Use of anti-viral therapies in hospitalised COVID-19 patients in the United Arab Emirates: a cost-effectiveness and health-care resource use analysis

Severe Acute Respiratory Syndrome Corona Virus type 2 (SARS-CoV-2) emerged at the end of 2019 as a novel coronavirus disease (COVID-19) which rapidly spread worldwide. As of the end of January 2022, there were an estimated 378 million cases of COVID-19 globally, with nearly six million deaths reported [1]. In the United Arab Emirates (UAE), the virus was first detected in March 2020, and there have been approximately 847,142 cases since the beginning of the pandemic, with about 2,248 deaths (January 2022) [2]. When assessed against the priority areas of strategic preparedness issued by the WHO, the UAE promptly deployed risk communication and public engagement, applied contact tracing and management and set up multiple field hospitals to ease the pressure on city-based hospitals and to aid faster response to COVID-19 [3]. In addition, the number of tests conducted per 1000 people in the UAE is one of the highest worldwide [3].

Patients infected with COVID-19 can experience various symptoms, some of which can lead to a substantial clinical burden and premature mortality. Disease progression can occur rapidly and in critical cases can result in mechanical ventilation (MV). Various treatments were developed to fight COVID, including anti-viral therapies and monoclonal antibodies. Early in the pandemic, one of the few approved therapies to treat moderate to severe hospitalized COVID-19 patients was remdesivir (RDV), an antiviral therapy which inhibits the viral RNA-dependent RNA polymerase (RdRp), essential for viral replication [4]. RDV (marketed as Veklury®) received emergency use authorization in May 2020 from the FDA, with full approval in adults in children over 12 years of age in October 2020 [5]. In Europe, RDV received conditional marketing authorization in July 2020 from the EMA and full approval in December 2021 [6]. In the UAE, RDV was granted full regulatory approval in June 2021 [7]. The regulatory approvals were primarily based on the pivotal ACTT-1 trial [4], which compared RDV plus standard of care (SOC) versus placebo plus SOC. The ACTT-1 trial results showed that treatment with remdesivir leads to a shorter time to recovery (15 vs. 10 days) with an increased recovery rate of 29%. In a post-hoc analysis, RDV was associated with a 70% reduction in mortality for patients who required low-flow oxygen support. The findings of the ACCT-1 trial have been supported by similar results from other trials [8] and in real-world comparative effectiveness studies [9, 10]. RDV has also shown continued activity against SARS-COV-2 variants [11]. Favipiravir (FAVI) is another antiviral which inhibits RdRp which at the time of writing (Feb 2022) was being tested against SARS-COV-2 in clinical trials, and several smaller studies had concluded mixed results as to FAVI effectiveness [12‐15]. No evidence was identified for the efficacy of FAVI on different variants.

As the pandemic develops and new evidence emerges, guidelines are constantly reviewed and revised. In the 2021 treatment guidelines from the UAE, treatment for hospitalized patients with COVID-19 is based on oxygenation status and risk of progression [16]. For patients with pneumonia or severe pneumonia who require supplemental oxygen, RDV or FAVI is recommended alongside standard of care (SOC)(i.e. dexamethasone or equivalent corticosteroid and other symptom-alleviation related drugs).

Given the strain that COVID-19 care can put on the health care system, it is important to ensure patients are receiving treatments that are delivering good value for money. This study is aims to investigate the cost-effectiveness of RDV + SOC for COVID-19 patients receiving low-flow oxygen in the UAE in line with clinical practice.

A health economic model (see Fig. 1) was developed in Microsoft Excel 2016 and conformed to the International Society for Pharmacoeconomic and outcomes Research (ISPOR) Modelling Good Practice Guidelines [17]. The model adopts a health system perspective. As SARS-COV-2 is an acute disease, and data on long-covid impacts are only just emerging, the model assumes a relatively short time horizon of one year. The model cycle length was two weeks, which reflects the primary clinical data as well as real-world information on hospital stays, with a majority of patients leaving hospital (or dying) within a 2-week period and most patients recovering or resolving by one month [18]. For this acute phase, the model adopts a decision tree approach, followed by a Markov model for the remaining time horizon, with cycle length continuing to be two weeks.

A cohort of patients was modelled via a decision tree for the patient’s potential time in hospital (time horizon 29 days) followed by a Markov model for the remaining time horizon (1 year). On day 15, patients transition into one of the following health states based upon their baseline ordinal score (OS) as defined in the ACTT-1 trial, with a minor modifications to reflect local clinical practice [4]: discharged (OS 1–3 plus a negative test), OS 1–3 – awaiting negative test, general ward with no supplemental oxygen (OS4), general ward with low-flow oxygen (OS5), ICU with non-invasive ventilation (NIV) (OS6), ICU with mechanical ventilation (MV) (OS7) or death (OS8). On day 29, patients have either died or assumed to have been discharged and are alive. After day 29, patients enter a Markov model, subject to age- and sex-adjusted mortality derived from UAE-specific lifetables [19]. Patients who are alive can be re-hospitalized post-discharge incurring hospital costs and outcomes associated with OS5 in either the decision tree or the Markov portion of the model.

The model was designed to estimate the number of deaths avoided, hospital days avoided (by ward, ICU and ICU + MV) as well as the associated costs for each arm. Long-term effects of COVID-19 are not considered as part of this model due to a lack of information on the impact of treatment. The model discounts both costs and outcomes at 3.5%, however, as the time horizon is one year, this is not applicable for this analysis. Half-cycle correction is also unnecessary as benefits and costs are almost entirely accrued within the first month with the exception of the natural background mortality in the post-acute Markov cycles which is equal for both arms.

