Determinants of end-of-life treatment preferences in chronic disease: implications for patient-centred tertiary prevention

Background Chronic diseases such as cardiovascular disease constitute an enormous burden to societies worldwide. Patient-centered care in such diseases includes adequate tertiary prevention. However, avoiding burdensome interventions that are inconsistent with patients’ values, particularly in the sensitive phase near the end of end of life is crucial. Evidence suggests that clinicians and surrogate decision-makers frequently misjudge patients’ values [1,2]. We evaluated determinants of preferences for treatment intensity and minimal intervention. Methods A 58-item questionnaire was disseminated in a tertiary hospital-based cohort of patients and visitors. Two items were assessed using a 5-point Likert scale: (1) “I would accept stressful therapies if they could prolong my life” and (2) “I want as little medical intervention as possible at the end of my life, even if this may shorten my life”. Associations with demographics and key predictors of pre-specified domains (e.g., money, family, environment, education) and fear of dying were examined using ordered logistic regression. Results Among the 1228 participants (mean age 63±15 years, 66% male), the following morbidities were common: cardiovascular disease (62%), malignant disease (9%), orthopaedic disease (36%), and psychiatric disease (10%). For item 1, responses clustered around neutrality, suggesting overall uncertainty (mean 3.1±1.2, no significant deviation from the midpoint, Figure A). Higher age and female sex were associated with lower willingness to accept stressful therapies, while greater importance assigned to family OR 3.28, 95% CI 1.31–8.20) and greater fear of dying (OR 1.52, 95% CI 1.38–1.68) were associated higher willingness (Figure B). For item 2, responses indicated overall agreement (p<0.0001 vs. midpoint, Figure C). Female sex (OR 1.36, 95% CI 1.05-1.77) and higher importance assigned to the environment (OR 2.70, 95% CI 1.50–4.89) were associated with greater agreement, while fear of death was associated with disagreement (Figure D). Interestingly, the importance assigned to health, religion, or money showed no meaningful agreement with either item. Conclusion End-of-life preferences are heterogeneous and only partly predictable. Our large-scale survey including >1200 participants showed that value domains and fear of dying appear relevant and should be explicitly addressed in tertiary prevention and advanced care planning to support individualized, shared decision-making and reducing preference–treatment mismatch.