Development of a decision-analytic model of stroke care in the United States and Europe

Abstract

Objective: Stroke places a huge burden on society in terms of premature death, disability, and costs of care. Increasingly, the cost-effectiveness of new interventions needs to be demonstrated before their widespread implementation. Clinical trials are unable to measure the long-term impact of such new interventions in stroke care, and a modeling approach is necessary. The Stroke Outcome Model has been developed in four countries: France, Germany, the United Kingdom, and the United States as a flexible tool for this purpose. Method: The decision-analytic model represents the management of acute stroke and long-term care and prevention of recurrence for stroke survivors. The latter consists of semi-Markov state-transition processes, with health states defined by therapy, disability, and occurrence of further stroke. Sources of clinical data include trials, meta-analyses, and prospective cohort studies such as the Oxfordshire Community Stroke Project and the Northern Manhattan Stroke Study. Resource use data were obtained from published sources and expert clinician panels. Outcome measures used were strokes averted, life years, and quality-adjusted life-years gained. Results: The model has been used to undertake economic analyses of antiplatelet therapy for the prevention of recurrent strokes, and of stroke unit care and thrombolytic therapy in acute stroke. From a health- and socialcare perspective, new interventions were found to be cost saving or to provide health benefits at modest additional cost. Results were sensitive to the cost perspective, time horizon, baseline risk of stroke recurrence, and choice of effectiveness measure. Conclusion: The development of this model highlights the need for improved information on prognosis and resources used by stroke survivors and the importance of differentiating between economically distinct end points such as death, disabled survival and nondisabled survival, which may be combined as outcomes in clinical trials.