The base-case cohort contained 4,156 patients, with a mean age of 38.4 and 69% of the cohort was male. In the base case, the model estimated 158 deaths in the RDV + SOC arm, compared to 532 and 513 deaths in the SOC alone and FAVI + SOC arms. This means that the model estimated an incremental difference of 374 deaths compared to SOC alone and 355 deaths compared to FAVI + SOC.

The model estimated that using RDV + SOC avoided a total of 11,338 general ward days, 7,003 ICU days, and 5,451 ICU + MV days when compared to SOC alone (or FAVI + SOC) (Table 2). This equates to a reduction of approximately six days hospital stay per patient compared to SOC alone and FAVI + SOC. The model estimated that treatment with FAVI + SOC resulted in 19 fewer deaths than SOC alone. In this comparison, no hospital day reductions were seen as FAVI + SOC efficacy in reducing progression was assumed to equal SOC given the lack of clinical data.

The model estimated that the estimated acquisition cost of RDV was approximately USD 12,434,625 and the cost of FAVI was approximately USD 489,927. Due to the reduced length of stay, the difference in hospitalization costs between RDV + SOC and SOC alone were USD -26,230,587 and USD -26,251,033 versus FAVI + SOC. This resulted in a cost-savings associated with RDV + SOC of USD -13,795,962. As most of the savings arose from reducing the number of hospitalization days, the use of FAVI + SOC resulted in a total cost increase of USD 510,374. Table 2 provides an overview of the detailed cost results.

The cost-effectiveness results by outcome measure are provided in Table 3. RDV + SOC was less costly and more effective than SOC alone or FAVI + SOC, RDV + SOC was dominant across outcome measures. FAVI + SOC was more costly but had no effect on hospitalization days and consequently was dominated by SOC alone. FAVI + SOC was marginally more effective than SOC in terms of deaths avoided, and therefore an incremental cost-effectiveness ratio (ICER) of USD 27,529 was estimated. Assuming a cost-effectiveness threshold equivalent to GDP per capita in 2022 (approximately USD 36,000) [26] this could be considered cost-effective.

The model has shown that the treatment of hospitalized COVID-19 who need low-flow oxygen can benefit from anti-viral treatment and that these benefits can lead to substantial health-care savings. The model estimated that treating low-flow oxygen patients with RDV + SOC resulted in areduction in deaths (between 100 and 374 depending on the scenario) as well as hospital bed days regardless of the comparator, which ultimately results in substantial health-care savings (USD 13.8 million) over one-year. These findings were substantiated in all scenarios tested as well as the parameter uncertainty (one-way and probabilistic).

While the model results appear robust, the model is sensitive to the relative risk reduction of length of stay. As length of stay was derived from expert opinion, this is a source of uncertainty. Evidence from the real world on a cohort of patients in the population of interest could reduce this uncertainty. In addition, the data on FAVI + SOC is limited. No head-to-head data exists between FAVI + SOC and RDV + SOC, and trial data in the literature versus SOC shows mixed results. Therefore, any comparisons with this anti-viral should be interpreted with caution. Subsequent to the conduct of this study, FAVI + SOC is no longer recommended by the UAE guidelines after an examination of emerging evidence did not show a benefit over SOC alone.

As part of the validation of model results, it was apparent that the data from the ACTT-1 study had much higher rates of mortality than would be expected to be seen in the UAE based on local epidemiological data. While we have tried to control for this by conducting a scenario analysis, this artificial adjustment can lead to bias in the estimates even while it may better reflect the local population experience. It can also be noted that ignoring the potential mortality benefit entirely would not change the overall findings that the use of RDV + SOC compared to SOC alone can potentially be cost savings as the majority of the modelled benefits arise from preventing progression and reducing the overall time to discharge.

The model does not directly address the different variants associated with SARS-COV-2. However, evidence has shown that RDV remains effective against the different variants so far seen in the pandemic [11] and different real-world analyses have shown that despite differences in hospitalization rates over time, for those in hospital the length of stay and has largely remained similar over time [18], as has the progression for patients who are on low-flow oxygen. This indicates that the results of this model may be valid despite the different variants over time. Finally, the model does not consider the long-term consequences of COVID-19, which could result in an underestimation of the benefit of early anti-viral treatment. This is an area for future research.

This model has estimated the potential for substantial clinical and economic benefit in treating COVID-19 patients who require low-flow oxygen support across a wide variety of scenarios in the UAE. While the COVID-19 situation in the UAE has seen a relatively low incidence of hospitalizations and deaths compared to the rest of the world, during the peaks of the pandemics, hospital beds were at a premium. As new therapies emerge in an outpatient setting, such as the recent extension of the label for RDV [27] and nirmatrelvir + ritonavir [28], both of which saw significant reductions of hospitalization or death (87% and 89% respectively), there is a chance that earlier use of antivirals could keep many people out of hospital entirely. However, only RDV is currently used in patients who are hospitalized in the UAE. It should be noted that nirmatrelvir + ritonavir has an ongoing clinical trial in the inpatient setting. Therefore, there is likely a need for a reassessment of the cost-effectiveness of antiviral therapies for treatment of COVID-19 in the future.

As the COVID-19 pandemic continues, it is essential to have a wide variety of potential treatments that can reduce the burden of the disease on patients and the health care system. As new treatments emerge for both the outpatient and in-patient setting, it is essential to understand the best approach to maximise patient outcomes and reduce the health care system burden.