk population, which the PANORAMIC study defines as adults with at least one of the following medical conditions: chronic respiratory disease; chronic heart or vascular disease; chronic kidney disease; chronic liver disease; chronic neurological disease; severe and profound learning disability; Down’s syndrome; diabetes; immunosuppression (either caused by inherited immune disorders, or caused by disease or treatment); solid organ, bone marrow and stem cell transplant recipient; morbid obesity; severe mental illness; care home resident; or, judged to be clinically vulnerable.[2] Many of the medical conditions denoting high-risk status require pharmacological management. Therefore, patients eligible for treatment in the community considered in this appraisal are highly likely to be already receiving pharmacological treatments.

As there are no known clinically meaningful drug-drug interactions associated with the use of molnupiravir, it is unlikely to cause an adverse reaction when administered concomitantly with other medications.[3] An additional benefit of molnupiravir is that dose adjustment is not required for treatment administration or when used concomitantly with medications for comorbidities.

Conversely, other treatments are associated with several drug-drug interactions, including interactions with anticoagulants, anticonvulsants and antiarrhythmics,[4,5] which are common treatments for the comorbid conditions defining high-risk patients. These drug-drug interactions may prevent or complicate the use of treatments other than molnupiravir in a substantial proportion of patients at high-risk of COVID-19 and such implications have not been captured within the model.[6] Drug-drug interactions complicate the ability for pharmacists to easily prescribe additional medication; they must perform a full medication review, which is time consuming. Following medication review, a dose adjustment may be required, which complicates treatment for the patient, increasing the likelihood of dosing errors. These complications of drug-drug interactions come with unquantifiable costs for both the NHS and the patient. The extensive diagnosis and treatment pathway is demonstrated through the NHS England published assessment pathways and guides.[5,7] These elements are of particular importance in the UK, where the aging population is likely to have been diagnosed with comorbidities requiring polypharmacy.[8] If patients in the UK can be treated within the community with molnupiravir, an easy to administer drug with no known drug-drug interactions, then considerable time and resource use is saved compared with the treatment of other community drugs for high-risk patients with COVID-19.

Moreover, some patients may be ineligible for some COVID-19 treatments, such as nirmatrelvir/ritonavir, due to concomitant medication for comorbidities. Patients ineligible for a COVID-19 treatment because of an ongoing treatment regimen are more likely to have a worse prognosis than those treated with molnupiravir because they would need to pause their concomitant treatment to treat their COVID-19 with the prescribed intervention. As

4

Confidential

molnupiravir has no clinically meaningful drug-drug interactions, these patients can be treated for COVID-19 without interruption to treatment for comorbidities.

Since treatment with molnupiravir does not require dose adjustment or a pause of any concomitant medications, the likelihood of exacerbating comorbidities is reduced compared to treatments where there are drug-drug interactions. Moreover, there is not the chance of side effects caused by mismanaged drug-drug interactions. Exacerbation of underlying comorbid conditions as a result of suspending treatment and adverse effects arising from drug-drug interactions can be costly to manage and can negatively impact quality of life. Therefore, the costs avoided and the QALYs gained are larger with molnupiravir than products with drugdrug interactions.

Reduced pharmacy costs compared to treatments with proven drug-drug interactions Pharmacy prescription costs are not accurately captured within the model. Choosing an appropriate treatment for patients with comorbidities is complicated and requires careful consideration of their other medication. In particular, ritonavir (in the nirmatrelvir and ritonavir combination) is a potent CYP38 inhibitor and interactions with other medicines may lead to severe, life-threatening or fatal events.[9] This is especially true for patients age 65 and above, which accounts for a larger proportion of the UK population compared to other markets.[8,10,11] As such, a full medication review is required before prescribing nirmatrelvir and ritonavir, which will delay the administration of treatment.[9] As molnupiravir is simple to administer and there are no known drug-drug interactions with the use of the drug, pharmacy costs will be lower for patients receiving molnupiravir compared to other treatments. Additionally, patients will be able to initiate treatment faster, leading to a much simpler process for the patient.[12] Additional evidence is required detailing the most common comorbidities in patients with COVID-19 and a high-risk of hospitalisation.

The impact of drug-drug interactions should be captured within the model Finally, the Company believes that both costs and utilities associated with unmanaged drugdrug interactions should be included in the economic model. Unmanaged drug-drug interactions may result in adverse events, hospitalisation, prolonged time to recovery, and death. The NHS operates within fixed budgets for both labour and capital resources and managing the drug-drug interactions is a direct cost to the healthcare system.[13] Additionally, adverse events and increased patient management are likely to have a negative impact on patients’ utility. If a utility decrement was applied in the model to capture the impact of drugdrug interactions, molnupiravir will show a relatively greater benefit to the patients and the clinical and cost-effectiveness will be more balanced.

Therefore, additional prescribing requirement, such as pharmacist assessment time for community antivirals, should be captured for the purposes of decision making.

Omitting assessment of the logistics and resource use for community-based treatments from the economic evaluation disregards key benefits of molnupiravir.

On pages 10 and 40 the EAG report states:

“The modelling did not assess the logistics of treatment in the community, but the EAG notes that this could be a large factor in deciding which treatments could be preferred,

5

Confidential

as oral treatments could be more acceptable to patients and healthcare systems than treatments that are given intravenously or subcutaneously.”

Molnupiravir is easy to prescribe in the outpatient setting. It is not necessary for patients or clinicians to check a long list of concomitant medications before prescribing molnupiravir, and dose adjustments of concomitant medications are not required, both of which minimise the risk of pill confusion among patients taking molnupiravir.[14] Additionally, molnupiravir is given as a single oral tablet, which further reduces the risk of pill confusion.

Reduced GP and pharmacist time

Molnupiravir is an orally administered, outpatient treatment without any known drug-drug interactions.[15] Additionally, molnupiravir’s simple administration means that infectious patients are kept out of the hospital, thereby reducing the likelihood of community transmission of COVID-19.

Compared with molnupiravir, many of the other treatments for COVID-19 included within the economic model do not have a simple administration and do have clinically meaningful drugdrug interactions.[16] Dosing errors occurring from incorrect patient self-administration, incorrect electronic prescribing, incorrect handling within pharmacy and failure to take multiple tablets together when self-administering (for COVID-19 and comborbidities)[14,17] are all issues not currently captured within the economic model that would not be associated with use of molnupiravir.

Furthermore, the economic model does not capture the pharmacological assessment costs associated with community drugs, for example, renal and hepatic laboratory tests associated with the administration of nirmatrelvir/ritonavir are not considered. Moreover, IV administration in the community was assumed to be the same for all community IV treatments, which the company considers does not reflect clinical practice in the UK; remdesivir in particular requires an observation period following administration, which has been omitted from the model.[18] Further consideration is needed for the full costs associated with the administration of IV treatments for COVID-19 in the community. This results in an unbalanced assessment of the costs associated with treatments in the community and introduces uncertainty within the cost-effectiveness analysis.

