Biobehavioral phenotypes of chronic low back pain: Psychosocial subgroup identification using latent profile analysis.

OBJECTIVE

This study identifies distinct biobehavioral phenotypes among patients with chronic low back pain (cLBP) using Latent Profile Analysis (LPA).

METHODS

These phenotypes were derived from baseline data from two cohorts within the NIH HEAL BACPAC consortium: BACKHOME, a large nationwide e-cohort (N = 3025) utilized for model training, and COMEBACK as external test set, a deep phenotyping cohort (N = 450) utilized for generalization. The analysis incorporated variables including pain characteristics, psychosocial factors, lifestyle habits, and social determinants of health. Model fit was optimized via 10-fold cross-validation with 100 bootstraps and evaluated using Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Entropy(uncertainty).

RESULTS

Four classes were identified: Class 1 ("High Distress and Maladaptive Behaviors") displayed high levels of anxiety, depression, and fear avoidance. Class 2 ("Resilient and Adaptive Coping") exhibited low maladaptive behaviors and high pain self-efficacy. Class 3 ("Intermediate Maladaptive Patterns") represented moderate levels of psychological and behavioral challenges, while Class 4 ("Emotionally Regulated with High Pain Burden") demonstrated strong emotional regulation despite significant pain burden. Class sizes were 701, 413, 893, and 947 for the train set, and 127, 108, 95, and 68 for the test set, respectively. Fit metrics supported the model's performance and generalizability (BACKHOME (train set): AIC = 77 792, BIC = 78 338, Entropy = 0.82; COMEBACK(test set): AIC = 72 437, BIC = 73 880, Entropy = 0.81). Statistical analysis revealed significant differences between classes (P < .05) in key variables such as pain self-efficacy, fear avoidance, and emotional awareness, and changes in pain severity and health-related quality of life over time (P ≤ .001), indicating clinical utility.

CONCLUSIONS

Our findings highlight the heterogeneity of cLBP and suggest that tailored treatments targeting these distinct subgroups could improve clinical outcomes. This work advances our understanding of cLBP by providing a robust framework for identifying patient subgroups based on biobehavioral characteristics. Results underscore the value of LPA in uncovering clinically meaningful patterns in complex conditions like cLBP, paving the way for more personalized treatment approaches.