Abstract 16644: Independent Predictors of 30-Day Readmission After Acute Myocardial Infarction at a Large, Inner City Medical Center

Background: Acute myocardial infarction (AMI) accounts for one of the most common reasons for cardiovascular disease admission in the United States. Despite modern therapeutics, 15-20% of index admissions for AMI result in unplanned readmission within 30 days of discharge. In tailoring efforts to decrease AMI 30 day readmissions we sought to identify independent predictors for readmission. Methods: All 1771 patient discharges from Johns Hopkins Hospital and Johns Hopkins Bayview Medical Center in fiscal years 2016 and 2017 with a primary discharge diagnosis of AMI were included. The age, sex, race, insurance status, aggregated AHRQ comorbidity count, concomitant diagnosis of specific comorbidities, discharging hospital, discharging service, destination at discharge, diagnosis (NSTEMI vs. STEMI), procedures during index admission and length of index admission for each discharge were recorded and odds ratios with 95% confidence intervals for 30 day readmission to a Maryland hospital were calculated. If a patient was readmitted to a Hopkins hospital the reason for readmission was recorded. Results: Of the 1771 patient discharges over the study period, 273 were excluded because the patient was not eligible for readmission in CMS reporting algorithms (i.e. same or next day transfer, death), had a planned readmission, or was <18 years old. Of the remaining 1498 discharges, 264 (17.6%) were readmitted within 30 days. In a multivariable model, major independent predictors of 30 day readmission risk were high comorbidity burden (OR for 4-5 vs. 0-1 comorbidities: 3.27, 95% CI 1.55 - 6.88; OR for > 5 vs. 0-1 comorbidities: 6.86, 95% CI 3.2 -14.71) and leaving the hospital AMA (OR 3.77, 95% CI 1.63-8.72). In a separate model considering specific comorbidities, a concomitant diagnosis of CHF (OR 1.94, 95% CI 1.4-2.67), renal failure (OR 1.76, 95% CI 1.27-2.45), or depression (OR 1.61, 95% CI 1.12-2.33) were independent predictors of readmission risk. Of patients whose readmission diagnoses were known (n=143), the majority (n=75, 52.4%) were readmitted with cardiac diagnoses. The most prevalent cardiac readmission diagnoses were CHF (n=28), NSTEMI (n=25), and chest pain (n=15). Conclusions: In our large, inner city hospital system, patients with a greater number of comorbidities, in particular CHF, renal failure, or depression, were at significantly greater risk for 30 day readmission after AMI. Identifying high risk patient characteristics is an important first step in designing tailored, system-wide initiatives to reduce AMI readmissions.