Article Text

Quality indicators for systemic lupus erythematosus based on the 2019 EULAR recommendations: development and initial validation in a cohort of 220 patients
  1. Katerina Chavatza1,
  2. Myrto Kostopoulou2,
  3. Dionysis Nikolopoulos1,
  4. Ourania Gioti3,
  5. Konstantina Togia1,
  6. Laura Andreoli4,5,
  7. Martin Aringer6,
  8. John Boletis7,
  9. Andrea Doria8,
  10. Frederic A Houssiau9,
  11. David Jayne10,
  12. Marta Mosca11,
  13. Elisabet Svenungsson12,
  14. Angela Tincani4,
  15. George Bertsias13,
  16. Antonis Fanouriakis1,3,
  17. Dimitrios T Boumpas1,14
  1. 1 Rheumatology and Clinical Immunology, Medical School, National and Kapodistrian University of Athens, "Attikon" University Hospital of Athens, Athens, Greece
  2. 2 Department of Nephrology, “G. Gennimatas” General Hospital, Athens, Greece
  3. 3 Department of Rheumatology, "Asklepieion" General Hospital, Voula, Athens, Greece
  4. 4 Department of Clinical and Experimental Sciences, University of Brescia, Brescia, Italy
  5. 5 Unit of Rheumatology and Clinical Immunology, Spedali Civili, Brescia, Italy
  6. 6 Division of Rheumatology, Department of Medicine III, University Medical Center & Faculty of Medicine Carl Gustav Carus at the TU Dresden, Dresden, Germany
  7. 7 Nephrology Department and Renal Transplantation Unit, “Laikon” Hospital, National and Kapodistrian University of Athens, Medical School, Athens, Greece
  8. 8 Rheumatology Unit, Department of Medicine, University of Padova, Padova, Italy
  9. 9 Cliniques Universitaires Saint-Luc, Université catholique de Louvain, Brussels, Belgium
  10. 10 Department of Medicine, University of Cambridge, Cambridge, UK
  11. 11 Rheumatology Unit, Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy
  12. 12 Division of Rheumatology, Department of Medicine Solna, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden
  13. 13 Rheumatology, Clinical Immunology and Allergy, University Hospital of Heraklion, Heraklion, Crete, Greece
  14. 14 Laboratory of Autoimmunity and Inflammation, Biomedical Research Foundation of the Academy of Athens, Athens, Greece
  1. Correspondence to Dr Dimitrios T Boumpas, Rheumatology and Clinical Immunology, National and Kapodistrian University of Athens, Athens 12462, Greece; boumpasd{at}


Background Quality of care is receiving increased attention in systemic lupus erythematosus (SLE). We developed quality indicators (QIs) for SLE based on the 2019 update of European League Against Rheumatism recommendations.

Methods A total of 44 candidate QIs corresponding to diagnosis, monitoring and treatment, were independently rated for validity and feasibility by 12 experts and analysed by a modified Research and Development Corporation/University of California Los Angeles model. Adherence to the final set of QIs and correlation with disease outcomes (flares, hospitalisations and organ damage) was tested in a cohort of 220 SLE patients with a median monitoring of 2 years (IQR 2–4).

Results The panel selected a total of 18 QIs as valid and feasible. On average, SLE patients received 54% (95% CI 52.3% to 56.2%) of recommended care, with adherence ranging from 44.7% (95% CI 40.8% to 48.6%) for diagnosis-related QIs to 84.3% (95% CI 80.6% to 87.5%) for treatment-related QIs. Sustained remission or low disease activity were achieved in 26.8% (95% CI 21.1% to 33.2%). Tapering of prednisone dose to less than 7.5 mg/day was achieved in 93.6% (95% CI 88.2% to 97.0%) while 73.5% (95% CI 66.6% to 79.6%) received the recommended hydroxychloroquine dose. Higher adherence to monitoring-related QIs was associated with reduced risk for a composite adverse outcome (flare, hospitalisation or damage accrual) during the last year of observation (OR 0.97 per 1% adherence rate, 95% CI 0.96 to 0.99).

Conclusion We developed QIs for assessing and improving the care of SLE patients. Initial real-life data suggest face validity, but a variable degree of adherence and a need for further improvement.

