Advertisement

Risk Factors for Failed Nonsurgical Treatment Resulting in Surgery on Thumb Carpometacarpal Arthritis

Published:April 06, 2021DOI:https://doi.org/10.1016/j.jhsa.2021.02.009

      Purpose

      The thumb carpometacarpal (CMC) joint is the second most common site of osteoarthritis in the hand, yet reported symptoms and ultimate treatment decisions are not simply a function of radiographic appearance. This study aimed to determine the patient- and/or disease-related factors associated with patients undergoing surgical treatment of thumb CMC arthritis.

      Methods

      This retrospective cohort study analyzed 1,994 patients with thumb CMC arthritis treated at 2 institutions between February 2015 and November 2018. Patient demographic and clinical information was obtained from medical records to characterize treatment modalities before hand surgeon evaluation, mental and physical comorbidities, and Patient-Reported Outcomes Measurement Information System assessments. After bivariate analysis, a multivariable logistic regression model evaluated factors associated with undergoing thumb CMC surgery.

      Results

      This cohort was predominately female (70%) and white (91%), mean age at first appointment, 62 ± 10 years. A total of 170 patients underwent surgery for thumb CMC arthritis (9%) at a median of 114 days (interquartile range, 27–328) after the first visit. Patient-Reported Outcomes Measurement Information System Depression scores correlated with Pain Interference and Physical Function scores. A history of diagnosed depression or anxiety was associated with less perceived Physical Function at presentation. However, only prior contralateral thumb CMC surgery, younger patient age, and treating institution were associated with undergoing surgery in regression modeling.

      Conclusions

      Pain and functional limitations associated with thumb CMC arthritis are influenced by mental health comorbidities, but these factors do not predict surgical treatment. Instead, patients’ prior surgical experience and surgeon attitudes toward thumb CMC arthritis appear to have a strong influence on the odds of patients undergoing surgery for thumb CMC arthritis.

      Type of study/level of evidence

      Prognostic IV.

