Background Chronic low back pain (cLBP) is the leading cause of disability worldwide. Current treatments have minor or moderate effects, partly because of the idiopathic nature of most cLBP cases, the complexity of its presentation, and heterogeneity in the population. Explaining this complexity and heterogeneity by identifying subgroups of patients is critical for personalized health. Clinical decisions tailoring treatment to patients’ subgroup characteristics and specific treatment responses can improve health outcomes. Current patient stratification tools divide cases into subgroups based on a small subset of characteristics, which may not capture many factors determining patient phenotypes.

Methods and Findings In this study, we use an unsupervised machine learning framework to identify patient subgroups within a specialized back pain clinic and evaluate their outcomes. Our analysis identified 25 latent factors determining patient phenotypes and found three distinctive clusters of patients. The research suggests that there is heterogeneity in the population of patients treated in a specialty setting and that several factors determine patient phenotypes. Cluster 1 consists of those individuals with characteristics found to be protective of chronic pain: younger age, low pain medication prescription, high function, good insurance access, and low overlapping pain conditions. Individuals in Cluster 3 associate with older age and present with a higher incidence of chronic overlapping pain conditions, comorbidities, and pain medication use. Cluster 2 is an intermediate group.

Conclusions We quantify cLBP population heterogeneity and demonstrate how ML analytical workflow can be used to explain, in part, this heterogeneity in relation to outcomes. Notably, considering a data-driven approach from multi-domain data produces different subgroups than the STarT back screening tool, and the addition of other functional metrics at baseline such as global physical and mental function, and pain intensity, increases the variance explained in outcomes. Our study provides novel insights into the complex nature of cLBP and the potential for data-driven methods to identify clinically relevant subtypes.

The authors have declared no competing interest.

Research reported in this publication was supported by the National Institute Of Arthritis And Musculoskeletal And Skin Diseases of the National Institutes of Health under Award Number U19AR076737. AS acknowledges funding from National Science Foundation Grant DMS-2210206.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Ethics committee/IRB of the University of California San Francisco gave ethical approval for this work.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors