==== Front JAMA Netw Open JAMA Netw Open JAMA Network Open 2574-3805 American Medical Association 37378984 10.1001/jamanetworkopen.2023.20520 zoi230608 Research Original Investigation Online Only Orthopedics Improvements in Physical Function and Pain Interference and Changes in Mental Health Among Patients Seeking Musculoskeletal Care Association of Improvements in Physical Health With Mental Health Among Patients Seeking Musculoskeletal Care Association of Improvements in Physical Health With Mental Health Among Patients Seeking Musculoskeletal Care Zhang Wei PhD 1 Singh Som P. BA 2 Clement Amdiel BA 3 Calfee Ryan P. MD MSc 4 Bijsterbosch Janine D. PhD 1 Cheng Abby L. MD MPHS 5 1 Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Missouri 2 University of Missouri–Kansas City School of Medicine, Kansas City 3 Washington University School of Medicine, St Louis, Missouri 4 Division of Hand and Wrist, Department of Orthopaedic Surgery, Washington University School of Medicine, St Louis, Missouri 5 Division of Physical Medicine and Rehabilitation, Department of Orthopaedic Surgery, Washington University School of Medicine, St Louis, Missouri Article Information Accepted for Publication: April 23, 2023. Published: June 28, 2023. doi:10.1001/jamanetworkopen.2023.20520 Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2023 Zhang W et al. JAMA Network Open. Corresponding Author: Abby L. Cheng, MD, MPHS, Washington University School of Medicine, Campus Box 8233, 660 S Euclid Ave, St Louis, MO 63110 (ChengAL@wustl.edu). Author Contributions: Dr Cheng had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Zhang, Singh, Calfee, Bijsterbosch, Cheng. Acquisition, analysis, or interpretation of data: Zhang, Clement, Calfee, Cheng. Drafting of the manuscript: Zhang, Singh, Clement, Calfee, Cheng. Critical revision of the manuscript for important intellectual content: Zhang, Singh, Calfee, Bijsterbosch, Cheng. Statistical analysis: Zhang, Clement. Obtained funding: Cheng. Administrative, technical, or material support: Singh. Supervision: Calfee, Bijsterbosch, Cheng. Conflict of Interest Disclosures: None reported. Funding/Support: This study was funded by the National Institutes of Health (grant Nos. K23AR074520 and P50MH122351) and the Doris Duke Charitable Foundation. Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Meeting Presentation: The findings described in this manuscript have previously been presented at the third annual symposium of the International Musculoskeletal Mental and Social Health Consortium (I-MESH); March 30, 2023; virtual meeting. Data Sharing Statement: See Supplement 2. 28 6 2023 6 2023 28 6 2023 6 6 e232052012 2 2023 23 4 2023 Copyright 2023 Zhang W et al. JAMA Network Open. https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the CC-BY License. jamanetwopen-e2320520.pdf Key Points Question Among patients seeking musculoskeletal care, are improvements in physical function and pain interference associated with meaningful changes in symptoms of anxiety and depression? Findings In this cohort study of 11 236 US adults seeking musculoskeletal care, improvement by 2.3 population-level standard deviations (SD) or more on PROMIS Physical Function or 1.2 SD or more on PROMIS Pain Interference were required for any association with meaningful improvement in anxiety symptoms. Improvements in physical function and pain interference were not associated with meaningfully improved depression symptoms. Meaning These results suggest that clinicians and patients cannot assume that exclusively addressing the physical aspect of a musculoskeletal condition will improve symptoms of depression or potentially even anxiety. This cohort study of US adults seeking musculoskeletal care examines whether self-reported improvements in physical function and pain were associated with improvements in anxiety and depression symptoms. Importance Among patients seeking care for musculoskeletal conditions, there is mixed evidence regarding whether traditional, structure-based care is associated with improvement in patients’ mental health. Objective To determine whether improvements in physical function and pain interference are associated with meaningful improvements in anxiety and depression symptoms among patients seeking musculoskeletal care. Design, Setting, and Participants This cohort study included adult patients treated by an orthopedic department of a tertiary care US academic medical center from June 22, 2015, to February 9, 2022. Eligible participants presented between 4 and 6 times during the study period for 1 or more musculoskeletal conditions and completed Patient-Reported Outcomes Measurement Information System (PROMIS) measures as standard care at each visit. Exposure PROMIS Physical Function and Pain Interference scores. Main Outcomes and Measures Linear mixed effects models were used to determine whether improvements in PROMIS Anxiety and PROMIS Depression scores were associated with improved PROMIS Physical Function or Pain Interference scores after controlling for age, gender, race, and PROMIS Depression (for the anxiety model) or PROMIS Anxiety (for the depression model). Clinically meaningful improvement was defined as 3.0 points or more for PROMIS Anxiety and 3.2 points or more for PROMIS Depression. Results Among 11 236 patients (mean [SD] age, 57 [16] years), 7218 (64.2%) were women; 120 (1.1%) were Asian, 1288 (11.5%) were Black, and 9706 (86.4%) were White. Improvements in physical function (β = −0.14; 95% CI, −0.15 to −0.13; P < .001) and pain interference (β = 0.26; 95% CI, 0.25 to 0.26; P < .001) were each associated with improved anxiety symptoms. To reach a clinically meaningful improvement in anxiety symptoms, an improvement of 21 PROMIS points or