key: cord-0929670-0048lu4h
authors: Rafia, Rachid; Martyn-St James, Marrissa; Harnan, Sue; Metry, Andrew; Hamilton, Jean; Wailoo, Allan
title: A Cost-Effectiveness Analysis of Remdesivir for the Treatment of Hospitalized Patients With COVID-19 in England and Wales
date: 2022-02-20
journal: Value Health
DOI: 10.1016/j.jval.2021.12.015
sha: fb2801aa23c737de5460f5d126f33c49dbb03e9e
doc_id: 929670
cord_uid: 0048lu4h

OBJECTIVES: COVID-19 is associated with significant morbidity and mortality. This study aims to synthesize evidence to assess the cost-effectiveness of remdesivir (RDV) for the treatment of hospitalized patients with COVID-19 in England and Wales. METHODS: A probabilistic cost-effectiveness analysis was conducted informed by 2 large trials and uses a partitioned survival approach to assess short- and long-term clinical consequences and costs associated with COVID-19 in a hypothetical cohort of hospitalized patients requiring supplemental oxygen at the start of treatment. Given that it is uncertain whether RDV reduces death, 2 analyses are presented, assuming RDV either reduces death or does not. Published sources were used for long-term clinical, quality of life, and cost parameters. RESULTS: Under the assumption that RDV reduces death, the incremental cost-effectiveness ratio for RDV is estimated at £11 881 per quality-adjusted life-year gained compared with standard of care (SoC) (probabilistic incremental cost-effectiveness ratio £12 400). The probability for RDV to be cost-effective is 74% at a willingness-to-pay threshold of £20 000 per quality-adjusted life-year gained. RDV was no longer cost-effective when the hazard ratio for overall survival compared with SoC was >0·915. CONCLUSIONS: Results from this study suggest that using RDV for the treatment of hospitalized patients with COVID-19 is likely to represent a cost-effective use of National Health Service resources at current willingness-to-pay threshold in England and Wales, only if it prevents death. Results needs to be interpreted caution as vaccination was introduced and the SoC and evidence available have also evolved considerably since the analysis is conducted.

COVID-19 is caused by the novel coronavirus severe acute respiratory syndrome coronavirus 2 and causes atypical pneumonia. 1 The disease is transmitted by inhalation or contact with infected droplets, and the incubation period is estimated to range from 2 to 14 days. 2 The symptoms are usually fever, cough, breathlessness, and fatigue, among others. 2 All populations are susceptible to severe acute respiratory syndrome coronavirus 2, with the elderly and people with underlying diseases or low immune function being more likely to become severe cases. 1 In response to the public health emergency, the National Health Service (NHS) England issued an interim commissioning policy (first published on the July 7, 2020, and updated on the November 12, 2020) for the use of remdesivir (RDV) in England and Wales, 3 for the treatment of hospitalized patients with COVID-19 requiring supplemental oxygen at the start of treatment (as per its European Medicines Agency marketing authorisation 4 ), typically defined as those on low-flow oxygen (LFO), high-flow oxygen (HFO), or other noninvasive ventilation (NIV).

The efficacy of RDV in hospitalized patients with COVID-19 on supplemental oxygen is described in a number of studies that are heterogeneous in terms of the included populations, study design, and outcomes. The Adaptive Covid-19 Treatment Trial 1 (ACTT-1) 5 is a double-blind randomized controlled trial (RCT) comparing RDV with placebo in the United States and reported a treatment effect on mortality (hazard ratio [HR]) of 0$30 (95% confidence interval [CI] 0$14-0$64) in hospitalized patients with LFO and 1$02 (95% CI 0$54-1$91) in patients with HFO or NIV at baseline. SOL-IDARITY 6 was conducted after ACTT-1 and is an unblinded, multicenter RCT across 30 countries in patients who were hospitalized with COVID-19. ACTT-1 compared RDV with local standard of care (SoC) and reported a rate ratio of death of 0$85 (95% CI 0$66-1$09) for the subgroup of patients requiring any supplemental oxygen (but not mechanically ventilated). Wang in a double-blind RCT and reported a death rate ratio of 0$81 (95% CI 0$21-3$07) in hospitalized patients with LFO and 1$40 (95% CI 0$20-9$52) in patients with HFO or NIV at baseline. Therefore, there is considerable uncertainty in the effectiveness of RDV in preventing death in the overall supplemental oxygen population. These RCTs were also conducted at different phases during the pandemic making any direct comparison challenging. NHS England's interim commissioning for RDV 3 was produced in response to the public health emergency and the rapid need for effective treatments to help reduce morbidity and mortality posed by COVID-19. This decision was not based on economic consideration. Therefore, it is unclear whether RDV represents a costeffective use of NHS resources in England and Wales for the treatment of hospitalized patients with COVID-19 on supplemental oxygen at entry.

