Abstract:Background: There is mounting evidence that objective measures of pathophysiology do not correlate well with symptom intensity. A growing line of inquiry identifies statistical combinations (so-called "phenotypes") of various levels of distress and unhelpful thoughts that are associated with distinct levels of symptom intensity and magnitude of incapability. As a next step, it would be helpful to understand how distress and unhelpful thoughts interact with objective measures of pathologic conditions such as the radiologic severity of osteoarthritis. The ability to identify phenotypes of these factors that are associated with distinct levels of illness could contribute to improved personalized musculoskeletal care in a comprehensive, patient-centered model. Questions/purposes: (1) When measures of mental health are paired with radiologic osteoarthritis severity, are there distinct phenotypes among adult patients with hip and knee osteoarthritis? (2) Is there a difference in the degree of capability and pain self-efficacy among the identified mental health and radiologic phenotypes? (3) When capability (Patient-reported Outcomes Measurement Information System Physical Function [PROMIS PF]) is paired with radiographic osteoarthritis severity, are there distinct phenotypes among patients with hip and knee osteoarthritis? (4) Is there a difference in mental health among patients with the identified capability and radiologic phenotypes? Methods: We performed a secondary analysis of data from a study of 119 patients who presented for musculoskeletal specialty care for hip or knee osteoarthritis. Sixty-seven percent (80 of 119) of patients were women, with a mean age of 62 ± 10 years. Seventy-six percent (91 of 119) of patients had knee osteoarthritis, and 59% (70 of 119) had an advanced radiographic grade of osteoarthritis (Kellgren-Lawrence grade 3 or higher). This dataset is well-suited for our current experiment because the initial study had broad enrollment criteria, making these data applicable to a diverse population and because patients had sufficient variability in radiographic severity of osteoarthritis. All new and returning patients were screened for eligibility. We do not record the percentage of eligible patients who do not participate in cross-sectional surveys, but the rate is typically high (more than 80%). One hundred forty-eight eligible patients started the questionnaires, and 20% (29 of 148) of patients did not complete at least 60% of the questionnaires and were excluded, leaving 119 patients available for analysis. We measured psychologic distress (Patient Health Questionnaire-2 [PHQ-2] and Generalized Anxiety Disorder-2 questionnaire [GAD-2]), unhelpful thoughts about pain (Pain Catastrophizing Scale-4 [PCS-4]), self-efficacy when in pain (Pain Self-Efficacy Questionnaire-2), and capability (PROMIS PF). One of two arthroplasty fellowship-trained surgeons assigned the Kellgren-Lawrence grade of osteoarthritis based on radiographs in the original study. We used a cluster analysis to generate two sets of phenotypes: (1) measures of mental health (PHQ-2, GAD-2, PCS-4) paired with the Kellgren-Lawrence grade and (2) capability (PROMIS PF) paired with the Kellgren-Lawrence grade. We used one-way ANOVA and Kruskal-Wallis H tests to assess differences in capability and self-efficacy and mental health, respectively. Results: When pairing measures of psychologic distress (PHQ-2 and GAD-2) and unhelpful thoughts (catastrophic thinking) with the grade of radiographic osteoarthritis, six distinct phenotypes arose. These groups differed in terms of capability and pain self-efficacy (for example, mild pathology/low distress versus average pathology/high distress [PROMIS PF, mean ± standard deviation]: 43 ± 6.3 versus 33 ± 4.8; p = 0.003). When pairing the degree of capability (PROMIS PF) with the Kellgren-Lawrence grade, four distinct phenotypes arose. Patients in three of these did not differ in terms of disease severity but had notable variation in the degree of limitations. Patients with these radiologic and capability phenotypes differed in terms of distress and unhelpful thoughts (for example, moderate pathology/low capability versus mild pathology/high capability [PHQ-2, median and interquartile range]: 3 [1 to 5] versus 0 [0 to 0]; p < 0.001). Conclusion: Statistical groupings ("phenotypes") that include both measures of pathology and mental health are associated with differences in symptom intensity and magnitude of incapability and have the potential to help musculoskeletal specialists discern mental and social health priorities. Future investigations may test whether illness phenotype-specific comprehensive biopsychosocial treatment strategies are more effective than treatment of pathology alone. Level of evidence: Level III, prognostic study.

Identifying phenotypes in persons with temporomandibular disorders using latent class analyses

Heterogeneity in response to treatment across tinnitus phenotypes

Catastrophizing Has a Better Prediction for TMD Than Other Psychometric and Experimental Pain Variables

An exploratory data-driven approach to classify subgroups of patients with temporomandibular disorders based on pain mechanisms

Disease Phenotypes in Refractory Musculoskeletal Pain Syndromes Identified by Unsupervised Machine Learning

Is there a superiority of multimodal as opposed to simple therapy in patients with temporomandibular disorders? A qualitative systematic review of the literature

Are There Distinct Statistical Groupings of Mental Health Factors and Pathophysiology Severity Among People with Hip and Knee Osteoarthritis Presenting for Specialty Care?

Clinical Phenotypes of Hip Osteoarthritis Across the Biopsychosocial Spectrum

Premorbid and concurrent predictors of TMD onset and persistence

Cause-Effect Relationships between Painful TMD and Postural and Functional Changes in the Musculoskeletal System: A Preliminary Report

Prevalence of temporomandibular disorder in children and adolescents with juvenile idiopathic arthritis – a Norwegian cross- sectional multicentre study

Polymorphisms in the HTR2A and HTR3A Genes Contribute to Pain in TMD Myalgia

Treatment options used in the management of people with temporomandibular disorders by Australian dentists and physiotherapists

Identifying latent subgroups in the older population seeking primary health care for a new episode of back pain – findings from the BACE-N cohort

Oral Health–related Quality of Life of Patients with Acute and Chronic Temporomandibular Disorder Diagnostic Subtypes

Correlates of jaw functional limitation, somatization and psychological distress among different temporomandibular disorder diagnostic subtypes

Longitudinal trends in temporomandibular joint disorder symptoms, the impact of malocclusion and orthodontic treatment: A 20-year prospective study

Characteristics of painful temporomandibular disorders and their influence on jaw functional limitation and oral health‐related quality of life

The Role of Mental Health and Behavioral Disorders in the Development of Temporomandibular Disorder: A SWEREG-TMD Nationwide Case-Control Study

Moving towards a taxonomy of cognitive impairments in epilepsy: application of latent profile analysis to 1178 patients with temporal lobe epilepsy

Identification of multimodal mental health signatures in the young population using deep phenotyping