more (95% CI, 20-23 points) on Physical Function or 12 points or more (95% CI, 12-12 points) on Pain Interference would be required. Improvements in physical function (β = −0.05; 95% CI, −0.06 to −0.04; P < .001) and pain interference (β = 0.04; 95% CI, 0.04 to 0.05; P < .001) were not associated with meaningfully improved depression symptoms. Conclusions and Relevance In this cohort study, substantial improvements in physical function and pain interference were required for association with any clinically meaningful improvement in anxiety symptoms, and were not associated with any meaningful improvement in depression symptoms. Patients seeking musculoskeletal care clinicians providing treatment cannot assume that addressing physical health will result in improved symptoms of depression or potentially even sufficiently improved symptoms of anxiety. ==== Body pmcIntroduction Physical and mental health have a complex bidirectional relationship. There is a high prevalence of comorbid physical and mental health impairment from conditions such as musculoskeletal disorders, anxiety, and depression, and patients seeking musculoskeletal care who have coexisting symptoms of anxiety and depression report more physical limitations and worse pain interference (ie, “consequences of pain on relevant aspects of a person’s life...including hindered engagement with social, cognitive, emotional, physical, and recreational activities”1) than those who do not.2,3,4,5,6,7,8,9,10,11,12,13 Mental health treatment for anxiety and depression can secondarily improve patients’ physical function and reduce their pain-related limitations, but the reverse phenomenon has not been as well investigated.14,15,16,17,18,19 That is, among patients seeking care for musculoskeletal conditions, there is mixed evidence regarding whether treatment of physical conditions is associated with spontaneous improvement in mental health symptoms.20,21,22 Currently, musculoskeletal clinicians and patients often focus on the treatment of physical concerns, with the hope that mental health–related symptoms will naturally improve as physical health improves.23,24 This practice may in part be related to unique barriers to accessing mental health care, such as societal stigma regarding mental illness, a lack of financial accessibility to mental health care, and a global shortage of mental health professionals.25,26,27 Furthermore, the structure of medical training is such that clinicians who subspecialize in treating physical impairments do not routinely receive training in addressing the mental health–related contributors to and sequelae of physical and pain-related impairments.24,28 Musculoskeletal clinicians also have discrepant opinions regarding whether addressing patients’ mental health falls within their professional role.24,28,29 Some musculoskeletal clinicians are interested in additional resources to better address mental health within the musculoskeletal care setting,30 but acquisition of these resources has remained challenging without widespread agreement regarding the need for this investment.28,29,31 A better understanding of the associations between physical health and mental health changes can guide musculoskeletal clinicians in: (1) the importance they place on addressing mental health–related symptoms as a component of their patient care, and (2) how they counsel patients regarding expectations of their symptom trajectory as their physical impairment is addressed. The goal of this study was to determine whether, among patients seeking musculoskeletal care, self-reported improvements in physical function and pain interference are each associated with meaningful improvements in self-reported symptoms of anxiety and depression. We hypothesized that clinically meaningful improvement in physical function and pain interference would each be associated with significantly improved symptoms of anxiety and depression. Methods This retrospective cohort study included patients who presented to a tertiary academic medical center between June 22, 2015, and February 9, 2022. All study data were extracted from the electronic medical record. Institutional review board approval was granted by Washington University in St Louis with a waiver of informed consent because this was a retrospective study of existing data. Study reporting follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies. Participants All study participants were adults aged 18 years or older who sought evaluation and management of 1 or more musculoskeletal conditions at an outpatient clinic of the study institution’s orthopedic department. Prior to each clinic evaluation all patients completed Patient-Reported Outcomes Measurement Information System (PROMIS) Computer Adaptive Test (CAT) Anxiety version 1.0, Depression version 1.0, Physical Function version 1.2, and Pain Interference version 1.1 measures as part of standard clinical care. Patient visits were excluded from consideration if they were missing scores for any of these measures. To capture time-varying associations between these PROMIS variables of interest while also maximizing statistical power, our primary cohort included a consecutive sample of patients who had between 4 and 6 eligible clinic visits during the study period (eMethods in the Supplement 1). Exposure and Outcome Measures Our exposures of interest were patients’ level of physical function and pain interference over time, which were quantified using their PROMIS CAT Physical Function version 1.2 and PROMIS CAT Pain Interference version 1.1 scores from each clinic visit, respectively.8,32,33,34 Our outcomes of interest were patients’ symptoms of anxiety and depression, which were quantified using their PROMIS CAT Anxiety version 1.0 