Influence of Audio Speech Rate and Source Text Difficulty on Health Information Comprehension and Retention

Abstract

Health literacy is crucial for patients to make informed healthcare decisions. Although text has historically been the main form of health information dissemination, people rely increasingly on audio-delivered information, e.g., through smart speakers. In this study, we evaluate the effects of audio speech rate and text difficulty on audio information comprehension and retention. We created audio snippets from easy and difficult text and conducted a study on Amazon Mechanical Turk (AMT). Audio speech rate and source text difficulty are the independent variables and perceived difficulty (measured with a Likert scale) and comprehension and retention (measured with AI-generated multiple-choice questions and free recall of information) are the dependent variables. Audio created from difficult source text was perceived as more difficult and comprehension was also lower than for audio from easy text. Speech rate also influenced information comprehension and retention of information: a higher speech rate (+60% faster audio speech rate) lowered the comprehension of health information by 38% compared to a moderate speech rate.