      Key words

      The thumb carpometacarpal (CMC) joint is the second most common site of osteoarthritis in the hand. Prevalence of thumb CMC arthritis increases with age,
      • Wilder F.V.
      • Barrett J.P.
      • Farina E.J.
      Joint-specific prevalence of osteoarthritis of the hand.
      ,
      • Becker S.J.
      • Briet J.P.
      • Hageman M.G.
      • Ring D.
      Death, taxes, and trapeziometacarpal arthrosis.
      and radiographic changes consistent with the disease have been reported to be as high as 85% in individuals aged 71 to 80 years and in 100% of women aged 91 and older.
      • Becker S.J.
      • Briet J.P.
      • Hageman M.G.
      • Ring D.
      Death, taxes, and trapeziometacarpal arthrosis.
      However, only 4% of individuals with radiographic thumb CMC arthritis will report symptoms to a physician.
      • Hwang R.W.
      • Ring D.
      Pain and disability related to osteoarthrosis of the trapeziometacarpal joint.
      Because a minority of people with thumb CMC arthritis receive medical treatment,
      • Zhang Y.
      • Niu J.
      • Kelly-Hayes M.
      • Chaisson C.E.
      • Aliabadi P.
      • Felson D.T.
      Prevalence of symptomatic hand osteoarthritis and its impact on functional status among the elderly: the Framingham Study.
      ,
      • Lawrence J.S.
      • Bremner J.M.
      • Bier F.
      Osteo-arthrosis: prevalence in the population and relationship between symptoms and x-ray changes.
      researchers have worked to understand factors associated with seeking such intervention. Radiographic severity does not predict symptom severity or the amount of functional impairment.
      • Hwang R.W.
      • Ring D.
      Pain and disability related to osteoarthrosis of the trapeziometacarpal joint.
      ,
      • Zhang Y.
      • Niu J.
      • Kelly-Hayes M.
      • Chaisson C.E.
      • Aliabadi P.
      • Felson D.T.
      Prevalence of symptomatic hand osteoarthritis and its impact on functional status among the elderly: the Framingham Study.
      In patients with thumb CMC arthritis, both depression and upper-extremity comorbid conditions contribute to worse patient-reported function.
      • Calfee R.
      • Chu J.
      • Sorensen A.
      • Martens E.
      • Elfar J.
      What is the impact of comorbidities on self-rated hand function in patients with symptomatic trapeziometacarpal arthritis?.
      As such, depressive symptoms, anxiety, and catastrophic thinking predict a substantial portion of the variation in Disabilities of the Arm, Shoulder, and Hand DASH for patients with thumb CMC arthritis.
      • Lozano-Calderon S.A.
      • Souer J.S.
      • Jupiter J.B.
      • Ring D.
      Psychological differences between patients that elect operative or nonoperative treatment for trapeziometacarpal joint arthrosis.
      Surgery is indicated when symptoms of pain and functional limitations persist after a trial of nonsurgical treatment
      • Weiss A.C.
      • Goodman A.D.
      Thumb basal joint arthritis.
      ,
      • Wilkens S.C.
      • Menendez M.E.
      • Ring D.
      • Chen N.
      QuickDASH score is associated with treatment choice in patients with trapeziometacarpal arthrosis.
      ; thus, it would stand to reason that this decision may be influenced by many biopsychosocial factors. Such factors, including the experience of pain and depression, are now readily studied with the Patient-Reported Outcomes Measurement Information System (PROMIS). These measures were developed by the National Institutes of Health with the goal of improving precision of patient-reported outcome measures. They use computer adaptive testing, selecting questions from an item bank based on an individual’s response to prior questions, allowing for a reduced number of questions compared with traditional static instruments. The t score for each PROMIS instrument has a mean of 50 and an SD of 10.
      Factors associated with pursuing other elective orthopedic surgeries include sex,
      • James D.
      • Young A.
      • Kulinskaya E.
      • et al.
      Orthopaedic intervention in early rheumatoid arthritis: occurrence and predictive factors in an inception cohort of 1064 patients followed for 5 years.
      ,
      • Nystad T.W.
      • Husum Y.S.
      • Furnes O.N.
      • Fevang B.S.
      Incidence and predictive factors for orthopedic surgery in patients with psoriatic arthritis.
      age,
      • Nystad T.W.
      • Husum Y.S.
      • Furnes O.N.
      • Fevang B.S.
      Incidence and predictive factors for orthopedic surgery in patients with psoriatic arthritis.
      ,
      • Churchill L.
      • Malian S.J.
      • Chesworth B.M.
      • et al.
      The development and validation of a multivariable model to predict whether patients referred for total knee replacement are suitable surgical candidates at the time of initial consultation.
      , tobacco use,
      • Oladeji L.O.
      • Dreger T.K.
      • Pratte E.L.
      • et al.
      Total knee arthroplasty versus osteochondral allograft: prevalence and risk factors following tibial plateau fractures.
      disease duration,
      • Pantos P.G.
      • Tzioufas A.G.
      • Panagiotakos D.B.
      • Soucacos P.N.
      • Moutsopoulos H.M.
      Demographics, clinical characteristics and predictive factors for total knee or hip replacement in patients with rheumatoid arthritis in Greece.
      ,
      • Conaghan P.G.
      • D'Agostino M.A.
      • Le Bars M.
      • et al.
      Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study.
      pain intensity,
      • Churchill L.
      • Malian S.J.
      • Chesworth B.M.
      • et al.
      The development and validation of a multivariable model to predict whether patients referred for total knee replacement are suitable surgical candidates at the time of initial consultation.
      ,
      • Conaghan P.G.
      • D'Agostino M.A.
      • Le Bars M.
      • et al.
      Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study.
      history of injections,
      • Churchill L.
      • Malian S.J.
      • Chesworth B.M.
      • et al.
      The development and validation of a multivariable model to predict whether patients referred for total knee replacement are suitable surgical candidates at the time of initial consultation.
      ,
      • Knutsen E.J.
      • Calfee R.P.
      • Chen R.E.
      • Goldfarb C.A.
      • Park K.W.
      • Osei D.A.
      Factors associated with failure of nonoperative treatment in lateral epicondylitis.
      and body mass index.
      • Churchill L.
      • Malian S.J.
      • Chesworth B.M.
      • et al.
      The development and validation of a multivariable model to predict whether patients referred for total knee replacement are suitable surgical candidates at the time of initial consultation.
      However, in one cohort of 72 patients, only younger patient age was associated with pursuing surgery for the thumb CMC joint. That study did not investigate medical comorbidities or prior treatments for the condition.
      • Lozano-Calderon S.A.
      • Souer J.S.
      • Jupiter J.B.
      • Ring D.
      Psychological differences between patients that elect operative or nonoperative treatment for trapeziometacarpal joint arthrosis.
      A recent study also identified that thumb CMC surgery was more likely after injection in patients who had advanced radiographic arthritis, were smokers, or had prior ipsilateral hand surgery.
      • Ostergaard P.J.
      • Hall M.J.
      • Dowlatshahi A.S.
      • Harper C.M.
      • Rozental T.D.
      Thumb carpometacarpal arthritis: prognostic indicators and timing of further intervention following corticosteroid injection.
      The primary aim of this study was to determine patient and clinical factors, at the time of presentation to the surgeon for this condition, that were associated with failure of nonsurgical treatment of thumb CMC arthritis, defined as persistent symptoms resulting in surgical treatment.

