Derivation of severity index for rheumatoid arthritis and its association with healthcare outcomes

Burden of rheumatoid arthritis among US Medicare population: co-morbidities, health-care resource utilization and costs

Objectives

The study aimed to assess the burden of RA among the US Medicare population (aged ≥65 years) by comparing co-morbidities, health-care resource utilization (HCRU) and costs against matched non-RA Medicare patients.

Methods

Data were obtained from the Medicare fee-for-service claims database from 2010 to 2013. RA Medicare patients were identically matched with Medicare patients without RA (controls) based on demographics. Bivariate analyses were conducted to examine differences between cohorts for co-morbidities, HCRU and costs. A generalized linear model was used to test relationships between patient-level characteristics, HCRU and costs.

Results

The study population included 115 867 RA patients and 115 867 age-, sex-, race- and region-matched non-RA controls. Mean age was 75.2 years; 79.4% were female. Co-morbidities were greater in RA vs non-RA patients [Charlson Co-morbidity Index (excluding RA): 1.86 vs 1.00; P < 0.0001]. All-cause annual HCRU was greater in RA vs non-RA patients. Total annual health-care costs were ∼3-fold higher in RA vs non-RA patients ($20 919 vs $7197, respectively; P < 0.0001) with the major driver of costs in the RA cohort being outpatient costs. Approximately half of the overall costs in the RA cohort were RA related ($11 587). After controlling for differences in patient characteristics and co-morbidities between cohorts, the adjusted total mean annual costs for RA patients were still more than twice those of non-RA patients ($16 374 vs $6712; P < 0.0001).

Conclusions

Among US Medicare patients, those with an RA diagnosis had a significantly greater burden of co-morbidities, HCRU and costs compared with a matched cohort without RA.

Models solely using claims-based administrative data are poor predictors of rheumatoid arthritis disease activity

BACKGROUND

This study developed and validated a claims-based statistical model to predict rheumatoid arthritis (RA) disease activity, measured by the 28-joint count Disease Activity Score (DAS28).

METHOD

Veterans enrolled in the Veterans Affairs Rheumatoid Arthritis (VARA) registry with one year of data available for review before being assessed by the DAS28, were studied. Three models were developed based on initial selection of variables for analyses. The first model was based on clinically defined variables, the second leveraged grouping systems for high dimensional data and the third approach prescreened all possible predictors based on a significant bivariate association with the DAS28. The least absolute shrinkage and selection operator (LASSO) with fivefold cross-validation was used for variable selection and model development. Models were also compared for patients with <5 years to those ≥5 years of RA disease. Classification accuracy was examined for remission (DAS28 < 2.6) and for low (2.6-3.1), moderate (3.2-5.1) and high (>5.1) activity.

RESULTS

There were 1582 Veterans who fulfilled inclusion criteria. The adjusted r-square for the three models tested ranged from 0.221 to 0.223. The models performed slightly better for patients with <5 years of RA disease than for patients with ≥5 years of RA disease. Correct classification of DAS28 categories ranged from 39.9% to 40.5% for the three models.

CONCLUSION

The multiple models tested showed weak overall predictive accuracy in measuring DAS28. The models performed poorly at predicting patients with remission and high disease activity. Future research should investigate components of disease activity measures directly from medical records and incorporate additional laboratory and other clinical data.

An external validation study reporting poor correlation between the claims-based index for rheumatoid arthritis severity and the disease activity score

INTRODUCTION

We conducted an external validation study to examine the correlation of a previously published claims-based index for rheumatoid arthritis severity (CIRAS) with disease activity score in 28 joints calculated by using C-reactive protein (DAS28-CRP) and the multi-dimensional health assessment questionnaire (MD-HAQ) physical function score.

METHODS

Patients enrolled in the Brigham and Women's Hospital Rheumatoid Arthritis Sequential Study (BRASS) and Medicare were identified and their data from these two sources were linked. For each patient, DAS28-CRP measurement and MD-HAQ physical function scores were extracted from BRASS, and CIRAS was calculated from Medicare claims for the period of 365 days prior to the DAS28-CRP measurement. Pearson correlation coefficient between CIRAS and DAS28-CRP as well as MD-HAQ physical function scores were calculated. Furthermore, we considered several additional pharmacy and medical claims-derived variables as predictors for DAS28-CRP in a multivariable linear regression model in order to assess improvement in the performance of the original CIRAS algorithm.

RESULTS

In total, 315 patients with enrollment in both BRASS and Medicare were included in this study. The majority (81%) of the cohort was female, and the mean age was 70 years. The correlation between CIRAS and DAS28-CRP was low (Pearson correlation coefficient = 0.07, P = 0.24). The correlation between the calculated CIRAS and MD-HAQ physical function scores was also found to be low (Pearson correlation coefficient = 0.08, P = 0.17). The linear regression model containing additional claims-derived variables yielded model R(2) of 0.23, suggesting limited ability of this model to explain variation in DAS28-CRP.

CONCLUSIONS

In a cohort of Medicare-enrolled patients with established RA, CIRAS showed low correlation with DAS28-CRP as well as MD-HAQ physical function scores. Claims-based algorithms for disease activity should be rigorously tested in distinct populations in order to establish their generalizability before widespread adoption.

Severity index for rheumatoid arthritis and its association with health care costs and biologic therapy use in Turkey

OBJECTIVE

This study aimed to apply the previously validated severity index for rheumatoid arthritis (SIFRA) to prevalent rheumatoid arthritis (RA) groups in Turkey and determine the effect of RA severity on health care costs and biologic use.

METHODS

This retrospective study used the Turkish national health insurance database MEDULA (June 1, 2009-December 31, 2011). Prevalent RA patients were required to be age 18 to 99, have two RA diagnoses at least 60 days apart and be continuously enrolled 1 year prior to (baseline period) and post (follow-up period) index date, which was the first RA claim during the identification period (June 1, 2010-December 31, 2010). SIFRA was calculated for the baseline period. Total health care costs and biologic use were examined for the follow-up period. The chi-square test was used to determine the association between SIFRA score terciles and outcomes. Generalized linear models were applied to determine health care costs while multivariate logistic regression determined the effect of SIFRA on outcome measures for biologic use.

RESULTS

A total of 1,920 patients were identified. The mean SIFRA score was 14.21, and 7.05 (49.57%) of the mean composed of clinical and functional status variables, followed by 6.32 (44.47%) for medications, 0.48 (3.40%) for radiology and laboratory findings, and 0.32 (2.25%) for extra-articular manifestation. There was a significant variation in scores across cities. After controlling for age, gender, region, and comorbidity index, patients in the high SIFRA tercile were 5.16 times more likely to be prescribed biologics (p<0.001, confidence interval [CI]: 3.46-7.69), and incurred annual health care costs that were €2,091 higher (p<0.001, CI: €1,557 - €2,625) than those in the low SIFRA score tercile.

CONCLUSION

RA severity varies throughout Turkey and is a significant determinant of health care costs and biologic therapy use. Therefore, future comparative effectiveness studies should include the severity measure in their analysis.