Be Careful What You Ask For: Effects of Benefit Descriptions on Diabetes Patients' Benefit-Risk Tradeoff Preferences.

BACKGROUND: As more studies report on patient preferences for diabetes treatment, identifying diabetes outcomes other than glycated hemoglobin (HbA1c) to describe effectiveness is warranted to understand patient-relevant, benefit-risk tradeoffs. OBJECTIVE: The aim of the study was to evaluate how preferences differ when effectiveness (glycemic control) is presented as long-term sequela (LTS) risk mitigation rather than an asymptomatic technical marker (HbA1c). METHODS: People with type 2 diabetes and using insulin (n = 3160) were randomly assigned to four self-administered, discrete-choice experiments that differed by their presentation of effectiveness. Epidemiologic reviews were conducted to ensure a close approximation of LTS risk relative to HbA1c levels. The relative importance of treatment benefit-risk characteristics and maximum acceptable risk tradeoffs was estimated using an error-component logit model. Log-likelihood ratio tests were used to compare parameter vectors. RESULTS: In total, 1031 people responded to the survey. Significantly more severe hypoglycemic events were accepted for a health improvement in terms of LTS mitigation versus HbA1c improvement (0.7 events per year; 95% confidence interval [CI]: 0.4-1.0 vs. 0.2 events per year 95% CI: -0.02 to 0.5) and avoidance of treatment-related heart attack risk (1.4 severe hypoglycemic events per year; 95% CI: 0.8-1.9 vs. 1 event per year; 95% CI: 0.6-1.3). This finding is supported by a log-likelihood test that rejected at the 0.05 level that respondent preference structures are similar across the different experimental arms of the discrete-choice experiment. CONCLUSION: We found evidence that benefit descriptions influence elicited preferences for the benefit-risk characteristics of injectable diabetes treatment. These findings argue for using carefully defined effectiveness measures to accurately take account of the patient perspective in benefit-risk assessments.

Duke Scholars

Author F. Reed Johnson Population Health Sciences