Population Variations in Rheumatoid Arthritis Treatment and Outcomes, Northern California, 1998-2009

Physician Prescribing Patterns and Risk of Future Long-Term Opioid Use Among Patients With Rheumatoid Arthritis: A Prospective Observational Cohort Study

OBJECTIVE

To identify the extent to which opioid prescribing rates for patients with rheumatoid arthritis (RA) vary in the US and to determine the implications of baseline opioid prescribing rates on the probability of future long-term opioid use.

METHODS

We identified patients with RA from physicians who contributed ≥10 patients within the first 12 months of participation in the Corrona RA Registry. The baseline opioid prescribing rate was calculated by dividing the number of patients with RA reporting opioid use during the first 12 months by the number of patients with RA providing data that year. To estimate odds ratios (ORs) for long-term opioid use, we used generalized linear mixed models.

RESULTS

During the follow-up period, long-term opioid use was reported by 7.0% (163 of 2,322) of patients of physicians with a very low rate of opioid prescribing (referent) compared to 6.8% (153 of 2,254) of patients of physicians with a low prescribing rate, 12.5% (294 of 2,352) of patients of physicians with a moderate prescribing rate, and 12.7% (307 of 2,409) of patients of physicians with a high prescribing rate. The OR for long-term opioid use after the baseline period was 1.16 (95% confidence interval [95% CI] 0.79-1.70) for patients of low-intensity prescribing physicians, 1.89 (95% CI 1.27-2.82) for patients of moderate-intensity prescribing physicians, and 2.01 (95% CI 1.43-2.83) for patients of high-intensity prescribing physicians, compared to very low-intensity prescribing physicians.

CONCLUSION

Rates of opioid prescriptions vary widely. Our findings indicate that baseline opioid prescribing rates are a strong predictor of whether a patient will become a long-term opioid user in the future, after controlling for patient characteristics.

Racial and ethnic differences in medication use among beneficiaries of social security disability insurance with rheumatoid arthritis

OBJECTIVE

To determine racial/ethnic differences in the use of conventional synthetic or biologic disease-modifying anti-rheumatic drugs (csDMARDs or bDMARDs, respectively) and long-term glucocorticoids (GC) or opioids among beneficiaries of the Social Security Disability Insurance (SSDI) with rheumatoid arthritis (RA) and <65 years old.

METHODS

Serial cross-sectional analyses of Centers for Medicare and Medicaid Services claims data (2007, 2011, and 2014) for individuals <65 years old with RA receiving SSDI Medicare and Medicaid, no longer working because they were considered disabled. Generalized estimating equation models were used to determine whether the proportion of patients who used csDMARD, bDMARD, long-term GC, and long-term opioids differed by race/ethnicity.

RESULTS

There were 12,931; 15,033; and 15,599 participants in 2007, 2011, and 2014, respectively. The overall use of csDMARD without bDMARD among beneficiaries of the SSDI were 31.1%, 30.3%, and 29.2%; 50.2%, 51.7%, and 53.8% used bDMARDs; 37.6%, 36.1%, and 34.4% used long-term GC; and 61.1%, 63.8%, and 63.7% used long-term opioids in years 2007, 2011, and 2014 respectively. The use of csDMARDs without bDMARDs was higher and the use of bDMARDs was lower among Blacks compared to Whites (adjusted absolute difference: +3.0%, +5.0%, and +3.3% for csDMARDs without bDMARDs and -4.6%, -5.7%, and -4.0% for bDMARDs in 2007, 2011, and 2014, respectively; all p<0.05). The use of bDMARDs was higher among Hispanics compared to Whites (adjusted absolute difference: +7.1%, +7.3%, and +7.5% in 2007, 2011, and 2014, respectively; all p<0.05). Long-term GC use was lower among Hispanics than among Whites only in year 2014 (absolute percentage point difference of -4.2%); no other difference in long-term GC use was identified. Whites were the patients with the highest use of long-term opioids (more than two third in each calendar year).

CONCLUSION

Racial and ethnic differences exists in regards to the treatment of RA among beneficiaries of the SSDI. These findings suggest that this already vulnerable population of patients with RA can also have a racial and ethnic disparity that can contribute to additional disease burden and that should be examined in order to inform future interventions or even inform future policy changes to the SSDI.

Retrospective Analysis of the Impact of Adalimumab Initiation on Corticosteroid Utilization and Medical Costs Among Biologic-Naïve Patients with Rheumatoid Arthritis

INTRODUCTION

Treatment guidelines recommend low-dose corticosteroids as short-term therapy among rheumatoid arthritis (RA) patients. However, it may be difficult to wean/eliminate steroids once initiated. Initiation of more effective therapies such as biologics may help to taper corticosteroid use. The objective was to examine the impact of adalimumab (ADA) initiation on steroid utilization and non-drug medical costs among patients with RA.

