Background: Hospitalized patients often require post-acute care (PAC) after discharge ranging from long-term acute care hospitals (LTACH) to home health services. However, patients waiting for discharge to PAC are at risk for delayed discharges [1], which can increase in-hospital complications, increase costs, and decrease access to hospital care for other patients. [2, 3] Disposition prediction has been attempted, but there are large gaps in evidence regarding its effectiveness. [4] We have previously found that the health confidence score (HCS) is associated with hospitalized general medicine patients’ hospital length of stay. [5, 6] We aim to understand if health confidence can be a simple and efficient way to predict hospital disposition.
Methods: This prospective cohort study was performed at an urban academic medical center from January 21, 2022, to October 23, 2022. The HCS was administered to patients as early as possible in their hospital stay as a part of a larger ongoing survey of all inpatients admitted to the hospitalist service. The survey data were combined with visit characteristics and discharge disposition, which were obtained from the electronic medical record. Population characteristics were defined by frequencies or mean and standard deviation (SD). We used bivariate analysis with ANOVA to establish the basic association. Multinomial logistic regression (with discharging to home without PAC as the base group) was used to assess the relationship of HCS with discharge disposition. Statistical significance was defined by p< 0.05.
Results: A total of 2,797 patients completed the HCS, of whom 55% identified as female, 70% identified as Black/African American, and 87% did not identify with Hispanic/Latino ethnicity. The mean age was 60 years (SD 17, range 19-105). During the study period, the mean HCS was 9.19 (SD 2.68, range 0-12) and 61% of patients had an above-average HCS, which we defined as an HCS≥9. For discharge disposition, 57% of patients were discharged to home without PAC, 14% to home with home health (HH) services, 13% to a skilled nursing facility (SNF), 3% to acute rehabilitation (AR) facilities, 1% against medical advice (AMA) and < 1% to LTACH or home/facility-based hospice. HCS was significantly different (p< 0.01) between the groups of patients discharged to different PAC types with average HCS highest for AMA (9.54) and home without PAC (9.50) while lowest for LTACH (7.78) and Hospice (6.35). In multinomial logistic regression, compared to discharge home without PAC, an increase in HCS of 1 point was associated with a 0.95 relative risk ratio (p< 0.01) of discharging to Home with HH, 0.88 (p< 0.01) of discharging to AR, 0.84 (p< 0.01) of discharging to home without PAC vs SNF, 0.80 (p=0.05) of discharging to LTAC and 0.70 (p< 0.01) of discharging to Hospice. HCS was not associated with the risk of discharging AMA vs discharge home without PAC.
Conclusions: To the best of our knowledge, this is the first study examining the relationship between HCS and hospital disposition. We found that patients with higher HCS at admission were less likely to have a discharge disposition with PAC services. One possible explanation for this could be that patients with greater health confidence are better able to care for themselves. Regardless, our findings suggest that HCS may have clinical utility in identifying hospitalized patients who may need PAC, potentially improving triage for case management services and planning for timely hospital discharge.