Benefits of treatment in the community

Patients receiving molnupiravir may be suffering burdensome symptoms from their COVID-19 infection, including fever, body aches, diarrhoea, nausea, fatigue and exhaustion.[19] As such, community treatment is likely to be favoured by patients to reduce disruption to their lives.[20] Similarly, patients with mild symptoms of COVID-19 will also prefer community treatment as they can remain at home with limited disruption to their lives. As an orally administered community treatment, molnupiravir enables patients to rest and recover in the comfort of their own homes, rather than travelling to the hospital for IV treatment. The ability to rest and recover at home would bring additional quality of life benefits to patients with COVID-19.

Reduced requirement for hospital administration

The proven effectiveness of molnupiravir in a community setting allows for a variety of patients to be treated at home, including those with complex additional needs, such as the elderly and those with disabilities who may otherwise struggle to receive treatment due to difficulties getting to the hospital.[9] In addition, these patient groups are more likely to have

6

Confidential

comorbidities and experience drug-drug interactions. Molnupiravir has no known drug-drug interactions and a demonstrated manageable safety profile as an outpatient treatment option.[3]

Removal of geographic barriers to treatment

For those with geographic barriers, if an oral treatment option that requires limited pharmacist assessment for potential drug-drug interactions is not available, transport will be required for these patients, which results in additional costs to the NHS and/or families and increases risk of COVID-19 onward transmission to NHS Ambulatory staff.[9] While the company considers that it is not possible to fully capture the risk of transmission of COVID-19 to ambulance staff within the economic model, its omission means that the cost-effectiveness is overestimated for the drugs the use of which would be potentially associated with this risk.

Overall, omitting the logistics of treatment in the community from the economic evaluation, such as consequences of drug-drug interaction in patients with comorbidities, and the increased risk of COVID-19 transmission in the transporting of patients receiving community drug that require further pharmacological assessment, disregards a key benefit of molnupiravir compared to other available COVID-19 treatments. Therefore, the costeffectiveness of molnupiravir relative to SoC is underestimated.

Alternative methods to parametrise QALYs are available and are more robust than using values derived from proxy conditions.

Page 53 of the EAG report states that:

“Due to the nature of this rapid assessment, no formal systematic review of the literature was conducted to identify the most appropriate utility values.”

The utility decrements used in the model were sourced from a paper by Rafia et al. which estimated the cost-effectiveness of remdesivir.[21] The authors sourced these values for the study from proxy conditions (recurrent Clostridium difficile infection and influenza) instead of the values being derived from patients with COVID-19. As such, uncertainty exists as to the generalisability of utilities for proxy conditions to patients with COVID-19.

A formal systematic review of the literature would have ensured that all clinically relevant data had been reviewed and assessed, and that the utility values used in the model were obtained from the most relevant data sources. An alternative method to quantify HRQoL in the model would be to conduct a utility study by developing a series of vignettes to describe the range of health states that characterise different levels of severity of COVID-19. The general public would then complete the EQ-5D-3L, a generic, preference-based HRQoL tool, for these health states, acting as proxies on behalf of patients. In the absence of utility values identified from a systematic literature review, the vignette approach would be more robust than assuming alternative diseases are a suitable proxy for COVID-19, aligning with the hierarchy of preferred HRQoL evidence outlined in the NICE health technology evaluations manual (2022).[22]

HRQoL benefits of patients in the community are underestimated

The economic model does not currently distinguish patients who are not hospitalised and experience no limitations of activities (health state 1) from patients who are not hospitalised and who do still experience limitations of activities, home oxygen requirements or both (health

7

Confidential

state 2). Therefore, a quality-adjusted life-year (QALY) decrement is not applied to patients who receive treatment in the community setting and go on to experience limitations of activities or require oxygen at home. Additionally, in patients who are hospitalised, a QALY decrement is not applied to patients who have been discharged from hospital but are still experiencing limitations of activities or require oxygen at home.

This does not capture the full health benefit of molnupiravir treatment. For most COVID-19 signs and symptoms, sustained improvement or resolution was more likely, and worsening progression of signs or symptoms was less likely, in the molnupiravir group than in the placebo group.[15] Therefore, lack of distinction between health state 1 and health state 2 eliminates the ability to characterise the full benefit of molnupiravir relative to SoC.

The therapeutics considered in the assessment were centrally procured as part of the Government’s response to the COVID-19 pandemic. Therefore, the list price of some COVID-19 therapeutic agents included in the EAG report are currently unknown and robust cost-effectiveness conclusions cannot be drawn. Page 50 of the EAG report states that:

“…list price for remdesivir, tocilizumab, baricitinib, lenzilumab, and sotrovimab. However, list prices were not available for molnupiravir, casirivimab/imdevimab, and nirmatrelvir/ritonavir. NICE requested that placeholder prices be used which were estimated from an Institute for Clinical & Economic Review report (dated March 2022) for molnupiravir and nirmatrelvir/ritonavir, and that the price for sotrovimab was used for casirivimab/imdevimab.”

Currently, the economic model uses indicative prices for some interventions including molnupiravir and nirmatrelvir/ritonavir. As noted within the assessment report, these were extracted from previous work conducted by the Institute for Clinical & Economic Review (ICER) report, which sourced inputs from US sources to inform the cost-effectiveness of alternative interventions in the US setting.

The UK list prices are not currently available for molnupiravir and nirmatrelvir/ritonavir.[23,24] In the absence of UK list prices, the Company is concerned that proxies from the ICER report may not be applicable. Additional generalisability issues arise as current treatment acquisition costs (list prices), used within the EAG report may include USA Government volume discounts that may be unavailable in other countries.[25]

The Company is aware that alternative list price proxies from other markets are available.[26,27] The Company also notes that stock of some antivirals were bought by the DHSC as part of the Government’s response to the COVID-19 pandemic. To ensure efficient use of stock already acquired and to reflect value already in the system, this ‘sunk cost’ stock should be factored into the economic model at £0 until this stock is exhausted.

The direct cost to the NHS of sickness amongst the NHS workforce due to COVID-19 exposure should be considered

Front-line healthcare workers are at a higher risk of contracting COVID-19 than the general public due to the patient-facing nature of the role. Staff sickness among health care providers will result in significant direct costs to the health service due to staff absenteeism and

8

Confidential

significant indirect costs from delayed or cancelled treatments. Preventing hospitalisation in high-risk patients with COVID-19 will reduce the transmission of COVID-19 to front-line healthcare workers, which would consequently result in the reduction of the costs associated with covering staff absences and delayed or cancelled treatments. These costs are borne by the NHS and so should inherently be included in the cost-effectiveness analysis as part of the NHS and PSS perspective.

Moreover, preventing hospitalisation in patients with COVID-19 reduces the risk of transmission to the front-line healthcare workers. As a treatment that is delivered entirely in the community, molnupiravir can reduce the exposure of the NHS workforce to COVID-19. Moreover, the phase II-III trial, MOVe-OUT,[15] found that molnupiravir reduced the risk of hospitalisation in at-risk unvaccinated adults compared to placebo.[15] Therefore, when patients are treated with molnupiravir in the community, front-line healthcare workers are at a reduced risk of contracting COVID-19. This key benefit of molnupiravir has been excluded from the economic model.

The EAG should adopt or present an analysis from a societal viewpoint to account for the impact COVID-19 treatments have on productivity, absenteeism, indirect costs, and utilities despite the deviation from NICE evaluation methods considering the wider implications COVID-19 infection.