and Depression version 1.0 scores from each clinic visit, respectively.35,36 PROMIS is a set of self-reported measures that was developed by the National Institutes of Health to measure multiple domains of health, irrespective of a person’s underlying medical conditions.37 Scores for each PROMIS measure are normalized to a representative sample of the general US population, with a mean score of 50 and standard deviation of 10. A higher score represents more of the domain being assessed, such that a high score on PROMIS Physical Function is favorable, whereas a low score on PROMIS Pain Interference, Anxiety, and Depression is favorable. Clinically meaningful effect sizes for symptom improvement were defined as at least 3.0 points for PROMIS Anxiety and 3.2 points for PROMIS Depression, which, based on previous studies, are the minimum clinically important differences among patients with musculoskeletal pain that also exceed the standard error of measurement for each PROMIS CAT at the study institution.38,39 Confounding Variables Patients’ age, self-reported gender, and self-reported race (based on categories defined by the US Census Bureau) were also available in the medical record and were included in all statistical models to account for potential influences of confounders. Race was included as a variable because it has previously been associated with self-reported physical and mental health in some contexts. Generalizability Analyses A second cohort was analyzed to assess the generalizability of our findings to patients who had relatively fewer visits to the study institution’s orthopedic department. Potential reasons for fewer visits include, but were not limited to, musculoskeletal conditions that were less chronic or more responsive to initial treatment, conditions that were less severe or complex, conditions that warranted referral for alternative treatment, and social factors such as patient transportation barriers. This generalizability cohort consisted of a consecutive sample of patients who had only 3 clinic visits during the 6-year study period (as compared with our primary cohort, which had 4 to 6 visits). These patients were not included in our primary cohort because the smaller number of visits is not ideal to capture time-varying associations between the measures of interest. Statistical Analysis Linear mixed effects models (LMM) were used for all analyses. LMM is a widely used statistical approach that incorporates fixed effects (ie, nonvarying coefficients of explanatory variables across all patients) and random effects (ie, varying coefficients within each patient) to estimate variance in longitudinal data with a multilevel or hierarchical structure. Here, we used LMM to estimate random effects of the grouping factors patient and visit (both as categorical variables) to account for individual-level variance (eg, different between-visit time intervals across patients) and assessment-level variance (eg, varying injury characteristics associated with the total number of clinic visits during the study period), respectively. Patients’ data from all clinic visits (ie, up to 6 visits for the primary cohort) were used for estimation. See eMethods in Supplement 1 for additional model specifications. Separate models were run for PROMIS Anxiety and PROMIS Depression as the dependent variable (outcome measure) and for physical function and pain interference as the independent variable (exposure measure). First, we tested for a main effect of PROMIS Physical Function on PROMIS Anxiety and then on PROMIS Depression, with adjustment for confounding variables including age, gender, and race (categorized as White vs non-White because the sample was predominantly White). Because anxiety and depression symptoms are frequently comorbid and highly correlated,40 we also adjusted for PROMIS Depression and PROMIS Anxiety scores as covariates in the respective models in order to account for shared variance and to test for the main effect of physical function specific to either disorder. Calculated β coefficients below 0.06 were considered to indicate nonsignificance because they would represent changes in PROMIS Anxiety or Depression scores that would not reach our predefined minimum clinically meaningful thresholds, even if the greatest possible score change in PROMIS Physical Function or Pain Interference was achieved. Next, to focus on patients who, according to our hypothesis, would be most likely to achieve clinically meaningful improvement in symptoms of anxiety and depression, we repeated all analyses on the subgroup of patients whose physical function meaningfully improved (ie, 5-point score increase in PROMIS Physical Function between their first and last clinic visit) during the study period (eFigure in Supplement 1). While the precise threshold for meaningful improvement varies based on the patient population and is not universally agreed upon, 5 points was chosen as the cutoff value for this study because it corresponds to a moderate effect size change (ie, 0.5 standard deviations) and has repeatedly been found to represent a clinically important difference across various orthopedic patient populations.38,41,42 We then repeated all statistical procedures using PROMIS Pain Interference as the exposure measure, rather than PROMIS Physical Function. Separate models were run for each of these 2 measures without adjusting for their shared variance because: (1) although physical function and pain interference are closely related, they are conceptually different constructs, and (2) the purpose of this study was to evaluate for longitudinal associations between these physical health constructs and mental health symptoms, but not necessarily to identify the specific influence