INTRODUCTION: Chronic low back pain (cLBP) is a common and heterogeneous condition that encompasses multiple patient-specific pathophysiological mechanisms, making accurate phenotyping a challenge. Wearable accelerometers, through continuous monitoring of daily physical activity (PA), offer valuable insights into individuals’ daily activity patterns, sedentary behaviors, and lifestyle characteristics that have the potential to contribute to identification of distinct cLBP phenotypes. This study aims to: (1) investigate the characteristics of 24-hour PA patterns based on accelerometry-derived step counts in individuals with cLBP; (2) examine the association between these PA patterns and patient-reported outcomes (PROs); and (3) explore clustering of individuals based on PA patterns to identify potential phenotypes of cLBP.
METHODS: This study utilizes data from the ongoing comeBACK study (1U19AR076737-01, NIH Back Pain Consortium-BACPAC), which seeks to inform a precision medicine approach to treating cLBP. Over 400 individuals with cLBP were recruited, and this preliminary analysis includes data from 283 participants (mean age 56.8 years, 57.0% female). Each participant wore an Actigraph GT3X+ sensor on their right hip for at least 7 days. The data were analyzed to determine 24-hour step accumulation patterns, and individual clusters were identified using K-means clustering based on activity pattern distribution percentages. Nonparametric comparisons of PA and PROs (PROMIS-29 Physical Function and Pain Interference) across clusters were performed using the Kruskal-Wallis test and Mann-Whitney U tests with Bonferroni adjustment. Spearman correlation analysis was used to assess associations between PA pattern distribution and PROs.
RESULTS: Eight distinct 24-hour activity accumulation patterns were identified: Inactive (17.3%), Low Active (25.3%), Morning Active (9.8%), Mid-Day Active (20.5%), Evening Active (5.6%), All Day Active (6.4%), Bi-Phase Active (14.5%), and Highly Active (0.6%). Based on the distribution of these 24-hour activity patterns within each participants’ 7-day sample, participants were grouped into one of five clusters: Inactive (N=53), Low Active (N=65), Bi-Phase Active (N=38), Mid-Day Active (N=68), and Variable Active Pattern (N=59). The Inactive cluster exhibited the lowest step count, poorest physical function, and highest pain interference compared to other clusters. In contrast, the Mid-Day and Variable Active Pattern clusters showed above-average step counts, higher physical function, and lower pain interference. Correlation analysis indicated that a higher percentage of Inactive patterns within the 7-day sample is associated with less favorable pain-related PROs. Conversely, a higher percentage of the Mid-Day Active pattern is associated with more favorable pain-related PROs.
DISCUSSION: We identified distinct 24-hour PA patterns among individuals with cLBP and explored their relationship with patient-reported physical function and pain interference. These findings suggest that clustering based on PA patterns could contribute to cLBP phenotyping, and identify targets for tailored physical activity interventions for cLBP management.