An epidemiological and healthcare utilisation study of rheumatoid arthritis in an adult population in the US

Summary

This study estimated the incidence of rheumatoid arthritis (RA) in a US population comprising adults from the MarketScan Research Database®*. The incidence population included subjects with no RA-related medical claims or treatment history during 2001–2002. Among the incidence population, patients with RA in 2003 made up the newly diagnosed RA population and patients with RA for more than 3 years comprised the longstanding RA population.

The age- and sex-adjusted RA incidence based on the 2002 US population was 0.08% (0.06% in males; 0.11% in females); the age- and sex-adjusted prevalence was 0.73% (0.43% in males; 0.98% in females). The prevalence-to-incidence ratio was 9.1 (7.2 in males; 8.9 in females). Treatment with non-steroidal anti-inflammatory drugs, disease-modifying antirheumatic drugs and biologicals, respectively, was documented in 36.7, 17.7 and 2.5% of newly diagnosed patients within 12 months after initial diagnosis and in 68.2, 87.3 and 34.4%, respectively, of patients with longstanding RA.

Management of existing RA cases remains a challenge, with only a small proportion of new cases receiving standard treatment.

Keywords: :

• rheumatoid arthritis
• incidence
• prevalence

Introduction

It has been widely referenced in the literature that the prevalence of rheumatoid arthritis (RA) is approximately 1% among adults in the US1. However, few epidemiological studies have been conducted to investigate the incidence of RA.

The healthcare costs attributable to RA have been estimated in several studies2–4, with direct healthcare costs being two to three times higher than average costs for individuals of similar age and gender5, and also being disproportionately higher in women than men6. However, since most studies have been conducted on patients with established RA, and healthcare utilisation and costs rise with patient age and disease duration5, information regarding the use of healthcare resources in patients with early RA is limited.

Similarly, treatment patterns have also been evaluated in patients with established RA7. In a review of 10 studies involving direct costs of RA, medications were second only to hospital care in contributing to RA-related costs, accounting for more than 25% of the total costs. These costs include the cost of medications as well as costs associated with monitoring and treatment of adverse events5. In patients with established RA, disease-modifying antirheumatic drugs (DMARDs) account for two-thirds of the total drug cost, and non-steroidal anti-inflammatory drugs (NSAIDs) account for the remainder5,8,9. Again, limited data are available on treatment patterns among patients with early RA.

A number of new medications have been approved for the treatment of RA in the past few years. Given that these medications, including inhibitors of tumour necrosis factor, interleukins, cytokines and COX-2, are more costly than drugs used in previous decades, it is relevant to assess treatment patterns in patients with early RA. In this analysis, the prevalence and incidence of RA were estimated and drug treatment and healthcare costs between newly diagnosed patients and those with longstanding RA in the US population were compared using data from a healthcare claims database.

Patients and methods

Data source

The MarketScan Research Database®* is a longitudinal healthcare claims database that includes patients covered by employer-sponsored health plans for eligible employees, early retirees and dependants. The study population also included retirees with an employer-sponsored Medicare supplemental plan for individuals over 65 years of age. The majority of patients were from the south and north-central regions of the US, with an initial population of 4.3 million patients. Inpatient and outpatient service information, including diagnoses (9th Revision of the International Classification of Diseases, Clinical Modification [ICD-9-CM] format), procedures (Current Procedural Terminology, 4th Edition [CPT-4] and Healthcare Common Procedure Coding System [HCPCS] formats), prescription drugs and the total payment including co-payment for each service or prescription, are recorded from healthcare claims forms from the service providers. Additional data elements include demographic variables (e.g. age, gender, geographic region) and start and stop dates for plan enrolment. All patients who met the selection criteria specified below were included in the analyses.

Selection criteria and measures

Adults (aged >17 years) continuously enrolled from the 1st January 2001 to the 31st December 2004 with medical and drug benefits were selected. Patients with claims for RA were identified using the ICD-9 code 714.x. Prescriptions for DMARDs, NSAIDs and biologicals (infliximab, etanercept, adalimumab and anakinra) were identified using published National Drug Codes or J-codes (infliximab). Data regarding over-the-counter medications were not available and were therefore not included in this analysis.