of each construct on mental health symptoms. Meaningfully improved pain interference was defined as a 5-point decrease in PROMIS Pain Interference.38,43,44 All statistical procedures were also repeated in follow-up tests using the generalizability cohort of patients who only had 3 clinic visits during the study period (as compared with our primary cohort which had 4 to 6 visits). The sample size of each cohort was determined by the availability of eligible patients. Given the large available sample size, the relatively small proportion of patients with missing PROMIS scores were excluded from analysis based on the study eligibility criteria. P values were derived from F tests using Satterthwaite methods.45 To account for multiple comparisons, false discovery rate (FDR) corrections were applied to all the models using the Benjamini-Hochberg approach.46 Hypothesis testing was 2-sided, and a P value < .05 after FDR corrections was considered statistically significant. All statistical analyses were conducted using R version 4.10 (R Project for Statistical Computing). Linear mixed effects models were run using the lmerTest package. Results Demographics Of 87 772 patients who were evaluated at the study institution during the study period, 11 236 were eligible for inclusion in the primary cohort (51 569 total visits) (Figure 1). This cohort had a mean (SD) age of 58 (16) years; 1288 (12%) were Black, 9706 (86%) were White, and 7218 (64%) were women. At baseline, the cohort reported similar symptoms of anxiety (mean [SD] PROMIS Anxiety score, 48.2 [9.1]) and depression (PROMIS Depression score, 46.0 [8.6]) compared with the general US population. Patients who achieved meaningfully improved physical function (1672 patients) and/or pain interference (1391 patients) during the study period reported similar baseline symptoms of anxiety and depression than those who did not achieve meaningfully improved physical function and/or pain interference (Table 1). Figure 1. Inclusion Flowchart PROMIS indicates Patient-Reported Outcomes Measurement Information System. Table 1. Patient Characteristics Characteristics Patients, No. (%) Primary cohort (N = 11 236) Physical function improved (n = 1672)a Pain interference improved (n = 1391)a Age, mean (SD), y 58 (16) 57 (16) 58 (15) Gender Men 4018 (36) 605 (36) 463 (33) Women 7218 (64) 1067 (64) 928 (67) Race American Indian or Alaska Native 26 (<1) 1 (<1) 2 (<1) Asian 120 (1) 15 (<1) 15 (1) Black or African American 1288 (12) 190 (11) 161 (12) White 9706 (86) 1455 (87) 1203 (87) Other Pacific Islander 15 (<1) 2 (<1) 1 (<1) Multiracial 39 (<1) 5 (<1) 1 (<1) Unable to answer 1 (<1) 0 1 (<1) Declined 41 (<1) 4 (<1) 7 (<1) Baseline PROMIS scores, mean (SD)b Anxiety 48.2 (9.1) 49.3 (9.3) 50.7 (9.2) Depression 46.0 (8.6) 46.8 (8.9) 47.9 (9.0) Physical Function 40.7 (8.5) 33.8 (7.4) 37.0 (7.8) Pain Interference 58.8 (7.8) 62.2 (7.7) 65.3 (6.4) Final PROMIS scores, mean (SD)b Anxiety 60.6 (9.0) 58.3 (9.1) 56.3 (9.8) Depression 54.2 (9.3) 52.1 (9.2) 51.3 (9.0) Physical Function 34.7 (8.1) 43.0 (7.8) 41.2 (8.1) Pain Interference 65.3 (7.6) 60.3 (8.0) 55.3 (7.5) Time from baseline to final follow-up, mean (SD), d 419 (192) 374 (170) 603 (151) a Improved was defined as a 5-point favorable change in patients’ Patient-Reported Outcomes Measurement Information System (PROMIS) scores from the first to last clinic visit during the study period (ie, 5-point score increase in PROMIS Physical Function, 5-point score decrease in PROMIS Pain Interference). b Baseline and final mean PROMIS scores for each PROMIS domain were statistically significantly different between each of the improved cohorts (ie, “physical function improved” and “pain interference improved”) and the remaining patients who did not meet the improved criteria (all P < .001). Associations With Anxiety For the primary cohort after adjusting for age, gender, race, and depression symptoms, improvements in physical function (β = −0.14; 95% CI, −0.15 to −0.13; P for FDR < .001) and pain interference (β = 0.26; 95% CI, 0.25 to 0.26; P for FDR < .001) were each associated with statistically and meaningfully improved anxiety symptoms (Figure 2; Table 2). To reach a clinically meaningful improvement in anxiety symptoms of at least 3.0 PROMIS Anxiety points, an associated improvement of 21 or more (95% CI, 20 to 23) PROMIS points on Physical Function or 12 or more (95% CI, 12 to 12) points on Pain Interference than would be expected (calculated as 3.0/β). Figure 2. Projected Patient-Reported Outcomes Measurement Information System (PROMIS) Mental Health Scores as a Function of PROMIS Physical Health Scores Projected scores were adjusted for age, gender, race, and shared variance between PROMIS Depression and Anxiety. Shading indicates 95% CIs. Table 2. Models Testing for Associations Between PROMIS Physical Function and Pain Interference With PROMIS Anxiety Variable Physical function model Pain interference model β (95% CI) P value β (95% CI) P value Intercept 29.83 (29.24 to 30.42) <.001 10.45 (9.89 to 11.01) <.001 PROMIS Physical Function −0.14 (−0.15 to −0.13) <.001 NI NI PROMIS Pain Interference NI NI 0.26 (0.25 to 0.26) <.001 PROMIS Depression 0.58 (0.58 to 0.59) <.001 0.54 (0.54 to 0.55) <.001 Age (per year) −0.02 (−0.02 to −0.01) <.001 −0.01 (−0.02 to −0.01) <.001 Gender (man) −0.03 (−0.20 to 0.13) .70 −0.12 (−0.28 to 0.03) .13 Race (White) −2.23 (−2.46 to −2.00) <.001 −1.79 (−2.01 to −1.57) <.001 Visit 2 2.44 (2.32 to 2.57) <.001 2.29 (2.16 to 2.41) <.001 Visit 3 3.92 (3.79 to 4.05) <.001 3.65 (3.53 to 3.78) <.001 Visit 4 5.68 (5.55 to 5.81) <.001 5.28 (5.15 to 5.41) <.001 Visit 5 6.50 (6.33 to 6.67) <.001 6.05 (5.88 to 6.22) <.001 Visit 6 7.59 (7.32 to 7.85) <.001 7.01 (6.75 to 7.27) <.001 Random