Predicting GLP-1 Discontinuation From Pharmacy Claims.
Level 5 - mechanism / opinion, no new human data
Expert commentary and framework proposal without new empirical data
PubMed 42663544 · doi:10.37765/ajmc.2026.89993
What was done
This commentary outlines a pragmatic, claims-based framework to address high real-world discontinuation rates of glucagon-like peptide-1 receptor agonists (GLP-1 RAs). The authors propose using routine pharmacy claims to track new-start cohorts, calculate metrics like proportion of days covered, identify early warning signs (such as refill delays, extended starter-dose use, high out-of-pocket costs, and proxy signals for gastrointestinal adverse events), and deploy rule-based alerts or predictive models for targeted clinical outreach.
What was found
No new empirical data were collected or analyzed. The authors reference existing real-world persistence data noting that approximately one-third of individuals using GLP-1 RAs for obesity remain on therapy at 1 year, while for type 2 diabetes cohorts, nearly 50% discontinue by 12 months and approximately 70% discontinue by 24 months.
Why it matters
Real-world non-persistence limits the clinical benefits and cost-effectiveness of GLP-1 therapies. A structured claims-based monitoring framework could help payers and health systems intervene early to support adherence through side-effect management and cost navigation.
Limits
The proposed framework and predictive metrics are purely conceptual and were not prospectively or retrospectively validated in a patient cohort. The abstract provides no data on model accuracy, intervention efficacy, operational costs, or actual impact on health outcomes.
Cited by
- supports Approximately 70% of Americans prescribed GLP-1 agonist medications discontinue taking the drug within two years.