Advance care planning preferences and associated factors in patients with multiple chronic conditions: a systematic...

Patients with multiple chronic conditions (MCC) often experience complex illness trajectories, functional decline, repeated healthcare encounters, and increasing need for future decision-making support. Advance care planning (ACP) may help align future care with patients' values and preferences, but evidence on ACP preferences and associated factors in this population remains fragmented. This review aimed to synthesize evidence on ACP preferences and associated factors among patients with MCC.

A systematic review with exploratory network meta-analysis of ACP preference-related domains was conducted in accordance with PRISMA guidance and registered in PROSPERO (CRD420251165443). PubMed, Embase, Web of Science, Cochrane Library, CNKI, Wanfang, VIP, and China Biomedical Literature Service System were searched from inception to December 2025. Studies published in English or Chinese that reported ACP-related outcomes among adults with MCC or multimorbidity were included. For mixed-population studies, MCC-specific or multimorbidity subgroup data were extracted whenever available. Separate exploratory Bayesian network meta-analyses were conducted for conceptually distinct preference domains where extractable data were available. Associated factors were synthesized narratively. Study quality was assessed using the Newcastle-Ottawa Scale and Joanna Briggs Institute checklist.

Twenty-five studies were included. The included studies varied substantially in population definition, illness trajectory, care setting, ACP outcome measurement, and availability of MCC-specific data. In exploratory ranking analyses, durable power of attorney for health care (DPAH) showed a relatively high ranking probability among ACP format categories (SUCRA = 83.76%). For end-of-life treatment preferences, do-not-resuscitate orders (SUCRA = 80.06%) and forgoing life-sustaining treatment (SUCRA = 72.15%) showed relatively high ranking probabilities. For place-of-care or place-of-death preferences, absence of a documented or expressed preference showed a high ranking probability (SUCRA = 89.12%), which should be interpreted as lack of recorded or reported preference rather than as an active preference not to express a preference. Associated factors were grouped into individual, clinical and functional, family and social, healthcare provider and care-setting, and cultural, legal, and health-system domains. Because of heterogeneity in populations and outcome definitions, ranking results should be interpreted as exploratory and not as evidence of clinical superiority or universal preference.

ACP among patients with MCC is heterogeneous and shaped by interacting clinical, functional, family, cultural, legal, and health-system factors. Exploratory network findings may help describe reported ACP preference patterns, but they should not be interpreted as comparative effectiveness or universal preference evidence. Future studies should use standardized definitions of MCC and ACP outcomes, distinguish informal communication from formal documentation, and report MCC-specific data in mixed illness populations.