104\u2005Predictors of readmission in lupus compared to general medicare patients

Abstract

Background Medicare measures readmissions within 30 days of hospitalization as a marker of care quality across several conditions, although beyond hospital quality, neighborhood disadvantage is a known predictor of readmission. While not a reporting condition, a 2010 study noting 27% lupus readmissions led to our interest in examining readmission in lupus versus general Medicare including the impact of neighborhood disadvantage. Methods Using a 20% US national Medicare sample, we identified all patients with inpatient hospitalizations 2013–14 to compare odds of 30\u2009day readmission among patients with SLE to the general Medicare population. Inclusion required age >18\u2009y.o., 12 mos. Medicare Part A/B (30d. post discharge), and being alive at discharge. Covariates included age, gender, race, ethnicity, prior Medicaid, comorbidity, prior hospitalization, length of stay, hospital volume, medical school affiliation, rural urban commuter area, and area deprivation index (ADI)a measure of neighborhood disadvantage derived from patients census block group (www.neighborhoodatlas.medicine.wisc.edu). Analysis used multivariable logistic regression clustered by patient. Results Nearly 1.4\u2009million acute Medicare hospitalizations were analyzed including 10\u2009868 lupus hospitalizations. Patients with lupus were younger (55% of hospitalizations in 18-<65\u2009y.o.) and 89% were female. Patients with lupus were over twice as likely to be Black or Hispanic (29% vs 12%; 5% vs 2% respectively). Among lupus patients, 46% had history of Medicaid, 65% were Medicare eligible due to disability, and 15% were eligible due to end stage renal disease (ESRD), compared to 30% Medicaid, 32% disability, and 3% ESRD among general Medicare. Lupus readmissions ranged from 20% within lower disadvantage to 27% within the greatest decile of ADI neighborhood disadvantage, compared to 17%–21% in general Medicare regardless of ADI decile (Figure 1). Before adjustment, patients with SLE were 60% more likely to be readmitted than non-lupus Medicare patients (Unadj. OR 1.61 (1.51, 1.71)). Lupus remained a readmission risk in adjusted models (Adj. OR 1.28 (1.20, 1.36)). SLE patients with 30\u2009day readmissions were more likely to be <65\u2009y.o., Black, to have had prior Medicaid or hospitalization, and other comorbidities; significant interactions were seen between ADI quintile and Medicaid predicting readmission.Abstract 104 Figure 1 30-day readmission by neighborhood disadvantage (ADI decile) Conclusions Among nationally representative Medicare hospitalizations, having lupus was a major readmission risk factor. Hospitalized patients with lupus who lived in a neighborhood with greatest disadvantage had higher 30\u2009day readmissions, impacting nearly 1 in 3 patients. Future studies are needed to examine readmission causes and interventions to systematically reduce readmissions among vulnerable hospitalized patients with lupus. Funding Source(s): None