Use of Rheumatologic Testing in Patients Who Eventually Receive a Diagnosis of Rheumatoid Arthritis

Abstract

Objectives:

Anti-cyclic citrullinated peptide antibody (ACPA) has excellent specificity and prognostic value in patients with early rheumatoid arthritis (RA). The American College of Rheumatology included ACPA in their 2010 classification criteria for RA, but we hypothesize that primary care physicians (PCPs) underuse ACPA, even when clinical suspicion for RA is high. We aimed to describe their use of diagnostic testing in patients who were referred to a rheumatologist and eventually diagnosed as having RA.

Methods:

In this retrospective cohort study, a systematic abstraction tool was used to review the medical records of patients seen between January 1, 2010 and June 15, 2014 in two rheumatology clinics: one private practice and one community health center associated with an academic medical center. For purposes of hypothesis generation, we compared the characteristics of patients with and without testing using unpaired t tests or Fisher exact tests.

Results:

We identified 173 patients with RA referred from 141 different PCPs: 82.7% were women with a mean ± standard deviation (SD) age of 55.5 ± 18.6 years. ACPA and rheumatoid factor were ordered in 28.9% (95% confidence interval 22.6–36.2) and 41.0% (95% confidence interval 33.9–48.6) of patients, respectively. Imaging was underused. Almost half (45.7%, or 37/81) of the patients with documented symptom duration had a delay of at least 1 year before referral; however, ACPA utilization was not associated with the delay to treatment initiation.

Conclusions:

Most PCPs failed to order diagnostic tests for RA before referring a patient with polyarthritis who eventually received a diagnosis of RA. We also observed delays in diagnosis, with half of the patients waiting >1 year from symptom onset to diagnosis. These findings suggest educational efforts for PCPs should focus on emphasizing earlier diagnostic workups, especially ACPA, in patients suspected to have RA.

Rheumatoid arthritis (RA) is a chronic debilitating inflammatory disease that affects 1% of adults, significantly transforming quality of life and socioeconomic productivity, and uncontrolled RA imposes both physical and emotional limitations.¹ The healthcare and socioeconomic costs associated with RA complications and comorbidities are high.^1–4 Anti-cyclic citrullinated peptide antibody (ACPA) is a diagnostic tool that predicts an increased risk of progression from undifferentiated arthritis to RA. Prior research has demonstrated that in patients with early RA, the prevalence of ACPA is between 34% and 48%, whereas the prevalence of rheumatoid factor (RF) positivity is 50% to 66%.^5–9 In contrast, patients who do not receive a diagnosis of RA have an ACPA prevalence of 3% to 9% and an RF positivity estimated between 7% and 13%.^5–9 Thus, the presence of ACPA may be a harbinger for the development of RA, making it a valuable diagnostic tool (the reported positive predictive value is 96.6%).^10–12

Various interactions with the environment and genetic factors contribute to ACPA formation and disease progression.^10,11,13 Although the differential diagnosis of polyarthritis is broad and includes RA, crystal arthritis (calcium pyrophosphate dihydrate deposition disease and gout), spondyloarthritis (including psoriatic arthritis and reactive arthritis), connective tissue disease–associated inflammatory arthritis, and osteoarthritis, ACPA is a particularly specific test (90%–95%).^5–9 Its presence allows the early identification of RA, which leads to earlier remission, fewer disease complications, and fewer socioeconomic consequences for the patient.^7,12,14–20 Because of this the American College of Rheumatology and the European League Against Rheumatism included ACPA in their 2010 RA classification guidelines.²¹ Although this recommendation is now nearly a decade old, primary care physicians (PCPs) may still be unaware of the importance of initiating evaluation with ACPA and RF when there is a clinical suspicion of RA. We aimed to describe the proportion of patients who presented to a rheumatology clinic with the results of diagnostic testing (ACPA, RF, and imaging). We hypothesized that PCPs underuse ACPA and other diagnostic tests for RA, which could contribute to delays in the diagnosis and treatment of RA.

Methods

We conducted a retrospective chart review examining the primary care diagnostic workup of patients who were seen sequentially between January 1, 2010 and June 15, 2014 in two rheumatology clinics: one private practice (PP) and one community health center (CHC) associated with an academic medical center. There were no restrictions on ordering any tests by practices, and all of the diagnostic tests were available for PCPs to order during the study period. We used chart abstraction to collect demographics, insurance type, referring provider specialty, diagnostic tests, functional status, symptom duration, and treatments before and after referral. We used descriptive statistics to compare the characteristics and outcomes of patients who had ACPA performed before rheumatologist visit to those who did not. We used counts and percentages, means and standard deviations, and percentile distributions to describe patient characteristics (we calculated mean age). Test proportions were calculated with exact 95% confidence intervals (CIs). We explored associations between ACPA testing and various patient or provider characteristics using unpaired t tests (continuous) or Fisher exact tests (categorical), but these analyses should be considered exploratory because they were not the primary outcome of the study. All of the analyses were carried out using SAS statistical software (version 9.3, SAS Institute, Cary, NC). The Baystate Medical Center institutional review board determined that the project did not constitute human subjects research.

