This mixed-methods evaluation tested an Advance Care Planning programme combining role-play, personal-value reflection, and two Circle of Learning sessions with an entire cohort of 50 first-year internal-medicine residents. Longitudinal analysis used 34 complete cases through 15 months, and 10 residents were interviewed. Knowledge rose from 7.74 to 8.65 out of 9, alongside several attitude and confidence improvements. Interviewees said role-play bridged theory and practice but identified a need for ongoing workplace support. The evidence concerns learner readiness, not observed conversation quality or patient outcomes.
Key findings
- Among 34 complete cases, knowledge increased from 7.74 to 8.65 out of 9 (p
- Several attitude and confidence domains improved, including perceived barriers, legal perspective, and communication hesitation. Change in attitude correlated with change in confidence at r=0.493 (p=0.003), which does not establish causal direction.
- Ten interviews produced three themes: role-play bridged theory and practice, value reflection catalysed empathy, and clinical realities required longitudinal support.
- Of the full 50-resident cohort, 34 contributed complete longitudinal data—a 68% complete-case proportion that may differ from residents lost to follow-up.
Why this matters globally
Serious-illness communication training is a shared health-system gap. Combining simulation, value reflection, and longitudinal learning circles offers a model for testing elsewhere, but scale-up should use stronger comparisons and measure observed behaviour, ACP documentation, patient and family experience, and unintended effects—not learner scores alone.
Thai researcher contribution
All seven authors have Thai affiliations through Ramathibodi, Hatyai Hospital, or the Compassionate Communities foundation; Saran Thanapluetiwong is corresponding author. Ramathibodi approved the study (MURA2023/406), and the work received no external funding. The Early Version does not expose individual CRediT roles, so no further task-level attribution is assigned.
Limitations to consider
This was a single-group before–after study without randomisation or a control, so programme effects cannot be separated from accumulating clinical experience or testing effects. It involved one 50-resident cohort, with 34 complete cases, creating attrition and selection concerns. Knowledge had a nine-point ceiling and high baseline; attitudes and confidence were self-reported; interviews included 10 participants. No observed patient conversations, ACP-document quality, or patient outcomes were measured. Trial registration was retrospective, and the accepted peer-reviewed Early Version remains subject to editing before the Version of Record.