The economic evaluation adopts a perspective which considers ‘ how the estimated efficacy for interventions used in hospital and for those at high-risk in the community impacted on patient health’ . Only direct costs to the health care system are included in the economic model: drug acquisition costs, administration costs and unit costs associated with hospitalisation. The model does not include indirect costs such as loss of productivity due to time off work.

COVID-19 increases the risk of sickness absence, both for physical and mental health, in the general workforce.[28–30] Furthermore, patients that contract COVID-19 may suffer from longterm sequalae, where the signs and symptoms of COVID-19 continue well beyond the first suspected COVID-19 infection, which can have profound effects for productivity.[31] Patients who contract COVID-19, whether they require hospitalisation or remain in the community, are required to take time off work to recover and prevent spreading the infection to co-workers. This results in absenteeism and considerable productivity losses at work which the current EAG model and report do not consider. The long-term costs and effects of COVID-19 are further heightened when considering long COVID as individuals experience extended periods off from work in recovery from poor health.

The phase II-III trial, MOVe-OUT,[15] found that progression of signs and symptoms of COVID19 was less likely in the molnupiravir group than in the placebo group.[15] As such, patients treated with molnupiravir are less likely to suffer long-term sequalae of severe COVID-19, which will enable them to re-join the workforce more quickly and increase their overall productivity.

Adopting a societal perspective which considers all direct and indirect costs and effects attributed to the intervention, would optimise the decision-making process. This would result in a more robust analysis, demonstrating the broader cost-effectiveness of interventions.

9

Confidential

Furthermore, the NICE manual 2022 states that under exceptional circumstances, the scope may adopt a broader perspective for costs.[22]

Even if the existing clinical data gaps are filled, the everchanging treatment landscape is a barrier to collecting and providing relevant data.

As highlighted in the EAG report, the evidence-base for COVID-19 is rapidly changing as new data emerge. This is paired with an everchanging environment in which the SoC, the percentage of people who have had a vaccination, and the dominant SARS-CoV-2 variant are variable. As such, this may limit the relevance of any conclusions made by the Appraisal Committee in the future. The Company want to ensure that the guidance issued by the Appraisal Committee remains appropriate and relevant to potential changes in the future.

The effects of molnupiravir on dominant SARS-CoV2 variants

Regarding variability in the dominant SARS-CoV-2 variant, molnupiravir has shown therapeutic value against the sublineages BA.2.12.1, BA.4, and BA.5 of SARS-CoV-2 omicron variant in an in vitro study.[32] Using molnupiravir, IC50 was higher by a factor of 1.1 in the BA.2.12.1 subvariant, 1.2 in the BA.4 subvariant and 1.5 in the BA.5 subvariant.[32]

The effects of molnupiravir on vaccinated patients

Molnupiravir has also shown meaningful benefits when administered to fully vaccinated patients with COVID-19. Page 25 of the EAG report states that:

“ The vaccine roll-out in England has provided protection that was not available to patients recruited to early studies, similarly, there is likely to be an increased level of protection associated with prior infection. Ideally there would be attempts to establish the impact of different circumstances on the observed clinical effectiveness of interventions in studies, although this was not possible within the timescales of the project. As such, the EAG had to make a simplistic assumption that none of the changes were treatment effect modifiers, and that given this, the relative benefits observed in the studies were transportable and could be applied to the estimated outcomes for patients with COVID19 in England in Summer 2022. ”

A recent study carried out in Italy has demonstrated meaningful benefits of molnupiravir within a fully vaccinated population. The early clinical experience demonstrated that out of 145 fully vaccinated patients enrolled with a mild to moderate breakthrough infection of COVID-19 treated with molnupiravir between January 2022 and February 2022, only 4 patients required hopsitalisation (2.7%) at day 30, no patient developed severe COVID-19, no patient was admitted to the ICU, and no patient died during the follow-up period.[33] Despite no comparator arm, the study shows that molnupiravir continues to provide a viable treatment option in a vaccinated population for high-risk patients when administered within the first 5 days of symptoms onset.[33] Similarly, a UK study found that, in a cohort of vaccinated and unvaccinated patients, patients with COVID-19 who were treated with molnupiravir returned a negative polymerase chain reaction (PCR) test three days faster than those treated with placebo; similar efficacy was observed in the vaccinated and unvaccinated cohort.[34]

The presented clinical data demonstrate that molnupiravir is expected to remain effective regardless of the vaccination status or variant of COVID-19, with the evidence base

10

Confidential

suggesting that it remains more effective than SoC. This evidence base will continue to mature over time.

11

Confidential

References

1. Risk factors for COVID-19 death revealed in world’s largest analysis of patient records to date. LSHTM https://www.lshtm.ac.uk/newsevents/news/2020/risk-factors-covid-19-death-revealedworlds-largest-analysis-patient-records.

2. PANORAMIC Active GP Sites. Google My Maps

https://www.google.com/maps/d/viewer?mid=1fV1C91Aj4XtRUg1jwPL0oMlDUM8br6mJ.

3. Food and Drug Administration. Center for Drug Evaluation and Research Antimicrobial Drugs Advisory Committee Meeting Briefing Document Molnupiravir Oral Treatment of COVID-19. (2021).

4. University of Liverpool. COVID-19 drug interactions. https://www.covid19druginteractions.org/checker (2022).

5. NHS England. UK Interim Clinical Commissioning Policy: Therapies for symptomatic nonhospitalised patients with COVID-19. (2021).

6. National Institute of Health. The COVID-19 Treatment Guidelines Panel’s Statement on Potential Drug-Drug Interactions Between Ritonavir-Boosted Nirmatrelvir (Paxlovid) and Concomitant Medications. (2021).

7. NHS England. Assessment, monitoring and management of symptomatic COVID-19 patients in the community. (2022).

8. The World Bank. Population ages 65 and above (% of total population) - United Kingdom | Data. https://data.worldbank.org/indicator/SP.POP.65UP.TO.ZS?locations=GB.

9. National Institute for Health and Care excellence. COVID-19 rapid guideline: Managing COVID-19. (2022).

10. Puenpatom, A. et al. Prevalence of potential drug-drug interactions with ritonavir-containing COVID-19 therapies. (2022).

11. Assessing the proportion of the Danish population at risk of clinically significant drug-drug

• interactions with new oral antivirals for early treatment of COVID-19 | Elsevier Enhanced Reader. https://reader.elsevier.com/reader/sd/pii/S1201971222003915?token=29E0FA20CD97565A17A01

12

Confidential

• 1C2C69FB2DA4CF949B41CABF5B7CBEB753AC6EBA8E569DC0A0E96AA5C6D2DCECC7172 37AA95&originRegion=eu-west-1&originCreation=20220728122432 doi:10.1016/j.ijid.2022.06.059.

12. Giammaria, D. & Pajewski, A. Can early treatment of patients with risk factors contribute to managing the COVID-19 pandemic? J Glob Health 10 , 010377–010377 (2020).

13. Awortwe, C. & Cascorbi, I. Meta-analysis on outcome-worsening comorbidities of COVID-19 and related potential drug-drug interactions. Pharmacol Res 161 , 105250–105250 (2020).

