Quality care in seniors with new-onset rheumatoid arthritis: A Canadian perspective


  • The opinions, results, and conclusions are those of the authors and are independent from the funding sources. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred.



To estimate the percentage of seniors with rheumatoid arthritis (RA) receiving disease-modifying antirheumatic drugs (DMARDs) within the first year of diagnosis.


We assembled an incident RA cohort from Ontario physician billing data for 1997–2006. We used a standard algorithm to identify 24,942 seniors with RA based on ≥2 billing codes ≥60 days apart but within 5 years. Drug exposures were obtained from pharmacy claims data. We followed subjects for 1 year, assessing if they had been exposed (defined as ≥1 prescription) to 1 or more DMARDs within the first year of RA diagnosis. We assessed secular trends and differences for subjects who had received rheumatology care (defined as ≥1 rheumatology encounter) versus those who had not.


In total, only 39% of the 24,942 seniors with new-onset RA identified over 1997–2006 were exposed to DMARD therapy within 1 year of diagnosis. This increased from 30% in 1997 to 53% in 2006. Patients whose care involved a rheumatologist were more likely to be exposed to DMARDs than those who had no rheumatology care. In 2006, 67% of subjects receiving rheumatology care were exposed to DMARDs versus 21% of those with no rheumatology care.


Improvements in RA care have occurred, but more efforts are needed. Subjects receiving rheumatology care are much more likely to receive DMARDs as compared to those with no rheumatology care. This emphasizes the key role of rheumatologists.


It is well known that for patients with rheumatoid arthritis (RA), delays in initiating therapy are associated with negative health outcomes (1). Existing guidelines recommend early and aggressive treatment (2). The Ontario Biologics Research Initiative (OBRI) represents a collaboration of rheumatologists, patients, decision makers, and health care researchers. The OBRI focuses on the real-world effectiveness and safety of drug therapies for RA; part of the work of the OBRI to date has been the evaluation of existing practice patterns in order to identify areas in need of improvement. Much of the OBRI analyses have made use of Ontario's administrative health care databases, which capture the health service encounters of all Ontario residents.

The prevalence of RA increases with age, and although early reports suggested that seniors with RA may have a more benign course than their younger counterparts, more recent studies report that the outcome in seniors with RA may be no better, or even worse, than those who are younger (3–5). Therefore, it is important that seniors be offered prompt and appropriate care, including access to rheumatologists and disease-modifying antirheumatic drugs (DMARDs).

The main objective of the current study was to estimate the percentage of seniors with RA exposed to DMARDs within the first year of diagnosis.


We assembled an incident RA cohort age ≥65 years using the administrative health databases of the Ontario Health Insurance Plan (OHIP) for January 1997 to March 2007. In the province of Ontario, all 13 million residents are covered by universal public health insurance. The OHIP physician billing database contains information regarding physician services, including the service provided and the patient's diagnosis, coded using International Classification of Diseases, Ninth Revision (ICD-9) codes. Medication exposures were determined using the pharmacy claims database of the Ontario Drug Benefit Program, which covers residents who are ages ≥66 years or receiving social assistance. We used a previously published algorithm to identify RA patients within the OHIP billing data, based on at least 2 billing code diagnoses of RA (ICD-9 code 714) more than 60 days apart but within 5 years. We further required patients to have at least one prescription for a glucocorticoid, DMARD, or biologic response modifier during the entire study period (6, 7). In an attempt to exclude potential prevalent cases of RA migrating into the province, individuals who did not have contact with Ontario's health care system within the 5 years prior to diagnosis were excluded.

We followed subjects with incident RA (by our definition) for a period of 1 year, assessing whether subjects received at least 1 DMARD (methotrexate, hydroxychloroquine, sulfasalazine, azathioprine, leflunomide, or cyclophosphamide) prescription within the first year of RA diagnosis. We assessed secular trends and differences for subjects who had received rheumatology care (defined as 1 or more rheumatologist encounter within the billing data) versus those who had not.

