Characterizing longitudinal post-acute rehabilitation utilization and outcomes during the first year after stroke in Medicare beneficiaries

PROJECT SUMMARY/ABSTRACT I am seeking a Mentored Research Scientist Development Award (K01) in order to acquire the necessary training, practical experience, and knowledge to become a leading independent investigator focused on improving the value of rehabilitation after stroke through an understanding of person and system level

variability. This award will provide the necessary support to achieve my Career Development Aims related to 1) health services research, 2) the application of machine learning approaches, 3) scientific communication, and 4) leadership and management skills. To achieve these goals, I have assembled a Mentoring Committee

comprised of: Dr. Amit Kumar (primary mentor), a health services researcher and expert in Medicare data; Dr. Julio Facelli (co-mentor), a biomedical informatician and recognized expert in the application of machine learning clustering approaches to healthcare data; Dr. Angela Presson (co-mentor), an epidemiologist with

expertise in predictive modeling wit machine learning approaches. I have also created an Advisory Committee of individuals with expertise in stroke care, rehabilitation, and health services research. Individuals with stroke experience multiple care transitions from acute hospitals to post-acute settings and

between post-acute settings. Post-acute rehabilitation (PAR) is essential to recovery after stroke, yet many individuals receive inadequate and interrupted PAR, which contributes to poor continuity of care, inefficient care, and poor outcomes. To improve PAR utilization, we must transformation our healthcare system to

improve the value of PAR after stroke. However, because past work has only focused on PAR utilization cross sectionally, we lack the foundational understanding of longitudinal PAR utilization that is needed to reshape post-acute stroke care. My Scientific Aims directly address this gap and are to: 1A) identify subgroups of

longitudinal PAR utilization patterns during the first year after stroke, 1B) characterize subgroups of PAR utilization with patient-, facility-, and community-level factors, and 2) determine the predictive strength of PAR utilization on one-year outcomes after stroke. The proposed work is significant because it will provide the required understanding of PAR utilization after

stroke that is needed to improve the continuity of care and the value of rehabilitative care. The proposed work is innovative because it focuses on longitudinal PAR and long-term outcomes, uses cutting edge approaches to access and analyze Medicare data, and incorporates nationally representative data sources. The expected

outcome will 1) lay the groundwork for subsequent R01 submissions examining the impact of PAR utilization on functional outcomes and validating models of PAR utilization and outcomes and 2) facilitate my transition to independence.