Comparative Effectiveness of Alternative Levels of Stroke

University Of Pennsylvania, Philadelphia PA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by the applicant): The policy relevance of this work is critical. Without evidence of effectiveness, multidisciplinary inpatient rehabilitation access will likely be curtailed or eliminated to the detriment of people with potentially reducible functional limitations and disabilities related to stroke. Changes in both Veterans Health Administration (VHA) and Medicare policies are shifting care away from high level inpatient multidisciplinary rehabilitation services without empirical evidence either supporting or not supporting lower levels of service. There are few, if any, sources of standardized data in the private sector available to address the benefits of alternative levels of rehabilitation. Data from the Department of Veterans Affairs (VA) will be applied to gain insights about optimal disability management strategies for Medicare and Medicaid. We propose to compare 2 different levels of inpatient rehabilitation (acute rehabilitation services received while patients are still hospitalized for stroke). The first level is consultation where patients are seen by rehabilitation professionals while remaining on non-rehabilitation bed services within the hospital. The second level is higher intensity specialized. multidiciplinary rehabilitation services which occur when patients are admitted onto a specialized rehabilitation bed unit within the hospital. Primary outcomes will include functional recovery and home discharge from the acute hospital stay. One-year survival, long-term care placement, rehospitalization, and total health care costs will be secondary outcomes. The study is a retrospective observational comparative effectiveness study using data merged from 14 large administrative databases compiled from 153 VA Medical Centers. Methods will include a series of observational studies on an estimated cohort of over 15,000 stroke patients using propensity score matching and instrumental variable analyses to adjust for selection bias, and mixed models to account for patient clustering within facilities. These methods will be applied to compare the outcomes of veterans who receive consultation rehabilitation, specialized rehabilitation, and no evidence of inpatient rehabilitation. We will identify the types of patients most likely to receive any form of acute rehabilitation, and among that group those most likely to receive specialized rehabilitation. This research is expected to greatly expand the limited body of empirical knowledge of relevance of acute rehabilitation following stroke. Such evidence-based knowledge will be essential to guide future practice and policies in the VHA and private sectors alike.