Discrepancies in conservative treatment for thumb carpometacarpal arthritis: a comparison between different specialities and patient characteristics

Abstract

Non-surgical treatment is successful in controlling pain and preventing disease progress in treating thumb carpometacarpal arthritis. We used Optum’s de-identified Clinformatics^® Data Mart Databases between 2015 and 2018 to conduct a study of the patient and provider characteristics associated with three types of non-surgical treatment (hand therapy, splinting and corticosteroid injection) prior to surgery. In this population-based cohort study, we found that non-surgical providers were more likely to provide three different types of non-surgical treatments, as compared with hand surgeons. In addition, women and patients with comorbid conditions, including carpal tunnel syndrome, obesity, chronic pain and depression, were less likely to exhaust the available non-surgical management options for thumb carpometacarpal arthritis. Therefore, we suggest that these specific patient populations can potentially benefit from additional non-surgical treatments that may delay or obviate surgery for this disease. These groups are target populations for future efforts to ensure that all patients receive equitable care.

Level of evidence:

II

Introduction

Non-surgical treatments of the carpometacarpal (CMC) arthritis of the thumb provide adequate pain relief for the majority of patients, particularly in the early stages of disease (Gillis et al., 2011; Spaans et al., 2015), and are offered for thumb CMC arthritis before considering surgical management. This approach is supported by the European League Against Rheumatism (EULAR) 2018 treatment guidelines, which recommend surgery for thumb CMC arthritis only if pain persists following non-pharmacologic treatment (Kloppenburg et al., 2019). In the United States, no surgical treatment guidelines exist for thumb CMC arthritis; the American College of Rheumatology strongly recommends splinting, but does not mention the role of surgical management in its guidelines (Kolasinski et al., 2020). Various studies have demonstrated that non-surgical treatments, such as hand therapy and splinting, can delay or obviate the need for surgical management (Berggren et al., 2001; Gravas et al., 2019). Despite this evidence, considerable provider variation exists in management of thumb CMC arthritis (Deutch et al., 2018; O’Brien and McGaha, 2014; Parker et al., 2020), from a choice of non-surgical regimens to decisions for surgical referral (Becker et al., 2015).

For many patients with thumb CMC arthritis, non-surgical treatment is successful in controlling pain and preventing conversion to surgical management. In this population-based study, we aim to investigate whether patient and provider characteristics influence the receipt of non-surgical treatment for thumb CMC arthritis. More specifically, we sought to examine associations between patient and provider characteristics and receipt of one, two or three types of non-surgical treatment: hand therapy, splinting and corticosteroid injection.

Methods

Data source and study population

We used Optum’s de-identified Clinformatics^® Data Mart Databases between 2015 and 2018 to conduct our study. Clinformatics^® Data Mart contains insurance claims from a large national payer and includes de-identified inpatient and outpatient encounters for over 140 million enrolees in the United States. This study was deemed not regulated from the Institutional Review Board, and the need for informed consent was waived. Patients age 18 years or older with a new diagnosis of thumb CMC arthritis were included, as identified via International Classification of Disease, Tenth Revision, Clinical Modification diagnosis codes (ICD-10) (Figure S1, available online).

Outcome

Our primary outcome measure was receipt of steroid injection, splinting or hand therapy, which were identified using current Procedural Terminology (CPT) codes and Healthcare Common Procedure Coding System (HCPCS). We considered each type of non-surgical treatment to be a unique category and relevant encounters were determined using both an ICD-10 code for thumb CMC arthritis and CPT/HCPCS code for the non-surgical treatment (Table 8S1, available online). Therefore, patients in our study could receive a maximum of three distinct types of non-surgical treatment, and a minimum of one type. The cumulative number of distinct types of treatment received was calculated for each patient using CPT/HCPCS codes associated with the treatments between our study timeframe of 2015 to 2018.

Explanatory variables

Variables of interest included provider-level and patient-level characteristics, sociodemographic data including age, sex, Elixhauser comorbidity score (based on ICD-10 codes), race, education level and insurance plan. Insurance plan was grouped into two distinct groups: fee-for-service (physicians are reimbursed for each service they provide) and managed care (physicians are reimbursed for a prespecified quantity and selection of services). Additionally, we identified patients with the following conditions that are commonly associated with thumb CMC arthritis: carpal tunnel syndrome, obesity, depression and chronic pain (Bakri and Moran, 2015; Dias et al., 2007; Haara et al., 2004; Tsehaie et al., 2019). Provider speciality included hand surgery (orthopaedic surgery, plastic surgery and general surgery), pain specialist/physical medicine and rehabilitation, medicine (medicine subspeciality, primary care and internal medicine) and other.

