Chronic Pain Challenge: A Statistical Machine-learning Method for Chronic Pain Assessment

Abstract

Aim: The objective of Chronic Pain Challenge project is designing and construction of a machine-learning system to calculate the dynamic changes to the chronic pain risk score of an individual based on various weighted health behaviors. Materials and methods: The visual analog scale (VAS) and Oswestry Disability Index (ODI) ratings of 218 subjects were studied for dynamic changes based on three weighted health behaviors, physical exercise, nutrition, and depression in order to predict their individual and cumulative impact on severity of chronic pain. The predictive function was used to produce confidence and prediction intervals for the calculation of new VAS and ODI scores using supervised and unsupervised machine-learning algorithms and R programing language for statistical computation. Results: This 9 months research study resulted in the development of innovative design and construction of a machine-learning program that accurately predicted the changes to standardized tests, such as VAS and ODI based on weighted values for depression score (DS), nutrition score (NS) and physical activity score (PAS). The testing of both extreme and moderate ranges of health behavior values in a variety of subjects and comparison against simple weightage confirmed the accuracy and validity of the program. Conclusion: Chronic Pain Challenge program is a valid and accurate method in predicting chronic pain risk of an individual based on the engagement in various health behaviors. The Chronic Pain Challenge program can predict and prevent progression of chronic pain and disability by global education and empowerment, thereby disrupting the current health care model with the emerging and accelerating technology. Clinical significance: The Chronic Pain Challenge program is an innovative statistical machine-learning program for chronic pain predictability based on individual's health behavior patterns.