Objective. To investigate the incidence and predictors of dyspnea on exertion among subjects with rheumatoid arthritis (RA).Methods. We investigated dyspnea on exertion using a prospective cohort, the Brigham RA Sequential Study (BRASS). Clinically significant dyspnea on exertion was defined as a score of ≥ 3 (unable to ambulate without breathlessness or worse) on the validated Medical Research Council (MRC) scale (range 0-5). We analyzed subjects with MRC score < 3 at BRASS baseline and at ≥ 1 year of follow-up. The MRC scale was administered annually. We determined the incidence rate (IR) of dyspnea on exertion. We used Cox regression to estimate the hazard ration (HR) for dyspnea on exertion occurring one year after potential predictors were assessed.Results. We analyzed 829 subjects with RA and no clinically significant dyspnea on exertion during a mean follow-up period of 3.0 years (SD 1.9). At baseline, mean age was 55.7 years (SD 13.6), 82.4% of subjects were female, and median RA duration was 8 years. During follow-up, 112 subjects (13.5%) developed incident dyspnea on exertion during 2476 person-years of follow-up (incidence rate 45.2 per 1000 person-years). Independent predictors of incident dyspnea on exertion were older age (HR 1.03 per year, 95% CI, 1.01-1.04), female sex (HR 2.22, 95% CI, 1.14-4.29), mild dyspnea (HR 2.62, 95% CI, 1.60-4.28), and worsened Multi-Dimensional Health Assessment Questionnaire score (HR 2.36 per unit, 95% CI, 1.54-3.60). Methotrexate use, RA disease activity, and seropositivity were not associated with incident dyspnea on exertion after accounting for other dyspnea risk factors.Conclusion. Dyspnea on exertion occurred commonly in patients with RA. Older women with impaired physical function were especially vulnerable to developing dyspnea on exertion.

BackgroundPrior studies have demonstrated challenges in developing and validating claims-based algorithms that accurately predict RA disease activity.1 2 The ability to adjust for and predict RA disease activity would be a powerful epidemiological tool for studies that lack direct disease activity measures such as the DAS28.ObjectivesWe used machine-learning methods to incorporate claims and electronic medical record (EMR) data to develop models to predict DAS28 (CRP) as a continuous measure, and to distinguish moderate-to-high disease activity from low activity/remission.MethodsWe identified 300 adults (≥18 years of age) with RA enrolled in a single academic centre cohort with ≥1 year of linked Medicare insurance claims preceding a DAS28 (CRP) measurement between 2006 and 2010. Of these, 95 had Medicare Part D pharmacy data. From claims we included demographics, co-morbidities, joint replacement surgery, physical therapy visits, numbers of RA-related codes, laboratory values and imaging studies, and healthcare utilisation. For those with Part D pharmacy data we included medications (steroids, analgesics, DMARDs) and switches between drugs. From the EMRs we obtained smoking status, BMI, blood pressure, medication use, laboratory values for seropositivity (RF or anti-cyclic citrullinated peptide antibodies), haematocrit, ESR and CRP. We constructed models with claims only, claims with medications and claims with EMR data. We examined these models with DAS28 (CRP) as a continuous measure and as a binary outcome (moderate/high activity vs low activity/remission). We used adaptive least absolute shrinkage and selection operator (LASSO), which avoids model overfitting by penalising large coefficients and selects a subset of variables by shrinking some coefficients to zero. We used adjusted R2 to compare continuous model fit and C-statistics to compare binary models.ResultsIn models that included DAS28 as a continuous measure, using claims alone explained 11% of the DAS28 variability. Adding medications and EMR data to claims improved the adjusted R2 by 6% (table 1). In models that included DAS28 as a binary outcome (moderate/high activity vs low activity/remission), our claims-only model yielded a C-statistic of 0.68, which increased to 0.79 after inclusion of medications and EMR data.Abstract OP0010 – Table 1Model Fit Statistics for Continuous DAS28 (CRP) (Adjusted R2) and Binary Categories (Moderate/High vs Low/Remission; C-Statistic)*Model 1: claims onlyModel 2: claims+Medicare medicationsModel 3: claims+Medicare and EMR medicationsModel 4: EMR data**Model 5: claims+Medicare medications +EMR data** Adjusted R20.110.120.140.160.17C-statistic0.680.740.770.760.79*n=300 except for Model 2 (n=95)**EMR data includes medications, laboratory tests, BMI, blood pressure and smoking statusConclusionsIncorporating medications, EMR data and laboratory values into a claims-based index did not significantly improve the ability to predict DAS28 scores as a continuous measure. However, models that include claims, medications and EMR...