The objective of this study is to assess the cost-effectiveness of RDV in England and Wales within its current marketing authorisation 4 and interim NHS commissioning policy, from a health services perspective. 3 

A decision-analytical model was constructed in Microsoft Excel to assess the short-term (during the hospitalization episode) and long-term (after hospital discharge) clinical consequences and costs associated with COVID-19 in a hypothetical cohort of hospitalized patients requiring supplemental oxygen at the start of treatment (defined as patients on LFO or HFO or NIV) in England and Wales.

The intervention assessed is RDV (Veklury®) 200 mg on day 1, followed by RDV 100 mg maintenance up to 5 days, in line with current NHS England interim commissioning for RDV. 3 The comparator is established clinical management (previous routine use of tocilizumab [TCZ] and sarilumab [SAR]) with or without corticosteroids. RDV is assessed as an adjunct to SoC.

The structure of the decision-analytical model is depicted in Figure 1 . and uses what is commonly referred to as a partitioned survival/area under the curve (AUC) approach, composed of 3 main mutually exclusive health states; (1) discharged from hospital and alive, (2) hospitalized with or without COVID-19, and (3) death from any cause (COVID-19 or because of other causes).

Patients enter the model when hospitalized with COVID-19 and requiring supplemental oxygen (LFO, HFO or NIV) at the start of treatment. During their hospital stay (defined as the AUC between the time to death and discharge curves), patients are further separated into 5 hospitalization subhealth states based on their hospitalization/oxygen requirement according to the clinical status ordinal scale as defined in the ACTT-1 trial 5 (ie, no care because of COVID-19, no oxygen, LFO, HFO or NIV, and invasive ventilation), each associated with different cost and utility impact. Movements between hospitalization health states are not explicitly modeled. Instead, the partitioned model deals with health state occupancy within each time intervals.

Due to their short duration, not everyone in trials for COVID-19 treatments will have been discharged or died at the time of outcome assessment. Therefore, time to death is extrapolated parametrically beyond the trial duration using a daily cycle length up to 70 days (35 days in the base-case), followed (weekly cycle length) by an elevated risk of death 8 compared with background mortality up to week 52 (varied between 6 months and 2 years in sensitivity analysis), and unadjusted background mortality used thereafter. Time to discharge is extrapolated parametrically until everyone either dies or is discharged.

Discharged patients are assumed to have a reduced quality of life (QOL) (52 weeks from model entry in the base-case) to reflect emerging evidence on the effect of COVID-19 after hospital discharge. [9] [10] [11] Emerging evidence also suggests that some patients are at an elevated risk of multiorgan dysfunctions 8 (such as respiratory diseases, diabetes, cardiovascular, liver and kidney diseases) and may require long-term management/monitoring. 12 These potential impacts are included in this economic model as an average one-off cost and quality-adjusted life-year (QALY) loss per patient discharged.