effects σ2 22.95 NA 21.46 NA τ00,id 12.76 NA 12.15 NA ICC 0.36 NA 0.36 NA Total patients, No. 11 236 NA 11 236 NA Observations, No. 51 569 NA 51 569 NA Marginal R2/conditional R2a 0.561/0.718 NA 0.585/0.735 NA Abbreviations: ICC, intraclass correlation coefficient; NA, not applicable; NI, not included in the model; PROMIS, Patient-Reported Outcomes Measurement Information System; σ2, random effect variance; τ00,id, random intercept for each individual patient. a R2 indicates the total variance in the data that is explained by fixed effects alone (marginal R2) and by fixed and random effects together (conditional R2). Among the subgroup of patients who achieved meaningfully improved physical function or pain interference during the study period (of at least 5 PROMIS points), these associations were similar (physical function: β = −0.11; 95% CI, −0.13 to −0.09; P for FDR < .001; pain interference: β = 0.22; 95% CI, 0.19 to 0.24; P for FDR < .001) (eTable 1 in Supplement 1). However, compared with the primary cohort, a somewhat greater associated improvement of 27 or more (95% CI, 23 to 33) PROMIS points on Physical Function or 14 or more (95% CI, 13 to 16) points on Pain Interference would be expected in patients who achieve a clinically meaningful improvement in anxiety symptoms of at least 3.0 PROMIS points. Associations With Depression For the primary cohort after adjusting for age, gender, race, and anxiety symptoms, improvements in physical function (β = −0.05; 95% CI, −0.06 to −0.04; P for FDR < .001) and pain interference (β = 0.04; 95% CI, 0.04 to 0.05; P for FDR < .001) were each associated with statistically but not meaningfully improved depression symptoms (Figure 2, Table 3). That is, to reach a clinically meaningful improvement in depression symptoms of at least 3.2 PROMIS Depression points, an associated improvement of 64 or more (95% CI, 53 to 80) PROMIS points on Physical Function or 64 or more (95% CI, 64 to 80) points on Pain Interference would be expected (calculated as 3.2/β), which was not possible based on the actual score ranges of these PROMIS measures. Table 3. Models Testing for Associations Between PROMIS Physical Function and Pain Interference With PROMIS Depression Variable Physical function model Pain interference model β (95% CI) P value β (95% CI) P value Intercept 17.34 (16.71 to 17.96) <.001 12.65 (12.08 to 13.22) <.001 PROMIS Physical Function −0.05 (−0.06 to −0.04) <.001 NI NI PROMIS Pain Interference NI NI 0.04 (0.04 to 0.05) <.001 PROMIS Anxiety 0.64 (0.64 to 0.65) <.001 0.64 (0.63 to 0.65) <.001 Age (per year) −0.01 (−0.02 to −0.01) <.001 −0.01 (−0.01 to −0.01) <.001 Gender (man) −0.17 (−0.32 to −0.01) .03 −0.21 (−0.36 to −0.06) .006 Race (White) 0.61 (0.30 to 0.73) <.001 0.66 (0.48 to 0.88) <.001 Visit 2 −0.20 (−0.33 to −0.06) .005 −0.18 (−0.32 to −0.04) .01 Visit 3 −0.29 (−0.43 to −0.15) <.001 −0.27 (−0.41 to −0.13) <.001 Visit 4 −0.46 (−0.61 to −0.31) <.001 −0.44 (−0.59 to −0.29) <.001 Visit 5 −0.45 (−0.64 to 0.26) <.001 −0.42 (−0.61 to −0.23) <.001 Visit 6 −0.26 (−0.55 to −0.03) .08 −0.24 (−0.53 to −0.06) .11 Random effects σ2 26.39 NA 26.40 NA τ00,id 9.38 NA 9.51 NA ICC 0.26 NA 0.26 NA Total patients, No. 11 236 NA 11 236 NA Observations, No. 51 569 NA 51 569 NA Marginal R2/conditional R2a 0.514/0.641 NA 0.511/0.641 NA Abbreviations: ICC, intraclass correlation coefficient; NA, not applicable; NI, not included in the model; PROMIS, Patient-Reported Outcomes Measurement Information System; σ2, random effect variance; τ00,id, random intercept for each individual patient. a R2 indicates the total variance in the data that is explained by fixed effects alone (marginal R2) and by fixed and random effects together (conditional R2). Among the subgroup of patients who achieved meaningfully improved physical function or pain interference during the study period (of at least 5 PROMIS points), these associations were unchanged (physical function: β = −0.03; 95% CI, −0.05 to −0.02; P for FDR < .001; pain interference: β = 0.04; 95% CI, 0.02 to 0.06; P for FDR < .001) (eTable 2 in Supplement 1). Specifically, the models for this subgroup still suggest that it would not be possible to reach a clinically meaningful improvement in depression symptoms of 3.2 PROMIS points due only to associated improvements in physical function or pain interference. Generalizability Analysis Compared with our primary cohort of patients who had at least 4 clinic visits during the study period, no meaningful differences in baseline characteristics (eTable 3 in Supplement 1) or clinically significant associations of physical with mental health were observed in our generalizability cohort of patients who only had 3 visits during the 6-year study period. eResults, eTable 4, and eTable 5 in Supplement 1 present these findings. Discussion In this large cohort of patients who sought care for 1 or more musculoskeletal conditions, large improvements in physical function and pain interference were associated with modestly improved anxiety symptoms but were not associated with any meaningful improvement in depression symptoms. These study findings were consistent, even when considering: (1) all patients vs only patients who achieved meaningfully improved physical function or pain interference during the study period, and (2) patients with 4 to 6 clinic visits vs only 3 visits during the study period. There is strong evidence that (1) preexisting anxiety and depression symptoms worsen physical health–related outcomes after musculoskeletal treatment, and (2) improving anxiety and depression symptoms can contribute to improved physical function and reduced pain.10,14,15,16,17,18,19,47,48,49 However, our results suggest that the converse is not true. That is, improved physical