Objective The first 3 months after symptom onset represent an important therapeutic window for rheumatoid arthritis (RA). This study investigates the extent and causes of delay in assessment of patients with RA in eight European countries.
Method Data on the following levels of delay were collected from 10 centres (Berlin, Birmingham, Heraklion, Lund, Prague, Stockholm, Umeå, Vienna, Warsaw and Zurich): (1) from onset of RA symptoms to request to see healthcare professional (HCP); (2) from request to see HCP to assessment by that HCP; (3) from initial assessment by HCP to referral to rheumatologist; and (4) from referral to rheumatologist to assessment by that rheumatologist.
Results Data were collected from 482 patients with RA. The median delay across the 10 centres from symptom onset to assessment by the rheumatologist was 24 weeks, with the percentage of patients seen within 12 weeks of symptom onset ranging from 8% to 42%. There were important differences in the levels underlying the total delays at individual centres.
Conclusions This research highlights the contribution of patients, professionals and health systems to treatment delay for patients with RA in Europe. Although some centres have strengths in minimising certain types of delay, interventions are required in all centres to ensure timely treatment for patients.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Early treatment of rheumatoid arthritis (RA) is associated with improved outcomes, with the first 3 months after symptom onset representing a therapeutic window.1,–,5 Disease-modifying antirheumatic drugs (DMARDs) are typically initiated by rheumatologists. Delays in assessment by rheumatologists reflect a composite of delays on the part of: (1) the patient seeking medical advice; (2) the initial healthcare professional (HCP; typically a general practitioner) seeing the patient once an appointment has been sought; (3) the initial HCP referring to a rheumatologist; and (4) the rheumatologist seeing the patient following a referral.
In Birmingham, UK, the median delay from symptom onset to assessment by the rheumatologist was 23 weeks, with the largest component being delay on the part of the patient in seeking medical advice (median 12 weeks).6 Similar data were obtained from an independent UK study.7 In the UK, a lack of awareness of RA and its presenting symptoms explained much of the delay in seeking help.8 9 The aim of this study was to quantify delays in assessment of patients with RA across several European countries and to identify whether the principal reasons for delay varied between countries.
Data were collected from consecutively presenting patients fulfilling 1987 American College of Rheumatology classification criteria for RA10 at initial assessment by a rheumatologist in 2009/10. Participating centres were from Germany, the UK, Greece, Sweden, the Czech Republic, Austria, Poland and Switzerland. Details of the clinics from which the patients were recruited and the strategies in place to facilitate rapid assessment at each centre are shown in table 1 in the online supplement. Data were collected on: (1) age; (2) gender; (3) time from symptom onset to request to see a HCP; (4) time from request to see a HCP to assessment by that HCP; (5) specialty of initial HCP of contact; (6) time from initial assessment by a HCP to referral to a rheumatologist; (7) time from referral to a rheumatologist to assessment by the rheumatologist. Data for (3), (4) and (6) were obtained by the rheumatologist from the patient's history, as in our previous study.6 Data for (7) were obtained from the dates of the referral letter and clinic appointment.
Research ethics committee approval was obtained in the Czech Republic. In the other countries, regulatory authorities determined that the collection of these anonymous data constituted a service evaluation for which approval from a research ethics committee was not necessary.
Comparisons across the same levels of delay at different centres used the Kruskal–Wallis test. Dunn's test (which corrects for multiple comparisons) was then used to compare each level of delay between pairs of centres. The different levels of delay within each centre were compared using the Friedman test. This only includes cases with values for all four delays. In some instances a value was not available. For example, when a patient went directly to a rheumatologist, the only relevant delays were the time from symptom onset to a request to see a HCP and the time from referral (in this case self-referral) to assessment by a rheumatologist; such cases were analysed separately.
Patient details are shown in table 1. There were no significant differences in the age (p=0.089, Kruskal–Wallis test) or gender (p=0.57, χ2 test) of patients at the 10 units. In all centres except Heraklion, the most frequent initial HCP of contact was a primary care physician although, in all centres, an alternative was occasionally consulted.