14. Dosing errors common with Paxlovid. Reactions Weekly 1916 , 3–3 (2022).

15. Jayk Bernal, A. et al. Molnupiravir for Oral Treatment of Covid-19 in Nonhospitalized Patients. N Engl J Med 386 , 509–520 (2022).

16. US Food and Drug Administration. PAXLOVID Patient Eligibility Screening Checklist Tool for Prescribers 07182022. US Food and Drug Administration 7 (2022).

17. Medication Safety Issues with Newly Authorized PAXLOVID. Institute For Safe Medication Practices https://www.ismp.org/alerts/medication-safety-issues-newly-authorized-paxlovid (2022).

18. EMA. Remdesivir. Summary of Product Characteristics.

19. NHS. Coronavirus (COVID-19) symptoms in adults. https://www.nhs.uk/conditions/coronaviruscovid-19/symptoms/main-symptoms/ (2022).

20. COVID-19 medicines delivery unit (CMDU) - Overview. Guy’s and St Thomas’ NHS Foundation Trust https://www.guysandstthomas.nhs.uk/our-services/covid-19-medicines-delivery-unit-cmdu.

21. Rafia, R. et al. A Cost-Effectiveness Analysis of Remdesivir for the Treatment of Hospitalized Patients With COVID-19 in England and Wales. Value Health 25 , 761–769 (2022).

22. NICE health technology evaluations: the manual. 181. 23. Medicines Complete.

• https://www.medicinescomplete.com/#/content/bnf/_988423144#DMD40251311000001101.

• 24. Medicines Complete. Paxlovid.

https://about.medicinescomplete.com/#/content/bnf/_272860093?hspl=paxlovid#DMD4032501100 0001107.

13

Confidential

25. Pfizer Inc. Pfizer to Provide U.S. Government with 10 Million Treatment Courses of Investigational Oral Antiviral Candidate to Help Combat COVID-19. https://www.pfizer.com/news/pressrelease/press-release-detail/pfizer-provide-us-government-10-million-treatment-courses (2021).

26. SOTO, Á. Spain has only administered two per cent of its Paxlovid anti-Covid pills that cost 238 million euros. SUR (2022).

27. The Pharmaceutical Benefits Scheme. NIRMATRELVIR (&) RITONAVIR. https://www.pbs.gov.au/medicine/item/12996B.

28. van der Plaat, D. A. et al. Impact of COVID-19 pandemic on sickness absence for mental ill health in National Health Service staff. BMJ Open 11 , e054533 (2021).

29. Hanly, P., Ahern, M., Sharp, L., Ursul, D. & Loughnane, G. The cost of lost productivity due to premature mortality associated with COVID-19: a Pan-European study. The European Journal of Health Economics 23 , 249–259 (2022).

30. Faramarzi, A., Javan-Noughabi, J., Tabatabaee, S. S., Najafpoor, A. A. & Rezapour, A. The lost productivity cost of absenteeism due to COVID-19 in health care workers in Iran: a case study in the hospitals of Mashhad University of Medical Sciences. BMC Health Services Research 21 , 1169 (2021).

31. NHS England. Supporting colleagues affected by Long COVID. NHS England https://www.england.nhs.uk/supporting-our-nhs-people/support-now/supporting-long-covid/.

32. Takashita, E. et al. Efficacy of Antibodies and Antiviral Drugs against Omicron BA.2.12.1, BA.4, and BA.5 Subvariants. N Engl J Med (2022) doi:10.1056/NEJMc2207519.

33. Vena, A. et al. Early Clinical Experience with Molnupiravir for Mild to Moderate Breakthrough COVID-19 among Fully Vaccinated Patients at Risk for Disease Progression. Vaccines 10 , (2022).

34. Khoo, S. H. et al. A Randomised -Controlled Phase 2 trial of Molnupiravir in Unvaccinated and Vaccinated Individuals with Early SARS-CoV-2 . http://medrxiv.org/lookup/doi/10.1101/2022.07.20.22277797 (2022) doi:10.1101/2022.07.20.22277797.

14

Confidential

Checklist for submitting comments

• Consultation responses must not be longer than 10 pages , excluding the references.

• Use this comment form and submit it as a Word document (not a PDF).

• • Complete the disclosure about links with, or funding from, the tobacco industry. • Combine all comments from your organisation into 1 response. We cannot accept more than 1 set of comments from each organisation.

• Do not paste other tables into this table – type directly into the table.

• Please underline all confidential information, and separately highlight information that is submitted under ‘commercial in confidence’ in turquoise, all information submitted under ‘academic in confidence’ in yellow. If confidential information is submitted, please also send a second version of your comments with that information replaced with the following text: ‘academic/commercial in confidence information removed’. See the Process and methods manual (section 5.4) for more information.

• Do not include medical information about yourself or another person from which you or the person could be identified.

• • Do not use abbreviations

• Do not include attachments such as research articles, letters, or leaflets. For copyright reasons, we will have to return comments forms that have attachments without reading them. You can resubmit your comments form without attachments, it must send it by the deadline.

• If you have received agreement from NICE to submit additional evidence with your comments on the pro-forma response document, please submit these separately.

Note: We reserve the right to summarise and edit comments received during consultations, or not to publish them at all, if we consider the comments are too long, or publication would be unlawful or otherwise inappropriate.

Comments received during our consultations are published in the interests of openness and transparency, and to promote understanding of how recommendations are developed. The comments are published as a record of the comments we received, and are not endorsed by NICE, its officers or advisory committees.

15

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Executable Model

Therapeutics for people with COVID-19 [ID4038]

The economic model enclosed and its contents are confidential and are protected by intellectual property rights, which are owned by ScHARR . It has been sent to you for information only. It cannot be used for any other purpose than to inform your understanding of the appraisal. Accordingly, neither the model nor its contents should be divulged to anyone other than those individuals within your organisation who need to see to them to enable you to prepare your response. Those to whom you do show the documents must be advised they are bound by the terms of the Confidentiality Agreement Form that has already been signed and returned to the Institute by your organisation.

You may not make copies of the file and you must delete the file from your records when the appraisal process, and any possible appeal, are complete. If asked, you must confirm to us in writing that you have done so. You may not publish it in whole or part, or use it to inform the development of other economic models.

The model must not be re-run for purposes other that the testing of its reliability .

Please set out your comments on reliability in writing providing separate justification, with supporting information, for each specific comment made. Where you have made an alteration to the model details of how this alteration was implemented in the model (e.g. in terms of programme code) must be given in sufficient detail to enable your changes to be replicated from the information provided. Please use the attached pro-forma to present your response.

Please prepare your response carefully. Responses which contain errors or are internally inconsistent (for example where we are unable to replicate the results claimed by implementing the changes said to have been made to the model) will be rejected without further consideration.

Results from amended versions of the model will only be accepted if their purpose is to test robustness and reliability of the economic model. Results calculated purely for the purpose of using alternative inputs will not be accepted.

No electronic versions of the economic model will be accepted with your response.

Responses should be provided in tabular format as suggested below (please add further tables if necessary).