  • lupus erythematosus
  • systemic
  • autoimmune diseases
  • glucocorticoids
  • quality indicators
  • health care

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information.

Statistics from

Request Permissions

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.

Key messages

What is already known about this subject?

  • Systemic lupus erythematosus (SLE) is a multisystem disease with considerable morbidity whose care is complicated by its extreme clinical heterogeneity. The European League Against Rheumatism (EULAR) has developed evidence-based and expert opinion-based recommendations for the management of various aspects of SLE. Quality indicators (QIs), a popular tool to measure the degree of quality of care received by patients, have been proposed for SLE, but for the most part they were not based on a comprehensive systematic literature review (SLR).

What does this study add?

  • This is the first comprehensive set of QIs in SLE based on an extensive SLR of the various aspects of SLE, performed as part of the EULAR recommendations for SLE. This study further capitalises on this work by developing QIs to detect potential gaps in SLE care and facilitate the implementation of the guidelines. Initial real-life data suggest a variable degree of adherence to the recommendations and identify areas for further improvement.

How might this impact on clinical practice or future developments?

  • These QIs can be used towards assessing and improving patient care. QIs may facilitate the implementation of the EULAR recommendations by creating a checklist to be used towards detecting gaps in lupus care and facilitating efforts towards closing them.


Quality of healthcare is defined as ‘the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge’ (Institute of Medicine 1999).1 The definition applies both to healthcare practitioners and to all settings of care (hospitals, nursing homes and physicians’ offices). Measurement of quality can help to identify problems caused by overuse, underuse or misuse of health resources.

Quality indicators (QIs) is a popular tool to measure the degree of quality of care received by patients. QIs are quantitative measures related to the structures, processes or outcomes of care,2 3 derived from guidelines, systematic literature reviews (SLR) or expert panel consensus, through the use of a systematic approach representing the current standard of care. In contrast to most guidelines or recommendations, QIs pertain to measurable aspects of healthcare, describing exactly what to do, when to do it and who is responsible for doing it, with respect to disease management and monitoring.4 5

Systemic lupus erythematosus (SLE) is a multisystem disease with considerable morbidity due to both the disease per se and the complications of chronic treatment.6 Care in SLE is complicated by the profound clinical heterogeneity and differences among individual patients. During the last two decades, the European League Against Rheumatism (EULAR) has developed evidence-based and expert opinion-based recommendations for the management of various aspects of SLE, including general SLE, renal and neuropsychiatric disease, and women’s health including fertility and pregnancy.7–9 These recommendations were recently updated.10–12 In addition to these recommendations, other initiatives such as the treat-to-target in SLE have also highlighted the importance of a multifaceted care targeting remission or low disease activity.13

Towards improving patient care, detect potential gaps in SLE care, and facilitate the implementation of the guidelines, herein we sought to develop QIs based on the 2019 update of the EULAR recommendations for SLE, and perform an initial validation in one academic centre.


Overview of the development of preliminary criteria and selection of the final set

QIs were developed using an adaptation of the Research and Development Corporation (RAND)/University of California Los Angeles (UCLA) modified Delphi method, a structured systematic approach that combines the best available evidence from an SLR with the collective expert opinion that has been shown to be valid in similar applications.14–16 During a two-round process, a panel of experts assessed the validity and feasibility of the proposed indicators. The final set of QIs was used to evaluate the quality of care in an SLE cohort of 220 patients and to explore possible associations with disease outcomes (figure 1).

Figure 1

Overview of the development of preliminary criteria and selection of the final set of quality indicators. EULAR, European League Against Rheumatism; QIs, quality indicators; RAND, Research and Development Corporation; SLE, systemic lupus erythematosus; UCLA, University of California Los Angeles.