Results

We identified 173 patients with RA, referred from 141 different PCPs. The Table demonstrates baseline patient characteristics. Of the 173 patients, 75.7% were seen at the PP, and 82.7% were women with a mean ± standard deviation age of 55.5 ± 18.6 years. The Figure shows laboratory tests ordered by the referring provider. Referring clinicians ordered ACPA in only 28.9% (95% CI 22.6–36.2) and RF in 41.0% (95% CI 33.9–48.6) of patients. Ordering of ACPAs varied nonsignificantly by presenting site (26.2% CHC vs 29.8% PP, P = 0.70). Radiological imaging was documented for 32.4% (n = 56) of patients, with x-ray being the most frequent (92.9%, or 52/56) and magnetic resonance imaging a distant second (12.5%, or 7/56). Functional ability was documented in only 2.3% (n = 4) of referred patients (95% CI 0.9–6.1).

Table.

Baseline patient characteristics of patients with RA

	Overall (N = 1173)	ACPA performed prereferral			ACPA or RF performed prereferral
		Yes (n = 50)	No (n = 123)	P-diff	Yes (n = 78)	No (n = 95)	P-diff
Age, y, mean (SD)	55.5 (18.6)	52.8 (19.2)	56.6 (18.3)	0.22	53.2 (17.0)	57.4 (19.7)	0.13
Male sex, no. (%)	30 (17.3)	10 (20.0)	20 (16.3)	0.66	15 (19.2)	15 (15.8)	0.55
Provider level, no. (%)
Internal medicine	119 (68.8)	40 (80.0)	79 (64.2)		51 (65.4)	68 (71.6)
Family medicine	15 (8.7)	4 (8.0)	11 (8.9)		7 (9.0)	8 (8.4)
	16 (9.3)	3 (6.0)	13 (10.6)		10 (12.8)	6 (6.3)
Subspecialty/multispecialty
NP or PA	16 (9.3)	3 (4.0)	14 (11.4)		9 (11.5)	7 (7.4)
Self-referred	7 (4.1)	1 (2.0)	6 (4.9)	0.35	1 (1.3)	6 (6.3)	0.23
Payer, no. (%)
Medicaid	76 (43.9)	23 (46.0)	53 (43.1)		38 (48.7)	38 (40.0)
Medicare	37 (21.4)	8 (16.0)	29 (23.6)		11 (14.1)	26 (27.4)
Private	34 (19.7)	11 (22.0)	23 (18.7)		18 (23.1)	16 (16.8)
Medicaid and Medicare	20 (11. 6)	7 (14.0)	13 (10.6)		9 (11.5)	11 (11.6)
Medicare and private	6 (3.5)	1 (2.0)	5 (4.1)	0.77	2 (2.6)	4 (4.2)	0.24
Location, no. (%)
CHC	42 (24.3)	8 (16.0)	34 (27.6)		13 (16.7)	29 (30.5)
PP	131 (75.7)	39 (78.0)	92 (74.8)	0.70	62 (79.5)	69 (72.6)	0.37

CHC, community health center associated with an academic medical center; NP, nurse practitioner; PA, physician assistant; PP, private practice; RA, rheumatoid arthritis; RF, rheumatoid factor.

Fig.

Tests ordered in all RA patients (N = 173) by referring provider. Error bars represent 95% confidence intervals. ACPA, anti-cyclic citrullinated peptide antibody; ANA, anti-nuclear antibody; Anti-dsDNA, anti-double-stranded DNA antibody; Anti-SSA/Ro, Sjogren syndrome type A antigen; Anti-SSB/La, Sjogren syndrome type B antigen/lupus La protein; CBC, complete blood count; CMV, cytomegalovirus; CRP, C-reactive protein; EBV, Epstein-Barr virus; ESR, erythrocyte sedimentation rate; LFT, liver function test; RA, rheumatoid arthritis; RF, rheumatoid factor; UA, undifferentiated arthritis.

We collected some additional data that will be useful for hypothesis generation for future studies. Most patients (86.7%, n = 150) had a disease-modifying antirheumatic drug initiated immediately on presentation to a rheumatologist. The proportion of patients waiting ≥1 month for treatment did not vary significantly by ACPA testing (14.3% if ACPA was ordered vs 19.8% if ACPA was not ordered, P = 0.63); however, almost half (45.7%, or 37/81) of the patients with documented symptom duration had a gap of at least 1 year between their first symptom and referral to a rheumatologist.