July 2022

Issue 1 The implications of drug-drug interactions for treatments administered in the community has not been considered

Issue 2 The full cost of providing treatments in the community has not been appropriately characterised

Issue 3 The utility decrements for health states have been proxied from recurrent Clostridium difficile infection and influenza rather than COVID-19

Issue 4 The direct cost to the NHS of sickness amongst the NHS workforce is not considered

Issue 5 The societal perspective should be considered

Issue 6 Health state 1 is not considered in patients who are hospitalised, whereas health state 2 is not considered in patients who remain in the community

Description of problem Description of proposed amendment Result of amended model or expected impact on the result (if applicable) The economic model does not currently distinguish The company propose that health state 1 is considered This proposed amendment would more patients who are not hospitalised and experience within the economic model such that patients in the accurately characterise the incremental costno limitations of activities (health state 1) from community who are not hospitalised, or who have been effectiveness of molnupiravir compared to patients who are not hospitalised and who do still discharged from hospital, have the opportunity to SoC within the community setting as experience limitations of activities, home oxygen transition to a health state that represents perfect, or molnupiravir is proven to reduce the requirements or both (health state 2). Therefore, a near perfect, health, with no lasting symptoms or longprogression of signs and symptoms of quality-adjusted life-year (QALY) decrement is not term sequelae of COVID-19. COVID-19.[11] As such, patients treated with applied to patients who receive treatment in the molnupiravir are less likely to suffer long-term Including health state 1 within the model then allows community setting and go on to experience sequalae of severe COVID-19. for a utility decrement to be applied to health state 2 to limitations of activities or require oxygen at home. accurately capture the reduced quality of life within this Additionally, in patients who are hospitalised, a health state in the outpatient setting. QALY decrement is not applied to patients who have been discharged from hospital but are still experiencing limitations of activities or require oxygen at home.

This does not capture the full benefit of molnupiravir treatment. For most COVID-19 signs and symptoms, sustained improvement or resolution was more likely and worsening progression of signs or symptoms was less likely in the molnupiravir group than in the placebo group.[11] Therefore, lack of distinction between health state 1 and health state 2 eliminates the ability to characterise the full benefit of molnupiravir relative to SoC.

Issue 7 Rates of long COVID are assumed to be 100% and 10% amongst patients who are hospitalised and remain in the community, respectively, regardless of treatment received.

References:

1. University of Liverpool. COVID-19 drug interactions. https://www.covid19-druginteractions.org/checker (2022).

2. National Institute of Health. The COVID-19 Treatment Guidelines Panel’s Statement on Potential Drug-Drug Interactions

   • Between Ritonavir-Boosted Nirmatrelvir (Paxlovid) and Concomitant Medications. (2021).

3. National Institute for Health and Care excellence. COVID-19 rapid guideline: Managing COVID-19. (2022).

4. Awortwe, C. & Cascorbi, I. Meta-analysis on outcome-worsening comorbidities of COVID-19 and related potential drug-drug

   • interactions. Pharmacol Res 161 , 105250–105250 (2020).

5. Food and Drug Administration. Center for Drug Evaluation and Research Antimicrobial Drugs Advisory Committee Meeting Briefing Document Molnupiravir Oral Treatment of COVID-19. (2021).

6. EMA. Remdesivir. Summary of Product Characteristics.

7. EMA. Paxlovid Summary of Procut Characteristics.

8. Rafia, R. et al. A Cost-Effectiveness Analysis of Remdesivir for the Treatment of Hospitalized Patients With COVID-19 in

   • England and Wales. Value Health 25 , 761–769 (2022).

9. Faramarzi, A., Javan-Noughabi, J., Tabatabaee, S. S., Najafpoor, A. A. & Rezapour, A. The lost productivity cost of

absenteeism due to COVID-19 in health care workers in Iran: a case study in the hospitals of Mashhad University of Medical Sciences. BMC Health Services Research 21 , 1169 (2021).

10. van der Plaat, D. A. et al. Impact of COVID-19 pandemic on sickness absence for mental ill health in National Health Service staff. BMJ Open 11 , e054533 (2021).

11. Jayk Bernal, A. et al. Molnupiravir for Oral Treatment of Covid-19 in Nonhospitalized Patients. N Engl J Med 386 , 509–520

• (2022).

12. Evans, R. A. et al. Physical, cognitive, and mental health impacts of COVID-19 after hospitalisation (PHOSP-COVID): a UK

• multicentre, prospective cohort study. The Lancet Respiratory Medicine 9 , 1275–1287 (2021).

National Institute for Health and Care Excellence Centre for Health Technology Evaluation

Pro-forma Response

Assessment Report

Therapeutics for people with COVID-19 [ID4038]

Please read the checklist for submitting comments at the end of this form. We cannot accept forms that are not filled in correctly. Deadline for comments 5pm on Friday 29 July . Please submit via NICE Docs .

Please return to: NICE DOCS

Please comment on the assessment report

Executive summary

The key issues we’d like to highlight from the assessment report are as follows:

• The definition for patients at high-risk was aligned to the PANORAMIC clinical study except for age which is currently excluded from this assessment. This proposed definition is overly restrictive and should not be based on a single source of evidence, the consequence of this is the exclusion of at-risk patients who would potentially benefit from treatment, including body mass index (BMI) > 25 kg/m2, smoking, hypertension, patients on cancer treatments, healthcare workers and unpaid carers. The omission of age as an independent risk factor also goes against the published evidence base and we suggest that this be included within the risk criteria

• Thorough assessment of any indirect comparison and clinical effectiveness evidence is not possible due to the lack of methodological detail presented. This omission presents a significant limitation when assessing the robustness and suitability of the data used to inform the modelling. Detailed methodology should be presented in line with standard requirements set out in NICE guidance.

• Whilst we acknowledge a pragmatic approach to this assessment was required due to time limitations, we believe it is critical that a probabilistic sensitivity analysis PSA be conducted to assess the full impact of uncertainties within the model and would request that the assessment group present this analysis for further interpretation.

• Transmission value should be captured in the model as has been done for other infection disease models appraised by NICE.

Comments on Assessment Report: Therapeutics for people with COVID-19 [ID4038]

The Assessment Group (AG) were commissioned to assess the clinical effectiveness of eight treatments, including nirmatrelvir/ritonavir (Paxlovid®), within their proposed marketing authorisations for treating people with coronavirus disease 2019 (COVID19) in hospital and in the community. Given the current indication for Paxlovid®, this response focuses on the second population presented in the AG report: people who are at high risk of hospitalization due to COVID-19.

1. Comments on the decision problem

The definition of patients at high risk considered in the decision problem aligns with the population of the PANORAMIC study,[1] which enrols patients aged ≥ 50 years OR aged 18–49 years with any known underlying chronic health condition considered to make them clinically vulnerable. However, the definition applied in the appraisal excludes age ≥ 50 years as a high-risk factor. Furthermore, the PANORAMIC study criteria do not include other established risk factors for severe COVID-19, including

Please return to: NICE DOCS

body mass index (BMI) > 25 kg/m[2] , smoking, hypertension, patients on cancer treatments, healthcare workers and unpaid carers. The consequence of the current definition is the exclusion of at-risk patients who would potentially benefit from treatment. The use of a single definition of high risk does not seem reflective of the appropriate population and we request that the full evidence base quantifying risk factors be considered, above and beyond those derived from the PANORAMIC study.

