Photo Credit: PhonlamaiPhoto
The following is a summary of “Haves and have-nots: socioeconomic position improves accuracy of machine learning algorithms for predicting high-impact chronic pain,” published in the October 2024 issue of Pain by Morris et al.
A lower socioeconomic position (SEP) was linked to a higher risk of developing chronic pain, experiencing more severe pain, and suffering more significant pain-related disability.
Researchers conducted a retrospective study to determine the SEP features with high-impact chronic pain, compare the relative importance to established correlates, and to examine the differences by race and sex.
They employed 3 machine learning algorithms to investigate questions among adults in the 2019 National Health Interview Survey. Gradient boosting decision trees demonstrated the highest accuracy and discriminatory power in predicting high-impact chronic pain.
The results showed that different dimensions of SEP, such as material resources (for example, the ratio of family income to the poverty threshold) and employment factors (such as working status in the past week and the number of employed adults in the family), were significant predictors of high-impact chronic pain. The analysis compared the factors between non-Hispanic Black and White adults, as well as between men and women, for non-Hispanic Black adults, body mass index and homeownership/rental status held greater importance, while for non-Hispanic White adults, the number of working adults and housing stability were more critical. Among women, anxiety severity, body mass index, and smoking were significant, whereas for men, housing stability and the frequency of anxiety and depression symptoms were more relevant.
Investigators concluded that machine learning algorithms had the potential to advance health equity research by identifying the critical predictors of high-impact chronic pain across diverse populations.
Source: journals.lww.com/pain/abstract/9900/haves_and_have_nots__socioeconomic_position.736.aspx
