Surrogate Decision-Making for Older Adults With and Without Dementia.

Importance: End-of-life care for older individuals in the US frequently includes interventions that may be burdensome and inconsistent with patient preferences. Few experimental studies have assessed potentially modifiable factors associated with surrogate recommendations. Objectives: To assess whether experimentally varied patient and decision-context factors are associated with participant treatment recommendations for seriously ill older adults. Design, Setting, and Participants: In this survey study, a nationally representative sample of US adults aged 18 years or older who were members of the Ipsos KnowledgePanel survey platform were surveyed from December 8 to 19, 2022. Data were analyzed from April 2023 to July 2026. Exposure: Randomized vignettes with varying dementia status, advance directive status and wishes, treatment framing, physician recommendations, hospital norms for care of similar patients, patient sex, and the surrogate's relationship to the patient. Main Outcomes and Measures: Whether participants recommended that surrogates choose life-sustaining treatments (LSTs) vs comfort-focused care for a hypothetical hospitalized patient aged 85 years who was severely ill, currently unconscious, and functionally dependent before admission. Results: Of 6019 individuals who started the survey, 5965 (99.1%) completed at least 1 vignette response (mean [SD] respondent age, 53.7 [16.9] years; 3079 [51.6%] female), including 2004 caregivers of an adult and 3175 with chronic illness, a mobility limitation, and/or recent emergency or urgent care use. When holding other randomized vignette features constant, participants recommended that surrogates choose LSTs for 15.6% (95% CI, 14.0%-17.2%) of patients with dementia without an advance directive compared with 41.0% (95% CI, 38.8%-43.1%) of those with an advance directive requesting LSTs and 7.6% (95% CI, 6.4%-8.8%) of patients with an advance directive requesting comfort care only. For patients who did not have dementia, participants recommended that surrogates choose life-extending treatment for 38.9% (95% CI, 36.7%-41.0%), 66.3% (95% CI, 64.2%-68.3%), and 14.4% (95% CI, 12.8%-15.9%) of patients in the respective advance directive groups. Conclusions and Relevance: In this survey study of US adults, respondents were less likely to recommend that surrogates choose LSTs for patients with dementia than for patients without dementia even when advance directives requested all care possible. These findings highlight the importance of preparing patients and surrogates for care decisions through explicit discussions of goals of care and decision-making roles.