Application of Innovation Tournament methodology to inform de-implementation strategies: lessons learned in addressing mammography overscreening in older women

De-implementation is a critical, understudied area within implementation science, particularly de-implementation strategy development. Aligned with most national guidelines, de-implementation of routine mammography screening among older women (without considering patient preferences, health, life expectancy), is a compelling area for advancing de-implementation science. While previously applied for implementation, this research seeks to advance understanding of Innovation Tournaments (InTs) as a participatory methodology to inform de-implementation strategy development.

Building on formative qualitative research and informed by User-Centered Design (UCD), we conducted an InT to generate, confirm, and refine de-implementation strategies among purposively selected providers and administrators across two healthcare systems in New York City. InT development began 07/2021 and launched 11/2022-05/2023. A key challenge in co-creating, launching, and maintaining the InT was difficulty communicating about de-implementation, especially given differences in opinion about guidelines across provider specialties. To address this complexity, we collaborated with experts in InTs, shared decision-making, healthcare informatics, communication, creative content development, and UCD. Based on input during the UCD process and to facilitate engagement across participants with varying opinions, we expanded the InT to assess ideas and opinions about mammography guidelines for older women and appropriateness of de-implementation, in addition to assessing de-implementation strategy ideas. We conducted descriptive analyses of InT forum engagement data and content analysis of forum posts/comments.

Of 47 participants who entered the InT website, 36 engaged in the forum (voted, commented, or posted). Participants voted 256 times on posted ideas/opinions/strategies and comments, submitted 71 comments, and posted 13 new ideas/opinions/strategies. Most participants engaged in the forum during their initial visit to the InT, with limited engagement over time. Content analysis of posts/comments identified de-implementation strategy categories largely aligned with the ERIC taxonomy (e.g., patient education, provider training/reminders), though system-level strategies specific to de-implementation emerged.

Findings advance de-implementation methods by indicating importance of formative qualitative research and UCD to inform InT development and engagement (particularly if stakeholders disagree on de-implementation appropriateness). Unique considerations include: how best to seed InT content; challenges in generating new strategies that meaningfully expand upon formative qualitative research findings; and time and resources needed to develop, elicit and sustain stakeholder engagement over time.