The mean age and sex distribution at entry (Table 1 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] are taken from a UK study among 47 780 hospitalized patients with COVID-19 and discharged alive. 8 Time to death in patients initiated on SoC (previous use of RDV)

The time to death in patients for the SoC arm is taken from a large UK trial; the first RECOVERY trial that compared usual care versus corticosteroids (dexamethasone). 21 Since this trial was conducted, clinical practice has changed with corticosteroids widely used. Therefore, a weighted spline model (3 knots) was constructed from the pseudo-individual patient data (IPD; reconstructed based on a published algorithm 22 ) that assumes that 90% of patients are on corticosteroids (dexamethasone arm from RECOVERY 21 ), with the remaining not on dexamethasone (usual SoC from RECOVERY 21 ). A spline model with 3 knots was selected after visual inspection and statistical tests; with the addition or removal of knots not materially changing the fit within the observed or short-term extrapolation in this study. Scenario analyses are conducted using the control arm from SOLIDARITY 6 or the control arm from the RECOVERY TCZ trial 23 in people with progressive COVID-19.

Time to discharge in patients initiated on SoC (previous use of RDV)

The time to discharge for the SoC arm is approximated by adjusting a spline model (3 knots) estimated from pseudo-IPD from the control arm from SOLIDARITY 6 using an HR we calculated so that the proportion of patients predicted to be discharged and alive at day 28 matches the proportion reported in the RECOVERY trial 21 

The effectiveness of RDV at reducing death in patients on supplemental oxygen at entry is very uncertain and therefore 2 analyses are conducted. The base-case uses the point estimate from the onoxygen population subgroup from SOLIDARITY 6 (ratio of death rate 0$85; 95% CI 0$66-1$09). A secondary analysis is presented assuming an HR of 1 (eg, RDV does not have an impact on survival) as the treatment effect is not statistically significant despite the large sample size for this subgroup in SOLIDARITY (n = 3639). 6 Similar findings were observed in Wang et al 7 (2020). A significant treatment effect was also reported in patients with LFO in ACTT-1, but not in HFO. 5 The treatment effects for time to discharge are estimated from the published data (IPD reconstructed using a published algorithm 22 ) from SOLIDARITY 6 using piecewise HRs (day 0-4, 5-9, 10-14, 15-19, 201) calculated simultaneously in a model that is stratified by time group.

The distribution of hospital care type in patients initiated on RDV is shown in Figure 2 . The distribution at baseline and at day 14 are informed by the distribution (derived from the ordinal scale of clinical status) from the ACTT-1 trial 5 and Goldman et al 24 (RDV for 5 and 10 days). We then assumed that the proportion of patients on invasive ventilation, NIV and no longer on oxygen increased linearly from day 0 to 14 and was carried forward beyond day 14 informed by the trend observed in ACTT-1 for RDV. 5 The treatment effect for the probability of ventilation in patients initiated on SoC (compared with RDV) is taken from an unpublished UK study 25 for the base-case, and assumed to be same as RDV for the secondary analysis assuming no survival difference.

The unadjusted rate of mortality for the general population is taken from the England and Wales life table 2017 to 2019. 26 Between the end of extrapolation (day 35 in the base-case) and week 52, patients hospitalized with COVID-19 are assumed to be, on average, at an elevated risk of death of 7$7 (95% CI 7$2-8$3) compared with the general population based on the rate ratio reported by Ayoubkhani et al 8 (2021) in the UK.

We adopt a health service perspective. Unit costs are summarized in Table 1 . 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] The unit costs per hospital bed day, according to the intensity of care required as measured by the ordinal scale of severity, are taken from NHS reference costs. 18 The list price for RDV is taken from the BNF. 16 The number of RDV doses is taken from an unpublished UK study. 25 Only drug costs for corticosteroids are included for SoC and calculated from the electronic market information tool 17 based on the weighted average of relevant formulations.

As the population entering the model is hospitalized, additional administration costs are likely to be minimal and are therefore not included in this economic evaluation.

