Associating Health-Related Quality-of-Life Score with Time Uses to Inform Productivity Measures in Cost-Effectiveness Analysis

• Published in: PharmacoEconomics Vol. 41; no. 9; pp. 1065 - 1077
• Main Authors: Jiao, Boshen, Basu, Anirban
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2023; Springer; Springer Nature B.V
• Subjects: Algorithms; Cost analysis; Cost benefit analysis; Econometrics; Economic aspects; Economic research; Health Administration; Health aspects; Health care expenditures; Health Economics; Health surveys; Households; Intervention; Labor market; Labor productivity; Medical care, Cost of; Medicine; Medicine & Public Health; Methods; Original Research Article; Patients; Pharmacoeconomics and Health Outcomes; Productivity; Prostate cancer; Public Health; Quality of life; Quality of Life Research; Society; Time use; Valuation; United States; United States > US
• ISSN: 1170-7690; 1179-2027
• DOI: 10.1007/s40273-023-01246-x

Background The Second Panel on Cost Effectiveness in Health and Medicine recommended that cost-effectiveness analyses (CEA) explicitly incorporate the valuation of productive time from a societal perspective. We developed a new approach to capture productivity impacts in CEA without direct evidence on these impacts by associating varying levels of health-related quality-of-life (HrQoL) score with different time uses in the United States. Methods We conceptualized a framework that estimates the association between HrQoL score with productivity through time uses. We used the American Time-Use Survey (ATUS) from year 2012–2013, when data on a Well-Being Module (WBM) was additionally collected alongside ATUS. The WBM measured the quality of life (QoL) score using a visual analog scale. To operationalize our conceptual framework, we employed an econometric approach that addressed three technical issues in the observed data: (i) distinction between overall QoL and HrQoL, (ii) correlation across different categories of time use and the share structure of time-use data, and (iii) reverse causality between time uses and HrQoL score in a cross-sectional setting. Furthermore, we developed a metamodel-based algorithm to summarize the numerous estimates from the primary econometric model efficiently. Finally, we illustrated the use of our algorithm to calculate productivity and time spent seeking care costs in an empirical CEA of a prostate cancer treatment. Results We provide the estimates of the metamodel algorithm. Incorporating these estimates into the empirical CEA reduced the incremental cost-effectiveness ratio by 27%. Conclusion Our estimates can facilitate the inclusion of productivity and time spent seeking care in CEA as recommended by the Second Panel.