Abstract WP13: Derivation and Validation of Utility Weights for the Modified Rankin Scale From the AcT Thrombolysis Trial

Ayooluwanimi Okikiolu,Olayinka Arimoro,Nishita Singh,Fouzi Bala,Ayoola Ademola,Mohammed A Almekhlafi,Aravind Ganesh,Brian H Buck,Richard H Swartz,Michael D Hill,Bijoy K Menon,Tolulope Sajobi

DOI: https://doi.org/10.1161/str.55.suppl_1.wp13

IF: 10.17

2024-02-01

Stroke

Abstract:Introduction: Utility-weighted mRS (UW-mRS) is increasingly being adopted as an important clinical endpoint in stroke trials as it offers some gains in statistical efficiency and power when estimating treatment effects. The widely used utility weights for estimating UW-mRS, derived from a systematic review of acute stroke trials, are prone to substantial variability across and between mRS categories. This study aims to derive utility weights for UW-mRS by directly mapping 5-item EuroQoL (EQ-5D-5L) responses to mRS scores. Methods: Data were collected from all patients included in Alteplase compared to Tenecteplase (AcT) trial. Quality of life at 90 days post-randomization was assessed using EQ-5D-5L. Health utilities (EQ-5D index) were estimated using the time trade-off approach based on Canadian norms and imputed as zero for patients who died. Using a predictive linear model, utility weights were derived by regressing the ordinal mRS on the utilities. Model performance was measured using R 2 and root mean square error (RMSE) after 5-fold cross-validation. We compared the distribution of the UW-mRS scores using the systematic review and model-based weights. Results: Of the 1503 acute stroke patients who completed the EQ-5D-5L questionnaire at 90 days, 717 (47.7%) were female, and the median (interquartile range [IQR]) age was 74.0 (20.0). The median (IQR) estimated health utility was 0.81 (0.52). The utility weights for mRS categories 0-6 were 0.93, 0.90, 0.80, 0.67, 0.41, 0.24, and 0.00. The model performed well after 5-fold cross-validation (R 2 = 0.89; RMSE = 0.12). The model-based weights had higher mean and median UW-mRS scores but smaller variability across mRS levels than those from the systematic review-based weights. There was no statistically significant difference in the median UW-mRS scores across both treatment groups for each utility weight type. Conclusion: The estimation of UW-mRS scores largely depends on the choice of weights used, which depends on sample heterogeneity in treatment interventions and stroke severity of the cohorts. Model-based approaches result in reduced variability between and across mRS levels. Future research will seek to validate these utility weights externally in an independent cohort of patients with stroke.

peripheral vascular disease,clinical neurology