health is not necessarily associated with meaningfully improved anxiety or depression symptoms. Other emerging research regarding the association between improved physical health and subsequent symptoms of anxiety and depression has been mixed.20,21,22 For example, a cohort of young adults reported a decrease in utilization of mental health treatment after they underwent hip surgery,21 but in contrast, symptoms of depression did not meaningfully change after a different cohort of somewhat older patients underwent orthopedic treatment for a hand condition.22 It is possible that some musculoskeletal treatments may improve symptoms of anxiety and depression in the short-term for patients who, for instance, may have situational or state (rather than trait) anxiety or depression.50 This could especially be true for elective orthopedic procedures which have a high success rate and are typically only performed on people who are generally healthy at baseline. In contrast to previous studies, this study was not designed to assess the effectiveness of a specific musculoskeletal treatment for a specific patient population. Rather, this study was designed to identify broad associations between physical and mental health trajectories over a long time period, regardless of a patient’s precise musculoskeletal condition(s) or structure-based treatment (eg, surgery, injection, physical therapy, etc). Our findings suggest that over the long-term, musculoskeletal clinicians should be aware that improvements in physical function and/or pain interference are not necessarily associated with meaningful and sustained improvements in symptoms of anxiety or especially depression. Furthermore, because preexisting symptoms of anxiety and depression are associated with worse orthopedic outcomes, we advocate for musculoskeletal clinicians to be equipped with the training and referral resources to address mental health as part of patient counseling and the musculoskeletal treatment plan.28 Limitations Although this study had key strengths including the large sample size and relatively long follow-up duration of over 6 years, there were also several limitations. First, this was an observational study, which limits the causality we can attribute to the associations between physical and mental health that we identified. Second, the patient cohort had limited racial diversity. Third, additional sociodemographic and clinical variables were not available to be included as possible confounders (eg, financial considerations, social support, ethnicity, clinical diagnosis of anxiety and/or depression, musculoskeletal and mental health related interventions, and medical comorbidities).20 It is also possible that other patient, diagnosis, and treatment characteristics could affect the associations we identified (eg, traumatic vs degenerative conditions, spine vs peripheral joint conditions, definitive vs palliative treatment intent, etc). Conclusions This large cohort study suggests that over the course of several years, improvements in physical function and pain interference may be associated with improvement in symptoms of anxiety but not of depression. Furthermore, substantial improvements in physical function and pain interference are expected for patients to reach clinically meaningful associations with improvement in anxiety symptoms. Therefore, musculoskeletal clinicians and patients cannot assume that exclusively structure-based treatment of a musculoskeletal condition will necessarily result in improved symptoms of depression or potentially even anxiety. We advocate for clinicians to thoughtfully and intentionally address the mental health–related contributors to, and sequelae of, musculoskeletal conditions when counseling patients and creating person-centered treatment plans. Further investigation is needed to identify methods of addressing mental health in the context of musculoskeletal care that are both feasible and effective. Supplement 1. eMethods. eResults. eFigure. PROMIS Physical Health Score Changes Among Patients Who Achieved Meaningful Improvement eTable 1. Main Effects of PROMIS Physical Function and Pain Interference on PROMIS Anxiety Among Patients With Clinically Improved Physical Function and Pain Interference, Respectively eTable 2. Main Effects of PROMIS Physical Function and Pain Interference on PROMIS Depression Among Patients With Clinically Improved Physical Function and Pain Interference, Respectively eTable 3. Patient Characteristics of the Generalizability Cohort Who Had Three Clinic Visits During the Study Period eTable 4. Main Effects of PROMIS Physical Function and Pain Interference on PROMIS Anxiety in the Generalizability Cohort eTable 5. Main Effects of PROMIS Physical Function and Pain Interference on PROMIS Depression in the Generalizability Cohort Click here for additional data file. Supplement 2. Data Sharing Statement Click here for additional data file. ==== Refs References 1 National Institutes of Health. Patient-Reported Outcomes Measurement Information System. February 23, 2021. Accessed February 12, 2023. https://staging.healthmeasures.net/images/PROMIS/manuals/PROMIS_Pain_Interference_Scoring_Manual.pdf 2 Choi KW, Chen CY, Stein MB, ; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium. Assessment of bidirectional relationships between physical activity and depression among adults: a 2-sample mendelian randomization study. JAMA Psychiatry. 2019;76 (4 ):399-408. doi:10.1001/jamapsychiatry.2018.4175 30673066 3 Hannerz H, Holtermann A, Madsen IEH. Musculoskeletal pain as a predictor for depression in the general working population of Denmark. Scand J Public Health. 