Using Separate Single-Outcome Risk Presentations Instead of Integrated Multioutcome Formats Improves Comprehension in Discrete Choice Experiments
Matthew J. Wallace,E. Hope Weissler,Jui-Chen Yang,Laura Brotzman,Matthew A. Corriere,Eric A. Secemsky,Jessie Sutphin,F. Reed Johnson,Juan Marcos Gonzalez,Michelle E. Tarver,Anindita Saha,Allen L. Chen,David J. Gebben,Misti Malone,Andrew Farb,Olufemi Babalola,Eva M. Rorer,Brian J. Zikmund-Fisher,Shelby D. Reed
DOI: https://doi.org/10.1177/0272989x241258466
2024-06-22
Medical Decision Making
Abstract:Medical Decision Making, Ahead of Print. IntroductionDespite decades of research on risk-communication approaches, questions remain about the optimal methods for conveying risks for different outcomes across multiple time points, which can be necessary in applications such as discrete choice experiments (DCEs). We sought to compare the effects of 3 design factors: 1) separated versus integrated presentations of the risks for different outcomes, 2) use or omission of icon arrays, and 3) vertical versus horizontal orientation of the time dimension.MethodsWe conducted a randomized study among a demographically diverse sample of 2,242 US adults recruited from an online panel (mean age 59.8 y, s = 10.4 y; 21.9% African American) that compared risk-communication approaches that varied in the 3 factors noted above. The primary outcome was the number of correct responses to 12 multiple-choice questions asking survey respondents to identify specific numbers, contrast options to recognize dominance (larger v. smaller risks), and compute differences. We used linear regression to test the effects of the 3 design factors, controlling for health literacy, graph literacy, and numeracy. We also measured choice consistency in a subsequent DCE choice moduleResultsMean comprehension varied significantly across versions (P < 0.001), with higher comprehension in the 3 versions that provided separated risk information for each risk. In the multivariable regression, separated risk presentation was associated with 0.58 more correct responses (P < 0.001; 95% confidence interval: 0.39, 0.77) compared with integrated risk information. Neither providing icon arrays nor using vertical versus horizontal time formats affected comprehension rates, although participant understanding did correlate with DCE choice consistency.ConclusionsIn presentations of multiple risks over multiple time points, presenting risk information separately for each health outcome appears to increase understanding.HighlightsWhen conveying information about risks of different outcomes at multiple time points, separate presentations of single-outcome risks resulted in higher comprehension than presentations that combined risk information for different outcomes.We also observed benefits of presenting single-outcome risks separately among respondents with lower numeracy and graph literacy.Study participants who scored higher on risk understanding were more internally consistent in their responses to a discrete choice experiment.
health care sciences & services,medical informatics,health policy & services