Statistical analyses

Descriptive statistics are presented for both patient and provider characteristics associated with receipt of different types of non-surgical treatment. Data are presented with mean and 95% confidence interval (CI) for most variables. We compared sociodemographic and clinical patient characteristics among the different non-surgical measures using the chisquared test for categorical variables and one-way ANOVA for continuous variables. We then used multinomial logistic regression to examine the associations among different regimens of non-surgical treatment with patient and provider characteristics.

Results

Patient and provider characteristics

A total of 35,899 patients received non-surgical treatment for newly diagnosed thumb CMC arthritis from 2015 through 2018 by 17,755 unique providers located in the United States. Of the total cohort, 10,248 (29%) were men and 25,651 (71%) were women, with a mean age of 72 years (range 21–93). The majority (26,772 patients, 75%) of patients received one distinct type of non-surgical treatment, compared with those who received two and three distinct types (4990 patients, 14% and 4137 patients, 11%, respectively). Table 1 provides the selected descriptive analysis of patient characteristics associated with number of distinct non-surgical treatments received. Among providers, 7954 (45%) were in medicine, 2246 (13%) in pain management, 1284 (7%) in hand surgery and 6271 (35%) in others. The median provider volume, which reflects the total number of patients each provider treated non-surgically within the timeframe of the study was 1.1 (IQR 2.1).

Table 1.

Number of patients associated with different conservative treatments received.

	Total number of patients	One non-surgical treatment	Two non-surgical treatments	Three non-surgical treatments
Mean age (SD)	72 (11)	72 (11)	72 (11)	72 (11)
Sex a
Male	10,248 (29%)	7390	1601	1257
Female	25,651 (71%)	19,382	3389	2880
Obesity	12,062 (34%)	9357	1428	1277
Carpal tunnel syndrome	8425 (24%)	6646	890	889
Depression	11,064 (31%)	8666	1260	1138
Chronic pain	10,114 (28%)	8127	1018	969
Elixhauser comorbidity score a	1388 (3.9%)	916	302	170
1–3	3580 (10%)	2463	643	474
4–7	5803 (16%)	4201	914	688
8+	25,128 (70%)	19,192	3131	2805

^ap-values of all these comparisons are p < 0.01, p-values refer to association among one, two and three types of non-surgical treatments and the patient characteristic of interest.

Further analyses of data to account for the differences in non-surgical treatment

After controlling for patient and provider characteristics, comorbid conditions and gender had the greatest effect on receipt of types of non-surgical treatment (Table 2). Patients with carpal tunnel syndrome, obesity, depression and chronic pain were less likely to receive two or three-types of non-surgical treatment. This was most pronounced in patients with a chronic pain diagnosis (two types: relative risk reduction (RRR), p < 0.01) (Table 2). Having chronic pain reduced the probability of receiving two types of conservative treatment by 5% and receiving three types of conservative treatment by 2%.

Table 2.

Multinomial logistic regression of patient and provider characteristics associated with number of distinct non-surgical treatments.^a

	Two types of non-surgical treatment			Three types of non-surgical treatment
	RRR	(95% CI)	p-value	RRRb	(95% CI)	p-value
Age	0.99	(0.99, 0.99)	<0.01	1.00	(0.99, 1.00)	0.02
Gender
Male	1.16	(1.09, 1.24)	<0.01	1.09	(1.02, 1.18)	0.01
Female (referencea)
Covariate diagnoses
Obesity	0.79	(0.74, 0.85)	<0.01	0.87	(0.81, 0.93)	< 0.01
Carpal tunnel syndrome	0.71	(0.66, 0.77)	<0.01	0.88	(0.81, 0.95)	<0.01
Depression	0.79	(0.73, 0.85)	<0.01	0.85	(0.79, 0.92)	<0.01
Chronic pain	0.65	(0.60, 0.70)	<0.01	0.75	(0.69, 0.81)	<0.01
Encounter type
Managed care	1.21	(1.10, 1.33)	<0.01	0.98	(0.88, 1.09)	0.74
Fee-for-service (reference)
Provider speciality
Medicine	1.47	(1.26, 1.72)	<0.01	1.36	(1.15, 1.61)	<0.01
Pain management	0.74	(0.61, 0.90)	<0.01	0.90	(0.74, 1.09)	0.29
Other	1.40	(1.20, 1.63)	<0.01	1.29	(1.10, 1.52)	<0.01
Hand surgeryb (reference)
Provider volume (natural log)	1.01	(1.00, 1.03)	0.09	0.99	(0.98, 1.01)	0.45

^aOverall, reference group is receipt of one type of non-surgical treatment.

