Abstract
BACKGROUND: Early identification of the palliative phase in patients with chronic obstructive pulmonary disease (COPD) and heart failure (HF) remains challenging. Timely identification enables recognition of palliative care needs and the initiation of comprehensive palliative care, yet it often occurs too late. Because most patients with COPD and HF spend the majority of their disease trajectory at home, recognition within community care is essential to anticipate needs and maintain quality of life. Globally, only 14% of people needing palliative care receive it. In primary care, existing screening tools are seldom used, leaving identification dependent on professional judgment. Machine learning (ML) offers a promising approach for detecting patterns in routinely collected health data to support earlier, more consistent identification of palliative care needs. This study aimed to identify and prioritize signs and symptoms indicating the transition to the palliative phase in patients with COPD or HF to inform the development of a ML-model supporting primary home-based care nurses in timely recognition of palliative care needs. METHODS: This mixed-methods study, in two home-based care organizations, comprised four phases: (1) semi-structured interviews with nine home-based nursing professionals; (2) rapid review of 13 studies on COPD and HF palliative signs and symptoms; (3) two multidisciplinary focus groups with 18 professionals; and (4) a survey among 115 home-based nursing professionals to validate and prioritize signs and symptoms. Qualitative data were thematically analysed, and survey data were analysed descriptively. RESULTS: Deterioration was identified as a multidimensional process across physical, psychological, social, and spiritual domains, accompanied by increasing care needs. Physical deterioration involved worsening symptoms and functional decline, while psychological and social domains reflected anxiety, withdrawal, and growing dependence, often noted by informal caregivers. The spiritual domain encompassed existential distress and loss of meaning. Increasing care needs were reflected in higher service use, unplanned interventions, and hospital admissions. Findings across study phases were integrated to prioritize key signs and symptoms for home-based nursing practice. CONCLUSIONS: This study identified and prioritized signs and symptoms used by home-based nurses to recognize the palliative phase in patients with COPD and HF. These findings form the basis for a practice-oriented ML model supporting timely identification of palliative care needs and comprehensive home-based palliative care. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12904-026-02064-6.