A weight prediction algorithm developed from clinical trials accurately predicted real-world weight loss (21% at one year) from semaglutide using self-reported data from a digital patient support app.
21% weight loss at 1 yearReal-world semaglutide users in a digital support program lost an average of 22 kg, accurately predicted by the algorithm
What the researchers found
The prediction algorithm accurately forecast individual weight loss in real-world semaglutide users with bias of 0.7-1.4 kg at 6 months and -0.6 to 0.6 kg at 1 year, with AUC 0.74-0.95 for categorical weight loss prediction.
Why it matters
Personalized weight loss predictions could help patients set realistic goals and help clinicians identify early non-responders who might benefit from treatment adjustment.
How the study worked
Application of exposure-response weight prediction model (from semaglutide RCTs) to 1,797 WegovyCare app users, with model validation at multiple timepoints using self-reported dosing and body weight.
What this study cannot tell us
Self-reported data may be inaccurate. App users are self-selected (likely more motivated). 81% women limits generalizability. Algorithm assumes consistent dosing adherence.
How to read the evidence
Validation study of a prediction model in real-world data. Strong prediction accuracy but limited to self-selected app users.
When this study was published
Published in 2025.
The bigger picture
This bridges clinical trial pharmacology with digital health, showing how prediction algorithms from trials can be deployed in real-world digital tools to personalize GLP-1 therapy.
Questions still open
- Can this algorithm identify non-responders early enough to switch therapy?
- Would the algorithm work for tirzepatide and other GLP-1 drugs?
- Could integrating wearable device data improve prediction accuracy?
Common questions
Can an app predict how much weight I'll lose on semaglutide?
How much weight do people actually lose on semaglutide in real life?
Read the original research
Predicting Long-Term Weight Loss Using Self-Reported, Digitally Collected, Real-World Data After Initiation of Semaglutide for Overweight or Obesity.
Advances in therapy
Citation
Færch, Kristine; Gomes, Mikel M; Bramming, Maja; Sørensen, Mads R; Strathe, Anders. (2026). Predicting Long-Term Weight Loss Using Self-Reported, Digitally Collected, Real-World Data After Initiation of Semaglutide for Overweight or Obesity.. Advances in therapy. https://doi.org/10.1007/s12325-026-03507-5