The Anatomy of a Failed DiGA: Taxonomic Insights from Withdrawn Reimbursable Digital Therapeutics

Germany's Digital Health Application (DiGA) pathway offers a fast-track route to reimbursement for digital therapeutics, yet a share of initially listed applications are subsequently withdrawn from the directory.

This study examines which software features are associated with sustained listing versus delisting, addressing a gap in the literature on unsuccessful DiGA trajectories.

We compared 35 permanently listed and 17 delisted DiGAs (n = 52) using qualitative feature coding following the Gioia methodology and the framework of Weimar et al. (2025). Feature prevalence at the 2nd Order theme and Aggregate Dimension levels was compared using Fisher's Exact Test.

Overall feature breadth was identical across groups. Two themes were significantly more prevalent among delisted DiGAs: Reminders and Notifications and Reports and User Data Export. No significant differences were found at the Aggregate Dimension level or across indication groups.

The findings suggest that continued reimbursement is associated with feature composition rather than quantity. Delisted DiGAs more often incorporated engagement- and infrastructure-oriented features, whereas listed DiGAs more often included components closer to the therapeutic mechanism.

Feature composition, rather than breadth, may differentiate sustained from withdrawn DiGAs, offering preliminary orientation for developers navigating the pathway and motivation for a more granular taxonomic framework in future research.