The prevalence of RA was estimated using eligible patients in 2003. To estimate the incidence of RA in 2003, the denominator population included subjects who had no RA-related medical claims, as well as no RA-related treatment history with DMARDs, NSAIDs or biologicals in the previous 2 years (2001–2002), and the numerator included patients who had newly diagnosed RA in 2003 among the incidence population. The age and gender distribution of the general US population for 2002 was used to estimate the overall age- and sex-adjusted prevalence and incidence rates, and 95% confidence intervals were constructed using the assumption that the number of cases per year follows a Poisson distribution.

Healthcare utilisation and drug treatment patterns for newly diagnosed patients within 12 months of disease onset were evaluated. The cost analysis included total medical costs (hospitalisation, outpatient service, emergency room visit, specialist visit and outpatient hospital service) and prescription drug costs. For comparison, a group of RA patients who had RA claims continuously for more than 3 years (from 2001 to 2004) were identified as patients with longstanding RA. Healthcare utilisation and drug treatment patterns for these patients were derived from 2004 records.

Results

Prevalence and incidence of rheumatoid arthritis

Among the 2,188,410 subjects who did not have RA in the previous 2 years, the incidence of RA was 0.10% (0.07% in males; 0.13% in females), with a US population age- and sex-adjusted rate of 0.08% (0.06% in males; 0.11%in females) and a female-to-male incidence ratio of 1.8. Among the 2,214,280 subjects in the 2003 database, the prevalence of RA was 0.91% (0.53% in males; 1.24% in females), with a US population age- and sex-adjusted rate of 0.73% (0.43% in males; 0.98% in females) and a female-to-male prevalence ratio of 2.3 (Table 1). The prevalence-to-incidence ratio was 9.1 (7.2 in males; 8.9 in females). Both the prevalence and incidence of RA increased with age; a significant increase was observed after 50 years of age (Figures 1 and 2, respectively).

Table 1. Estimated prevalence and incidence of RA in the study population.

	Male	Female	Total
Adults enrolled from the 1st January 2001 to the 31st December 2003	1,026,574	1,187,706	2,214,280
Prevalence of RA in 2003 (n (%))	5,486 (0.53)	14,687 (1.24)	20,173 (0.91)
US population sex/age-adjusted RA prevalence (%) (95% CI)	0.43 (0.42–0.44)	0.98 (0.96–1.0)	0.73 (0.72–0.74)
Adults enrolled from 2001 to 2003 and without RA in 2001–2002	1,019,388	1,169,022	2,188,410
Incidence of RA in 2003 (n (%))	690 (0.07)	1,515 (0.13)	2,205 (0.10)
US population sex/age-adjusted RA incidence (%) (95% CI)	0.06 (0.06–0.07)	0.11 (0.10–0.12)	0.08 (0.08–0.08)

CI, confidence interval; RA, rheumatoid arthritis.

Figure 1. . Estimated prevalence of rheumatoid arthritis in 2003 in the study population by age and gender.

Figure 2. . Estimated incidence of rheumatoid arthritis in 2003 in the study population by age and gender.

Healthcare costs

Among the patients with newly diagnosed RA, the average total medical cost over 12 months was $7,254. The average pharmacy-related cost was $2,131 (Table 2).

Table 2. Healthcare costs among patients with newly diagnosed RA and longstanding RA*

	Newly diagnosed RA (N = 2,205)	Longstanding RA (>3 years) (n = 9,930)
Female gender (%)	68.71	74.76
Mean (sd) age (years)	56.6 (13.7)	59.1 (13.1)
Total medical costs ($)^†
Mean	7,254.33	10,796.24
sd	19,388.77	22,838.28
Median	2,354.46	3,659.46
Emergency room costs ($)
Mean	34.64	49.38
sd	116.95	195.58
Median	0.00	0.00
Hospitalisation costs ($)
Mean	2,325.00	3,549.80
sd	11,883.46	17,143.46
Median	0.00	0.00
Outpatient visit costs ($)
Mean	613.26	853.61
sd	553.88	1,161.10
Median	494.00	645.13
Outpatient hospital costs ($)
Mean	13.81	14.03
sd	95.62	102.47
Median	0.00	0.00
Specialist visit costs ($)
Mean	43.77	139.33
sd	127.36	722.32
Median	0.00	0.00
Other medical costs ($)
Mean	4,223.85	6,190.09
sd	11,955.52	12,597.01
Median	1,585.97	2,411.35
Pharmacy costs ($)
Mean	2,131.30	12,120.96
sd	3,111.68	16,372.08
Median	1,221.26	3,853.62
Total costs ($)
Mean	9,390.14	22,917.20
sd	20,239.24	27,642.27
Median	4,292.09	13,819.44

sd, standard deviation; RA, rheumatoid arthritis.