      Materials and Methods

      The study was approved by the institutional review boards at both institutions.
      This retrospective cohort study initially gathered a dataset of 2,077 patients with a diagnosis of symptomatic thumb CMC arthritis seen at 1 of 2 academic medical centers between February 2, 2015, and November 27, 2018. Manual chart review of the records with an International Classification of Diseases, Ninth Revision, Clinical Modification or International Classification of Diseases, 10th Revision, Clinical Modification code for thumb CMC arthritis confirmed the diagnosis in 1,994 patients (n = 83 excluded). Date of birth, age at appointment, sex, race, appointment date, scheduling provider, PROMIS scores (Physical Function, Depression, and Pain Interference) and Current Procedural Terminology codes for procedures were automatically extracted from the electronic health record. Current Procedural Terminology codes used to identify surgical treatment were 25210, 25447, 26841, and 26842.
      • Yuan F.
      • Aliu O.
      • Chung K.C.
      • Mahmoudi E.
      Evidence-based practice in the surgical treatment of thumb carpometacarpal joint arthritis.
      Variables recorded from manual chart review included prior diagnosis of anxiety or depression documented in the medical record during the index office visit or in a prior visit to any provider in the electronic health record; the presence or absence of bilateral symptoms; history of contralateral surgery for thumb CMC arthritis; insulin-dependent or non–insulin dependent diabetes mellitus; depression or anxiety treated with medications; current opioid use; present, past, or no smoking history; previous hand therapy; orthotic use; thumb CMC joint corticosteroid injection to either side; number of prior surgeries; and number of prior orthopedic surgeries. Duration of symptoms and occupation were considered variables but were not documented consistently enough in the records to include in this retrospective study.

      Statistical methods

      We used descriptive statistics to characterize the patient sample. Bivariate testing examined the influence of mental health variables on pain and perceived function at presentation. Bivariate testing for the association with failure of nonsurgical treatment was carried out on all variables. Chi-square or Fisher exact test was used for categorical variables and Mann-Whitney U test was used for continuous variables. Variables associated with failure of nonsurgical treatment at the P < .1 significance level on bivariate testing were force-entered into a multivariable logistic regression model. Treating institution was also entered into the multivariable model because the patient samples demonstrated some demographic differences (Table E1, available on the Journal’s Web site at www.jhandsurg.org) and because surgeon-based variation in thumb CMC arthritis treatment has been documented.
      • Deutch Z.
      • Niedermeier S.R.
      • Awan H.M.
      Surgeon preference, influence, and treatment of thumb carpometacarpal arthritis.
      ,
      • Ochtman A.E.
      • Guitton T.G.
      • Buijze G.A.
      • et al.
      Trapeziometacarpal arthrosis: predictors of a second visit and surgery.
      Because 170 individuals underwent operative treatment, it would be appropriate to include up to 17 variables in the final model.
      • Bagley S.C.
      • White H.
      • Golomb B.A.
      Logistic regression in the medical literature: standards for use and reporting, with particular attention to one medical domain.
      Statistical significance was set at α = 0.05 for the logistic regression model. Clinically relevant differences on PROMIS domains were set at 3 or more points.

      Lee DJ, Calfee RP. The minimal clinically important difference for PROMIS Physical Function in patients with thumb carpometacarpal arthritis [published online ahead of print October 18, 2019]. Hand (N Y). https://doi.org/10.1177/1558944719880025.

      ,
      • Kazmers N.H.
      • Hung M.
      • Bounsanga J.
      • Voss M.W.
      • Howenstein A.
      • Tyser A.R.
      Minimal clinically important difference after carpal tunnel release using the PROMIS platform.

