Effects of an electronic health record-based mobility assessment and automated referral for inpatient physical therapy on patient outcomes: A quasi-experimental study.

Abstract

To assess the effectiveness of a hospital physical therapy (PT) referral triggered by scores on a mobility assessment embedded in the electronic health record (EHR) and completed by nursing staff on hospital admission. EHR and billing data from 12 acute care hospitals in a western Pennsylvania health system (January 2017-February 2018) and 11 acute care hospitals in a northeastern Ohio health system (August 2019-July 2021). We utilized a regression discontinuity design to compare patients admitted to PA hospitals with stroke who reached the mobility score threshold for an EHR-PT referral (treatment) to those who did not (control). Outcomes were hospital length of stay (LOS) and 30-day readmission or mortality. Control variables included demographics, insurance, income, and comorbidities. Hospital systems with EHR-PT referrals were also compared to those without (OH hospitals as alternative control). Subgroup analyses based on age were also conducted. We identified adult patients with a primary or secondary diagnosis of stroke and mobility assessments completed by nursing (n=4859 in PA hospitals, n=1749 in OH hospitals) who completed their inpatient stay. In the PA hospitals, patients with EHR-PT referrals had an 11.4 percentage-point decrease in their 30-day readmission or mortality rates (95% CI -0.57, -0.01) relative to the control. This effect was not observed in the OH hospitals for 30-day readmission (β=0.01; 95% CI -0.25, 0.26). Adults over 60 years old with EHR-PT referrals in PA had a 26.2 percentage-point (95% CI -0.88, -0.19) decreased risk of readmission or mortality compared to those without. Unclear relationships exist between EHR-PT referrals and hospital LOS in PA. Health systems should consider methodologies to facilitate early acute care hospital PT referrals informed by mobility assessments.