Predictors of Advance Care Planning in Older Women: The Nurses' Health Study

BACKGROUND/OBJECTIVES: Relatively little is known regarding predictors of advance care planning (ACP) in former nurses. We aimed to evaluate potential predictors of ACP documentation and discussion. DESIGN: Cross-sectional study, 2012-2014. SETTING: Nurses' Health Study. PARTICIPANTS: A total of 60,917 community-dwelling female nurses aged 66 to 93 years living across the United States. MEASUREMENTS: Based on self-reports, participants were categorized as having (1) only ACP documentation, (2) ACP documentation and a recent ACP discussion with a healthcare provider, or (3) neither. Multivariable log-binomial models were used to estimate prevalence ratios (PRs) and 95% confidence intervals (CIs) of the two separate ACP categories vs those with neither. We evaluated various demographic, health, and social factors. RESULTS: The large majority (84%) reported ACP documentation; 35% reported a recent ACP discussion. Demographic factors such as age and race were associated with both ACP categories. In multivariable analyses, race was most strongly associated: compared with whites, African Americans were 27% less likely (PR = 0.73; 95% CI = 0.69-0.78) to report ACP documentation alone and 41% (PR = 0.59; 95% CI = 0.54-0.66) less likely to report documentation with discussion. Additionally, health/healthcare-related characteristics were more strongly associated with ACP documentation plus discussion. Women with functional limitations (PR = 1.15; 95% CI = 1.10-1.20), women who were recently hospitalized (PR: 1.10; 95% CI = 1.08-1.12) or women who had seen a physician for health symptoms (PR = 1.43; 95% CI = 1.35-1.52) or screening (PR = 1.40; 95% CI = 1.32-1.49) were more likely to report having both ACP documentation and discussion. Social factors showed limited relationships with ACP documentation only; for documentation plus discussion, being widowed and living alone was associated with higher prevalence (PR = 1.21; 95% CI = 1.19-1.24) and having little emotional support was associated with lower prevalence (PR = 0.84; 95% CI = 0.81-0.86). CONCLUSIONS: Among older nurses, most of whom reported having documented ACP, 35% reported recent patient-clinician ACP discussions, indicating a major participatory gap in an element critical to ACP effectiveness. Even in nurses, African Americans reported less ACP documentation or discussion.