      Results

      For the 1,994 patients included in the final analysis, baseline characteristics are presented in Table 1. This cohort was predominately female (70%) and white (91%), mean age at first appointment of 62 ± 10 years. A total of 170 patients had surgical treatment of thumb CMC arthritis during the study period (9%). Median time from the first appointment with a hand surgeon for thumb CMC osteoarthritis to surgery was 114 days (interquartile range, 27–328 days). Median time from that first hand surgeon appointment to the final date of available data was 580 days (interquartile range, 260–924 days).
      Table 1Demographic Data for Entire Population (n = 1,994)
      VariableFrequency (%) or Mean (SD)
      Age at first appointment, y62 (10)
      Female1,397 (70)
      Race
       Caucasian1,819 (91)
       Non-Caucasian175 (9)
      Diabetes
       Absent1,707 (86)
       Non–insulin dependent209 (11)
       Insulin-dependent78 (4)
      Smoke
       Never1,204 (60)
       Prior645 (32)
       Current145 (7)
      Anxiety366 (18)
      Depression492 (25)
      Medication for depression or anxiety616 (31)
      PROMIS
       Depression47 (9)
       Pain Interference59 (6)
       Physical Function45 (8)
      Bilateral symptoms755 (38)
      Contralateral surgery86 (4)
      Opioid use166 (8)
      Opioid use for thumb27 (1)
      Therapy or bracing or injection974 (49)
      Prior surgeries, n2.8 (34)
      Prior orthopedic surgeries, n1 (2)
      At presentation, PROMIS Depression scores correlated with Pain Interference scores (r = –0.38; P < .05) and Physical Function scores (r = –0.37; P < .05). A history of diagnosed depression or anxiety was associated with clinically relevant lower perceived Physical Function at presentation (Table 2). Diagnosed depression and anxiety were also associated with greater PROMIS Pain Interference, but this was not clinically relevant.
      Table 2Association Between Anxiety and Depression Diagnoses and Mean PROMIS Scores (SD)
      Values are rounded to nearest whole number. Three-point threshold for clinically relevant difference not met.
      Promis DomainNo AnxietyDiagnosed AnxietyP Value
      PROMIS Physical Function46 (8)41 (8)<.05
      PROMIS Pain Interference59 (6)60 (6)<.05
      No DepressionDiagnosed Depression
      PROMIS Physical Function46 (8)42 (8)<.05
      PROMIS Pain Interference58 (6)
      Values are rounded to nearest whole number. Three-point threshold for clinically relevant difference not met.
      61 (6)
      Values are rounded to nearest whole number. Three-point threshold for clinically relevant difference not met.
      <.05
      Values are rounded to nearest whole number. Three-point threshold for clinically relevant difference not met.
      On bivariate analysis, factors that met the threshold for inclusion in the multivariable logistic regression model for surgical treatment included age at first appointment, institution, bilateral symptoms, history of contralateral surgery for thumb CMC arthritis, race, smoking status, nonsurgical treatment before surgeon evaluation (ie, therapy, bracing, steroid injection), and number of prior orthopedic surgeries. A history of anxiety, depression, PROMIS Depression scores, PROMIS Physical Function scores, and PROMIS Pain Interference scores were not associated with surgery (Table 3).
      Table 3Bivariate Assessment for Association With Failure of Nonsurgical Treatment
      VariableNonsurgical (n = 1,824) (Frequency [%] or Mean [SD])Operative (n = 170) (Frequency [%] or Mean [SD])P Value
      Age at first appointment, y62 (10)61 (9).07
      Female1,286 (71)111 (65).15
      Race.09
       Caucasian1,658 (91)161 (95)
       Non-Caucasian166 (9)9 (5)
      Diabetes.69
       Absent1,565 (86)142 (84)
       Non–insulin dependent188 (10)21 (12)
       Insulin-dependent71 (4)7 (4)
      Smoke<.05
       Never1,117 (61)87 (51)
       Prior579 (32)66 (39)
       Current128 (7)17 (10)
      Anxiety332 (18)34 (20).56
      Depression450 (25)42 (25).99
      Medication for depression or anxiety565 (31)51 (30).79
      PROMIS score
       Depression46 (9)47 (10).21
       Pain Interference59 (6)60 (6).38
       Physical Function45 (8)47 (10).21
      Bilateral symptoms675 (37)80 (47)<.05
      Contralateral thumb surgery62 (3)24 (14)<.05
      Opioid use154 (8)12 (7).53
      Opioid use for thumb25 (1)2 (1)>.99
      Therapy or bracing or injection843 (46)141 (77)<.05
      Prior surgeries, n2.80 (3)3.15 (4).59
      Prior orthopedic surgeries, n1.2 (2)1.6 (2)<.05
      On multivariable logistic regression analysis, prior contralateral thumb CMC surgery, younger patient age, and treating institution were associated with failure of nonsurgical treatment. Bilateral symptoms, smoking status, nonsurgical treatment before surgeon evaluation, and number of prior orthopedic surgeries were not associated with progression to thumb CMC surgery (Table 4).
      Table 4Results of Multivariable Logistic Regression for Dependent Outcome of Surgical Treatment Provided
      VariableOdds Ratio95% Confidence IntervalP Value
      Institution3.782.55–5.60<.05
      Contralateral thumb surgery3.822.15–6.79<.05
      Higher age, y0.980.96–1.00<.05
      Therapy or bracing or injection1.210.71–2.08.49
      Caucasian race1.950.93–4.05.08
      Former smoker1.280.89–1.84.18
      Current smoker1.350.74–2.47.32
      Bilateral symptoms1.030.73–1.47.87
      Prior orthopedic surgeries, n0.990.91–1.08.82

