Hospital readmissions are frequent and costly, and excess risk-adjusted readmissions at a hospital may reflect poor quality of care. Reducing hospital readmissions has been a focus for quality improvement in VA in order to decrease health care costs. Beginning in 2012, the Center for Medicare and Medicaid Services implemented the Hospital Readmissions Reduction Program to reduce reimbursements to hospitals with excessive readmissions. Heart Failure (HF) and chronic obstructive pulmonary disease (COPD) are among the targeted conditions. Despite a growing focus on reducing readmissions and costs for readmissions, organizational factors to reduce readmissions are not well understood.
The study assessed readmission rates and costs for HF and COPD, two of the most common conditions for hospitalization in VA. The study had two aims:
1. To assess variation in hospital-level readmission rates and readmission costs among VA inpatients.
2. To identify VA hospital-level organizational factors associated with systematic variation in readmission rates.
The study used a retrospective cohort study design. The data sources included VA administrative datasets, non-VA medical care data, Medicare claims and facility surveys of clinical practices for HF and COPD. The study sample included two cohorts: 1) HF cohort: 39,158 patients with an index admission in 128 VA hospitals for HF during FY2007-2009, and 2) COPD cohort: 33,640 patients with an index admission in 130 VA hospitals for COPD during FY2009-2011. Facilities with less than 25 patients were excluded from the facility level analysis. The primary outcomes were 30-day all cause readmission for Aim 1 and 30-day HF (or COPD) specific readmission for Aim 2. Readmissions included the subsequent admissions to VA, non-VA hospitals via fee-basis, or Medicare fee-for-service. We assessed healthcare costs in the one year after the index admission, including VA, fee-basis, and Medicare costs. Organizational factors, which were self-reported by Chief of Medicine or Chief of Cardiology or Pulmonology, included clinical structures and practices (such as availability of cardiologists/
pulmonologists; HF or COPD focus clinics, programs and activities, standardized orders, performance measures, and clinical reminders), as well as, discharge practices and follow-up (such as assessment of quality measures, patient education, follow-up calls, and scheduling of a follow-up visit).
To estimate risk adjusted facility-level all cause readmission rates, we estimated logistic regression models with hospital-level fixed effects to predict readmission rates at each hospital while holding patient characteristics constant at the national average. To examine the association between organizational factors and 30-day disease-specific readmission, we used mixed-effects logit regression that controlled for patient demographics and comorbidity and clustering patient-level observations by VA hospital.
For the HF cohort, the 30-day, 60-day, 90-day and 1-year all cause readmission rates were 22%, 33%, 41% and 68%, respectively. The average number of readmissions per patient during the 1-year follow-up period was 1.8 (1.4 admissions in VA and 0.4 admissions in Medicare). The average cost of healthcare during the year after index admission was $46,783 with 61% of the cost from inpatient care. In total, the HF cohort incurred $1.8 billion in one year following the index admission with 84% of the cost from VA, including fee-basis. At the facility level, the median adjusted 30-day all cause readmission rate was 21%, ranging from 12% to 29% (interquartile range 19% - 25%).
The 30-day HF-specific readmission rate was 11%. Adjusted results show that standardized orders for outpatients (Odds ratio (OR)=0.82, p=0.004), a written contract signed by the patient at the time of discharge (OR=0.77, p=0.038), and having a HF follow-up clinic (OR=0.87, p=0.047) were associated with a lower odds of HF-specific readmission. Factors associated with increased risk of readmission included standardized heart failure exercise program (OR=1.15, p=0.001) and routine scheduling of the first post-discharge follow-up visit more than 4 weeks after discharge (OR=1.20, p=0.040) compared to facilities scheduling the follow-up visit within 2 weeks.
For the COPD cohort, the 30-day, 60-day, 90-day and 1-year all cause readmission rates were 18%, 28%, 35% and 62%, respectively. The average number of readmissions per patient during the 1-year follow-up period was 1.6 (1.2 admissions in VA and 0.4 admissions in Medicare). The average cost of healthcare during the year after index admission was $44,629 with 71% of the cost from inpatient care. In total, the COPD cohort incurred $1.5 billion in one year following the index admission with 90% of the cost from VA, including fee-basis. At the facility level, the median adjusted 30-day all cause readmission rate was 16%, ranging from 7% to 25% (interquartile range 14% - 18%).
The 30-day COPD-specific readmission rate was 8%. After controlling for patient characteristics, organizational factors associated with significantly lower risk for readmission included having standardized patient education for COPD (OR = 0.85, p=0.004) and having hospitalists practice at the facility (OR = 0.83, p=0.012). Factors associated with increased risk for readmission included availability of pulmonary rehabilitation (OR =1.13, p=0.031) and admitting more than 250 COPD patients per year (OR = 1.15, p=0.018).
This study shows that the clinical and economic burden of HF and COPD inpatients is substantial. There was wide variation in 30-day all cause readmission for HF and for COPD across VA hospitals. This study identified important clinical practices that could potentially reduce HF-specific and COPD-specific readmissions. For HF, the clinical practices associated with a lower odds of readmission included standardized outpatient orders for managing HF patients, a written contact signed by patient at the time discharge, timely scheduling of a post-discharge follow-up visit, and having a HF-specific HF follow-up clinic. For COPD, the clinical practices associated with lower risk for readmission included having standardized patient education and having hospitalists practice at the facility. Further research on the unexpected positive relationships between cardiac exercise programs and HF-specific readmission as well and pulmonary rehab programs and COPD-specific readmissions is needed.
External Links for this Project
Grant Number: I01HX000299-01A2
None at this time.