Factors Influencing Informed Consent Preferences in Digital Health Research: Survey Study of Prospective Participants.

Background: Readability is important to consider when developing informed consent communications for prospective research participants, but not the most important consideration. Other factors to consider relate to learning preferences and literacy needs of people recruited to participate in research, as these factors can influence understanding of consent communications. To promote understanding among prospective participants, researchers should take a human-centered approach to develop consent communications.

Objective: This study aims to explore how factors related to readability, topic, and participant demographic characteristics play into preferences for digital health research consent material. These factors are important to consider as not attending to some details that matter to a specific subgroup of prospective participants may systematically exclude people from research.

Methods: People eligible to participate in a digital health study were recruited to review 31 paragraph length sections of a consent form, referred to as "text snippets," for an existing institutional review board-approved digital health study. Participants (N=79) were surveyed and asked to choose between 2 variations of the text snippets, either indicating a preference for the institutional review board-approved original or a version that was modified to improve readability.

Results: A slim majority of participants provided feedback about the snippets (n=44; 55%). Our qualitative analysis of the feedback found that participants preferred shorter snippets, in general, but the snippets also elicited new questions not addressed by the original consent material. This observation is supported by our quantitative analysis, which found that when the character length of the original was longer, participants were less likely to prefer the original (P<.001) and more likely to prefer the modified text by a factor of 1.20 times (P=.04), and particularly for snippets explaining study risks (P=.03). Our analysis also found significant differences in participant demographic characteristics. For example, older participants tended to prefer the original more than younger participants, by a factor of 1.95 times (P=.004). The results present illustrative examples of how factors related to sex, age, physical activity, and ethnicity all play into preference for consent communication.

Conclusions: The findings point toward new ways of evaluating informed consent communication: (1) for responsiveness to specific prospective participant populations, and (2) effectiveness at eliciting informed questions from people considering participation. We discuss how creating partnerships with prospective participants to prototype informed consent materials, specifically study procedures and risks, can be a way to identify those details before launching a study. Furthermore, future research should go beyond "readability" to explore alternate measures of evaluating consent materials, such as the likelihood that the consent material and communication procedures will elicit "informed questions" for the research team.