We performed logistic regression analyses to determine whether demographics (age, sex, urban versus rural residence, socioeconomic status [SES], calendar year, and clinical factors [comorbidity, proxies for disease severity]) were associated with DMARD use, after accounting for whether patients saw a rheumatologist. SES was defined as the patient's neighborhood income quintile from the Statistics Canada Census. The measure of comorbidity was the Deyo-Charlson comorbidity index, derived from hospital discharge abstracts (8, 9). For patients who did not receive inpatient care during the period of analysis, a variable for “missing” was included as a separate Charlson category. As an additional measure of comorbidity, we counted the number of prescription drugs each patient received in the preceding year (10).

Proxies for disease severity included prescriptions for nonsteroidal antiinflammatory drugs and systemic corticosteroids in the first year of RA diagnosis, as well as the presence of extraarticular features of RA at cohort entry.

A sensitivity analysis was performed to explore the potential effects of disease severity on DMARD use. A similar regression model was constructed that included frequency of physician visits in the first year (measured as a continuous variable) as a proxy for disease severity, replacing the binary variable of contact with a rheumatologist.

Finally, we performed secondary analyses to assess whether DMARD use increased significantly after the emergence of a national consensus statement. In 2004, the Canadian Rheumatology Association (CRA) convened an expert panel regarding optimal therapy in early RA. This led to the development and dissemination of a consensus statement, which reinforced the importance of early RA treatment with DMARDs (11). Secondary analyses examined whether DMARD prescription in early RA increased in the period following as compared to before the development and dissemination of this consensus statement.


We identified 25,141 seniors with new-onset RA (according to our definition) over 1997–2006. We excluded 199 who had no contact with the health care system in the 5 years prior to their RA diagnosis, for a total study sample of 24,942. The majority of subjects (17,027 [68.3%]) were women, and the mean ± SD age at cohort entry was 74.6 ± 6.01 years. A significant minority (4,038 [16.2%]) resided in rural areas. The Charlson comorbidity index values were 1 or greater in 6.5%, with the majority of subjects having no hospitalization during the period of interest.

Table 1 compares those who received care from a rheumatologist versus those who did not, in terms of demographics and clinical features. Seniors who saw a rheumatologist were more likely to live in urban (versus rural) areas, and were more likely to have ≥1 extraarticular feature of RA. In the 24,942 seniors with new-onset RA identified over 1997–2006, 39% (n = 9,737) were exposed to DMARD therapy within 1 year of diagnosis. This increased from 30% in 1997 to 53% in 2006 (Figure 1).

Table 1. Demographics and clinical features of seniors for those referred versus those not referred*
 Overall (n = 24,942)Seen by a rheumatologist in the first year after cohort entry
No (n = 10,793)Yes (n = 14,149)
  • *

    Values are the number (percentage) unless otherwise indicated. RA = rheumatoid arthritis.

Women17,027 (68.3)7,262 (67.3)9,765 (69)
Age, mean ± SD years74.6 ± 675 ± 6.2874.2 ± 5.76
Income quintile   
 Missing82 (0.3)37 (0.3)45 (0.3)
 14,911 (19.7)2,303 (21.3)2,608 (18.4)
 25,234 (21)2,292 (21.2)2,942 (20.8)
 35,134 (20.6)2,243 (20.8)2,891 (20.4)
 44,710 (18.9)2,010 (18.6)2,700 (19.1)
 54,871 (19.5)1,908 (17.7)2,963 (20.9)
Rural4,038 (16.2)2,125 (19.7)1,913 (13.5)
Calendar year of entry   
 19972,758 (11.1)1,466 (13.6)1,292 (9.1)
 19982,879 (11.5)1,490 (13.8)1,389 (9.8)
 19992,748 (11)1,431 (13.3)1,317 (9.3)
 20002,662 (10.7)1,222 (11.3)1,440 (10.2)
 20012,437 (9.8)1,030 (9.5)1,407 (9.9)
 20022,344 (9.4)951 (8.8)1,393 (9.8)
 20032,299 (9.2)883 (8.2)1,416 (10)
 20042,270 (9.1)823 (7.6)1,447 (10.2)
 20052,328 (9.3)818 (7.6)1,510 (10.7)
 20062,217 (8.9)679 (6.3)1,538 (10.9)
Charlson comorbidity index   
 02,529 (10.1)1,130 (10.5)1,399 (9.9)
 1965 (3.9)398 (3.7)567 (4)
 2394 (1.6)181 (1.7)213 (1.5)
 3+253 (1)114 (1.1)139 (1)
 Missing20,801 (83.4)8,970 (83.1)11,831 (83.6)
No. of drugs in the year prior to entry, mean ± SD9.2 ± 5.79 ± 5.79.4 ± 5.6
Presence of extraarticular features of RA4,088 (16.4)1,544 (14.3)2,544 (18)
No. of physician visits in the year prior to entry, mean ± SD13.8 ± 9.5112.9 ± 9.3314.5 ± 9.59
Figure 1.

