The following is a summary of “Problem of Pain in Rheumatology: Variations in Case Definitions Derived From Chronic Pain Phenotyping Algorithms Using Electronic Health Records,” published in the March 2024 issue of Rheumatology by Falasinnu et al.
Researchers conducted a retrospective study to analyze and contrast various case definitions for chronic pain. They aimed to offer insights into potential misclassifications in research due to constraints of electronic health records and administrative claims data, thereby enhancing the precision of case identification.
They assessed the prevalence of various case definitions for chronic pain (N = 3042) among individuals with autoimmune rheumatic diseases. Chronic pain prevalence was calculated using 15 distinct combinations of pain scores, diagnostic codes, analgesic medications, and pain interventions.
The results showed that chronic pain prevalence varied across different pain phenotyping algorithms. The lowest prevalence rates were observed in unimodal algorithms: 15% for analgesic medications, 18% for pain scores, 21% for pain diagnostic codes, and 22% for pain interventions. A well-validated algorithm reported a prevalence of 37%. Chronic pain prevalence also increased as the number of modalities increased in multimodal algorithms (bimodal to quadrimodal). The highest prevalence (47%) was found in the quadrimodal algorithm combining pain scores, diagnostic codes, analgesic medications, and pain interventions. This quadrimodal algorithm overestimated chronic pain by 10% compared to the well-validated algorithm.
Investigators concluded that significant variations in chronic pain estimates depend on phenotyping methods, highlighting the need for standardized approaches to improve data accuracy.
Source: jrheum.org/content/51/3/297
