OBJECTIVES: COVID-19 is an acute respiratory illness caused by the SARS-CoV-2 virus. Severe disease can require hospitalization and respiratory support, which can cause longer-term morbidity. Literature on patient health-related quality of life (HRQoL) is limited, particularly for severe disease (where patients may be hospitalized and/or unconscious) as HRQoL collection can be challenging or unethical. A vignette study was designed to estimate utility scores for COVID-19 disease states.

METHODS: UK adults (n=500) completed the EQ-5D-5L questionnaire for eight vignettes as patient proxies via an online survey in September 2021. Recruitment was stratified by age, sex and ethnicity to reflect the UK demographic distribution. Vignettes described different hospitalization and disease severity permutations in a qualitative framework, informed by a large UK COVID-19 infection survey. Vignette content was validated by medical experts and in a pilot study. The EQ-5D-5L crosswalk algorithm was used to estimate EQ-5D-3L utility index scores as per the National Institute for Health and Care Excellence recommendations.

RESULTS: The highest utility was reported for the state reflecting full recovery with no long-term sequelae (mean [standard deviation]: 0.87 [0.16]), followed by baseline utility pre-infection (0.73 [0.22]). The ICU state had the lowest utility (-0.37 [0.22]). Mild and moderate outpatient states reported similar utility (0.32 [0.25] vs 0.35 [0.25]). The severe general hospital ward state had lower utility than the high dependency unit state (-0.19 [0.29] vs -0.09 [0.30]). Low utility was also reported for the state reflecting recovery with long-term sequelae (0.26 [0.28]).

CONCLUSIONS: Utility values were estimated by a representative UK population sample, and vignettes were designed according to best practice guidelines. Our findings provide a reasonable alternative to patient-reported utility values, addressing gaps where patient data collection is unfeasible. Vignette studies are naturally limited by reliance on hypothetical disease states and contextualization that cannot capture all aspects of patient experience.