Will the biopsychosocial model of medicine survive in the age of artificial intelligence and machine learning?

Albert F G Leentjens*, Stephen Leslie Smith

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Background: the biomedical model of medicine was replaced by the biopsychosocial model in order to better accommodate psychological and social aspects of illness. The introduction of machine learning techniques provides the perspective of truly personalized medicine. This poses new challenges to our medical model.
Aim: to explore the implications of personalized medicine for the biopsychosocial model.
Methods: scholarly reflection.
Results: The ability of machine learning technology to integrate a wide diversity of data makes it possible to develop predictive models for presentation, course and treatment response in individual patients. Such models are based on individual risk factors and protective factors that may have diverging influences in different individuals. In a medical model adjusted to accommodate the possibilities of personalized medicine, it should be possible to highlight the importance and impact of each single factor in each individual patient. At present, the
biopsychosocial model is not well prepared for this.
When adopting machine learning technology in clinical practice, new skills and expertise will be required from physicians. They should be able to weigh and explain algorithms supported decisions to their patients. Moreover,
new research should be designed in such a way that data will be suited for machine learning and can be integrated with existing databases in order to increase their size and scope.
Conclusion: Currently, the biopsychosocial model is not well prepared to accommodate the possibilities of personalized medicine. Adaptations are needed to deal with the highly individual aspects of the patient's disease.
Original languageEnglish
Article number111207
Number of pages4
JournalJournal of Psychosomatic Research
Early online date13 Mar 2023
Publication statusPublished - 1 May 2023

Bibliographical note

This is an author-produced version of the published paper. Uploaded in accordance with the publisher’s self-archiving policy.

Cite this