Health utility values are summarized in Table 1 . 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] Utility values are age-adjusted as patients get older based on Ara and Brazier 27 , with the baseline utility value pre-COVID-19 estimated from the mean age at entry, adjusted by a decrement in utility taken from Ara and Brazier 27 to reflect increased comorbidities for patients with COVID-19 compared with the general population. 8, 28 During the hospitalization episode, decrements in utility values are applied (subtracted) to the baseline, taken from the published literature. 14, 15 As with the assignment of costs, these utility decrements align with the degree of care required while in the hospital as indicated by the ordinal scale. Following hospital discharge patients with COVID-19 have a reduced QOL, 8 with QOL returning to pre-COVID-19 baseline after 52 weeks.

Increased risks of multiorgan dysfunction and monitoring-one-off cost and QALY decrement at the point of discharge A one-off cost and QALY loss per patient discharged (Table 1 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] ) is applied in the economic model to reflect the elevated risk of multiorgan dysfunction after COVID-19 hospitalization (assumed to last a year). These are calculated from the rates reported by Ayoubkhani et al 8 (2021) in the UK in patients hospitalized with COVID-19 compared with matched-controls and assumptions on costs and QALY loss (Appendix Table 1 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2021.12.015).

Increased monitoring/follow-up is assumed to occur in the first year only. An average one-off cost (Table 1 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] ) is applied at the point of discharge calculated based on the assumption that discharged patients require on average 2 chest x-ray and 6 GP electronic consultations, and unit costs from Stokes et al 19 (2016) and Curtis and Burns 20 (2020).

In accordance with the National Institute of Clinical Excellence reference case, 29 patients are followed over a lifetime horizon, an NHS/personal social services perspective is used and costs and benefits are discounted at 3$5% per annum.

Results are presented both deterministically and probabilistically to take account of the simultaneous effect of uncertainty relating to model parameter values. A total number of 1000 simulations were performed to obtain sufficient precision. Base-case results are also presented as cost-effectiveness planes and costeffectiveness acceptability curves.

The treatment effect for RDV on overall survival (OS) is a key driver and highly uncertain. Consequently, a threshold analysis is conducted, with results presented in terms of net monetary benefits to determine the point at which RDV is no longer costeffective at a willingness to pay (WTP) of £20 000 under our base-case assumptions. Threshold analysis to determine the costeffective price was also conducted. A range of sensitivity and scenario analyses are also conducted to test the robustness of results to key input parameters and assumptions (Appendix Table 2 in Supplemental Materials found at https://doi.org/10.1 016/j.jval.2021.12.015).

The deterministic and probabilistic results are presented in Table 1 . 6, 8, 9, [13] [14] [15] [16] [17] [18] [19] [20] For the probabilistic analysis, under the assumption that RDV reduces death, the model estimates total discounted costs associated with RDV to be £12 758 compared with £9393 for established clinical management, an incremental cost of £3365. The total discounted QALYs for RDV are estimated to be 6$62 compared with 6$35 for patients treated with established clinical management, an incremental QALY gain of 0$27. The probabilistic incremental cost-effectiveness ratio (ICER) is £12 400 per QALY gained.

We found that 89$3% of Monte Carlo simulations comparing RDV with established clinical management were in the northeast quadrant (more costly and more effective) of the costeffectiveness plane. RDV was more costly but less effective in 10$7% of case (northwest quadrant [Fig. 3] ). Cost-effectiveness acceptability curves demonstrate that RDV has a 74% probability of being a cost-effective treatment option at a £20 000 per QALY gained WTP threshold.

When it is assumed that RDV does not reduce death (Table 2) , RDV is predicted to lead a very small increase in QALYs (0$00002), at an incremental cost of £1666, leading to a high ICER (.£1 million per QALY). Approximately half of Monte Carlo simulations (52$2%) were in the northeast quadrant (more costly and more effective) of the cost-effectiveness plane, with RDV having a 0% probability of being a cost-effective treatment option at a £20 000 per QALY gained WTP threshold.

At its current list price, the threshold analysis (Fig. 3) shows that RDV is no longer cost-effective at a WTP of £20 000 per QALY gained when the HR is .0$915.

