Related Experiment Video
Updated: Jun 5, 2025
14:56
An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP
Published on: January 27, 2010
21.3K
Prediction of Multisite Pain Incidence in Adolescence Using a Machine Learning Approach: A 2-Year Longitudinal Study
Laura Joensuu1, Ilkka Rautiainen1, Arto J Hautala1
1Faculty of Sport and Health Sciences University of Jyväskylä Jyväskylä Finland.
Health Science Reports
|December 11, 2024
Summary
Multisite pain in adolescents is predicted by various factors, not just one. Machine learning identified numerous predictors, with more identified in girls than boys, highlighting complex contributing elements.
Area of Science:
- Adolescent Health
- Pain Research
- Machine Learning in Medicine
Background:
- Multisite pain is a common and serious issue in adolescents.
- It is linked to negative physical, psychological, and social outcomes.
Purpose of the Study:
- To predict the incidence of multisite pain in adolescents.
- To identify sex-specific predictors of multisite pain using machine learning.
Main Methods:
- A 2-year longitudinal study of 410 Finnish adolescents (12.5 years old).
- Utilized machine learning (random forest, AdaBoost, support vector classifier) with extensive baseline data.
- Defined multisite pain as self-reported pain in at least three sites, excluding disease-related pain.
Main Results:
- 16% of boys and 28% of girls developed whole-body multisite pain.
- Machine learning models achieved a maximum AUC of 0.78.
- Identified up to 33 predictive variables in girls and 13 in boys.
Conclusions:
- Multisite pain in adolescents arises from a combination of factors.
- Machine learning effectively identified a wide range of predictors, with sex-specific patterns observed.