Age as a risk factor

The Joint Committee on Vaccination and Immunisation (JCVI) has advised that all adults aged 50 years and over should be included in the autumn COVID-19 booster vaccination program.[2] Additionally, during initial vaccine roll out, prioritisation was based in part upon age, independent of other factors. In light of this continued prioritisation of the older population for vaccination, it seems contradictory to now exclude this population from the high-risk bracket when considering treatment for COVID-19. Further consideration should be given to the proposed marketing authorisations for treatments assessed in the community setting; not considering age as a high-risk factor would potentially exclude a portion of the licensed population from this NICE appraisal. We suggest that further consideration be given to the inclusion of age ≥ 50 years as an independent risk factor in this assessment.

There is a substantial UK and international evidence base that has demonstrated that age is an independent risk factor for developing severe COVID-19, requiring hospitalisation and death.[3-7] Notably, the living risk prediction algorithm for hospitalisation and death (QCOVID) includes age as an independent risk factor and has been validated in a large UK cohort.[8] Similarly, the International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) risk score for predicting COVID-19 mortality includes age >50 as a risk factor and has been validated in a large cohort from 260 UK hospitals.[9] The CDC has linked age as a risk factor for severe COVID-19 in US populations of various vaccination status’, including data to show hospitalisation is 3x more likely in patients aged 50-59 compared to a reference group (aged 18-29), which increases to 10x more likely in patients aged 85+.[10-12] Romero-Starke et al . 2021, an SLR and meta-analysis of 11 studies assessing COVID-19 hospitalisation rates, found age to be an independent risk factor of hospitalisation (risk ratio: 1.034; 95% CI 1.021 - 1.048) and that risk increased continuously with age.[13] Vahey et al. 2021 also found a significant association between age as an independent risk factor and hospitalisation, specifically patients aged 65+ (adjusted odds ratio: 3.22; CI 1.20 - 7.97).[14] In addition to increased risk of hospitalisation, an association between age and increased severity and mortality among patients already hospitalised has been shown.[15-20]

Hypertension and smoking

Please return to: NICE DOCS

Whilst we believe the omission of age from the AG’s risk criteria to be the primary issue, we also note that there is substantial evidence implicating hypertension[3,4,14,21] and smoking[4,22] as risk factors of hospitalisation due to COVID-19; particular consideration should be given to the links between smoking and other lung conditions (COPD, lung cancer etc.) which are already considered risk factors within the appraisal. We suggest that these factors also be considered for inclusion within the high-risk criteria.

Overweight, underweight and obesity

Adults with excess weight are shown to be at a consistently elevated risk of severe disease and hospitalisation from the Centres for Disease Control (CDC’s) evidence review.[23] Additionally, data from the ‘second wave’ of the pandemic in England showed that patients with BMI reflective of being underweight (<18.5 kg/m2), overweight (> 25 kg/m2) and obese (> 30 kg/m2) have 10%, 24% and 93% higher odds of hospital admission compared to patients of a healthy weight.[24] We suggest that exclusion of these patients from the high-risk group be reconsidered.

Patients on cancer treatment and splenectomy patients

Patients currently receiving or having recently received cancer treatment can currently access treatment through the COVID medicines delivery unit (CMDU) for high-risk, non-hospitalised patients with COVID-19. However, it is unclear whether these patients would still be eligible under the PANORAMIC based definition of high risk. Considering this patient group is extremely vulnerable and at high risk, we request that the high-risk definition be updated to ensure alignment with CMDU criteria and avoid exclusion of at risk patients.[25]

People who have had a splenectomy, have been identified as being at high risk of severe COVID-19 and have hence been prioritised for vaccination.[26] However, they do not seem to be included in either the CMDU or the PANORAMIC eligibility list.

Healthcare workers (including care workers)

While this analysis is focusing on those at high risk of hospitalisation and severe COVID-19, we should not overlook the additional value community treatments can bring to the healthcare system. The COVID-19 pandemic has indeed exacerbated NHS capacity issues. The British Medical association estimated that the waiting list for consultant-led elective care has grown by 2.24 million from 4.24 million in March 2020 to 6.48 million in April 2022.[27] This has been driven in part by COVID-19 related hospitalisation taking up hospital beds but also by staff shortages. Staff absenteeism has increased, mainly driven by COVID-19 infections and requirement to isolate. Studies have shown that patient-facing healthcare workers are at increased risk of infection due to on-the-job exposure with an odds ratio of between 1.5 and 2.5 compared to non-patient facing NHS staff[28] , and that during the second

Please return to: NICE DOCS

wave of the pandemic in England, the odds ratio of infection for NHS staff working on an inpatient ward or emergency department were between 1.5 and 2.1 (SIREN study).[29] In England alone, from March 2020 to February 2022 a monthly average of 279,214 FTE days were lost due to COVID-19.[30] The last 2 peak losses were in January 2021 (637,734 FTE days) and January 2022 (862,085 FTE days). Data from Wales and Scotland shows a similar pattern with staff absence due to COVID-19 sickness ranging from 0.3%-2.5%.[31,32] Multiple other studies from across the globe present further evidence of the increased risk of developing severe COVID-19 among healthcare workers and the impact of this to healthcare systems.[33-36] We believe community treatments have a key role to play in reducing duration of staff absence to ensure sufficient resourcing on the health and care system. This would help to enable the clearing of the care backlog and alleviate winter pressures.

Unpaid Carers

Unpaid carers provide care to those affected by disability, physical or mental health illness, or frailty, presenting a significant cost saving to the NHS every day.[37] We believe the potential costs of absenteeism of unpaid carers due to severe COVID-19 should not be overlooked. Community treatment of unpaid carers could reduce the duration of absenteeism, and in turn reduce the cost of emergency temporary care for those with often complex care needs.

2. Derivation of clinical effectiveness evidence

The AG adopted a pragmatic approach to identifying and collating evidence on COVID-19 treatments in the community setting in order to provide evidence for decision making in an efficient manner. To this end, third-party sources of evidence identification and synthesis were used, instead of conducting a systematic literature review (SLR) and network meta-analysis (NMA). Limited methodology is available, and no critical appraisal of these sources is presented by the AG.

When reviewing the methods for deriving the clinical evidence, Pfizer acknowledges the time constraints faced by the AG, while noting the inherent limitations in this approach.

2.1 Systematic literature review

The COVID-NMA initiative[38,39] was used as a third-party source to identify relevant trials. The appropriateness of the evidence could not be assessed due to limited methodology and the omission of absolute numbers of outcomes.

2.2 Indirect comparison

The clinical effectiveness section of the AG report (section 2) does not provide detailed methodology of the Indirect Treatment Comparison (ITC) analysis, and this is not readily available on any of the third party source websites. This omission

Please return to: NICE DOCS

presents a significant limitation to assessing the robustness and suitability of the data used to inform the economic modelling, which would not be acceptable in any standard NICE HTA. Detailed methodology should be presented by the AG to allow for a comprehensive review to address these various uncertainties.

The following considerations should be taken into account:

• The methodology for evidence synthesis requires clarification by the AG . It is unclear if a Network Meta Analysis (NMA) has been conducted by the AG or the third-party source, or if only extracted relative treatment effects from relevant studies are included. It appears that these relative treatment effects are compared naively, applied in the economic model using the RECOVERY trial as representation of SoC. However, this lack of clarity makes it difficult to assess whether the comparison was undertaken effectively and according to the relevant best practice guidelines, which would not be considered appropriate during a standard NICE HTA.