2021;49 (6 ):589-597. doi:10.1177/1403494819875337 31969068 4 Guglielmo D, Hootman JM, Boring MA, . Symptoms of anxiety and depression among adults with arthritis - United States, 2015-2017. MMWR Morb Mortal Wkly Rep. 2018;67 (39 ):1081-1087. doi:10.15585/mmwr.mm6739a2 30286053 5 Murphy LB, Sacks JJ, Brady TJ, Hootman JM, Chapman DP. Anxiety and depression among US adults with arthritis: prevalence and correlates. Arthritis Care Res (Hoboken). 2012;64 (7 ):968-976. doi:10.1002/acr.21685 22550055 6 Demyttenaere K, Bruffaerts R, Lee S, . Mental disorders among persons with chronic back or neck pain: results from the World Mental Health Surveys. Pain. 2007;129 (3 ):332-342. doi:10.1016/j.pain.2007.01.022 17350169 7 DeVeaugh-Geiss AM, West SL, Miller WC, Sleath B, Gaynes BN, Kroenke K. The adverse effects of comorbid pain on depression outcomes in primary care patients: results from the ARTIST trial. Pain Med. 2010;11 (5 ):732-741. doi:10.1111/j.1526-4637.2010.00830.x 20353408 8 Amtmann D, Cook KF, Jensen MP, . Development of a PROMIS item bank to measure pain interference. Pain. 2010;150 (1 ):173-182. doi:10.1016/j.pain.2010.04.025 20554116 9 Alattas SA, Smith T, Bhatti M, Wilson-Nunn D, Donell S. Greater pre-operative anxiety, pain and poorer function predict a worse outcome of a total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc. 2017;25 (11 ):3403-3410. doi:10.1007/s00167-016-4314-8 27734110 10 Cheng AL, Schwabe M, Doering MM, Colditz GA, Prather H. The effect of psychological impairment on outcomes in patients with prearthritic hip disorders: a systematic review and meta-analysis. Am J Sports Med. 2019;48 (10 ):2563-2571. doi:10.1177/0363546519883246 31829034 11 Knapp P, Layson JT, Mohammad W, Pizzimenti N, Markel DC. The effects of depression and anxiety on 90-day readmission rates after total hip and knee arthroplasty. Arthroplast Today. 2021;10 :175-179. doi:10.1016/j.artd.2021.06.013 34458530 12 Zalikha AK, Karabon P, Hajj Hussein I, El-Othmani MM. Anxiety and depression impact on inhospital complications and outcomes after total knee and hip arthroplasty: a propensity score-weighted retrospective analysis. J Am Acad Orthop Surg. 2021;29 (20 ):873-884. doi:10.5435/JAAOS-D-20-00721 34525481 13 Cancienne JM, Browning R, Werner BC. Patient-related risk factors for contralateral anterior cruciate ligament (ACL) tear after ACL reconstruction: an analysis of 3707 primary ACL reconstructions. HSS J. 2020;16 (suppl 2 ):226-229. doi:10.1007/s11420-019-09687-x 33380951 14 Archer KR, Devin CJ, Vanston SW, . Cognitive-behavioral-based physical therapy for patients with chronic pain undergoing lumbar spine surgery: a randomized controlled trial. J Pain. 2016;17 (1 ):76-89. doi:10.1016/j.jpain.2015.09.013 26476267 15 La Cour P, Petersen M. Effects of mindfulness meditation on chronic pain: a randomized controlled trial. Pain Med. 2015;16 (4 ):641-652. doi:10.1111/pme.12605 25376753 16 Kamper SJ, Apeldoorn AT, Chiarotto A, . Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350 :h444. doi:10.1136/bmj.h444 25694111 17 Baumeister H, Paganini S, Sander LB, . Effectiveness of a guided internet- and mobile-based intervention for patients with chronic back pain and depression (WARD-BP): a multicenter, pragmatic randomized controlled trial. Psychother Psychosom. 2021;90 (4 ):255-268. doi:10.1159/000511881 33321501 18 Beesdo K, Hartford J, Russell J, Spann M, Ball S, Wittchen HU. The short- and long-term effect of duloxetine on painful physical symptoms in patients with generalized anxiety disorder: results from three clinical trials. J Anxiety Disord. 2009;23 (8 ):1064-1071. doi:10.1016/j.janxdis.2009.07.008 19643572 19 Ashar YK, Gordon A, Schubiner H, . Effect of pain reprocessing therapy vs placebo and usual care for patients with chronic back pain: a randomized clinical trial. JAMA Psychiatry. 2022;79 (1 ):13-23. doi:10.1001/jamapsychiatry.2021.2669 34586357 20 Beleckas CM, Guattery J, Chamberlain AM, Khan T, Kelly MP, Calfee RP. Using patient-reported outcomes measurement information system measures to understand the relationship between improvement in physical function and depressive symptoms. J Am Acad Orthop Surg. 2018;26 (24 ):e511-e518. doi:10.5435/JAAOS-D-17-00039 30216242 21 Zacharias AJ, Lemaster NG, Hawk GS, . Psychological healthcare burden lessens after hip arthroscopy for those with comorbid depression or anxiety. Arthrosc Sports Med Rehabil. 2021;3 (4 ):e1171-e1175. doi:10.1016/j.asmr.2021.05.005 34430898 22 Crijns TJ, Bernstein DN, Gonzalez R, Wilbur D, Ring D, Hammert WC. Operative treatment is not associated with more relief of depression symptoms than nonoperative treatment in patients with common hand illness. Clin Orthop Relat Res. 2020;478 (6 ):1319-1329. doi:10.1097/CORR.0000000000001170 32097128 23 Borges G, Aguilar-Gaxiola S, Andrade L, . Twelve-month mental health service use in six countries of the Americas: a regional report from the World Mental Health Surveys. Epidemiol Psychiatr Sci. 2019;29 :e53. doi:10.1017/S2045796019000477 31452485 24 Vranceanu AM, Beks RB, Guitton TG, Janssen SJ, Ring D. How do orthopaedic surgeons address psychological aspects of illness? Arch Bone Jt Surg. 2017;5 (1 ):2-9.28271080 25 Weil TP. Insufficient dollars and qualified personnel to meet United States mental health needs. J Nerv Ment Dis. 2015;203 (4 ):233-240. doi:10.1097/NMD.0000000000000271 25816044 26 Mojtabai R, Olfson M, Sampson NA, . Barriers to mental health treatment: results from the National Comorbidity Survey Replication. Psychol Med. 2011;41 (8 ):1751-1761. doi:10.1017/S0033291710002291 21134315 27 Appelbaum PS, Parks J. Holding insurers accountable for parity in coverage of mental health treatment. Psychiatr Serv. 2020;71 (2 ):202-204. doi:10.1176/appi.ps.201900513 31722647 28 Cheng AL, Leo AJ, Calfee RP, Dy CJ, Armbrecht MA, Abraham J. What are orthopaedic patients’ and clinical team members’ perspectives regarding whether and how to address mental health in the orthopaedic care setting? a qualitative investigation of patients with neck or back pain. Clin Orthop Relat Res. 2022. doi:10.1097/CORR.0000000000002513 36480637 29 Reichman M, Bakhshaie J, Grunberg VA, Doorley JD, Vranceanu AM. What are orthopaedic healthcare professionals’ attitudes toward addressing patient psychosocial factors? a mixed-methods investigation. Clin Orthop Relat Res. 2022;480 (2 ):248-262. doi:10.1097/CORR.0000000000002043 34779793 30 Vranceanu AM, Bakhshaie J, Reichman M, Ring D. A Call for interdisciplinary collaboration to promote musculoskeletal health: the creation of the International Musculoskeletal Mental and Social Health Consortium (I-MESH). J Clin Psychol Med Settings. 