Abstract
BACKGROUND: Eligibility for thrombolysis as an acute stroke treatment is determined through the use of unenhanced noncontrast computed tomography (CT), time since stroke onset, and patient history. Assessing penumbral patterns, which can be examined only through the use of diagnostic technologies such as magnetic resonance imaging (MRI) and perfusion CT (CTP), may be able to better select patients for thrombolysis. However, trade-offs in terms of administration time and cost may affect the value of using these diagnostic studies. OBJECTIVE: We examined the trade-offs among patient selection via usual care with CT, usual care plus MRI using diffusion-weighted and perfusion imaging, and usual care plus CTP for their effect on costs and outcomes when diagnosing stroke and selecting candidates for thrombolysis in the United Kingdom. METHODS: A decision-analytic model was developed. Efficacy and utilities were obtained from published studies. Costs were obtained from standard UK costing sources and were supplemented with data from the published literature. Outcomes included a favorable outcome (modified Rankin Scale score <2), costs, life-years, quality-adjusted life-years, and incremental cost-effectiveness ratios. RESULTS: Compared with usual care selection, adding CTP or MRI to better select patients for thrombolysis reduced the number of patients receiving thrombolysis by 9 and 14.6 per 1000 patients treated, respectively, while improving favorable outcome (19.2 and 17.6 per 1000 patients treated, respectively). In both scenarios, costs were decreased slightly. Both CTP and MRI selection were cost saving (more efficacious and less costly) compared with unenhanced CT selection; CTP selection was found to dominate MRI selection. CONCLUSIONS: Adding diagnostic tests such as CTP and MRI to select UK patients for thrombolysis may be a good value for the money and may improve patient outcomes. If a preferred diagnostic test had to be chosen based on economic value, CTP might be the best compromise between unenhanced CT selection and MRI selection.