Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.

Descriptive Epidemiology of Adult Rheumatoid Arthritis in an Insurance Claims Database.

Crane1,  Martin M., Juneja3,  Maneesh, Aziz3,  Fayaz A., Kurrasch2,  Regina H., Chu2,  Myron E., Quattrocchi3,  Emilia, Manson4,  Stephanie

GlaxoSmithKline R&D, Research Triangle, NC
GlaxoSmithKline R&D, King of Prussia, PA
GlaxoSmithKline R&D
United Biosource


Our goal was to estimate the incidence and prevalence of rheumatoid arthritis (RA) in a US administrative medical claims database. Databases such as this one contain the majority of encounters and prescriptions for any disease that can be defined using an ICD9 code in an insured US population ("real-world" practice); however, the major disadvantages are lack of any clinical information, left-censoring (incomplete medical history)and right-censoring (truncated follow-up) of data, and under-representation of the elderly due to Medicare coverage at age 65 years.


The study population was defined as all patients >= 18 years having medical information for the entire 4.75-year period of 1-Jan-2004 to 30 Sept-2008 (N=4.65 million) in a health claims database (IMS LifeLink: Health Plan Claims Database, PharMetrics, Inc, Watertown, WA.) Prevalent RA was defined as having at least one occurrence of a code (ICD9 7140–7142) prior to 2006, and one occurrence during 2006, with the two occurrences at least 30 days apart. Incident RA was defined as no RA codes prior to 2006, and two codes at least 30 days apart during the interval 1-Jan-2006 to 31-March 2007. Polyarthritis or undifferentiated arthritis (7149) was not part of the algorithm; however, an "aggressive" incident subset was defined based on having >8 RA physician visits in the 12 months post diagnosis and no prior 7149 code.


Prevalence was 0.65% overall (0.92% in females and 0.36% in males) and incidence was 0.07% (0.10% female; 0.04% male). Median age at the occurrence of the first RA code was 57.0 in the incident and 55.0 years in the prevalent patients and, prior to 2006, 47.8% of the latter had received at least one prescription for methotrexate. Other connective tissue diagnoses (based on two occurrences 30 days apart) were present in 7.9% and 9.7% of the incident and prevalent patients respectively. In the 12 months after the first RA visit, 64.0% of incident and 55.1% of prevalent patients were prescribed a non-biologic DMARD and 20.4% and 22.1% were prescribed an anti-TNF. The "aggressive" subset was approximately 11% of all incident patients.


Prevalence rate (0.65%) was lower than in other cohorts (Mayo Clinic – 0.72%; UK Norfolk–~0.80%) that used ACR diagnostic criteria and fully captured RA in the aged. Incidence (0.07%) was greater than reported in either prior cohort (0.04% and ~0.025% respectively for Mayo Clinic, US and Norfolk, UK). Claims databases probably underestimate the true prevalence of RA but may better reflect the proportion of patients actively seeking care; incidence is likely to be overestimated due to left-censoring.

To cite this abstract, please use the following information:
Crane, Martin M., Juneja, Maneesh, Aziz, Fayaz A., Kurrasch, Regina H., Chu, Myron E., Quattrocchi, Emilia, et al; Descriptive Epidemiology of Adult Rheumatoid Arthritis in an Insurance Claims Database. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :58
DOI: 10.1002/art.27827

Abstract Supplement

Meeting Menu