*Costs for newly diagnosed RA were accumulated costs in 12 months from initial date of diagnosis, and 19 patients missed pharmacy costs (total costs and pharmacy costs were based on 2,186 patients); costs for longstanding RA were accumulated costs in 2004, and 11 patients missed all costs.

^†Total medical costs included hospitalisation, outpatient service, emergency room visit, specialist visit, outpatient hospital service and other medical costs.

Among the patients with longstanding RA (i.e., those cases diagnosed since 2001 or earlier), the average total medical cost over 12 months was $10,796. The total pharmacy cost was $12,121 (Table 2).

Treatment patterns

Approximately 36.7% of the newly diagnosed patients were treated with NSAIDs, 17.7% were treated with DMARDs and 2.5% were treated with biologicals within 12 months of the initial diagnosis (Table 3).

Table 3. RA-related prescriptions during the following year for newly diagnosed RA and longstanding RA*.

	Newly diagnosed RA (n = 2,205)	Longstanding RA (>3 years) (n = 9,930)
NSAIDs (%)	36.73	68.24
DMARDs (%)	17.73	87.36
>1 DMARD (%)	14.97	85.77
Biological agents (%)	2.45	34.42

RA, rheumatoid arthritis; NSAID, non-steroidal anti-inflammatory drug; DMARD, disease-modifying antirheumatic drug.

*Prescriptions for newly diagnosed RA were derived from data 12 months following the initial date of diagnosis; prescriptions for longstanding RA were data in 2004.

Among the patients with longstanding RA, approximately 68.2% of patients were treated with prescription NSAIDs, 87.4% were treated with DMARDs and 34.4% were treated with biologicals in 2004 (Table 3).

Discussion

RA is a chronic inflammatory condition of unknown cause that occurs with considerable variability among different populations. Patients with RA often experience progression to a chronic state of joint inflammation that is associated with significant functional disability and morbidity. Few epidemiological studies have been conducted to investigate the incidence of RA.

Healthcare claims databases contain anonymous patient-level data from an insured population and have been used extensively for outcomes research and pharmacoeconomic analyses. Specifically, the MarketScan claims database has been employed in research across a variety of disease states, including RA, cancer and anaemia10–12. Individuals in the MarketScan database are employees and their dependants, which could raise questions regarding the relevance of analysis results to the general US population. Interestingly, a comprehensive comparative analysis of the MarketScan and PharMetrics claims databases with regard to the prevalence and co-morbidities across various immune-mediated inflammatory diseases was recently completed10. The PharMetrics database derives from the insured participants of over 80 healthcare plans covering more than 55 million unique individuals in the US and has been considered nationally representative. The analyses demonstrated almost identical results between the two databases, and these results were similar to findings within the general US population10. Therefore, using the MarketScan claims database to estimate the prevalence and incidence of RA, and to analyse healthcare costs and treatment patterns in RA patients with newly diagnosed and established RA, might be an appropriate analytical option.

The occurrence of RA appears to vary among countries and areas of the world. For example, prevalence estimates tend to be lower in southern European and Asian countries compared with northern European and American countries. With regard to the US, the overall prevalence of definite RA among adults has been reported to be 1.0%, with rates for women being approximately 2.5 times higher than rates for men1. In the UK, the prevalence of RA is reported to be 0.8% among the adult population13. However, when comparing prevalence and incidence estimates, differences between the study populations in age and gender must be considered since female gender and older age are strong risk factors for RA. Indeed, in an evaluation of patients from Rochester (MN, USA) aged at least 35 years, Gabriel et al estimated the prevalence of RA to be 1.1% (0.74% in males; 1.31% in females)14, whilst evaluations involving younger cohorts (i.e. at least 16–20 years of age) yielded lower prevalence estimates (i.e. 0.18–0.85%)15. Based on the authors’ analysis of a US adult population at least 18 years of age, the age- and sex-adjusted prevalence of RA was 0.73% (0.43% in males; 0.98% in females). If the study population is limited to only those individuals at least 35 years of age, the age- and sex-adjusted prevalence rate would be 0.99% (0.59% in males; 1.33% in females), which is consistent with findings reported by Gabriel et al 14.

