Background/Purpose:

Observational registries provide a valuable complement to clinical trials yet might suffer from limited generalizability referent to the desired population of interest. We compared characteristics of older patients enrolled in a large U.S. rheumatoid arthritis (RA) registry compared to RA patients treated by non-participating rheumatologists throughout the U.S.

Methods:

We used national data from the U.S. Medicare program from 2006-2010 and linked it to the Consortium of Rheumatology Researchers of North America (CORRONA) registry. Linkage was accomplished using deterministic methods on multiple non-unique identifiers and achieved > 96% accuracy. RA patients were identified in Medicare on the basis of having at least one ICD-9 diagnosis code for RA (714.X) from a rheumatologist. Demographics, comorbidities, and drug utilization were examined in Medicare data during the most recent consecutive 12 month period with full fee-for-service coverage including pharmacy benefits. Characteristics of CORRONA RA patients were compared to characteristics of RA patients of non-CORRONA rheumatologists. Differences were quantified as standardized mean differences (SMDs), with values < 0.1 considered clinically irrelevant.

Results:

A total of 620,111 RA patients in Medicare treated by 3,834 non-CORRONA rheumatologists were identified and their characteristics compared with 5,643 RA patients treated by 282 CORRONA rheumatologists (Table). Mean age and sex were similar; CORRONA patients were more likely to be recruited in the Northeast and West. The prevalence of common comorbidities was comparable. During the 12m observation period, CORRONA RA patients were more likely to use biologics (22 vs. 12%) and somewhat less likely to use oral glucocorticoids (33 vs. 38%). Health services utilization was similar between CORRONA and non-CORRONA RA patients.

Conclusion:

Despite some differences in regional geographic representation, older RA patients enrolled in CORRONA were generally similar in their characteristics to older RA patients of other rheumatologists not participating in CORRONA, although were more likely to receive biologic therapies. Based upon these findings, results from this U.S. registry may be generalizable to a national U.S. RA population.

Table: Characteristics of Older RA Patients Enrolled in CORRONA vs. RA Patients of Non-CORRONA Rheumatologists

Patient Characteristic	CORRONA Patients	Non-CORRONA Patients	Standardized Mean Differences
Age, years Female sex, % Caucasian race Region, %    Northeast    Midwest    West    South Rural Residence, % Comorbidities, %    COPD    Diabetes    Myocardial infarction    History of cancer RA Medications, %    Oral glucocorticoids    Methotrexate    Any biologic Healthcare utilization    Number of physician visits, mean+- SD    Any hospitalization, %	72.3 (9.9) 74.6 89.5 30.2 25.1 12.7 32.1 21.7 13.5 18.0 1.3 10.5 33.3 30.9 22.3 13.9 +- 10.2 26.2	72.7 (11.1) 73.8 85.2 17.4 23.4 15.9 43.3 23.4 14.0 21.9 1.5 10.1 37.9 21.7 12.4 14.2 +- 10.9 28.9	0.03 0.02 0.18 0.038 0.303 0.091 0.233 0.06 0.01 0.08 0.02 0.01 0.11 0.21 0.27 0.02 0.06

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/generalizability-of-a-u-s-rheumatoid-arthritis-registry-a-comparison-of-participants-vs-non-participants-characteristics/