Background Decisional conflict is a commonly used measure in studies of medical decision-making. However in advance care planning (ACP) studies, which often focus on anticipatory decisions, using the Decisional Conflict Scale (DCS) requires pre-specified, hypothetical scenarios. Most existing ACP studies which use the DCS have used non-validated scenarios. In this pilot study we report on preliminary validation of three serious illness scenarios, designed for use with a culturally diverse sample of older adults as part of a multi-centre clinical trial.

Methods A multi-disciplinary investigator group developed a bank of serious illness scenarios, which were refined through consultation with service providers (e.g. aged care organisations), cultural community leaders and people with lived experience of life-limiting illness. Three scenarios, relating to ‘hospital treatment’, ‘care needs’ and ‘emergency treatment’ decision-making were selected for inclusion in a cross-sectional survey of older adults (administered in English, Italian and/or simplified Chinese). Additional measures included the Advance Care Planning Engagement survey (ACP-9), Hospital Anxiety and Depression Scale (HADS) and 10-item DCS, as well as demographics, health literacy, education and self-reported health. Multiple imputation was undertaken to address missing data.

Results Data were collected from 291 eligible participants (age 65–93 years, 16% spoke a language other than English). Fleisch-Kincaid grade reading level scores (English scenarios) were 7.3–11.8 years. Missing responses (10–13%) were not associated with participant gender, age, preferred language, or health literacy. Responses to the ‘emergency treatment’ scenario suggested floor effects. Those who had previously engaged in ACP had lower decisional conflict in response to the hospital treatment (p<.001) and emergency treatment scenarios (p<.001).

Conclusion Preliminary data from a culturally diverse sample of older Australian adults suggest that the ‘hospital treatment’ and ‘care needs’ scenarios are acceptable, comprehensible and have appropriate statistical properties. These scenarios will be tested longitudinally in future ACP trials.