Background: The 30-day readmission rate is a key performance indicator for hospitals and hospitalist groups alike. Studies have reported a correlation between scheduling follow-up appointments and reduced readmission rates in specific patient populations (e.g. Heart Failure, COPD). However, evidence regarding the effectiveness of scheduling post-discharge appointments in general medicine patients has been mixed.
Methods: We conducted a retrospective cohort study of adults discharged from the hospitalist service at an academic medical center between July 1st, 2021, and June 30th, 2022, to investigate the effect of scheduled follow-up appointments on the rate of unplanned readmission. To control for confounders, we also collected information on other patient characteristics predictive of rehospitalization risk (e.g., length of stay, LACE score).
Results: Of the 3624 patients meeting our inclusion criteria (adults discharged to home, not against medical advice), 50.2% (1819 of 3624) had a follow up visit with either a PCP or specialist documented in their discharge instructions; 12.2% (441 of 3624) experienced an unplanned readmission within 30 days of discharge. Patients with an appointment documented in their discharge instructions had a paradoxically higher readmission rate (14.5%; 263 of 1819 discharges) than patients discharged without an appointment (9.86%; 178 of 1805 discharges), Odds Ratio 1.54 (OR 95% CI 1.26 – 1.89, P < 0.01). This difference persisted even after stratification by Length of Stay (LOS): In patients with LOS > 5 days, and patients with LOS ≤ 5 days alike, a documented follow-up appointment was associated with a higher readmission rate, OR 1.73 (CI 1.09 – 2.72, p= 0.019) and OR 1.39 (CI 1.11-1.76, p= 0.005) respectively. However, stratifying patients by LACE score attenuates this effect. A documented follow-up appointment was not associated with a higher risk of readmission in either patients with high LACE Scores (LACE > 10) (OR 95% CI 0.97 – 1.66; P=0.08) or in patients with intermediate LACE Scores (LACE 5-10) (OR 95% CI 0.95 – 1.86; P=0.10). This may be explained by providers prioritizing appointments for patients with a higher perceived risk of readmission (using a binary logistic regression model, each unit increase in LACE increases odds of having a follow-up appointment by 1.10-fold [CI 1.07-1.11]). This results in a higher mean LACE for patients with follow up appointments (mean LACE of 10 vs. 8.8, t-test P < 0.01). Higher LACE scores in turn confer a higher risk of readmission (using a binary logistic regression model, OR for LACE was 1.20 [95% CI 1.16 -1.23], i.e., each unit increase in LACE increases odds of readmission 1.2-fold)
Conclusions: Stratifying by LACE score, a scheduled follow-up appointment does not affect the readmission rate of either high LACE, or intermediate LACE patients. Therefore, the paradoxical association between scheduling follow-up appointments and an increased risk of readmission in the general patient population may be due to the confounding effect of provider teams preferentially making appointments for patients with a higher perceived risk of readmission.