If no survival difference is assumed between RDV and SoC (Fig. 4) , the price per 100 mg vial for RDV needs to be ,£18$6 for RDV to be cost-effective under current WTP threshold.

In addition to the treatment effect for OS and list price for RDV, ICERs were affected the most by the model time horizon, the baseline curve for SoC (patients with progressive COVID-19 included in the RECOVERY TCZ trial), and inclusion of unrelated costs.

This is the first study to undertake an economic evaluation of the use of RDV for the treatment of hospitalized patients with COVID-19 in England and Wales. This study found that, although RDV reduces time to recovery, it is likely to be cost-effective only if it also prevents death. Nevertheless, the effectiveness of RDV at reducing death is highly uncertain within the overall supplemental oxygen population. Evidence suggests that patients requiring LFO are more likely to derive benefits from RDV than patients on HFO or NIV. Therefore, RDV is likely to represent a cost-effective use of NHS resources in patients with LFO only. The analysis was conducted during the second wave of the pandemic. Since the analysis was conducted, vaccination was introduced in the United Kingdom and the SoC and evidence available for RDV have evolved considerably. Therefore, results from this analysis should be interpreted in this context.

This study also sets out a framework to capture the key drivers of costs and benefits (mortality, duration of stay, and intensity of care required while in the hospital) that can be used to rapidly evaluate other treatments for hospitalized patients with COVID-19. This is important during a public health emergency because COVID-19 is associated with significant morbidity and mortality, and therefore, there is need to inform public policy rapidly and ensure that NHS resources are allocated efficiently and fairly.

A key strength of this study is that it addresses an important public health question for decision makers in the light of uncertain information, synthesizes many sources, and reflects uncertainty. This emphasizes the value of modeling given the difficulty to conduct clinical research in a rapidly changing environment. Key strengths of this study also include that it is based on 2 large open-label trials: the RECOVERY trial 21 and SOLIDARITY. 6 This study also focuses on patients treated with RDV on supplemental oxygen at the start of treatment (defined as LFO, HFO, or NIV) in line with its European marketing authorization and NHS England commissioning policy. 3, 4 The duration of hospitalization is calculated from the AUC of OS and time to discharge curves with outcomes extrapolated beyond the trial duration to account for the fact that a proportion of patients are neither dead nor discharged by day 28 in the respective trials. Albeit with some simplification because of limitations with the evidence base, this economic evaluation attempts to include some of the short-, medium-, and long-term effects associated with COVID-19 (elevated risk of death, reduced QOL, elevated risk of multiorgan dysfunctions, and increased monitoring) alongside movement among the different hospitalization oxygen requirements (with costs and QALYs calculated accordingly). Uncertainty in the model inputs and assumptions has also been explored extensively in a number of scenario and sensitivity analyses.

Results are not directly comparable with other published economic evaluation for RDV [30] [31] [32] given that there are a number of important differences in terms of population considered, comparators, methodology, and evidence used. Previous published economic evaluations 30-32 at the time this analysis was conducted used a decision-tree approach with patients considered either recovered or dead during the first month, followed by a Markov model or payoff. The population included in previous economic evaluations [30] [31] [32] was also broader including patients not on oxygen or on invasive ventilation at the start of treatment. This differs from the licensed indication for RDV in Europe, which specifies patients with pneumonia requiring supplemental oxygen. The duration of hospitalization is also calculated differently, derived from the OS and time to discharge extrapolated curves in our model, compared with a more simplistic approach in previous economic evaluation based on the median time to recovery in recovered patients and assumptions for those who died. 30 In our model, hospitalized patients are also able to move (implicitly) between different hospitalization/oxygen requirements with costs and QALYs calculated accordingly. In Campbell et al 30 (2020) for instance, hospitalization costs are calculated based on the oxygen support required as highest level of respiratory support (one-off cost not based on duration of hospitalization) and are disconnected from the starting health state and therefore only applicable for the overall ACTT-1 population. Compared with previous economic evaluations, 30-32 our model further considers that COVID-19 patients are, on average, at an elevated risk of death, reduced QOL, and multiorgan dysfunctions after discharge. None of these effects were included in previous published economic evaluations at the time this analysis was conducted. [30] [31] [32] As with any economic evaluation, there are limitations to be acknowledged. First, any economic evaluation for COVID-19 is challenging to conduct because of the rapidly changing environment. SoC is consistently evolving and has changed since the evidence used in this economic evaluation was published. Corticosteroids are now the SoC in England and Wales, and this has been reflected in the base-case. NHS England recently issued advice for the use of TCZ and SAR. Unlike RDV, TCZ and SAR have different mechanisms of action, and it is unclear to what to extent these treatments would be considered in patients who would have been otherwise eligible for RDV in England and Wales (and be appropriate comparators). Albeit limited to the supplemental oxygen subgroup from the respective source of evidence, this economic evaluation combines evidence from different studies that are heterogeneous in population, design, and outcomes. The level of patients' oxygen requirement was also not reported.

