791 - A Predictive Model of the Acute Care Transfer (ACT) Rate at an Academic Hospital

Objectives: Prior research found that risk factors for ACT vary by setting and patient population. Prior risk prediction models designed to apply to all patients in an IRF have achieved poor to acceptable predictive value, while better performance (C-statistic >0.8) was attained by models isolated to specific impairment groups. Our aim was to determine if a general risk prediction model for ACT attained sufficient predictive utility when applied across all patients requiring IRF stay.

Design: This was retrospective review of all patients admitted to IRF between 2016-2021. Variables included were selected based on predictors that were statistically significant in prior studies. Multivariable logistic regression of ACT was fitted to cases with no missing data, using all available covariates, and its predictive utility (C-statistic) was then checked for each impairment group with >100 patients.

Results: We identified 4,120 patients admitted to IRF during the study period, and analyzed the earliest admission for each patient. After excluding 718 cases missing data, we retained 3,402 cases for analysis (including 266 ACTs, 8%). Multivariable logistic regression on the cohort demonstrated fair predictive utility (C-statistic 0.74; 95% confidence interval [CI]: 0.71, 0.77) of the derived risk score, while applying the same risk score to individual rehab impairment groups (RIC) revealed highly heterogeneous C-statistics, from 0.63 (95% CI: 0.52, 0.74) in the cardiac rehab impairment group (RIC 9.x) to 0.87 (95% CI: 0.77, 0.97) in the major multiple trauma impairment group (RIC 14.x).

Conclusions: Our derivation of a comprehensive risk score for ACT based on all patients admitted to IRF demonstrated fair predictive utility in the overall cohort, but predictive utility differed strongly across impairment groups. Further work is needed to develop personalized prediction of ACT risk based on the intersection of impairment group, clinical status at IRF admission, and pertinent medical and surgical history.