828 - Evaluation of Home Time as a Patient-Centered Metric for Pneumonia Hospitalizations: A Retrospective Cohort Study of Medicare Fee-For-Service Beneficiaries

Background: The Centers for Medicare & Medicaid Services (CMS) uses hospital readmission to incentivize hospital care delivery for acute conditions including pneumonia. However, current CMS performance metrics do not account for the competing risk of mortality in the post-discharge period or during the hospital stay. Our objective was to assess home time within 30 days after discharge among pneumonia hospitalizations, as a patient-centered metric.

Methods: A retrospective observational study was conducted in a cohort of Medicare fee-for-service beneficiaries admitted between 01/01/2015 and 11/30/2017. Home time was the number of days spent alive, out of an acute care setting, skilled nursing facility, or a rehabilitation facility within 30 days of discharge. If a patient spends any part of a day in a care facility or died after discharge, then that day was not included in the calculation for home time. Hospital-level rates of risk-adjusted home time were calculated using multilevel regression models. We compared hospital performance on 30-day risk-standardized home time with its performance on 30-day risk standardized readmission rate (RSRR) and mortality rate (RSMR). Characteristics of hospitals with high and low risk-adjusted home-time were compared.

Results: Among 1.7 million pneumonia admissions admitted to 3,116 hospitals, the median 30-day risk-standardized home time was 20.5 days (interquartile range: 18.9-21.9 days). Hospital-level characteristics such as case volume, bed size, for-profit ownership, rural location of hospital, teaching status, and participation in the bundle payment program were significantly associated with home-time. RSRR (rho: -0.233, p< 0.0001) and RSMR (rho: -0.223, p< 0.0001) had weak, inverse correlations with home time. Using the home time metric, 35.5% of hospitals were reclassified as high performers compared with their average or poor performance on the RSRR or RSMR metric.

Conclusion: Home time is a novel, patient-centered, hospital-level metric that can be easily calculated using claims data, accounts for differences in post-discharge mortality and can be intuitively interpreted. Utilization of this metric could potentially have policy implications in assessing hospital performance on delivery of healthcare to pneumonia patients.