      Discussion

      During the 3.5-year study period, most patients with thumb CMC arthritis treated at our institutions were managed nonsurgically. The odds of thumb CMC surgery were primarily associated with patients’ contralateral thumb surgical experience and surgeon attitudes toward thumb CMC arthritis, because the incidence of surgery was 4% at one institution and 21% at the second institution.
      Because thumb arthritis surgery is an elective intervention to treat patient-reported pain and limited function, we hypothesized that comorbidities associated with heightened pain experience and worse PROMIS scores at presentation would influence the odds of surgery. However, neither a history of depression or anxiety, increased PROMIS Depression or Pain Interference scores, nor decreased PROMIS Physical Function scores were associated with patients undergoing surgery. This is counterintuitive and was not predicted. We believe that these patient-reported measures are capturing these patients’ symptomatology, because patients at each institution presented with nearly identical mean Pain Interference scores roughly 1 SD greater than population averages, and Physical Function scores 0.5 SD less than population averages. Yet, our findings mirrored those of Lozano-Calderon et al,
      • Lozano-Calderon S.A.
      • Souer J.S.
      • Jupiter J.B.
      • Ring D.
      Psychological differences between patients that elect operative or nonoperative treatment for trapeziometacarpal joint arthrosis.
      who reported that depression, pain catastrophizing, and pain anxiety correlated with Disabilities of the Arm, Shoulder, and Hand scores but were not associated with operative intervention in 72 patients with thumb CMC arthritis. It is possible that surgeons may approach the patient who has magnified pain, anxiety, or depression with a greater reluctance to offer surgery. That possibility was supported by a recent survey that found surgeons were more likely to offer surgery for thumb arthritis in patients with substantial pain, but were less likely to do so in the face of increased depressive symptoms.
      • Ottenhoff J.S.E.
      • Teunis T.
      • Janssen S.J.
      • Mink van der Molen A.B.
      • Ring D.
      Variation in offer of operative treatment to patients with trapeziometacarpal osteoarthritis.
      Alternatively, depressed patients may report more pain and functional limitations, but may remain reluctant to pursue surgical treatment.
      A history of contralateral thumb surgery was strongly associated with operative treatment of the thumb under treatment during our study. We expected this because most patients are satisfied with thumb CMC surgery and as Alqueza
      • Alqueza A.B.
      • Fostvedt S.
      • Emerson Floyd W.
      • Jupiter J.B.
      Patient satisfaction after bilateral thumb carpometacarpal osteoarthritis surgery.
      reported, 33/41 patients would be willing to undergo thumb CMC surgery on the other thumb after unilateral surgery. Furthermore, surgeons may not see value in prolonged nonsurgical treatment in patients who had received surgery for the same diagnosis in the contralateral hand. In addition, patient insurance status, need for assistance with transportation to appointments, therapy, or time away from work or personal activities may be a factor in who receives surgery, even if all patients were offered surgery. Patients whose life circumstances afforded them the opportunity to pursue surgery on the contralateral hand may remain the most likely ones with the means to pursue surgery at this time.
      Our modeling indicated substantial variation in operative rates between our centers. This influence of geographic location on treatment received replicates findings from studies of patients with arthritis and distal radius fractures.
      • Fanuele J.
      • Koval K.J.
      • Lurie J.
      • Zhou W.
      • Tosteson A.
      • Ring D.
      Distal radial fracture treatment: what you get may depend on your age and address.
      • Huetteman H.E.
      • Shauver M.J.
      • Malay S.
      • Chung T.T.
      • Chung K.C.
      Variation in the treatment of distal radius fractures in the United States: 2010 to 2015.
      • Barbour K.E.
      • Moss S.
      • Croft J.B.
      • et al.
      Geographic variations in arthritis prevalence, health-related characteristics, and management—United States, 2015.
      This may result from differences in patient populations, patient expectations, or surgeon approaches. Ottenhoff et al
      • Ottenhoff J.S.E.
      • Teunis T.
      • Janssen S.J.
      • Mink van der Molen A.B.
      • Ring D.
      Variation in offer of operative treatment to patients with trapeziometacarpal osteoarthritis.