Identification of potential indicators and rounds of voting

An inventory of candidate QIs was developed based on the 2019 EULAR recommendations for SLE and the corresponding SLR.7–12 17 A preliminary set of 44 QIs, which addressed seven distinct clinical domains and was graded according to the existing level of evidence, was evaluated by a panel of experts (nine rheumatologists and three nephrologists from five European countries) and one patient representative (see online supplemental appendix and online supplemental table 1). Every panel member was asked to rate each QI item for validity and feasibility, using a 9-point scale, with 9 representing the highest possible rating (definitions for validity and feasibility and voting instructions provided in online supplemental appendix). Panel members were also asked to comment on the draft QIs and suggest amendments as required. Following round 1 ratings, an analysis of the candidate QIs was performed, as described in the RAND/UCLA Method.14 15 For each candidate QI, the median rating, median absolute deviation, lower and upper limit inter-percentile range, were calculated. The Disagreement Index (DI) was calculated using the equations provided in online supplemental appendix. Median validity/feasibility scores of ≤ six were used to exclude QIs. The other measurements (deviation, agreement/disagreement) were used as additional information for the selection of the QIs for round 2 (online supplemental table 2). The moderator (DTB) and two other panel members (GB and AF) convened to discuss the results and revise the initial set, based on the ratings from round 1 (online supplemental table 3). Candidate QIs were modified, merged or eliminated accordingly, and the revised set was sent to the expert panel for round 2 of ratings. Experts were asked to rerate the QIs for validity and feasibility based on the same 9-point scale. Results were analysed and finalised based on the same principles used during round 1, reaching the final set of QIs (table 1).

Table 1

Final set of quality indicators (QIs) and rating during the second round of voting


To perform an initial validation of the proposed QIs, we used patients from the ‘Attikon’ lupus cohort, based in the largest tertiary hospital of western Attica and considered as a referral centre for patients with SLE.18 Patients from the cohort were included if they (1) fulfilled the EULAR/American College of Rheumatology (ACR) 2019 and/or 2012 Systemic Lupus International Collaborating Clinics (SLICC) classification criteria,19 20 (2) had at least 1 year of follow-up and (3) had at least four visits over the last year. Included patients were derived from an inception cohort (patients followed from January 2016 (year of establishment of the Attikon cohort) to date—60%) and a prevalent cohort (patients with SLE diagnosis before January 2016–40%)].

Chart review and patient interviews were performed retrospectively to assess patient eligibility for each QI (eg, only smokers were eligible for the corresponding QI regarding counselling for smoking cessation), adherence to each QI item, and to document disease outcomes of interest, specifically flares, SLICC/ACR damage index (SDI) and hospitalisations (due to cardiovascular events, infections and flares). Patient interviews were performed by physicians (KC, OG), with the completion of a patient-reported questionnaire, for measures that were not universally available in medical records (eg, counselling for fertility and sun protection).


Performance for each QI was assessed as ‘fulfilled’, ‘not fulfilled’ or ‘missing’, in order to assess the grade of adherence. Because most QIs consisted of more than one individual component, ‘fulfilled’ denoted that all components were met, ‘not fulfilled’ that at least one component of the corresponding QI was not met, and ‘missing’ that data were not recorded in the chart. As an example, QI1 would be labelled as ‘fulfilled’ if all recommended laboratory and serological tests were obtained at diagnosis (see QI1, table 2). By contrast, QI1 would be labelled as ‘not fulfilled’, if at least one of the recommended tests was not obtained at diagnosis.

Table 2

Adherence to 18 selected quality indicators (QIs)


Disease-related outcomes were recorded in all patients. Recorded outcomes included (1) flares, (including major flares), defined as a measurable increase in disease activity leading to therapeutic intervention,21 (2) SDI increase (an increase in the SDI score during the observation period), (3) adverse outcomes related to glucocorticoids (GC, that is, GC-related complications), (4) cardiovascular events or serious infections necessitating hospitalisation and (5) a composite adverse outcome (CAO), defined as occurrence of at least one of the following: flares, hospitalisations or SDI progression. The phenotype of SLE was categorised as mild, moderate or severe according to the British Isles Lupus Assessment Group 2004 classification of manifestations,22 combined with expert physician judgement (DTB and AF), as previously described.18

Statistical analysis

Descriptive statistics were used for continuous variables and mean/SD or median/IQR values were calculated as appropriate. Adherence to each QI was calculated as the number of patients who received the designated care (numerator) divided by the number of eligible patients for this particular QI (denominator). In addition, a patient-specific mean score was calculated as the number of QIs ‘fulfilled’ divided by the number of QIs eligible for each patient. Accordingly, the average delivered care for each domain was calculated as a composite score from the cases in which recommended care was successfully delivered divided by the number of eligibility events within each domain.