• The appropriateness of pooling this data from multiple studies should be thoroughly evaluated by the AG . Regardless of the methods used for evidence synthesis (naïve comparison, NMA or meta-analysis), a full assessment of the appropriateness of synthesis is required according to NICE best practice. A naïve comparison implicitly assumes that there are no underlying differences between studies in terms of study design, baseline characteristics or SoC, while an NMA is only appropriate if these differences are not considered to impact on outcomes. However, it does not appear that any assessment has been undertaken, although this is unclear.

• The results of the NMA require additional context, as these are not clearly presented or explained . This makes it difficult for readers to understand the results, which may lead to differing conclusions on the robustness of the results and suitability to inform decision making. In particular, no methodology for ranking the therapies is given, reducing the opportunity for comment on this aspect. It appears the probability rank is used to compare the relevant performance of treatments, but alternative methods such as the surface under the cumulative ranking curve (SUCRA) should be considered. These alternative methods may show more obvious patterns for different treatment performance by visual inspection.

It would also be useful to clarify how robustly the two living systems, COVID-NMA and metaEvidence, carry out the unattended NMA analysis based on the living systematic review; for example, whether any diagnostics checks for transitivity and consistency between direct evidence and indirect evidence are performed. As per guidelines, reproduction of the results from COVID-NMA and metaEvidence using the raw data from each trial and NICE published computer codes would go a distance to validating the two pre-existing systems utilised in this assessment.

Please return to: NICE DOCS

2.3 Appropriateness of clinical evidence for synthesis

Based on the assessments of the identification and synthesis of evidence, described above, the following points regarding the appropriateness of the clinical evidence used for evidence synthesis should be considered:

• The report acknowledges study heterogeneity attributable to changes in SoC, vaccination status, prior COVID-19 infections, and SARS-CoV-2 variant as a main factor of uncertainty in the review findings. Although a brief description of each included RCT is provided in Appendix 1, no details are given on those variables within each trial to permit assessment of the potential impact of this heterogeneity.

• In particular, the report states that it is assumed that “all relative treatment effects were transportable to different settings”. This meant that the naïve comparison included no assessment or adjustments were made to account for relevant covariates, including: SoC at the time of the clinical study; age; disease severity; vaccination status; history of SARS-CoV-2 infection; geographical location; dosage of relevant interventions. Given that COVID-19 severity and, in turn, patient outcomes, will be affected by these factors – particularly vaccination status, which is highly time-specific – this may have a significant impact on the relative treatment effects.

• It is acknowledged that there are data limitations across the COVID-19 evidence base when considering high-quality data sources, principally due to low clinical trial recruitment rates, which make adjustment unfeasible and lead to the aforementioned uncertainty. Regardless of the associated uncertainty, Pfizer agrees with the AGs approach to consider only the highest quality clinical evidence from randomised trials; broadening of the scope of evidence would only increase uncertainty further.

• Currently, the results report the relative risk (RR) for the following outcomes in outpatients: hospitalisation or death, and all-cause mortality at 28 days. However, the absolute number of the outcomes is omitted. In some studies, these outcomes are relatively rare in both arms (for example, the three casirivimab+imdevimab studies report 0/156, 1/838, 1/1,529 deaths in the treatment arm, and 0/158, 1/840, 3/1,500 deaths in the placebo arm), therefore the RR is very sensitive to event numbers and may be misleading. The raw data, including absolute numbers of outcomes, should be presented in order to aid interpretation.

3. Cost-effectiveness model approach

3.1 Design of cost-effectiveness model

Please return to: NICE DOCS

The AG model for evaluation of the high-risk, non-hospitalised (community setting) population does not fully capture all relevant aspects of COVID-19 natural history, treatment pathway, treatment effects and outcomes necessary for decision making. As a result, it does not fully explore the value of treatment in the community setting, so that cost-effectiveness outcomes are vastly underestimated.

In particular, the following omissions are of concern:

• The residual effect of community treatments once patients are hospitalised should be captured in the model: The AG economic model assumes that community treatments are followed by current SoC when a patient is hospitalised. While the treatment algorithms for COVID-19 are unclear, it is unlikely that community patients who are hospitalised would immediately transition to receive only current SoC. As stated in the report, community treatments are expected to have a residual effect, which should be captured in the economic model in the form of a long-term impact, reducing costs and improving outcomes. By contrast, patients with declining status are likely to receive the most effective treatment options for hospitalised patients, which are likely to include treatments recommended as a result of this MTA. Thus, this assumption does not reflect the likely treatment pathway for hospitalised patients who have received treatment in the community setting. However, Pfizer agree that the studies assessing community- and hospital-setting treatments are not interchangeable and evidence from each population should not be compared.

• Disease severity for hospitalised patients should impact on disease progression, mortality or subsequent costs : In the PSM module, different severity levels should be captured in separate health states/events. While the model does use an 8-point ordinal scale for severity, occupancy on this scale is used to associate with supplemental oxygen and other management requirements and the subsequent impact on health-related quality of life (HRQoL) and costs. However, it does not capture the impact on progression of the disease and other outcomes (i.e. hospitalisation, discharge, mortality) and subsequent costs.

• Viral load suppression should be reflected in the economic model: Although viral load suppression was included in the scope of this evaluation, it has not been included in the economic model. Viral load suppression is a key endpoint for many virologic diseases, with impacts on clinical outcomes and disease transmission, and is the surrogate endpoint of interest for economic models, including for HIV modelling.[40] As a result, this decision requires justification by the AG.

• The economic model should reflect the broader value of new antimicrobial treatment: Previous assessments of the value of novel antimicrobials have noted that there are additional attributes of value for new antimicrobials that are

Please return to: NICE DOCS

not typically included in standard HTA evaluations: diversity value; transmission value; enablement value; spectrum value; and insurance value.[41,42] These values remain relevant to novel antiviral therapies, particularly for COVID-19, where further outbreaks remain likely in the near future. In particular, it is essential that the following attributes are captured:

o Transmission value : The economic model does not capture the impact of treatments on preventing the transmission of the SARS-CoV-2 virus to noninfected people. The JCVI routinely includes transmission value in assessment of vaccines against other transmittable diseases. Whilst this would add a level of complexity in the model, it has long been established that the transmissible nature of infectious diseases is the critical characteristic that sets them apart from other diseases, with ISPOR-SMDM guidelines on incorporating transmission into economic models published in 2012.[43] Community treatments in particular have a crucial effect on transmission dynamics. The infection rate is a key consideration for the burden of disease of COVID-19, and we recommend that at least a simplistic approach is taken to include this component. The model could follow an approach similar to that used in other communicable disease areas such as HIV[40] , through the use of a multiplier to be applied to the cohort, simulating the growth of an "Infected" compartment. This multiplier would reflect the effect of treatments on transmission dynamics and the model could be calibrated for this parameter against published studies.