Comparisons of delays between centres
The median delay from symptom onset to assessment by a rheumatologist in all 482 patients was 24 weeks (range between centres 16–38 weeks; p<0.001, Kruskal–Wallis test; table 1). Looking at all patients together, there were no significant correlations between age and delays (1), (2), (3) or (4) or total delay (p=0.35, p=0.66, p=0.86, p=0.11, p=0.39, respectively; Spearman test) or differences between males and females in terms of delays (1), (2), (3) or (4) or total delay (p=0.10, p=0.11, p=0.64, p=0.68, p=0.053, respectively; Mann–Whitney test). Across centres, between 8% and 42% of patients were seen within 12 weeks of symptom onset (table 1). Differences between centres across level of delays are shown in table 2 (data on delay from request to see initial HCP to assessment by that HCP are not shown as this was very short for all centres except Heraklion). The shortest patient delays were in Berlin and Vienna (median 2 weeks). Patient delay was significantly shorter in Berlin than in all other centres except Stockholm, Vienna and Warsaw and was significantly longer in Birmingham and Heraklion than in several other centres (table 2, delay 1). Delays by the initial HCP in referring to a rheumatologist were shortest in Birmingham and Stockholm (median 2 weeks); these were significantly shorter than at several centres (table 2, delay 3). The shortest delay in rheumatologists seeing patients following referral was in Vienna (median 1 week) and Zurich (median 2 weeks); again these delays were significantly shorter than in several other centres (table 2, delay 4).
Comparisons of individual levels of delay within each centre
The relative contributions of individual levels of delay to total delay varied between centres (table 1), and comparisons within each centre are shown in table 3. In Birmingham and Heraklion, patient delay contributed most to overall delay and was significantly longer than initial HCP delay in referring and rheumatologist delay in assessing following referral (table 3). In all other centres apart from Berlin there was no significant difference between patient delay and initial HCP delay in referring to a rheumatologist (table 3). In Berlin the initial HCP delay was significantly longer than the patient delay (table 3). In Lund and Zurich, both these delays were significantly greater than the rheumatologist's delay in seeing the patient following referral. In Berlin, Vienna and Warsaw, patient delay was relatively short, other components accounting for the majority of the overall delay. For centres where patients consulted rheumatologists directly, patient and rheumatologist delays were compared specifically for these patients; no additional significant differences were found that had not been previously identified.
Across Europe, delays in the assessment of patients with RA by rheumatologists are unacceptably long. As highlighted in a recent report from the UK National Audit Office, tackling this should represent a priority for strategies to improve patient outcomes.11 Economic modelling within this report suggested that increasing the proportion of patients treated with DMARDs within 3 months of symptom onset from the current estimate of 10–20% could result in productivity gains for the UK economy of £31 million over 5 years. Furthermore, for the individual, reduced delay will have a major prognostic impact.2
To reduce overall delay, a detailed understanding of its components is required. The level of delay contributing the most to overall delay varied between centres. At some (Birmingham, Heraklion), patient delay was a key component while, in others (Berlin, Vienna), patients presented very quickly. Interestingly, a recent study from Leiden reported rapid presentation in Dutch patients.2 In most centres, delay on the part of the initial HCP in referring to a rheumatologist was an important contributor to overall delay; in seven of the 10 centres the median value for this delay was at least 8 weeks. Our findings highlight the type of delay that should be targeted in each of these centres.
Reducing patient and referral delays requires a detailed understanding of the underlying reasons. Studies such as those carried out in the UK8 9 and the USA12 to look at reasons for patient and HCP delay should be undertaken in other countries to identify local issues. Interventions to reduce patient delay have been developed for other chronic illnesses, although the success of these has been varied. A systematic review of interventions to reduce patient delay in cancer reported modest increases in cancer awareness.13 However, this was short-lived and there was only limited evidence that these interventions caused people to present earlier.13 In some centres the initial HCP of contact was frequently an orthopaedic surgeon. In addition to promoting rapid help-seeking behaviour, interventions in the populations served by these centres should encourage help-seeking from an appropriate source.
Our study has a number of limitations. The number of patients studied at each centre was relatively small and, except for Sweden, only one centre was included from each country. Also, we relied on patient memory to determine symptom onset, when they sought healthcare advice and when they were initially seen by a HCP. However, evidence suggests that this is an acceptable approach to data collection, with delays reported by patients agreeing with delays recorded in hospital notes at the aggregate level.14
This research highlights the contribution of patient behaviour, professional behaviour and health systems to assessment delay for people with RA in selected European rheumatology centres. Some centres have strengths in minimising delay at particular levels. However, in all cases the median time from symptom onset to assessment by a rheumatologist was ≥4 months. Intervention is required to ensure that more patients are assessed within the window during which they are most likely to respond to treatment—that is, the first 3 months.
The authors are grateful to Dr Peter Nightingale for statistical advice.
Funding This work has been supported by a grant from the European Community's Sixth Framework Programme (AutoCure). The Prague centre was supported by a grant NR 10065-4 from the Internal Grant Agency of the Ministry of Health of the Czech Republic.
Ethics approval This study was conducted with the approval of the research ethics committee in the Czech Republic. In the other countries the regulatory authorities determined that the collection of these anonymous data constituted a service evaluation for which approval from a research ethics committee was not necessary.
Provenance and peer review Not commissioned; externally peer reviewed.