Prevalence and Impact of Comorbid Widespread Pain in Adults With Chronic Low Back Pain

A Registry-Based Study

John C. Licciardone, DO, MS, MBA; Vishruti Pandya, MBBS, MHA


J Am Board Fam Med. 2020;33(4):541-548. 

In This Article


Patients for this historical cohort study were selected from the Pain Registry for Epidemiologic, Clinical, and Interventional Studies and Innovation (PRECISION Pain Research Registry) within Texas from April 2016 through January 2019. A registry overview has been published.[8] Registry patients were recruited using clinic advertisements, communications with health care providers, newspaper advertisements, and social media to direct potential participants to the online screening site. Registry eligibility criteria included 1) being 21 to 79 years of age, 2) having English-language proficiency, and 3) having subacute low back pain or CLBP. However, the present study included only patients who met the NIH-RTF case definition for CLBP.[4] The latter required patients to report low back pain duration of at least 3 to 6 months, with a corresponding pain frequency of at least half of the days during the past 6 months. Registry patients are currently required to have a physician for their low back pain, as no diagnostic testing or treatment is provided by the registry. Study patients were required to attend the baseline registry encounter in person; however, follow-up encounters may have been completed in person, telephonically, or online. Regardless of the manner in which encounters were completed, all data were self-reported by patients and entered using electronic case report forms that precluded missing item responses. Thus, missing data were attributable only to missed encounters or registry attrition. Enrolled patients with complete data at each subsequent quarterly encounter over 12 months were included in the study. Pregnant or institutionalized patients were excluded. The North Texas Regional Institutional Review Board approved the study.

The annual period prevalence of WP was assessed using the NIH-RTF minimum dataset item that measured the bothersomeness of "widespread pain or pain in most of your body" during the past 4 weeks ("not bothered at all," "bothered a little," or "bothered a lot") at each quarterly encounter.[4] Patients who consistently reported being "not bothered at all" over 12 months were considered to not have WP. Patients who consistently reported being "bothered a lot" over 12 months were considered to have persistent WP. The remaining patients were considered to have nonpersistent WP.

Baseline clinical measures included the patient-reported history of depression, Pain Catastrophizing Scale,[9] Pain Self-Efficacy Questionnaire,[10] numeric rating scale for low back pain intensity during the past 7 days, Roland-Morris Disability Questionnaire for back-related disability,[11] and Patient-Reported Outomes Measurement Information System (PROMIS-29)[12] Quality of life was measured using the SPADE cluster based on PROMIS-29 scales for sleep disturbance, pain interference with activities, anxiety, depression, and low energy/fatigue. The measures of low back pain intensity, back-related disability, and quality of life were repeated at quarterly intervals over 12 months. Contingency table methods and analysis of variance were used to compare WP groups at baseline. Multiple logistic regression was used to identify baseline factors associated with the presence of WP (either nonpersistent or persistent) over 12 months of follow-up. Predictor variables included patient sociodemographic characteristics, cigarette smoking status, CLBP duration greater than 5 years, history of low back surgery, history of 9 spinal or medical conditions, tercile levels of pain catastrophizing and pain self-efficacy (low, moderate, or high), and current use of opioids for low back pain. Repeated measures analysis of variance was used for longitudinal comparisons of the 3 WP groups with regard to clinical measures involving low back pain intensity, back-related disability, and quality of life. Statistical analyses were performed using the IBM SPSS Statistics software package (version 25, Armonk, NY). All hypotheses were assessed at the 0.05 level of statistical significance using 2-sided tests.