Background Weight gain is a common side effect in patients treated with olanzapine (N05AH03), contributing to increased risks of metabolic complications such as diabetes, cardiovascular disease, and reduced treatment adherence. However, personalised prevention strategies are currently lacking in day-to-day clinical practice. Identifying factors that can predict which patients are most likely to gain a significant amount of weight is therefore critical. Such insights could enable early risk prediction and guide clinicians in implementing personalized interventions to minimize side effects and improve long-term treatment outcomes.

Methods In this study, we developed explainable machine learning models using population-based Electronic Health Record data and compared the performance between logistic regression, decision trees and ensemble tree classifiers like XGboost to predict significant BMI increase of >5% in the next contact to the clinic while on olanzapine.

Results XGBoost model achieved the highest performance for predicting a >5% BMI increase on olanzapine with an AUROC of 0.72, surpassing logistic regression (AUROC = 0.69) and other evaluated classifiers. Model performance was consistent across sexes, but varied across age groups, with the highest performance for individuals 30-69 years of age (AUROC = 0.73) and lowest in individuals over 70 years of age (AUROC = 0.67). The SHAP analysis highlighted several key predictive features including prolonged intervals between follow-up visits, higher baseline BMI, younger age, shorter time since olanzapine initiation, cumulative hospitalization days, increasing olanzapine dosage, polypharmacy and especially concurrent use of sedatives as well as multiple prescriptions with anxiolytics.

Conclusion Our models validate prior known risk factors for olanzapine induced weight gain and further uncover previously unknown factors that influence >5% BMI increase on olanzapine. These findings underscore the need for continued research in this domain to establish effective preventive measures for individuals undergoing antipsychotic treatments.

S.R. is the founder and owner of BioAI and has received a research grant from Sidera Bio ApS. The remaining authors declare no conflicts of interest.

F.A. was funded by the Novo Nordisk Foundation Copenhagen Bioscience Ph.D. Program Grant Agreements No. NNF0078229 and NNF0078230. This work is supported by the Novo Nordisk Foundation (NNF23SA0084103) and by unrestricted grants from the Lundbeck Foundation (grant numbers R278-2018-1411 and R383-2022-285).

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

The Data Protection Agency of the Capital Region of Denmark and the Danish Society for Patient Safety (approval numbers: P-2020-101 and R-22002033), gave ethical approval for this work. All data processing was conducted in compliance with stringent data protection regulations to safeguard patient confidentiality. According to Danish regulations the analyses do not require informed consent.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

The project was approved by the Data Protection Agency of the Capital Region of Denmark and the Danish Society for Patient Safety (approval numbers: P-2020-101 and R-22002033), and all data processing was conducted in compliance with stringent data protection regulations to safeguard patient confidentiality. According to Danish regulations the analyses do not require informed consent. The data used in this study are not publicly available in order to comply with Danish data protection regulations and safeguard patient confidentiality. Access to the data may be granted to qualified researchers who meet the criteria for access to confidential health data, subject to approval by the relevant Danish authorities.