      investigated variation in the offering of operative treatment to patients with thumb CMC arthritis. Presenting a large group of surgeons with 16 patient scenarios, they found that practice region, likely a surrogate for surgeon attitude, was the single predictor of surgery technique. Differences in patient demographic variables existed between our centers. Although these could have contributed to a higher likelihood of surgery, we accounted for these differences with multivariable modeling. Most of these factors were rendered nonsignificant in the final model whereas institution as a variable remained significant. Thus, we postulate that institutional differences result from variation in surgeon and patient attitudes. Such variation is frequently noted to increase health care costs and indicate provision of care that is not evidence-based.
      • Huetteman H.E.
      • Shauver M.J.
      • Malay S.
      • Chung T.T.
      • Chung K.C.
      Variation in the treatment of distal radius fractures in the United States: 2010 to 2015.
      Our analysis of patient demographic risk factors showed mixed consistency with prior reports. In 239 patients from one center, advanced radiographic findings, current smoking, and a history of ipsilateral hand surgery predicted surgery after injection for thumb CMC arthritis.
      • Ostergaard P.J.
      • Hall M.J.
      • Dowlatshahi A.S.
      • Harper C.M.
      • Rozental T.D.
      Thumb carpometacarpal arthritis: prognostic indicators and timing of further intervention following corticosteroid injection.
      Although we queried contralateral thumb surgery, both studies indicate that prior hand surgery makes additional hand surgery more likely. We did not find that smoking increased the odds of surgery, but only 7% of the patients in this study currently smoked compared with 20% of patients in the other study, which may have obscured the impact of smoking in our study. Similar to the findings of Lozano-Calderon et al,
      • Lozano-Calderon S.A.
      • Souer J.S.
      • Jupiter J.B.
      • Ring D.
      Psychological differences between patients that elect operative or nonoperative treatment for trapeziometacarpal joint arthrosis.
      younger patient age was associated with surgery in our study. Finally, in our study, non-Caucasian patients comprised 9% of the nonsurgical group but only 5% of patients undergoing surgery. This observation may warrant further investigation, because Medicare data (2009–2014) demonstrate that rates of surgical intervention are lowest for racial and ethnic minorities and not corrected after the implementations of Accountable Care Organizations.
      • Schoenfeld A.J.
      • Sturgeon D.J.
      • Dimick J.B.
      • et al.
      Disparities in rates of surgical intervention among racial and ethnic minorities in Medicare Accountable Care Organizations.
      Our study had several limitations. We could capture only potential risk factors that were documented in the electronic health record. Second, we chose surgical treatment to define failure of nonsurgical treatment because it is objective and frequently used for this purpose, but this does not quantify persistent symptoms.
      • Knutsen E.J.
      • Calfee R.P.
      • Chen R.E.
      • Goldfarb C.A.
      • Park K.W.
      • Osei D.A.
      Factors associated with failure of nonoperative treatment in lateral epicondylitis.
      ,
      • Philippon M.J.
      • Briggs K.K.
      • Carlisle J.C.
      • Patterson D.C.
      Joint space predicts THA after hip arthroscopy in patients 50 years and older.
      Third, it is possible that some patients left our health systems for surgery elsewhere. Fourth, the surgeons providing care to patients in this study did not have a uniform approach to determining a failure of nonsurgical treatment before offering surgery. Finally, we did not include radiographic severity in our modeling because radiographs were not consistently taken during nonsurgical treatment. Notably, the recent study by Ostergaard et al
      • Ostergaard P.J.
      • Hall M.J.
      • Dowlatshahi A.S.
      • Harper C.M.
      • Rozental T.D.
      Thumb carpometacarpal arthritis: prognostic indicators and timing of further intervention following corticosteroid injection.
      found that radiographic staging was associated with surgery and we cannot refute or confirm that finding in the patients in the current study.
      Embarking on this study, we expected depressive symptoms to be associated with greater pain interference and worse patient-reported function. We expected these factors to predict elective surgery for the thumb CMC joint. However, our data suggest that variable surgeon approaches to thumb CMC arthritis more strongly influence who undergoes thumb CMC surgery than do magnitudes of patient-reported pain and functional limitation among symptomatic patients. Thus, it appears that predicting which patients undergo surgery is incompletely understood and may be influenced by a range of surgeon and patient attitudes. The variation between centers in our study points to potential value in further investigation focused on the cost-effectiveness and cost utility of treatment protocols for thumb CMC osteoarthritis.

      Acknowledgments

      The authors thank the University of Rochester Health Lab for their assistance with collection and reporting of patient-reported outcomes via UR VOICE, and Josh Atkinson, who assisted with the initial data extraction.

