Understanding the relationship between costs and the modified Rankin Scale: A systematic review, multidisciplinary consensus and recommendations for future studies.

Background and purpose: Cost-of-illness studies often describe a single aggregate cost of a disease state. This approach is less helpful for a condition with a spectrum of outcomes like stroke. The modified Rankin Scale is the most commonly used outcome measure for stroke. We sought to describe the existing evidence on the costs of stroke according to individual modified Rankin Scale categories. This may be useful in future cost effectiveness modelling studies of interventions where cost data have not been collected, but disability outcome is known. Methods: Systematic review of the published literature, searching electronic databases between 2004 and 2015 using validated search filters. Results were screened to identify studies presenting costs by individual modified Rankin Scale categories. Results: Of 17,782 unique identified articles, 13 matched all inclusion criteria. In only four of these studies were costs reported by modified Rankin Scale categories. Most studies included direct medical costs only. Societal costs were assessed in two studies. Overall, studies had a high methodological and reporting quality. The heterogeneity in costing methods used in the identified studies prevented meaningful comparison of the reported cost data. Despite this limitation, the costs consistently increased with greater severity (increasing modified Rankin Scale score). Conclusions: Few cost studies of stroke include information based on stroke recovery measured by individual modified Rankin Scale categories and the existing data are limited. To reliably capture this information, future studies are needed that preferably apply standardised costing methods to promote greater potential for use in cost-effectiveness analyses whereby direct collection of patient-level resource use has not been possible.