Kidney Week

Abstract: FR-PO0050

Applying the SHARE Framework to Evaluate Shared Decision-Making in KDpredict Implementation: A Qualitative Focus Group Study

Session Information

• Artificial Intelligence: Clinical Prediction, Decision Support, and Digital Kidney Care
  October 23, 2026 | Location: Exhibit Hall A, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Artificial Intelligence, Digital Health, and Data Science

• 300 Artificial Intelligence, Digital Health, and Data Science

Authors

• Naeem, Iffat, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Iffat Naeem, MD

• Donald, Maoliosa, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Maoliosa Donald

• Cheng, Michelle, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Michelle Cheng

• Hui, Maggie, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Maggie Hui

• Virani, Shane, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Shane Virani, MSc

• Rizai, Mastoora, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Mastoora Rizai, MPH

• Ravani, Pietro, University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Pietro Ravani, MD, PhD

• Elliott, Meghan J., University of Calgary Cumming School of Medicine, Calgary, Alberta, Canada

Meghan J. Elliott, MD

Background

Shared decision-making (SDM) gaps persist in chronic kidney disease (CKD) care. Risk prediction tools that estimate both kidney failure and mortality risk, such as KDpredict, may facilitate SDM, but little research has addressed how such tools support collaborative decision-making. This study applied the Agency for Healthcare Research and Quality’s SHARE framework (Seek, Help, Access, Reach, and Evaluate) to assess how KDpredict facilitates or hinders SDM from provider and patient/caregiver perspectives.

Methods

A qualitative study using focus groups to inform integration of KDpredict into Alberta, Canada's provincial electronic health record (EHR). Six focus groups included healthcare providers (n=12; nephrologists, primary care physicians, nurses) and patients/caregivers (n=13). Transcripts were analyzed using directed content analysis with deductive coding to SHARE domains and valence coding to classify barriers, facilitators, and dual-valence factors.

Results

For Seek, KDpredict gave providers confidence initiating conversations, while patients confirmed trust and transparency as prerequisites; however, both groups identified discomfort with mortality discussions as a barrier. For Help, personalized visuals enhanced understanding, though both groups noted interpretation difficulties; patients wanted tailored over generalized content. For Assess, mortality information was a dual-valence factor, motivating planning for some and causing distress for others; patients wanted scenario-based input to explore how choices affect outcomes. For Reach, both groups identified lack of actionable guidance as a barrier; team-based and caregiver involvement facilitated decisions. For Evaluate, both groups desired longitudinal tracking yet no systematic processes existed. Cross-cutting barriers included absent EHR/patient portal integration and role-specific training gaps, with clustering in early SHARE steps (Seek, Help).

Conclusion

Preliminary findings suggest KDpredict has strong potential as an SDM facilitator, but organizational and relational barriers need to be assessed across individual, clinical, and health system levels. Further integration of provider and patient/caregiver perspectives will help refine our understanding of how barriers and facilitators interact and inform targeted implementation recommendations