Policymakers often rely on pre-release polls to design and deploy public health apps. However, these apps often fail to realize their full potential because the actual size of the user population is significantly lower than expected, and the characteristics of those who use the app differ from those previously inclined to install it. This study aims to understand the success of public health apps by examining shifts between expected and actual user populations during app releases. Rooted in the context of public health apps, we identify factors that potentially intervene in the time between an initial poll of intentions and actual behavior. We delineate two deviations between stated installation intentions in polls and actual installations at an individual level, which have been commingled in literature thus far: inclined nonusers (Type I deviation) and disinclined users (Type II deviation). By introducing this distinction, we hypothesize that app releases can suffer from (1) a volume shift, which results in overly optimistic expectations of user numbers, and (2) a profile shift, which results in biased expectations of user profiles. Using data from a multi-wave survey, we find evidence for the postulated deviations. Our findings also unveil how contextualized factors from the health belief model can explain the occurrence and extent of these deviations. The findings contribute to information systems research that aims to predict technology adoption based on behavioral intentions. The results also offer actionable guidance for policymakers who rely on stated intentions in polls to tackle public health issues.
Read full abstract