End-Of-Life Outcomes in Nursing Homes: Association with Care Processes

Annually, approximately 500,000 nursing home (NH) residents die. Yet, little is known about the quality of end-of-life (EOL) care in NHs, and even less about EOL care processes. We examined the relationship between two EOL outcomes – pain and place of death - and EOL care process measures in NHs in New York State (NYS).

EOL care process measures - care delivery and staff communication with residents/families – were derived from NH survey data. These measures have been demonstrated to be psychometrically reliable and valid. Secondary data include: Minimum Data Set (MDS), Medicare beneficiary file; inpatient and hospice claims; and Online Survey Certification and Reporting System.

Study population consisted of long-term decedent NH residents. It included 20,930 decedents in 299 facilities, over a 2-year period (CY2006-2007). We estimated individual level logistic regression models with facility random effects. Key explanatory variables included measures of EOL care processes. For each outcome, risk factors derived from the MDS were included as independent variables. Facility characteristics, e.g. staffing, size, and ownership were also included. Residents in facilities with better care delivery have significantly (p=0.033) lower odds of pain (standardized OR=0.904), and residents in facilities with better communication have significantly (p=0.007) lower odds of in-hospital death (standardized OR=0.895). The proportion of residents who die in pain or in-hospitals continues to be high. Findings suggest that modest improvements in EOL care processes may result in significant improvements in quality of care. We provide empirical evidence for the feasibility of improving EOL quality of care for residents.