Persuasive Systems Design Features of Smartphone Apps for Psychosis: Systematic Review.

Background: It is unclear why some smartphone apps designed for people with psychosis are engaging, while others are not. One possible explanation is the apps' persuasive features and the operationalization and implementation of these features. Objective: This systematic review set out to quantify and describe the persuasive features used in smartphone apps for psychosis, investigate whether there was any association between persuasive features and attrition or adherence rates. and document the quality of the included apps. Methods: We searched electronic databases PsycINFO, PubMed, and Google Scholar for eligible papers published between the years of 2013 and 2025. Hand searches of reference lists were completed. Apps were selected if they were designed for people with psychosis and there were published empirical studies investigating the apps. Two reviewers from the review team (TT, JP, DK, and SD) independently screened papers and extracted data on adherence and attrition, as well as coded papers for evidence of persuasive features according to the persuasive systems design (PSD) model. Available data were synthesized descriptively and narratively. We attempted to access apps via app stores or by correspondence with the research team. Results: We found 22 apps for psychosis, with 30 associated published papers. The persuasive features were as follows: personalization (18 apps); reminders (15 apps); suggestions (11 apps); tunneling and self-monitoring (10 apps); reduction (9 apps); liking (8 apps); social role (6 apps); rehearsal, praise, and similarity (5 apps); rewards (4 apps); simulation, real-world feel, and social learning (3 apps); surface credibility and normative influence (2 apps); and trustworthiness, social comparison, and social facilitation (1 app). Expertise, authority, third-party endorsements, verifiability, cooperation, competition, and recognition were present in zero apps. Features in the categories of primary task support and dialogue support were well represented, while social support and system credibility support were underused. It was found that there was no association between the number of persuasive features and attrition; an association between persuasive features and adherence could not be assessed. The quality of the apps could not be judged due to 20 of the 22 apps being inaccessible either through the research papers' authors or through app stores. Conclusions: Our findings indicate that in psychosis apps there is potential to include a broader range of persuasive features, which might maximize engagement. Psychosis apps may benefit from incorporating more features that leverage the persuasive impact of having users interact (social support) and incorporating features that emphasize system credibility and trustworthiness. Further studies could determine whether an increase in the number of persuasive features will impact app engagement and which features are most impactful in this context.