Understanding Medicine Preferences of Older Adults: The Role of Messaging in a Multi-Methods Experimental Survey

Efforts to optimise medicine use often focus on starting evidence-based therapies or deprescribing inappropriate medicine, but not on the combination of both. We explored whether behavioural science-informed messaging could impact older adults' decisions on starting and stopping diabetes medicine.

A 15-min online survey was conducted using a hypothetical vignette among adults 65 years and older from Australia and the United States. Participants were randomised in two experiments: the first involved different framings of a primary care provider (PCP)'s rationale for starting a diabetes medicine, and the second focused on stopping the same medicine 10 years later. Preferences were measured using 6-point Likert scales. Data were analysed using descriptive and nonparametric statistics. We also conducted a descriptive thematic analysis of participants' free-text responses explaining their preferences.

Out of 2420 participants, 68.5% agreed with starting the medicine, whereas 64.4% agreed with stopping it (p = 0.009). The proportion of participants who agreed with the recommendation to start or stop the medicine did not differ by the messaging they received in the first experimental treatment. However, participants were more likely to agree with a recommendation to stop a medicine if they received messaging about the medicine's harms versus lack of benefit (69.7% vs. 59.1%, p < 0.001). Thematic analysis revealed that participants' trust in clinician expertise, their perceptions of the medicine (e.g., benefits, harms, necessity and burdens) and a hesitance to change influenced their decisions.

Two-thirds of participants agreed with the recommendation to start and stop a diabetes medicine regardless of messaging at medicine initiation. Subsequent messaging focused on the potential harm of continuing a medicine significantly increased participants' agreement with deprescribing compared to messaging about the lack of benefit.

Older adult community members and general practitioners contributed to the development of the survey and vignette, helping to refine the content and improve its clarity and relevance.