^bHand surgery is composed of providers belonging to orthopaedic surgery, general surgery and plastic surgery.

RRR: relative risk ratio.

For provider speciality, medicine speciality was associated with a significant increase in the receipt of two types and three types of non-surgical regimens compared with hand surgeons (Table 2). Patients treated by pain management specialists had a lower relative risk of receiving two types of non-surgical treatment compared with hand surgeons (p < 0.01). There was no significant association between provider volume and receipt of two types and three types of non-surgical treatment regimens (Table 2).

Discussion

Our population-level analysis found that certain groups of patients, for example, with diagnoses of depression and chronic pain, were less likely to receive more than one type of non-surgical treatment for thumb CMC arthritis. This could reflect the fact that patients with psychological comorbidities perceive worse pain and symptoms related to osteoarthritis (Sharma et al., 2016). In addition, pain catastrophizing behaviour and greater psychological distress have been identified as independent predictors of higher pretreatment pain levels in those patients (Hoogendam et al., 2019). Our study also demonstrated that women were significantly less likely to receive more than one type of non-surgical treatment compared with men. There is evidence that pain perception associated with musculoskeletal conditions may be worse for women than men (Ruau et al., 2012), and that women are also more likely than men to report worsening symptoms as hand osteoarthritis progresses (Allen et al., 2006). Other studies have shown that women are less likely to be offered diagnostic modalities, including radiographs for non-traumatic wrist pain, and are commonly treated with narcotic analgesia, highlighting disparities in diagnosis and treatment of musculoskeletal conditions between men and woman (Billig et al., 2018). We recommend that providers should offer multiple types of non-surgical treatments for thumb CMC arthritis, regardless of gender or comorbid conditions, as this may provide adequate symptom control and obviate the need for surgery.

Our study also showed that specific provider specialities were more likely to provide two or three types of non-surgical treatments for thumb CMC arthritis. Hand surgeons were less likely to offer multiple types of non-surgical treatment as compared with other specialities, which reflects a greater inclination to offer surgical treatment. The variation that we observed in provider treatment regimens, coupled with the lack of formal treatment guidelines in the United States for non-surgical management of this disease, underscores the need for uniform care delivery from the providers. Future guidelines should address which types of patients will benefit from the different non-surgical treatment regimens and provide guidance on the expected intervals between treatment and symptom resolution. These guidelines are needed to inform all providers across specialities.

This study has limitations inherent to the use of insurance claims data. Clinformatics^® Data Mart databases do not contain granular clinical data, including disease severity or pain symptomatology, which may affect choice of non-surgical treatment. We were also unable to provide insight as to why providers in different specialities, such as hand surgeons and internal medicine providers, have significantly different approaches in non-surgical treatment of thumb CMC arthritis. Our study relied on the ICD-10 code M18.X as an inclusion criterion for our cohort, which may have eclipsed patients with thumb CMC arthritis who did not receive a diagnosis with this degree of specificity. It is also possible that our cohort included some patients for whom the diagnosis of thumb CMC arthritis was not new. However, we expect this number to be minimal owing to the fact that thumb CMC arthritis is a chronic diagnosis and should remain in a patient’s medical record following the initial diagnosis. Lastly, the databases used contain insurance claims for one large private employer-sponsored insurer in the United States, thus we cannot generalize the results to the uninsured, other insurers or other countries.

Despite these limitations, our findings have implications for future efforts that seek uniform and equitable care for this disease. In this study, we identified women and patients with carpal tunnel syndrome, obesity, chronic pain and depression as less likely to receive multiple types of non-surgical treatment. These groups may represent a target population for future efforts to ensure that all patients are receiving equitable care. Moreover, robust clinical practice guidelines should be created and disseminated to encompass all providers who treat this disease to facilitate more uniform treatment.