Op-brhe120233 411..412

RA is common and costly for individuals, health services and societies. Current guidelines for the management of RA highlight the importance of early administration of DMARDs and the titration of therapy against disease activity to try and achieve remission. Reasonable evidence exists to show that early institution of DMARDs reduces the rate of progression of radiological damage in the long term [1]. In this issue of Rheumatology, Feldman et al. [2] report Canadian data suggesting that early consultation with a rheumatologist reduces the need for subsequent orthopaedic surgery. This finding is consistent with previous data that the early institution of DMARDs is associated with a lower risk of joint surgery [3] and adds to the concept that early intervention improves long-term outcomes. Patients studied by Feldman et al. [2] represent a very specific group: those in whom a rheumatologist had made a new diagnosis of RA and in whom the physician who referred the patient to rheumatology had also made a diagnosis of RA. The question in the title of the article is straightforward: ‘Early consultation with a rheumatologist for RA: does it reduce subsequent use of orthopaedic surgery?’ [2]. However, when interpreting the results it is important to recognize that the delay in consultation with a rheumatologist being studied represents only a proportion of the total delay from the onset of the patients’ RA symptoms. The journey that patients with RA follow before rheumatology assessment has ample potential for delays at multiple levels. Three important levels are delay on the part of the patient in seeking advice from the general practitioner (GP) (or other health care professional), delay on the part of the initial health care professional in making a referral to a rheumatologist and delay on the part of the rheumatologist in seeing the patient after the referral has been made [4]. Feldman et al. [2] specifically address the relationship between the risk of joint surgery and the interval from RA diagnosis by the referring physician to assessment by the rheumatologist (who presumably started DMARDs). The interval of delay studied in this article is clinically relevant and has important practical implications. The data provide strong evidence that once a physician has determined that a patient has RA, the quicker that patient is seen by a rheumatologist, the better the long-term outcome. However, why it took such a long time for patients in this study to be referred to a rheumatologist is unclear. In the RA patients who saw a rheumatologist late (49% of the 1051 RA patients), the median delay was 2 years from the time the initial physician diagnosed RA. Although this delay within the Quebec health care system is longer than reported in a recent European study, the delays at the level of the initial health care provider and rheumatologist were still quite long in some European centres [5]. The data thus emphasize the importance of the need to encourage rapid rheumatology assessment once RA is strongly suspected—something that will involve raising awareness of the importance of early DMARD treatment among those physicians with whom patients are likely to have initial contact and well as streamlining referral and assessment pathways. A rather overlooked aspect of this study relates to those 2839 patients in whom the rheumatologist made a diagnosis of RA but in whom the referring physician did not—and in whom the impact of delay was thus not assessed. Importantly, these patients comprised the largest proportion (73%) of patients diagnosed with RA by a rheumatologist in Quebec in 1995. Strategies to facilitate rapid referral and assessment once the referring physician has decided that their patient has RA will not help the situation for this majority in whom the referring physician has a lower index of suspicion. It is not surprising that in most cases the referring physician did not make a diagnosis of RA. The initial clinically apparent phases of RA often include a period of symptoms without joint swelling, followed by the development of an unclassified arthritis that only later evolves into a disease classifiable as RA [6]. The nearer the onset of symptoms that treatment is commenced, the better the outcomes for RA patients [7, 8]. The ability to accurately predict the future development of RA in the large numbers of patients for whom there is initial clinical uncertainty thus represents an important goal. A predictive algorithm has been validated for use in patients with unclassified arthritis [9] and data to allow the development of an equivalent algorithm in patients with ACPA-positive arthralgia are available [10]. However, these algorithms have their origins in hospitalbased populations. There is now an urgent need to develop algorithms for the prediction of RA for use in primary care that have their origins in primary care and that can guide investigation and referral by GPs. Once RA is strongly suspected, there is widespread consensus on the need for rapid referral and assessment, and the data from Feldman et al. [2] support this. However, even in the early consulters a significant proportion required orthopaedic intervention. The goalposts are now shifting and attention is focused on how patients can be captured at an even earlier stage in their journey. The fact that musculoskeletal symptoms are common and that