Percentages of patients with new-onset rheumatoid arthritis who received disease-modifying antirheumatic drugs (DMARDs; methotrexate, hydroxychloroquine, sulfasalazine, azathioprine, leflunomide, cyclophosphamide, and combinations) within the first year of diagnosis and who were seen by a rheumatologist within the first year of diagnosis.

Overall, 56% of patients saw a rheumatologist within the first year of RA diagnosis. This percentage increased over time, from 47% in 1997 to 69% in 2006. In 1997, the percentage of early RA patients prescribed a DMARD was 51% among those who saw a rheumatologist compared to 11% of those who had no such contact. In 2006, the same figures were 67% and 21%, respectively (Figure 1). Therefore, much of the increase in DMARD exposure from 1997–2006 was related to increasing exposure to rheumatologists.

In half of the cases, the initial DMARD prescribed was hydroxychloroquine (51.6%). The next most common first- use DMARD was methotrexate (36.3%), followed by sulfasalazine (6.3%). Combination therapy was prescribed in 5.8% of cases, with the most common combination therapy being hydroxychloroquine and methotrexate. In the other subjects, the first medication exposures were varied, but included (in descending order) azathioprine, leflunomide, and cyclophosphamide. Five individuals were identified as receiving a biologic response-modifying drug prior to their first DMARD prescription.

The results of the logistic regression analyses (Table 2) showed that seniors receiving care from a rheumatologist were 10 times more likely to receive a DMARD in the first year of RA diagnosis, as compared to those without rheumatology care. There was also less DMARD use among seniors with increasing age. As expected, DMARD use was correlated with steroid use.

Table 2. Odds of a senior receiving a disease-modifying antirheumatic drug within the first year of RA diagnosis from the multivariable regression for the effects of demographic and clinical factors*
 OR (95% CI)
  • *

    RA = rheumatoid arthritis; OR = odds ratio; 95% CI = 95% confidence interval; NSAID = nonsteroidal antiinflammatory drug.

  • We used information from the Statistics Canada Census on neighborhood income levels for the dissemination area in which each subject resided.

  • Derived from hospital encounters, based on administrative database records.

Age, years0.97 (0.97–0.98)
Women0.95 (0.89–1.02)
Income quintile (ref. = 1) 
 21.00 (0.91–1.09)
 31.07 (0.98–1.18)
 40.96 (0.87–1.05)
 50.98 (0.89–1.07)
Rural residence1.18 (1.09–1.28)
Calendar year of entry (ref. = 1997) 
 19981.08 (0.95–1.23)
 19991.07 (0.94–1.22)
 20001.11 (0.98–1.27)
 20011.21 (1.06–1.38)
 20021.22 (1.07–1.39)
 20031.50 (1.31–1.71)
 20041.80 (1.58–2.05)
 20052.00 (1.75–2.28)
 20062.11 (1.85–2.41)
Presence of extraarticular RA features1.11 (0.93–1.33)
Charlson comorbidity index (ref. = 0) 
 11.09 (0.85–1.41)
 21.00 (0.73–1.36)
 3+0.99 (0.89–1.09)
 Missing0.91 (0.84–0.99)
No. of drugs in the year prior to index0.99 (0.99–1.00)
NSAID use in year 11.07 (1.00–1.14)
Systemic steroid use in year 11.21 (1.14–1.28)
Rheumatologist visit9.99 (9.32–10.70)

Interestingly, prior to controlling for whether or not a patient had encountered a rheumatologist, the odds ratio (OR) for rural residence suggested less DMARD prescriptions for rural residents (0.92, 95% confidence interval [95% CI] 0.86–0.99). However, after adjusting for the lower rate of rheumatology encounters among subjects from rural areas, rural residents were relatively more likely to receive a DMARD prescription within the first year of RA diagnosis (OR 1.18, 95% CI 1.09–1.28).