Evidence is also constantly evolving. For instance, time to discharge had to be approximated because only the proportion of events at the end of trial was reported at the time of conducting study, but new data have now been published. Using the recently published KM was explored (Appendix Table 2 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2021.12.015). Therefore, the model needs to develop as more evidence is available on both the impact of treatments on COVID-19 and its long-term effect. Since this analysis was conducted, additional evidence of the effectiveness RDV has been published from a phase III, open-label, adaptive, multicenter RCT conducted in 48 sites in Europe (DisCoVeRy). 33 This study found that no clinical benefit was observed from the use of RDV in patients who were admitted to hospital for COVID-19 and requiring oxygen support.

The model uses an AUC approach that does not allow us to track individual patients, leading to assumptions being required. A cohort partitioned survival approach was chosen in the absence of individual patient level data from the key relevant trials/studies used in this economic evaluation and necessity to work with aggregate published data. This is a limitation because patients with COVID-19 admitted to the hospital are heterogeneous, with important factors affecting the progression of their disease. It was also not possible to conduct subgroup analysis in the absence of subgroup data reported in patients on supplemental oxygen at entry. Treatment effects from studies evaluating RDV for 10 days are used as a proxy for RDV for 5 days. The assumption of equivalence for survival is likely to be reasonable and is supported by Goldman et al 24 (2020) and Spinner et al 34 (2020) , albeit in a broader population. The trend for time to discharge is taken from SOLIDARITY where patients were treated up for to 10 days, which could have affected the decision to discharge patients. It is possible that patients on RDV in SOLIDARITY were kept longer at the hospital to finish the 10-day course.

Analyses are conducted at list prices. Any confidential discount offered to the NHS is not considered in this analysis. It is also unclear whether RDV reduces or increases ventilation/oxygen support because of mixed evidence.

Since this economic evaluation was conducted, 2 UK studiesthe PHOSP-COVID collaborative group 10 and ISARIC 11 -reported estimates on QOL before COVID-19 and for lower mean age. Therefore, it is unclear whether the additional decrement in utility associated with comorbidities included in this analysis was required and led to double counting. This was explored in a scenario analysis and led to an improvement in the ICER (Appendix Table 2 in Supplemental Materials found at https://doi.org/10.1 016/j.jval.2021.12.015).

Assumptions were required to capture the effect of COVID-19 in the medium to long term. It is unclear how long discharged COVID-19 patients are at an elevated risk of death or reduced QOL. It is also unclear whether multiorgan dysfunctions reported in the literature are acute (temporary) or chronic. The approach to capture costs and impact on QOL associated with multiorgan dysfunction is simplistic because of the heterogeneity in patients experience and does not take into account patients with post-COVID-19 syndrome and the recommendation for them to have rehabilitation. It also does not take account of other reported longterm effects on mental health that have been reported such as posttraumatic stress disorder. The duration and frequency of monitoring after discharge of patients with COVID-19 are also unknown and challenging to capture because of the heterogeneity in patient experience. The mortality, costs, and morbidity impact associated with readmission because of COVID-19 is not included separately to avoid double counting because evidence suggests that readmission because of COVID-19 occurs shortly after the initial hospitalization episode, typically within 5 to 10 days. 5, 35 The model conservatively assumes that patients initiated on RDV or SoC experience the same long-term outcomes, in the absence of evidence. Consequently, any short-term difference in survival will translate into commensurate gain in the long-term. It is possible that outcomes for patients initiated on RDV may be worse if more patients require ventilation because of the reduced death rate.