2022;29 (3 ):709-715. doi:10.1007/s10880-021-09827-8 34605997 31 Ring D. Priorities for advancing mental and social health among people presenting for care of musculoskeletal symptoms. J Clin Psychol Med Settings. 2023;30 (1 ):197-203. doi:10.1007/s10880-022-09865-w 35318572 32 Rose M, Bjorner JB, Becker J, Fries JF, Ware JE. Evaluation of a preliminary physical function item bank supported the expected advantages of the Patient-Reported Outcomes Measurement Information System (PROMIS). J Clin Epidemiol. 2008;61 (1 ):17-33. doi:10.1016/j.jclinepi.2006.06.025 18083459 33 Hung M, Clegg DO, Greene T, Saltzman CL. Evaluation of the PROMIS physical function item bank in orthopaedic patients. J Orthop Res. 2011;29 (6 ):947-953. doi:10.1002/jor.21308 21437962 34 Crins MHP, Terwee CB, Ogreden O, . Differential item functioning of the PROMIS physical function, pain interference, and pain behavior item banks across patients with different musculoskeletal disorders and persons from the general population. Qual Life Res. 2019;28 (5 ):1231-1243. doi:10.1007/s11136-018-2087-x 30600494 35 Pilkonis PA, Choi SW, Reise SP, Stover AM, Riley WT, Cella D; PROMIS Cooperative Group. Item banks for measuring emotional distress from the Patient-Reported Outcomes Measurement Information System (PROMIS): depression, anxiety, and anger. Assessment. 2011;18 (3 ):263-283. doi:10.1177/1073191111411667 21697139 36 Schalet BD, Pilkonis PA, Yu L, . Clinical validity of PROMIS Depression, Anxiety, and Anger across diverse clinical samples. J Clin Epidemiol. 2016;73 :119-127. doi:10.1016/j.jclinepi.2015.08.036 26931289 37 Cella D, Yount S, Rothrock N, ; PROMIS Cooperative Group. The Patient-Reported Outcomes Measurement Information System (PROMIS): progress of an NIH Roadmap cooperative group during its first two years. Med Care. 2007;45 (5 )(suppl 1 ):S3-S11. doi:10.1097/01.mlr.0000258615.42478.55 17443116 38 Lee AC, Driban JB, Price LL, Harvey WF, Rodday AM, Wang C. Responsiveness and minimally important differences for 4 patient-reported outcomes measurement information system short forms: physical function, pain interference, depression, and anxiety in knee osteoarthritis. J Pain. 2017;18 (9 ):1096-1110. doi:10.1016/j.jpain.2017.05.001 28501708 39 Kroenke K, Yu Z, Wu J, Kean J, Monahan PO. Operating characteristics of PROMIS four-item depression and anxiety scales in primary care patients with chronic pain. Pain Med. 2014;15 (11 ):1892-1901. doi:10.1111/pme.12537 25138978 40 Haug TT, Mykletun A, Dahl AA. The association between anxiety, depression, and somatic symptoms in a large population: the HUNT-II study. Psychosom Med. 2004;66 (6 ):845-851. doi:10.1097/01.psy.0000145823.85658.0c 15564348 41 Lee DJ, Calfee RP. The minimal clinically important difference for PROMIS physical function in patients with thumb carpometacarpal arthritis. Hand (New York, NY). 2021;16 (5 ):638-643. doi:10.1177/1558944719880025 42 Sandvall B, Okoroafor UC, Gerull W, Guattery J, Calfee RP. Minimal clinically important difference for PROMIS Physical Function in patients with distal radius fractures. J Hand Surg Am. 2019;44 (6 ):454-459.e1. doi:10.1016/j.jhsa.2019.02.015 30954311 43 Chen CX, Kroenke K, Stump TE, . Estimating minimally important differences for the PROMIS pain interference scales: results from 3 randomized clinical trials. Pain. 2018;159 (4 ):775-782. doi:10.1097/j.pain.0000000000001121 29200181 44 Amtmann D, Kim J, Chung H, Askew RL, Park R, Cook KF. Minimally important differences for Patient Reported Outcomes Measurement Information System pain interference for individuals with back pain. J Pain Res. 2016;9 :251-255. doi:10.2147/JPR.S93391 27175093 45 Satterthwaite FE. An approximate distribution of estimates of variance components. Biometrics. 1946;2 (6 ):110-114. doi:10.2307/3002019 20287815 46 Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B. 1995;57 (1 ):289-300. doi:10.1111/j.2517-6161.1995.tb02031.x 47 Browne JA, Sandberg BF, D’Apuzzo MR, Novicoff WM. Depression is associated with early postoperative outcomes following total joint arthroplasty: a nationwide database study. J Arthroplasty. 2014;29 (3 ):481-483. doi:10.1016/j.arth.2013.08.025 24090662 48 Rasouli MR, Menendez ME, Sayadipour A, Purtill JJ, Parvizi J. Direct cost and complications associated with total joint arthroplasty in patients with preoperative anxiety and depression. J Arthroplasty. 2016;31 (2 ):533-536. doi:10.1016/j.arth.2015.09.015 26481408 49 Harris AB, Marrache M, Puvanesarajah V, . Are preoperative depression and anxiety associated with patient-reported outcomes, health care payments, and opioid use after anterior discectomy and fusion? Spine J. 2020;20 (8 ):1167-1175. doi:10.1016/j.spinee.2020.03.004 32179156 50 Forrest SJ, Siegert RJ, Krägeloh CU, Landon J, Medvedev ON. Generalizability theory distinguishes between state and trait anxiety. Psychol Assess. 2021;33 (11 ):1080-1088. doi:10.1037/pas0001060 34398636