There are very few population-based epidemiological studies on the incidence of RA. The majority of studies carried out in northern European and American countries estimate a mean annual incidence of 0.02–0.05%16. In an evaluation by Doran et al17, who followed an inception cohort of residents (>18 years of age) from Rochester (MN, USA) from 1955 to 1995, the overall age- and sex-adjusted annual incidence of RA, as assessed by the American College of Rheumatology (ACR) 1987 criteria, was 0.045% (0.03% in males; 0.057% in females). In the Doran report, the incidence of RA fell progressively over the course of the four-decade study, from 61.2/100,000 in 1955–1964 to 32.7/100,000 in 1985–199417. In the authors’ analysis, the age- and sex-adjusted incidence of RA was higher than in these previous reports, i.e. 0.08% (0.06% in males; 0.11% in females). There are several possible explanations for the observed difference in RA incidence. A high-sensitivity case identification method was used (i.e. one or more healthcare claims for RA). If at least two RA claims for case identification had been a requirement, then the age- and sex-adjusted incidence would be 0.06% (0.04% in males; 0.07% in females). Diagnosis criteria could be another source for variation. Compared with the use of ACR criteria for RA diagnosis in the Doran report, the authors were unable to verify diagnosis criteria for each case of claimed RA owing to the anonymous nature of patient data in the claims database.

Both the prevalence and incidence of RA increased with age, with a significant increase observed after 50 years of age. However, the data show a continuous increase in incidence after 60 years of age and a decreasing prevalence of RA after 75 years. The high incidence of co-morbidities, such as cardiovascular disease and diabetes, seen in RA patients may lead to a higher mortality in the aged population, reflecting a decreased prevalence of underlying diseases such as RA.

This evaluation of healthcare costs indicated that for both new and established cases of RA, hospitalisation and pharmacy costs accounted for significant portions of the total costs. In a 2001 review of 10 studies on the direct costs of RA, Lubeck reported that hospitalisation costs were the most significant contributor to healthcare costs even though only approximately 10% of the RA population is hospitalised in a given year5. As such, treatment strategies aimed at controlling inflammation early in the disease process, thus decreasing disability, joint destruction and subsequent surgery/hospitalisation, could result in lower overall RA healthcare costs. Since factors influencing total costs of RA might help clinicians and decision makers direct resources appropriately, future prospective evaluations of RA incidence and prevalence should include an assessment of factors influencing these costs, including demographic, clinical, treatment and occupational variables.

Historically, RA had been viewed as a benign disease that was initially managed with NSAIDs; therapy with DMARDs was reserved for those with persistent disease despite treatment with NSAIDs. However, it is now appreciated that RA is an aggressive disease that causes erosive joint damage even within the first 2 years. As such, early and aggressive therapy with DMARDs and newer biological agents is now advocated to halt disease progression. In this analysis, among the patients with longstanding RA diagnosed since 2001, 87.4 and 34.4% of patients had received DMARDs and biological agents, respectively, during 2004. These data are consistent with recent findings suggesting the current standard of care for RA is treatment with methotrexate monotherapy, followed by methotrexate in combination with a biological or other DMARD7. Interestingly, the treatment data collected indicate that only a small proportion of new cases received standard treatment, with only 17.7 and 2.5% of new patients receiving DMARDs and a biological agent, respectively, within 12 months after initial diagnosis. Although data regarding the use of over-the-counter NSAIDs, which currently might comprise the major treatment option for this patient group, were not available for inclusion in this analysis, these findings suggest that newly diagnosed RA patients are not consistently receiving treatment aimed at controlling inflammation and stabilising radiographic progression early in the disease process.

In conclusion, although these study results need to be confirmed using medical records data owing to the inherent limitations of claims data, the findings support the belief that RA has a great impact on individuals and society because of its high prevalence, and associated large use of acute and long-term healthcare and support services2. The authors recommend further evaluation of factors influencing treatment patterns and overall healthcare costs, as such information might help clinicians and decision makers to direct healthcare resources.

Acknowledgement

The authors would like to thank Michelle Perate MS for her writing support.

This study was designed by staff members from Centocor, Inc., who analysed and interpreted the data, and wrote and submitted this manuscript for publication.