Discussion

Because nearly all workup of rheumatologic illness begins in the primary care setting, quality improvement efforts in the diagnosis of RA must focus on PCPs. The rates of use of ACPA and RF by PCPs in appropriate patients were previously unknown, however. This study examined their use of diagnostic tools, including ACPA, in patients with a high clinical suspicion for RA. We found that most referring PCPs failed to order diagnostic tests such as ACPA, RF, or x-rays before referring a patient to rheumatology. Although describing diagnostic delay was not the primary outcome of our study, we found that (in a small subset of patients in whom we knew the timing of symptom onset) approximately 50% patients waited ≥1 year to be referred to a rheumatologist, potentially missing the window for early diagnosis and treatment.^8,14,16,17 These findings suggest that failure to test (and, perhaps, failure to refer) could be one of the possible mechanisms that underlies delays in diagnosis of RA.^22–27

The relation between ordering ACPA and RF and diagnostic delay remains far from clear, however. For example, failure to order ACPA prereferral did not delay treatment initiation by the rheumatologist. This suggests that rheumatologists acted quickly in the presence or absence of the test and that once referred, treatment was initiated promptly. Any delay, therefore, occurred in the primary care setting while waiting for referral or appointment. Because we have limited information regarding symptom presentation in the primary care setting (we included patients from >140 primary care practices, making chart review of PCPs’ notes extremely difficult), we cannot determine whether PCPs were following current American College of Rheumatology/European League Against Rheumatism recommendations to increase the early diagnosis and treatment of RA.

This study is an important contribution to the literature because early diagnosis of RA is a situation in which drawing a relatively inexpensive laboratory test could save months or years of pain, suffering, expensive surgeries, complications, and lifelong disability.^14–16 Any contributor to diagnostic delay is important to identify and attempt to rectify. Future research should continue to attempt to determine how PCPs address complaints of polyarthritis. Quality improvement interventions also are needed (including education, audit and feedback, and electronic order sets) to increase PCPs’ awareness about the need to order diagnostic testing prereferral in patients with high clinical suspicion for RA. For example, customized referral forms (or order sets) specifying clinical examination findings and diagnostic workup for polyarthritis could be formulated by local rheumatologists. These would provide education to PCPs and allow rheumatologists to better triage referrals by severity. This is especially important in geographic areas where there is poor access to rheumatologists. Finally, education to improve knowledge and the clinical musculoskeletal skills of PCPs will help in detection of synovitis earlier in the disease course.

This study is limited by its small sample size and retrospective review of records from only two rheumatology clinics. Furthermore, our review of charts did not allow us to determine the timing of symptom onset (e.g., gradual vs rapid symptom onset), the presence of specific joint involvement because of variation in documentation, the urban/rural makeup of our population, or socioeconomic status of patients²⁸ (although a rough proxy of socioeconomic status is the proportion of patients seen in a CHC vs a PP). We were unable to obtain complete comorbidity information on the patients and did not have information about the referring physicians (e.g., years in practice, location of medical school or residency program), which could have influenced physician decisions to order ACPA. We also did not collect information on inflammatory markers (e.g., erythrocyte sedimentation rate/C-reactive protein). This is an important limitation because a lower level of inflammatory markers or small joint involvement may contribute to delays in diagnosis. These limitations have less relative importance, however, because our study was primarily descriptive and focused on the efforts of PCPs to use laboratory and radiologic testing to identify RA at an early stage.

Conclusions

Most PCPs failed to order diagnostic tests for RA before referring a patient with polyarthritis who eventually received a diagnosis of RA. In particular, PCPs failed to order ACPA in 61% of referred patients who eventually received a diagnosis of RA. Failure to order ACPA did not appear to significantly delay the initiation of RA treatment once the patient had been seen by a rheumatologist, and we could not determine the impact of testing (or lack of testing) on referrals. These findings suggest that educational efforts for PCPs and quality improvement efforts in primary care settings should focus on emphasizing earlier diagnostic workup, especially ACPA, in patients suspected to have RA.

Key Points.

• Rheumatoid arthritis (RA) is a disease that, when untreated, can cause disability and extraarticular complications, leading to death.

• Anti-cyclic citrullinated peptide antibody (ACPA) is a specific test that is useful in diagnosing early RA, but we found that ACPA is underused by primary care physicians (PCPs).

• PCPs ordered ACPA in fewer than one-third of patients who received an eventual diagnosis of RA.

• Educational efforts for PCPs and quality improvement efforts in primary care settings should focus on emphasizing earlier diagnostic workups, especially ACPA, in patients suspected to have RA.