The AbilityQuotient Dashboard: Outcomes of Implementing Patient-Specific Predictive Modeling in Inpatient Team Conference

ObjectiveRecent reports have highlighted the importance of data-driven decision making as it relates to precision medicine and the field of rehabilitation. One promising method to facilitate the integration of data into patient care involves the use of data warehousing to process and host patient data, analytics to produce useful results, and dashboarding technology to disseminate analytical results to care teams in a digestible and interpretable format. This report describes the implementation of a new composite rehabilitation outcome score (cROS), the AbilityQuotient, and predictive modeling into inpatient interdisciplinary conferences through a patient data dashboard and its impact on outcomes. Design SettingInpatient Rehabilitation Hospital Participants13,397 patients completing inpatient rehabilitation from January 1, 2019 to December 31, 2023 InterventionA patient centered, composite rehabilitation outcome score (cROS) and predictive modeling dashboard implemented into team conference Main Outcome MetricsSelf-care and mobility IRF-PAI Form GG change scores, length of stay pre- and post-dashboard implementation; GG change scores compared to weighted national averages; clinician survey regarding perspectives of dashboard use; GG item long term goal modifications and goal attainment as a measure of influence on clinical plan of care ResultsAfter implementation of the patient outcomes dashboard into routine care, IRF-PAI Form GG self-care scores rose by 2.09 points and corresponding mobility scores rose by 7.18 points despite a 2.29 day reduction in length of stay. Further exploration investigating these changes as they pertained to payor revealed that these benefits occurred irrespective of insurer. Reports comparing facility to national averages extracted from eRehabData, a national outcomes data system and registry, suggest that the use of the outcomes dashboard resulted in greater reductions in length of stay and greater improvements in functional outcomes during the 2019-2023 period compared to the previous period. A corresponding survey assessing clinical perceptions of dashboard implementation revealed that it facilitated tracking and summarizing patient progress, reinforced the use of outcome metrics, and was perceived as valuable in goal setting and adjustment. Clinicians modified self-care goals six times more frequently and patients met these goals 19% more of the time after introduction of the cROS while they changed mobility goals nine times more frequently and patients met these goals 21% more of the time after introduction. ConclusionThe incorporation of individual patient data and predictive modeling into rehabilitation patient care through use of a team conference dashboard has potential to provide an objective basis from which to perform precision rehabilitation. It also has the potential to impact outcome metrics improving value-based care and consequently deserves further study.