Abstract 118: Physician Specific Practice Patterns and 30-day Heart Failure Readmission Rates: Is One Way Better Than the Other?

Abstract

Listen

Translate article

Background: Heart failure (HF) admissions place a significant burden on hospital resource utilization. Management of inpatient HF remains heterogeneous, from daily diuresis to discharge strategies. We sought to evaluate if specific physician practices were associated with two important contributors to resource utilization, readmission and length of stay (LOS). Methods: Physician-specific discharges, LOS, and 30-day readmission rates from 7/1/2015-6/30/2016 were extracted from the EHR. Patients with “shock” as the primary diagnostic code were excluded. We created a survey to assess HF management strategies that was administered to all cardiologists and hospitalists at our institution; physicians were blinded to LOS and readmission data. Many factors potentially impacting LOS and discharge decisions were queried, including physical examination, medication use and kidney function. Results: We received 62 of 69 (90%) survey responses – 26 cardiologists and 36 hospitalists. Overall, 52% (32 of 62) were male with a median 7 years in practice [IQR 3-13]. We tracked HF utilization metrics in 58 physicians, encompassing 753 patient discharges in the preceding year. Total 30-day readmission rate was 17.0% (128 of 753). No significant differences were seen in discharge practices between readmission rate tertiles. Median of the median LOS was 4.7 days (IQR 4.0-5.8), and physicians above median LOS were compared to physicians below median LOS. Point estimates suggest physicians with lower median LOS targeted a daily net diuresis goal greater than 3L (40% vs 11% for those above median LOS, p = 0.1), and were less likely to hold beta blockers (17% vs 37%, p = 0.1) or to consider 24 hours of inpatient oral diuretics prior to discharge to be “very important” (16% vs 56%, p < 0.01). The pooled readmission rate of physicians below median LOS was not significantly different than those above median LOS (14.5% vs 16.2%, p = 0.6). Conclusions: We identified three specific HF management patterns that may be associated with lower LOS without a concomitant increase in 30-day readmission: (1) more aggressive daily diuresis, (2) continuation of beta blockade, and (3) not delaying discharge to observe 24 hours of an oral diuretic regimen. Further studies are warranted to confirm and quantify the impact of these changes on resource utilization.