In New York, 1 in 5 HIV Hospital Admissions Followed by Readmission Within 30 Days

Among HIV-positive adults admitted to a hospital in New York, 22% were readmitted within 30 days, according to a statewide analysis. Behavioral factors and comorbid diseases ranked among the stronger predictors of readmission in the study, which was published in the April 1 issue of JAIDS.

Hospital readmissions impose a considerable physical and psychological burden on people with HIV while escalating costs to hospitals and insurers. Before the arrival of effective antiretroviral therapy, relentless HIV disease progression accounted for most readmissions. In the current antiretroviral era, factors linked to readmission remain incompletely understood. Because identifying those factors may help prevent readmissions, researchers conducted this retrospective cohort analysis of HIV patient readmissions at all New York State hospitals.

The analysis focused on all 2012 hospital admissions recorded among HIV-positive adults in the all-payer, all-hospital database of the New York State Department of Health. The analysis excluded transfers from other hospitals, admissions to psychiatric and detoxification units and admissions for maintenance chemotherapy. The researchers identified readmissions within 30 days of discharge and calculated the population-wide readmission rate by dividing the number of readmissions by total index (initial) admissions. They used ICD-9 codes to identify 25 diagnoses selected as possible predictors of readmission.

Among 16,558 individuals with HIV admitted to the hospital, 27.5% were readmitted at least once in 2012. Age of the study group averaged 49.1 years, 60.5% were men, 50.6% were African American and 95.4% lived in urban areas. Among 23,544 statewide index admissions of people with HIV infection, 5121 (21.8%) resulted in readmissions within 30 days. An additional 2900 readmissions resulted in 1419 readmission chains, 44% of which involved three or more separate readmissions.

Logistic regression analysis identified two factors that independently lowered odds of hospital readmission: female gender (adjusted odds ratio [aOR] 0.88) and Hispanic ethnicity (aOR 0.78 versus Caucasian race). Two sociodemographic factors independently raised odds of readmission: Medicaid use (aOR 1.23) and unstable housing (aOR 1.51). Several clinical, behavioral or psychological factors also independently boosted odds of readmission: AIDS illness diagnosis (aOR 2.73), kidney disease (aOR 1.87), one to two comorbidities (aOR 2.8 versus none), three to four comorbidities (aOR 5.78), five or more comorbidities (aOR 11.54), alcoholism (aOR 1.40), depression (aOR 1.61) and psychosis (aOR 1.43).

Several hospital-use factors independently predicted readmission: leaving the hospital against medical advice (aOR 2.28), one or two inpatient visits in 2011 (aOR 1.39 versus none), two to four inpatient visits in 2011 (aOR 2.19), five or more inpatient visits in 2011 (aOR 2.77), two to four emergency department visits in 2011 (aOR 1.17 versus none) and five or more emergency department visits in 2011 (aOR 1.50).

One-third of readmissions occurred at hospitals different from the initial admitting hospital. A higher proportion of readmissions in New York City than elsewhere involved a different hospital (36% versus 24%, P < .01). Regression analysis identified two factors independently associated with a lower chance of readmission to a different hospital: kidney disease (estimate -0.11) and an AIDS illness diagnosis (-0.07). This analysis linked five factors to higher chances of readmission to a different hospital: illicit drug use (+0.64), having ever left the hospital against medical advice (+0.081), unstable housing (+0.23), two to four inpatient visits in 2011 (+0.10 versus none) and five or more inpatient visits in 2011 (+0.26).

The authors observe that readmission to a hospital different from the initial admitting facility represents "a fragmentation of care that has significant implications for care coordination and quality of care." They suggest that the links between behavioral factors and readmission should encourage hospitals to "consider targeted interventions among this subset of the population as part of comprehensive care and discharge planning." Referring to their finding that more comorbidities inflate odds of readmission, the researchers stress that "inadequate management of chronic conditions during inpatient care has been identified as a preventable cause of readmission" among people with HIV infection.