• Enablement value : Enablement value is defined as the value of the benefits of being able to perform medical procedures because of new antimicrobials. This was deemed as a significant area of value in the NICE HTA assessment for antimicrobials. A key value benefit of Paxlovid[®] is the ability to ensure patients are not in hospital, occupying resources (beds, staff, etc) that could be used to enable other procedures to go ahead. This has not been accounted for in the model. Where possible, and to align with the NICE HTA assessment for antimicrobials, the cost-effectiveness model should look to capture the benefits of releasing hospital resources, that would otherwise be used for treating infections to enable healthcare and procedures in other patients. Where this is not possible, due to time and resource constraints, or lack of data, additional considerations should be made to the level of enablement value that exits.

• Insurance value : this refers to the value of having effective therapies available in case of an increase in the prevalence of infections. In the case of COVID-19, where future outbreaks are likely in the near future, the value should be reflected in the economic model.

3.2 Cost-effectiveness model inputs

• It is critical than a PSA be conducted to assess the full impact of uncertainties within the model and would request that the AG present this

Please return to: NICE DOCS

analysis for further interpretation: The level of uncertainty associated with the majority of the efficacy measures is very high. Most efficacy measures were derived from small sample sizes and highly heterogeneous populations. For treatments in the community setting, the RRs for hospitalisation and death at 28 days have wide confidence intervals (CIs). For the RR for death at 28 days, all CIs crossed 1, except for molnupiravir and nirmatrelvir/ritonavir. As detailed in section 3.3 of this response, no probabilistic sensitivity analysis (PSA) has been presented, making it difficult to assess the impact of these uncertainties.

• For the decision tree, it is unclear how valid the admission rates used in the model are. For SoC, the model used an adjusted rate of admission for a general COVID-19 population, multiplied by 2, to account for the high-risk nature of the population, based on data used for the QCovid3 risk algorithm and clinical advice; however, it is unclear how the published data and advice were used to arrive at the multiplier. Nonetheless, we agree the admission data from Nyberg et al . is an appropriate source given the relative certainty that admissions are COVID-19 related. We also agree that it is important to consider the impact of a higher hospitalisation rate, as indicated by alternate data sources, through sensitivity analyses.

• Long COVID was defined in the model as at least 4 weeks of COVID symptoms. The model assumes that 10% of community patients not hospitalized will experience long COVID, based on the prevalence of long COVID in different populations based on vaccination status. Additionally, the AG estimated a mean duration of long COVID by fitting parametric distributions to published Office for National Statistics data. However, both elements fail to capture the differentiation between patients based on their vaccine history and severity of disease – where data allows, scenarios should be run to account for this variance in the population. We are not currently aware of treatments which impact the prevalence of long-COVID. It would be useful to test this assumption in scenario analyses and account for the potential residual effect that community treatments can have on the manifestation of long COVID.

• The AG should clarify how model inputs were identified and the criteria used to select the most appropriate data from the identified sources: The AG did not undertake an SLR to identify costs and healthcare resource use associated with the management of COVID-19 in the high-risk population. For infectious disease It may be more appropriate to model opportunity costs for hospitalisation (bed days)[44,45] – this should also be considered. No formal SLR was conducted to identify the most appropriate utility and disutility values to inform the economic model.

• We suggest more appropriate sources be considered for the utility decrements applied in the model: We acknowledge that EQ-5D data on COVID-19 are scarce and proxies would be acceptable. While an effort is made

Please return to: NICE DOCS

to use values from influenza, the use of HRQoL for hospitalisations not requiring interventions is not appropriate. Values used were obtained from patients with severe diarrhoea and colitis due to Clostridium difficile bacterial infection.

• More appropriate sources for unit costs by ordinal scale applied in the model should be considered : VC40Z, the code used for ordinal scale 4, corresponds to rehabilitation post respiratory related admission and would be more appropriate to use for long-COVID costs. We suggest non-intensive care unit (ICU) hospitalisations be approximated with ICD-10/HRG codes for other viral pneumonia (J12.8 [Other viral pneumonia], J12.9 [Viral pneumonia, unspecified], DZ11 [other viral pneumonia). Considering the length of stay due to COVID-19 is at least 2 days, the use of non-elective long stay costs would be most appropriate. This approach was used by the UK Health Security Agency (UKHSA) in their cost-effectiveness analysis of vaccination and social distancing measures.[46] The complexity of care level and number of interventions allow for an alignment with the ordinal scores for non-ICU admissions, for example DZ11R-V could be mapped to Ordinal score 3; DZ11N-Q could be mapped to Ordinal score 4 and DZ11K-M could be mapped to Ordinal score 5.[47]

3.3 Model outputs and analyses

It is essential that a PSA be conducted to assess which inputs are causing uncertainty in the model and address this uncertainty: Deterministic sensitivity analyses can only inform the range of values (extremes) expected based on uncertainty, but do not provide a clear picture of the distribution of results according to uncertainty. As many of the estimates used in the model are either based on small sample sizes, heterogeneous populations and/or assumptions, it is crucial that this is tested.

4. Conclusions

Whilst we appreciate the time constraints the AG had to work with during this assessment and the uncertainties that are to be expected as a result, we believe there are some critical issues that should be addressed before this assessment is used to inform the NICE technology appraisal. Primarily: thorough assessment of the current evidence base to define ‘high risk’, which is currently over restrictive and excludes at-risk patients who would benefit from treatment (older age, hypertension, smoking, health care workers, carers etc.), provision of detailed indirect comparison methodology to allow the appropriateness of the clinical evidence informing the model to be assessed, inclusion of a PSA to assess the impact of uncertainties within the model and incorporation of transmission value within the model.

Please return to: NICE DOCS

Checklist for submitting comments

• Consultation responses must not be longer than 10 pages , excluding the references.

• Use this comment form and submit it as a Word document (not a PDF).

• • Complete the disclosure about links with, or funding from, the tobacco industry. • Combine all comments from your organisation into 1 response. We cannot accept more than 1 set of comments from each organisation.

• Do not paste other tables into this table – type directly into the table.

• Please underline all confidential information, and separately highlight information that is submitted under ‘commercial in confidence’ in turquoise, all information submitted under ‘academic in confidence’ in yellow. If confidential information is submitted, please also send a second version of your comments with that information replaced with the following text: ‘academic/commercial in confidence information removed’. See the Process and methods manual (section 5.4) for more information.

• Do not include medical information about yourself or another person from which you or the person could be identified.

• • Do not use abbreviations

• Do not include attachments such as research articles, letters or leaflets. For copyright reasons, we will have to return comments forms that have attachments without reading them. You can resubmit your comments form without attachments, it must send it by the deadline.

• If you have received agreement from NICE to submit additional evidence with your comments on the pro-forma response document, please submit these separately.

Note: We reserve the right to summarise and edit comments received during consultations, or not to publish them at all, if we consider the comments are too long, or publication would be unlawful or otherwise inappropriate.

Comments received during our consultations are published in the interests of openness and transparency, and to promote understanding of how recommendations are developed. The comments are published as a record of the comments we received, and are not endorsed by NICE, its officers or advisory committees.

Please return to: NICE DOCS

References

1. PANORAMIC trial. Platform Adaptive trial of NOvel antiviRals for eArly treatMent of COVID-19 In the Community. 2022. Available at: https://www.panoramictrial.org/ [Accessed July 2022].

2. Department of Health and Social Care. Health and Social Care Secretary accepts JCVI advice on -