Goals‐of‐care Conversation Prevalence and Life‐Sustaining Treatment Preferences Among VA Nursing Home Residents with and without Dementia

Research Objective The Veterans Health Administration uses structured goals‐of‐care conversations (GoCCs) to document veterans’ preferences for life‐sustaining treatments (LST). Responses include information on decision making capacity, goals, and preferences for LST following cardiopulmonary and noncardiopulmonary events. Durable medical orders are generated to inform care decisions and are accessible to clinicians across the VA health care system. We conducted a descriptive analysis of GoCC documentation among veterans residing in VA Community Living Centers (CLC), with a focus on those diagnosed with dementia. This population of veterans is at high risk of life‐threatening events and is suited to have GoCCs. Study Design Data were pulled from the VA’s Corporate Data Warehouse (CDW). Inpatient data were used to identify veterans who resided in a CLC for one or more days between January 2018 and June 2019. Veterans were considered to have dementia if they had one or more dementia diagnoses during an inpatient stay or two or more face‐to‐face diagnoses at outpatient visits in the last two years. Completed GoCCs that were documented between one year prior to admission, or January 2018 if admitted earlier, and discharge, or June 2019 if discharged later, were analyzed. The prevalence of GoCC documentation, responses to the LST questions, and dementia status were cross‐tabulated. Population Studied Veterans in CLCs between January 2018 and June 2019 (N = 55 271). Principal Findings Among all CLC veterans, 29 460 (53%) had a documented GoCC. Dementia was indicated for 14 231 veterans (26%). Completed GoCC documentation was more prevalent for those with dementia than those without (59% vs 51%). Proxies provided consent for 74% of responses for veterans with dementia and 21% for veterans without dementia. Of the 29 460 CLC veterans with completed GoCC documentation, 43% were deceased within 6 months. Among veterans with completed GoCCs, veterans with dementia were more likely to document a preference of being comfortable (71% vs 57%) and less likely to document a preference of improving/maintaining function or prolonging life (40% and 14% vs 48% and 22%). The documentation of a preference to “attempt CPR for cardiopulmonary events” was less prevalent for veterans with dementia (20% vs 36%) as was a preference for the full scope of treatment for noncardiopulmonary events (27% vs 41%). Conclusions The high mortality rate following documented GoCCs for this population highlights the urgency with which documented care preferences are needed in order to provide goal concordant care; it also suggests that providers may be targeting high‐risk CLC veterans. The differences in LST goals documented for veterans diagnosed with dementia highlight how clinical conditions may shape goals and preferences and should be taken into consideration by providers. Since proxies responded for many more veterans with dementia than those without, it is unclear whether differences in preferences reflect those of the veterans or their proxies. Implications for Policy or Practice The documentation of LST preferences can capture meaningful patient preferences across different populations and can augment end‐of‐life care planning. Primary Funding Source Department of Veterans Affairs.