RDV is likely to represent a cost-effective use of NHS resources if it reduces death and does not increase ventilation requirement. RDV is likely to be more cost-effective in patients requiring LFO at entry only than those requiring more intensive HFO or NIV.

Supplementary data associated with this article can be found in the online version at https://doi.org/10.1016/j.jval.2021.12.015. .£1M, more than £1 million. *In this scenario, RDV the treatment effect for the probability of ventilation in patients initiated on SoC (compared with RDV) is assumed to be same as RDV. 

Coronavirus disease 2019 (COVID-19): current status and future perspectives

A review of coronavirus Disease-2019 (COVID-19)

Interim Clinical Commissioning Policy: remdesivir for patients hospitalised with COVID-19 (adults and children 12 years and older), version 2

Veklury summary of product characteristics

Remdesivir for the treatment of COVID-19 -final

Repurposed antiviral drugs for COVID-19 -interim WHO solidarity trial results

Remdesivir in adults with severe COVID-19: a randomised, double-blind, placebo-controlled, multicentre trial [published correction appears in Lancet

Epidemiology of post-COVID syndrome following hospitalisation with coronavirus: a retrospective cohort study

Postdischarge symptoms and rehabilitation needs in survivors of COVID-19 infection: a cross-sectional evaluation

Physical, cognitive and mental health impacts of COVID-19 following hospitalization (PHOSP-COVID): a multi-centre prospective cohort study

Long COVID in adults discharged from UK hospitals after COVID-19: a prospective multicentre cohort study using the ISARIC WHO Clinical Characterisation Protocol

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

Populating an economic model with health state utility values: moving toward better practice. Value Health

Impact of recurrent Clostridium difficile infection: hospitalization and patient quality of life

Impact of influenza on health-related quality of life among confirmed (H1N1)2009 patients

Drugs and pharmaceutical electronic market information tool (eMIT)

National cost collection for the NHS: archived reference costs

Are lower levels of red blood cell transfusion more cost-effective than liberal levels after cardiac surgery? Findings from the TITRe2 randomised controlled trial

Unit costs of health & social care 2020

Dexamethasone in hospitalized patients with COVID-19

Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves

Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): preliminary results of a randomised, controlled, open-label, platform trial

Remdesivir for 5 or 10 days in patients with severe Covid-19

Remdesivir effectiveness evaluation study

Using health state utility values from the general population to approximate baselines in decision analytic models when conditionspecific data are not available

Preexisting comorbidities predicting COVID-19 and mortality in the UK Biobank community cohort

National Institute for Health and Care

Alternative pricing models for Remdesivir and other potential treatments for COVID-19

Treatment of moderate to severe respiratory COVID-19: a cost-utility analysis

Cost-effectiveness of remdesivir and dexamethasone for COVID-19 treatment in South Africa

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, open-label trial

Effect of remdesivir vs standard care on clinical status at 11 days in patients with moderate COVID-19: a randomized clinical trial

COVID-19: outcomes of patients with confirmed COVID-19 re-admitted to hospital

We thank Ruth Wong, School of Health and Related Research, for conducting searches for the clinical and cost-effectiveness evidence. Ethical Approval: This research did not collect primary data from individuals, or have any access to individual level data and therefore did not require ethical approval.

The authors reported no conflicts of interest.Funding/Support: This project funded by the National Institute for Health and Care Excellence through the Decision Support Unit.Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The funder was able to comment on previous drafts of this article and provide comments for consideration by the authors.