PCV148 - A SYSTEMATIC LITERATURE REVIEW ON HEALTH STATE UTILITIES ASSOCIATED WITH NON-FATAL STROKE

Abstract

Cost-utility models require utilities, values representing the strength of preferences for various health states. This systematic literature review (SLR) aimed to identify and describe the utility values associated with non-fatal stroke reported in the available literature. An SLR was conducted to identify English publications in Embase®, MEDLINE® and the grey literature published in the last 10 years (2008-2017). Two reviewers independently conducted the screening, data abstraction, and the quality assessment with a third reviewer consulted in case of any discrepancies. Only primary studies reporting utilities for stroke in adult patients were included. Utilities were evaluated by geography, methodology, and severity. Sixty-five studies (sample size: 11 to 9,479) reported utilities across different geographies, including Europe (N=34), Asia (N=12), US (N=5), Australia (N=2), multinational (N=5) and other countries (N=4). Study country was not reported in 3 studies. EQ-5D was the most widely used instrument (N=51) to calculate utilities, followed by TTO (N=4), SF-6D (N=3), SG (N=3), HUI (N=3) and other scales (N=3). Average utilities ranged widely for stroke across all geographies (Europe: -0.19, 0.82; US: 0.33, 0.69; Asia: -0.23, 0.82; Australia: 0.46, 0.74; multinational: -0.48, 0.74; other countries: -0.09, 0.73; country NR: 0.34, 0.80) and methods of elicitation (EQ-5D: -0.48, 0.95; HUI: -0.23, 0.48; SF-6D: 0.66, 0.70; SG: 0.08, 0.93; TTO: 0.33, 0.90; other: 0.15, 0.83). Severity of stroke event was associated with lower utility values (severe: -0.48, 0.45; mild: 0.57, 0.95). Within studies, utilities were sensitive to time since stroke event. Utility values tend to reduce with age and female gender. The utility estimates for stroke varied greatly, likely due to differences in the assessment methodologies, patient populations, inconsistent definition of stroke and variability in time of assessment. This variability must be taken in consideration. However, the results from this SLR can be a valuable resource for future cost-utility analyses.