In terms of the subanalyses for time periods before and after the CRA guidelines, we did indeed observe a significant increase in DMARD use over 2004–2006 versus 2001–2003. The percentage of early RA patients exposed to DMARDs over 2001–2003 was 40% (95% CI 39–41%) compared to 50% (95% CI 49–51%) in 2004–2006. This substantial increase is statistically significant (OR 9%, 95% CI 10–11%). However, when we looked at early periods, the percentage of early RA patients exposed to DMARDs in 1998–2000 was 32% (95% CI 31–33%), indicating an increase during 1998–2000 versus 2001–2003, even before publication of the CRA consensus statement.


Earlier population-based assessments have demonstrated that many Canadians are not provided optimal RA therapy (12, 13); this seems especially true for older persons with RA. Our work emphasizes that even in the setting of universal health insurance, suboptimal treatment of RA occurs. Our results support an earlier study by Shipton et al, which suggested that the problem is at least in part due to poor access to specialty care (14). The problems do not lie only in the failure to establish therapy. Increasing age is associated with a greater risk of methotrexate discontinuation in RA (15).

What factors may have contributed to improved rheumatology care for seniors with RA in Ontario over our study interval? Some have indicated that clinical guidelines may indeed have a role (16). Our subanalyses suggest that appropriate drug therapy for older individuals with RA has improved since the CRA consensus statement was established. However, given the increase in DMARD use during 1998–2000 versus 2001–2003, awareness of the need for early therapy was apparently increasing even before the CRA consensus statement was published.

Our data also suggested an increasing recognition of the need for prompt referral of patients with early RA to a rheumatologist. This occurred in the absence of any increase in the per capita ratio of rheumatologists in Ontario. As of 2006, there were 152 rheumatologists in Ontario (representing 126 full-time equivalents as based on the American College of Rheumatology standard of at least 32 hours per week of direct clinical care [17]). The overall provincial per capita provision in 2006 was 1.20 rheumatologists per 100,000 population (1.00 full-time equivalents per 100,000), which is actually 25% less than the figures in 1997 (18). In recent years, there have been other rheumatology manpower trends in Ontario, including more female rheumatologists (25% in 1992 versus 37% in 2006), and a small but steady increase over time in average number of years of experience (14 in 1992 versus 18 in 2006) and the number of clinic half days worked per week (8.89 in 1997 versus 9.58 in 2006).

Despite improvements in care, our findings suggest that DMARD use in older RA patients remains suboptimal in Ontario. This is particularly true in patients who do not have access to a rheumatologist. Among these patients, fewer than 20% received a DMARD in the first year of diagnosis. We interpret this as emphasizing the important role rheumatologists play in RA care.

Administrative data are efficient tools for evaluating quality of care. However, they have several important limitations. Our analyses were limited to persons ages >65 years. In Ontario, analyses of younger individuals are difficult, as prescription drug insurance is continuous for seniors only. Also, administrative data preclude clinical confirmation of diagnoses and disease onset. Future work will include assessment of the validity of RA diagnosis coding in Ontario administrative data. It is possible that some of the RA patients we studied were prevalent rather than incident cases. However, the rates of DMARD use we observed are a problem regardless of whether cases were incident or prevalent, since recent data show that most patients with longstanding RA indeed require continuous DMARD treatment (19). We note that the 5 individuals who received a biologic drug prior to their DMARD may be indicative of prevalent cases misclassified as new onset according to our administrative definition. Alternatively, these individuals may have received the drug for another indication (e.g., inflammatory bowel disease or psoriasis), or they could represent rare severe cases of RA and/or special circumstances, in which early access to biologics was granted on an exceptional basis.

In conclusion, some improvements in RA care have occurred over time, but more efforts are needed, especially for vulnerable populations like seniors. Future OBRI research will further delineate practice patterns, the influence of funding, and the real-world effectiveness and safety of antirheumatic therapies.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Bernatsky had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Widdifield, Bernatsky, Paterson, Thorne, Cividino, Pope, Bombardier.

Acquisition of data. Gunraj.

Analysis and interpretation of data. Widdifield, Bernatsky, Paterson, Thorne, Cividino, Pope, Gunraj, Bombardier.


The authors would like to acknowledge contributions from members of the OBRI: Catherine Hofstetter, Anne Lyddiatt, and Annette Wilkins.