Health shocks and advance care planning among U.S. older adults: Evidence from HRS 2012-2022.

11093 Background: Advance care planning (ACP) allows individuals to discuss and document preferences for future medical care under serious illness, yet engagement among older adults remains limited. Experiences of a health shock, such as a diagnosis of cancer, major illnesses, or hospitalization, may prompt participation in ACP. This study aimed to examine the association between health shocks and subsequent ACP engagement and variation by key sociodemographic characteristics. Methods: We conducted a longitudinal cohort study using data from the Health and Retirement Study (HRS), a nationally representative survey of U.S. adults aged ≥65 years. Adults from 2012–2022 HRS with no baseline ACP who completed at least one follow-up wave were included. The exposure was a new health shock, defined as a new diagnosis of cancer, stroke, heart disease, or hospitalization. The outcome was new ACP engagement in subsequent waves. Generalized estimating equation models estimated adjusted odds ratios (ORs) of experiencing a health shock due to cancer diagnosis and due to other causes (vs. no health shock) on new ACP; and how such associations varied by respondents’ characteristics. Results: Among 8,553 person-wave observations (4,426 unique respondents), median age was 74 years (IQR 69–79); 52.9% were women, 24.3% lived alone, 27.6% experienced a health shock, and 3.11% reported a health shock from new cancer diagnosis. Health shocks were associated with higher odds of ACP engagement (OR = 1.41; 95% CI, 1.28–1.56). In the three-level shock model, both non-cancer shock (OR = 1.43; 95% CI, 1.28–1.59) and cancer shock (OR = 1.31; 95% CI, 1.01–1.70) were associated with higher odds of ACP compared with no shock. Older age, female gender, and higher education were associated with greater ACP engagement overall. Compared with non-Hispanic Whites, non-Hispanic Black (OR = 0.73; 95% CI, 0.64–0.84) and Hispanic (OR = 0.55; 95% CI, 0.47–0.65) had lower odds of ACP. Living alone was associated with higher ACP engagement (OR = 1.29; 95% CI, 1.14–1.45). However, when stratified by patient sociodemographic characteristics, the associations between health shock and ACP were similar across groups, except that for those living alone who showed larger odds of initiating a new ACP following health shock (OR = 1.76; 95% CI, 1.45-2.14) than those not living alone (OR = 1.31; 95% CI, 1.16-1.47). Conclusions: Health shocks were associated with new ACP engagement among older adults, with similar associations observed for cancer-specific and non-cancer health shocks. Alhtought disparities in ACP by sociodemographic characteristics persisted, health shocks were associated with similarly odds of initiating ACP across population subgroups. This highlights the importance of leveraging health shocks as clinical touchpoints for ACP discussions, and the need for targeted strategies to reduce inequities in ACP engagement outside of acute health events.