PMU92 - CAN THE EQ-5D-3L PREDICT POST-HOSPITALIZATION HEALTHCARE USE AND OUTCOMES?

Abstract

To examine whether EQ-5D-3L completed at the time of discharge from hospital can predict post-hospitalization healthcare use and outcomes. Data from a longitudinal observational study of adults discharged from hospitals in Alberta, Canada, was used. Each of the EQ-5D-3L dimensions was analyzed as a dichotomous variable. The EQ-5D-3L misery-index, index and VAS scores were also used. Outcomes, including outpatient visits, emergency department (ED) visits, re-hospitalizations, or death at 30-days and 90-days were treated as dichotomous. Logistic regression models adjusted for age, sex, income, Charlson comorbidities score, and frailty were used. Average age of patients (N=495) was 63 years (SD1.0) and 50.5% were female. At discharge, more than half (58.4%) reported problems in mobility, 28.3% in self-care, 62% in usual-activities, 62.4% in pain/discomfort, and 42.2% in anxiety/depression. The average misery index was 7.9 (2.0), index score was 0.70 (0.22), and VAS score was 63.7 (18.4). In adjusted analysis, there were no significant associations between EQ-5D-3L at discharge with any 30-days post-discharge outcomes. For 90-days post-discharge outcomes, mobility, self-care and pain/discomfort were not associated with any outcome. However, patients with problems in usual-activities were 1.7 times (95%CI 1.1,2.6) more likely to have an ED visit and 2.0 times (1.2,3.3) more likely to get re-hospitalized/die than their counterparts. Additionally, patients with moderate-severe anxiety/depression were less likely to have an outpatient visit (OR=0.55,95%CI 0.34,0.90), but 1.7 times (1.1,2.6) more likely to have an ED visit, and 1.7 times (1.1,2.8) more likely to either get re-hospitalized/die within 90 days. The misery-index was associated with all outcomes at 90 days. The index-score was significantly associated with ED visits, and the composite outcome. VAS-score was not associated with any outcomes within 90 days. The EQ-5D-3L, particularly the usual-activities and anxiety/depression dimensions, misery index and index score, could be used to predict patients’ post-hospitalization healthcare use and outcomes.