Back

Preferences for benefit and harm outcomes of GLP-1 receptor agonists in adults with overweight or obesity: a multinational best-worst scaling study

Moll, H.; Puhan, M. A.; Gerber, P.; Beuschlein, F.; Frenes, K.; Spanu, A.; Walter, J.; Yebyo, H. G.

2025-12-16 health economics
10.64898/2025.12.15.25342259 medRxiv
Show abstract

AimsWe aimed to elicit preferences for weight and harm outcomes of glucagon-like peptide-1 receptor agonist (GLP-1 RA) treatment. We examined heterogeneity across key patient subgroups to support tailored treatment decisions to identify patients most likely to benefit with minimal risks. Material and MethodsWe conducted a best-worst scaling survey of adults with overweight or obesity in 15 European countries, assessing 21 GLP-1 RA outcomesincluding 5% and 10% weight gain (to align with harms) and 19 adverse outcomes (eg. gastrointestinal events, hypoglycaemia, pancreatitis, gallbladder outcomes). Participants were provided lay descriptions of outcomes, rated seriousness on a visual analogue scale, and completed a best-worst scaling task. Preferences were estimated using a Bayesian hierarchical mixed multinomial logit model and rescaled from 0 (least worrisome) to 1 (most worrisome). Subgroup analyses were conducted based on sex, age, body mass index (BMI), and physical activity (36 subgroups). ResultsAmong 2112 participants (48.1% female; mean age 39.51 years; median BMI 30.0 kg/m2), the least concerning outcomes were eructation (0.04) and flatulence (0.05), and the most concerning were pancreatitis (0.42), followed by 10% weight gain (0.41), cholecystitis (0.35), and cholelithiasis (0.32). Other outcome weights ranged from 0.11 to 0.25. Preference patterns were similar for extremes but varied by sex, age, BMI, and physical activity. ConclusionsPreferences for GLP-1 RA outcomes varied across individuals and subgroups. Incorporating explicit patient preferences into routine care may better align GLP-1 RA prescribing with patient values and support individualised benefit-harm decisions.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.