METHODS

A retrospective analysis was conducted among adult RA patients initiating ADA as the initial biologic in the MarketScan Database (2012-2016). Study outcomes included whether oral/injectable steroids were used, daily dose, dosage categories (< 5 and ≥ 5 mg/day), number of steroid injections, and non-drug medical costs. Outcomes were compared 6 months pre- and post-ADA initiation. Mixed effects logistic, classical linear, multinomial logistic models, and linear model with a log link and gamma distribution were used to adjust for patient demographic and health characteristics.

RESULTS

The sample included 7404 ADA initiators. Compared to pre-ADA initiation, in the post-initiation period there was a reduction in proportions of patients using oral steroids (from 71.80 to 62.56%) and injectable steroids (from 34.91 to 29.88%), average daily dose of oral steroids (from 3.30 to 2.62 mg/day), patients with dose ≥ 5 mg/day (from 21.76 to 16.34%), number of injections (from 0.64 to 0.53), and non-drug medical costs (from $5356.30 to $5146.84) (P < 0.01). The multivariate analysis produced similar patterns. For example, post-ADA initiation, patients were less likely to use oral steroids [odds ratio (OR) 0.51; 95% confidence interval (CI) 0.47-0.56]; coefficient estimate for daily dose reduction was - 0.68 (95% CI - 0.81 to - 0.56); ratio estimate for medical costs was 0.91 (95% CI 0.86-0.97).

CONCLUSIONS

Among patients with RA, following ADA initiation, there is a reduction in steroid utilization and dosage, and non-drug medical costs. Prospective studies should be conducted to confirm this relationship in the future.

Patterns of glucocorticoid prescribing and provider-level variation in a commercially insured incident rheumatoid arthritis population: A retrospective cohort study

OBJECTIVE

Glucocorticoids are common in RA management despite an unfavorable, exposure-dependent risk profile impacted by patient and provider-level factors. Existing work describing glucocorticoid use in RA is not generalizable and does not adequately examine provider factors. We aim to describe how providers prescribe glucocorticoids to commercially insured, newly diagnosed RA patients in the United States.

METHODS

This was a retrospective cohort study which used the national Optum© administrative database. We identified 9221 adults ages 18-65 with RA diagnosed 2010-2014. We assessed glucocorticoid dispensing 3 months pre-diagnosis through 12months post-diagnosis ("study period"), cumulatively stratified by calendar quarter and prescriber specialty (rheumatologist, primary care, other). We examined prescribing variation among individual rheumatologists by dividing quarterly distribution of per-patient dose and days' supply into quartiles.

RESULTS

6717 (72.8%) patients filled ≥1 glucocorticoid prescription during the study period. 2890 (31.3%) patients received ≥3 months' supply, with median (IQR) daily dose 10 (6.6) mg/day and days' supply 189 (143) days. 52.6% of patients received glucocorticoids 1-3 months post-diagnosis; 29.2% received glucocorticoids 10-12 months post-diagnosis. Among glucocorticoid users post-diagnosis, quarterly median daily dose and days' supply were consistently ≥10 mg/day and ≥30 days, respectively. Rheumatologists prescribed most glucocorticoids, with median per-quarter daily dose and days' supply 10 mg/day and 43-60 days. Individual rheumatologists' prescribing varied widely across all quarters.

CONCLUSION

Among commercially insured incident RA patients, receipt of ≥10 mg/day prednisone equivalent for months is common, typically prescribed by rheumatologists, and persists a year post-diagnosis in 29.2% of patients. Glucocorticoid prescribing varies widely across rheumatologists. Further work is warranted to identify provider factors explaining variation in glucocorticoid prescribing, and assess how these affect health outcomes.

Use of machine learning techniques in the development and refinement of a predictive model for early diagnosis of ankylosing spondylitis

OBJECTIVE

To develop a predictive mathematical model for the early identification of ankylosing spondylitis (AS) based on the medical and pharmacy claims history of patients with and without AS.

METHODS

This retrospective study used claims data from Truven databases from January 2006 to September 2015 (Segment 1) and October 2015 to February 2018 (Segment 2). Machine learning identified features differentiating patients with AS from matched controls; selected features were used as inputs in developing Model A/B to identify patients likely to have AS. Model A/B was trained and developed in Segment 1, and patients predicted to have AS in Segment 1 were followed up in Segment 2 to evaluate the predictive capability of Model A/B.

RESULTS

Of 228,471 patients in Segment 1 without any history of AS, Model A/B predicted 1923 patients to have AS. Ultimately, 1242 patients received an AS diagnosis in Segment 2; 120 of these were correctly predicted by Model A/B, yielding a positive predictive value (PPV) of 6.24%. The diagnostic accuracy of Model A/B compared favorably with that of a clinical model (PPV, 1.29%) that predicted AS based on spondyloarthritis features described in the Assessment of SpondyloArthritis international Society classification criteria. A simplified linear regression model created to test the operability of Model A/B yielded a lower PPV (2.55%).

CONCLUSIONS

Model A/B performed better than a clinically based model in predicting a diagnosis of AS among patients in a large claims database; its use may contribute to early recognition of AS and a timely diagnosis.