JMIR Publications

ABSTRACT

Background:

Despite ubiquity of smartphones, there is little guidance for how to design mobile health apps to increase use. Specifically, knowing what features users expect, grab their attention, encourage use (via predicted use or through positive app evaluations), and signal beneficial action possibilities can guide and focus app development efforts.

Objective:

We investigated what features users expect and how the design (prototypicality) impacts app adoption.

Methods:

In an online survey, we elicited expectations, including presence and placement, for 12 app features. Following, participants (n=462) viewed two health apps (high prototypicality similar to top downloaded apps, vs. low prototypicality similar to research interventions) and reported willingness to download, attention, and predicted use of app features. Participants rated both apps (high and low) for aesthetics, ease of use, usefulness, perceived affordances, and intentions to use.

Results:

Most participants (92%, 425 of 462) expected features for navigation or personal settings (e.g., menu) in specific regions (e.g., top corners). Features with summary graphs or statics were also expected by many (86%, 395-396 of 462), with a center placement expectation. A feature to “share with friends” was least expected among participants (44%, 203 of 462). Features fell into four unique categories based on attention and predicted use, including: essential features with high (>50%) predicted use and attention (e.g., calorie trackers), flashy features with high attention but lower predicted use (e.g., links to specific diets), functional features with modest attention and low use (e.g., settings), and mundane features with low attention and use (e.g., discover tabs). When given a choice, most (75%) participants would download the high prototypical app. High prototypicality apps (vs. low) led to greater aesthetics, ease of use, usefulness, and intentions, all p<.001. Participants thought high prototypicality apps had more perceived affordances.

Conclusions:

Intervention designs that fail to meet a threshold of mHealth expectations will be dismissed as less usable or beneficial. Individuals who download health apps have shared expectations for features that should be there, as well as where these features should appear. Meeting these expectations can improve app evaluations and encourage use. Our typology should guide presence and placement of expected app features to signal value and increase use to impact preventive health behaviors. Features that will likely be used and are attention getting – essential flashy, and functional – should be prioritized in app development.