      Appendix A

      Table E1Patient Demographic Data According to Institution
      VariableInstitution 1 (n = 1,489) (Frequency [%] or Mean [SD])Institution 2 (n = 505) (Frequency [%] or Mean [SD])P Value
      Underwent surgery62 (4)108 (21)<.05
      Age at first appointment, y62 (9)62 (10).23
      Female1,041 (70)356 (71).81
      Race.55
       Caucasian1,355 (91)134 (9)
       Non-Caucasian464 (92)41 (8)
      Diabetes.28
       Absent1,284 (86)423 (84)
       Non–insulin dependent152 (10)57 (11)
       Insulin-dependent53 (4)25 (5)
      Smoke<.05
       Never943 (63)261 (52)
       Prior448 (30)197 (39)
       Current98 (7)47 (9)
      Anxiety260 (17)106 (21).08
      Depression341 (23)151 (30)<.05
      Medication for depression or anxiety444 (30)172 (34).07
      PROMIS Score
       Depression46 (9)49 (9)<.05
       Pain Interference59 (6)59 (7).46
       Physical Function45 (8)44 (8).08
      Bilateral symptoms530 (36)225 (45)<.05
      Contralateral surgery46 (3)40 (8)<.05
      Opioid use132 (9)34 (7).13
      Opioid use for thumb12 (0.8)15 (3)<.05
      Therapy or bracing or injection843 (46)141 (77)<.05
      Prior surgeries, n2 (3)4 (5)<.05
      Prior orthopedic surgeries, n1 (1)2 (2)<.05

      References

        • Wilder F.V.
        • Barrett J.P.
        • Farina E.J.
        Joint-specific prevalence of osteoarthritis of the hand.
        Osteoarthritis Cartilage. 2006; 14: 953-957
        • Becker S.J.
        • Briet J.P.
        • Hageman M.G.
        • Ring D.
        Death, taxes, and trapeziometacarpal arthrosis.
        Clin Orthop Relat Res. 2013; 471: 3738-3744
        • Hwang R.W.
        • Ring D.
        Pain and disability related to osteoarthrosis of the trapeziometacarpal joint.
        J Hand Microsurg. 2011; 3: 63-65
        • Zhang Y.
        • Niu J.
        • Kelly-Hayes M.
        • Chaisson C.E.
        • Aliabadi P.
        • Felson D.T.
        Prevalence of symptomatic hand osteoarthritis and its impact on functional status among the elderly: the Framingham Study.
        Am J Epidemiol. 2002; 156: 1021-1027
        • Lawrence J.S.
        • Bremner J.M.
        • Bier F.
        Osteo-arthrosis: prevalence in the population and relationship between symptoms and x-ray changes.
        Ann Rheum Dis. 1966; 25: 1-24
        • Calfee R.
        • Chu J.
        • Sorensen A.
        • Martens E.
        • Elfar J.
        What is the impact of comorbidities on self-rated hand function in patients with symptomatic trapeziometacarpal arthritis?.
        Clin Orthop Relat Res. 2015; 473: 3477-3483
        • Lozano-Calderon S.A.
        • Souer J.S.
        • Jupiter J.B.
        • Ring D.
        Psychological differences between patients that elect operative or nonoperative treatment for trapeziometacarpal joint arthrosis.
        Hand (N Y). 2008; 3: 271-275
        • Weiss A.C.
        • Goodman A.D.
        Thumb basal joint arthritis.
        J Am Acad Orthop Surg. 2018; 26: 562-571
        • Wilkens S.C.
        • Menendez M.E.
        • Ring D.
        • Chen N.
        QuickDASH score is associated with treatment choice in patients with trapeziometacarpal arthrosis.
        Hand (N Y). 2017; 12: 461-466
        • James D.
        • Young A.
        • Kulinskaya E.
        • et al.
        Orthopaedic intervention in early rheumatoid arthritis: occurrence and predictive factors in an inception cohort of 1064 patients followed for 5 years.
        Rheumatology (Oxford). 2004; 43: 369-376
        • Nystad T.W.
        • Husum Y.S.
        • Furnes O.N.
        • Fevang B.S.
        Incidence and predictive factors for orthopedic surgery in patients with psoriatic arthritis.
        J Rheumatol. 2018; 45: 1532-1540
        • Churchill L.
        • Malian S.J.
        • Chesworth B.M.
        • et al.
        The development and validation of a multivariable model to predict whether patients referred for total knee replacement are suitable surgical candidates at the time of initial consultation.
        Can J Surg. 2016; 59: 407-414
        • Oladeji L.O.
        • Dreger T.K.
        • Pratte E.L.
        • et al.
        Total knee arthroplasty versus osteochondral allograft: prevalence and risk factors following tibial plateau fractures.
        J Knee Surg. 2019; 32: 380-386
        • Pantos P.G.
        • Tzioufas A.G.
        • Panagiotakos D.B.
        • Soucacos P.N.
        • Moutsopoulos H.M.
        Demographics, clinical characteristics and predictive factors for total knee or hip replacement in patients with rheumatoid arthritis in Greece.
        Clin Exp Rheumatol. 2013; 31: 195-200
        • Conaghan P.G.
        • D'Agostino M.A.
        • Le Bars M.
        • et al.
        Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study.
        Ann Rheum Dis. 2010; 69: 644-647
        • Knutsen E.J.
        • Calfee R.P.
        • Chen R.E.
        • Goldfarb C.A.
        • Park K.W.
        • Osei D.A.
        Factors associated with failure of nonoperative treatment in lateral epicondylitis.
        Am J Sports Med. 2015; 43: 2133-2137
        • Ostergaard P.J.
        • Hall M.J.
        • Dowlatshahi A.S.
        • Harper C.M.
        • Rozental T.D.
        Thumb carpometacarpal arthritis: prognostic indicators and timing of further intervention following corticosteroid injection.
        J Hand Surg Am. 2020; 45: 986.e1-986.e9
        • Yuan F.
        • Aliu O.
        • Chung K.C.
        • Mahmoudi E.
        Evidence-based practice in the surgical treatment of thumb carpometacarpal joint arthritis.
        J Hand Surg Am. 2017; 42: 104-112.e1
        • Deutch Z.
        • Niedermeier S.R.
        • Awan H.M.
        Surgeon preference, influence, and treatment of thumb carpometacarpal arthritis.
        Hand (N Y). 2018; 13: 403-411
        • Ochtman A.E.
        • Guitton T.G.
        • Buijze G.A.
        • et al.
        Trapeziometacarpal arthrosis: predictors of a second visit and surgery.
        J Hand Microsurg. 2013; 5: 9-13
        • Bagley S.C.
        • White H.
        • Golomb B.A.
        Logistic regression in the medical literature: standards for use and reporting, with particular attention to one medical domain.
        J Clin Epidemiol. 2001; 54: 979-985
      1. Lee DJ, Calfee RP. The minimal clinically important difference for PROMIS Physical Function in patients with thumb carpometacarpal arthritis [published online ahead of print October 18, 2019]. Hand (N Y). https://doi.org/10.1177/1558944719880025.

