BACKGROUND: Little is known about the impact of non-cardiovascular disease (CVD) burden on 30- -day readmission in heart failure (HF) patients. The aim of the study was to assess the role of non-CVD burden on 30-day readmission in HF patients. \ METHODS: We analyzed the effect of non-CVD burden by frequency of ICD-9 code categories on readmis-sions of patients discharged with a primary diagnosis of HF. We first modeled the probability of readmis-sion within 30 days as a function of demographic and clinical covariates in a randomly selected training dataset of the total cohort. Variable selection was carried out using a bootstrap LASSO procedure with 1000 bootstrap samples, the final model was tested on a validation dataset. Adjusted odds ratios and confidence intervals were reported in the validation dataset. RESULTS: There were a total of 6228 HF hospitalizations, 1523 (24%) with readmission within 30 days of discharge. The strongest predictor for 30-day readmissions was any hospital admission in the prior year (p < 0.001). Cardiovascular risk factors did not enter the final model. However, digestive system diseases increased the risk for readmission by 17% for each diagnosis (p = 0.046), while respiratory diseases and genitourinary diseases showed a trend toward a higher risk of readmission (p = 0.07 and p = 0.09, respectively). Non-CVDs out-competed cardiovascular covariates previously reported to predict readmission. CONCLUSIONS: In patients with HF hospitalization, prior admissions predicted 30-day readmission. Diseases of the digestive system also increase 30-day readmission rates. Assessment of non-CVD burden in HF patients could serve as an important risk marker for 30-day readmissions.
BACKGROUND: Little is known about the impact of non-cardiovascular disease (CVD) burden on 30- -day readmission in heart failure (HF) patients. The aim of the study was to assess the role of non-CVD burden on 30-day readmission in HF patients. \ METHODS: We analyzed the effect of non-CVD burden by frequency of ICD-9 code categories on readmis-sions of patients discharged with a primary diagnosis of HF. We first modeled the probability of readmis-sion within 30 days as a function of demographic and clinical covariates in a randomly selected training dataset of the total cohort. Variable selection was carried out using a bootstrap LASSO procedure with 1000 bootstrap samples, the final model was tested on a validation dataset. Adjusted odds ratios and confidence intervals were reported in the validation dataset. RESULTS: There were a total of 6228 HF hospitalizations, 1523 (24%) with readmission within 30 days of discharge. The strongest predictor for 30-day readmissions was any hospital admission in the prior year (p < 0.001). Cardiovascular risk factors did not enter the final model. However, digestive system diseases increased the risk for readmission by 17% for each diagnosis (p = 0.046), while respiratory diseases and genitourinary diseases showed a trend toward a higher risk of readmission (p = 0.07 and p = 0.09, respectively). Non-CVDs out-competed cardiovascular covariates previously reported to predict readmission. CONCLUSIONS: In patients with HF hospitalization, prior admissions predicted 30-day readmission. Diseases of the digestive system also increase 30-day readmission rates. Assessment of non-CVD burden in HF patients could serve as an important risk marker for 30-day readmissions.
Entities:
Keywords:
heart failure readmission; non-cardiovascular disease burden; predictive model