JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) tools may be able to personalize advice about medications.

Objective:

To identify older adults’ perceptions of using AI tools when deciding to start or stop medications.

Methods:

We conducted a vignette-based online experiment in which participants aged ≥65 years from the United States were asked to report their likelihood of stopping a medication by source of information using a 6-point Likert scale (scale anchors 1=not at all likely and 6=extremely likely). Three medications were presented in a randomized order: aspirin (risk of bleeding), ranitidine (cancer-causing chemical), or simvastatin (lack of benefit with age). Five sources of information were presented: primary care provider (PCP), pharmacist, AI that connects with the electronic health record (EHR) and provides advice to the PCP for approval before sharing the recommendation (‘EHR-PCP’), AI with EHR access that directly provides advice (‘EHR-Direct’), and AI that asks questions to provide advice (‘Questions-Direct’) directly. We calculated descriptive statistics to identify participants who were extremely likely (score 6) to stop the medication and used logistic regression to identify demographic predictors of being likely (score 4-6) as opposed to unlikely (scores 1-3) to stopping a medication.

Results:

Older adults (n=1,245) more frequently reported being extremely likely to stop a medication when the recommendation came from a PCP [ranging from 60% (aspirin) to 69% (ranitidine)] compared to a pharmacist [ranging from 18% (simvastatin) to 29% (ranitidine)]. Older adults were extremely likely to stop a medication when recommended by AI [EHR-PCP: 15% (aspirin) to 23% (ranitidine); EHR-Direct: 10% (aspirin and simvastatin) and 17% (ranitidine); Questions-Direct: 10% (aspirin) to 16% (ranitidine). In adjusted analyses, characteristics that increased the likelihood of stopping a medication when recommended by AI included being Black or African American as compared to White (ranging from Questions-Direct: OR 1.28, 95% C.I. 1.06, 1.54 to EHR-PCP: OR 1.42, 95% C.I. 1.17, 1.73), higher self-reported health (ranging from EHR-PCP: OR 1.09, 95% C.I. 1.01, 1.18 to EHR-Direct: OR 1.13 95% C.I. 1.05, 1.23), higher confidence in using an EHR (ranging from Questions-Direct: OR 1.36, 95% C.I. 1.16, 1.58 to EHR-PCP: OR 1.55, 95% C.I. 1.33, 1.80), and higher confidence using applications (ranging from EHR-Direct: OR 1.38, 95% C.I. 1.18, 1.62 to EHR-PCP: OR 1.49, 95% C.I. 1.27, 1.74). Older adults with higher health literacy were less likely to stop a medication when recommended by AI (ranging from EHR-PCP: OR 0.81, 95% C.I. 0.75, 0.88 to EHR-Direct: OR 0.85, 95% C.I. 0.78, 0.92).

Conclusions:

These findings suggest that older adults have reservations about stopping a medication when it is recommended by AI. However, individuals who are Black or African American, have higher self-reported health, or higher confidence in using an EHR or applications may be receptive to AI-based medication recommendations. Clinical Trial: N/A