        • Kazmers N.H.
        • Hung M.
        • Bounsanga J.
        • Voss M.W.
        • Howenstein A.
        • Tyser A.R.
        Minimal clinically important difference after carpal tunnel release using the PROMIS platform.
        J Hand Surg Am. 2019; 44: 947-953.e1
        • Ottenhoff J.S.E.
        • Teunis T.
        • Janssen S.J.
        • Mink van der Molen A.B.
        • Ring D.
        Variation in offer of operative treatment to patients with trapeziometacarpal osteoarthritis.
        J Hand Surg Am. 2020; 45: 123-130.e1
        • Alqueza A.B.
        • Fostvedt S.
        • Emerson Floyd W.
        • Jupiter J.B.
        Patient satisfaction after bilateral thumb carpometacarpal osteoarthritis surgery.
        J Surg Orthop Adv. 2017; 26: 250-256
        • Fanuele J.
        • Koval K.J.
        • Lurie J.
        • Zhou W.
        • Tosteson A.
        • Ring D.
        Distal radial fracture treatment: what you get may depend on your age and address.
        J Bone Joint Surg Am. 2009; 91: 1313-1319
        • Huetteman H.E.
        • Shauver M.J.
        • Malay S.
        • Chung T.T.
        • Chung K.C.
        Variation in the treatment of distal radius fractures in the United States: 2010 to 2015.
        Plast Reconstr Surg. 2019; 143: 159-167
        • Barbour K.E.
        • Moss S.
        • Croft J.B.
        • et al.
        Geographic variations in arthritis prevalence, health-related characteristics, and management—United States, 2015.
        MMWR Surveill Summ. 2018; 67: 1-28
        • Schoenfeld A.J.
        • Sturgeon D.J.
        • Dimick J.B.
        • et al.
        Disparities in rates of surgical intervention among racial and ethnic minorities in Medicare Accountable Care Organizations.
        Ann Surg. 2019; 269: 459-464
        • Philippon M.J.
        • Briggs K.K.
        • Carlisle J.C.
        • Patterson D.C.
        Joint space predicts THA after hip arthroscopy in patients 50 years and older.
        Clin Orthop Relat Res. 2013; 471: 2492-2496