Patient Aid for Theory-Driven Health Communication (PATH): Translating Goal-Power Theory Into an Instrument and Generative AI Support for Palliative Care Communication

Effective health communication is essential to patient well-being, yet the field is constrained by major theoretical and methodological gaps. Existing theories tend to address isolated communication phenomena and lack an integrative framework spanning the full continuum of patient-provider interaction. As a result, many measurement instruments lack a clear theoretical foundation and demonstrate inconsistent psychometric quality. In parallel, most AI-supported communication tools are technology-driven rather than theory-driven, and few have undergone rigorous clinical evaluation. Palliative care represents a high-stakes clinical context in which communication quality is central to outcomes, making it an ideal setting to test theory-based generative artificial intelligence (GenAI) communication interventions.

Grounded in the Goal-Power Theory of Communication developed by the applicant, this four-year project adopts a mixed-methods design to develop and evaluate a theory-based communication measurement instrument and a GenAI-supported communication aid, NurseChat+. The project comprises three sequential phases: (1) instrument development and validation, (2) intervention development and pilot testing, and (3) clinical evaluation using a randomized controlled trial (RCT). Phase 1 aims to develop and psychometrically validate a theory-based instrument for assessing health communication quality. Instrument development and validation will follow the MEASURE framework, including expert content validation and field testing with approximately 500 participants recruited from clinical settings within the National Taiwan University Hospital system. Phase 2 aims to develop and preliminarily validate the NurseChat+ using a user-centered design approach. A pilot usability and feasibility study will be conducted using a one-group pretest-posttest design with a minimum of 20 healthy adults to assess acceptability, usability, and preliminary effects on communication self-efficacy. Phase 3 will conduct a parallel-group, superiority RCT with an embedded mixed-methods design in palliative care outpatient and home care settings at National Taiwan University Hospital. Eligible adult patients receiving palliative care will be randomly allocated in a 1:1 ratio to the intervention group (NurseChat+ plus usual care) or the control group (usual care alone). The target sample size is 240 participants (120 per group). Intervention participants will use NurseChat+ prior to scheduled medical encounters.

• System Usability Scale (SUS) Score of NurseChat+ [Срок оценки: Post-test (immediately after the interaction with GenAI communication aid).]
• Inventory of Technology Acceptance Model (ITAM) Score [Срок оценки: Post-test ( immediately after the interaction with GenAI communication aid)]
• Evaluation of User Experience via the User Experience Questionnaire (UEQ) [Срок оценки: Post-test (immediately after the interaction with GenAI communication aid)]
• Improvement in Symptom Burden via the Edmonton Symptom Assessment System (ESAS) [Срок оценки: Baseline (T0: within 12 hours prior to the scheduled consultation) and Post-test (T1: post-consultation follow-up within 24-72 hours).]
• Patient Overall and Communication Satisfaction via the Chinese Patients' Satisfaction Scale (C-PSS) [Срок оценки: Post-test follow-up (T1: 24-72 hours after the consultation via telephone interview).]
• Patient Satisfaction with Decision Making via the Satisfaction with Decision (SWD) Scale [Срок оценки: Post-test follow-up (T1: 24-72 hours after the consultation via telephone interview).]