To detect potential differences in adherence between patient subgroups, we applied three criteria: (1) disease duration (<2 years vs ≥2 years from diagnosis), (2) severity pattern (mild vs moderate vs severe disease) and (3) origin of cohort (inception vs prevalent). A separate analysis was performed to compare the adherence between various QI domains (diagnostic, treatment, monitoring). To compare mean values or the equality of distribution between different categories the one-way analysis of variance and the non-parametric Kruskal-Wallis test were used accordingly. Logistic regression models were performed to estimate the association between adherence to QIs and adverse disease outcomes that occurred in two different time frames (during the total duration of follow-up and during the last year of follow-up). All models were adjusted for age and disease duration. All tests were two tailed and p values less than 0.05 were considered statistically significant. Data management and statistical analyses were performed using STATA/MP V.13.1 (StataCorp).


Final set of QIs

Out of 44 initial candidate QIs (online supplemental table 3), three were removed due to a low median validity/feasibility score. For the remaining, agreement was reached in 31 QIs and disagreement in 10. In the former set, minor edits and discussion based on experts’ comments resulted in 4 QIs being included without change, 4 being retained with edits, 17 being merged to 7 separate QIs and 6 being rejected. Of the 10 QIs for which there was disagreement, one was retained without changes, two were retained with edits, four were excluded and three were merged with two previously formed QIs. This resulted in a revised set of 18 candidate QIs, which was available for Round 2 of rating (table 1). The 18 finally selected QIs were further divided into three categories: (1) Screening/diagnosis-related (QIs 1–5), (2) Treatment-related (QIs 6–10) and 3) Monitoring (QIs 11–18). More specifically, selected individual QIs pertain to diagnosis, monitoring, therapy and its targets, fertility and pregnancy and adjunct therapy, including prevention of cardiovascular disease (CVD) and osteoporosis, vaccination, counselling for smoking and sunscreen protection.

Adherence to QIs

The final set of 18 QIs was tested for adherence in all eligible patients of our cohort (N=220)(table 2). Characteristics of the cohort are shown in online supplemental table 4. On average, patients received 54% (95% CI 52.30% to 56.25%) of the indicated care. Complete laboratory work-up at diagnosis was performed in 48.5% (95% CI 39.8% to 57.1%), with antiphospholipid antibodies being the most frequently missed component (68.9%). Disease activity evaluation in at least three out of four visits and annual assessment of organ damage were completed in only 14.1% (95% CI 9.4% to 18.7%) and 28.6% (95% CI 22.6% to 34.6%), respectively. By contrast, lupus nephritis related QIs had excellent overall adherence (88%, 95% CI 66.7% to 96.4% for the use of ACE inhibitor/angiotensin receptor blocker, 100% for the use of immunosuppressive treatment), except for laboratory monitoring (48.5%, 95% CI 36.6% to 60.6%). Overall adherence rate was 50% (95% CI 38.5% to 61.5%) for reproductive health counselling, 62% (95% CI 49.9% to 72.7%) for pregnancy counselling and 91.4% (95% CI 86.8% to 94.4%) for sunscreen protection. Notably, preventive measures for comorbidities had generally low to moderate adherence. More specifically, overall adherence rates for cardiovascular risk modification and vaccination QIs (at least one of the available pneumococcal vaccines in combination with influenza vaccine) were 40.5% (95% CI 34.1% to 47.1%) and 47.7% (95% CI 41.2% to 54.4%), respectively. Regarding osteoporosis prevention and treatment, the corresponding QI was fulfilled in 45.5% (95% CI 38.9% to 52.1%) of patients. A total of 73% of eligible patients had bone mineral density measurement performed at baseline and 58.6% at follow-up (every 2 years), while 60% of patients belonging in the high fracture risk group received antiresorptive treatment; almost 75% (74.3%) received calcium and vitamin D. Of note, 63.8% of patients on GC received calcium and vitamin D protection.

In a subgroup analysis, patients with severe disease were more likely to receive the indicated care (57.2%) compared with patients with moderate (53.9%) or mild (49.3%) disease (p=0.006). Similarly, higher adherence rates were observed in patients with short (<2 years) vs longer (≥2 years) disease duration (54.8% and 49.3% respectively, p=0.02). No significant differences were observed between the inception and the prevalent cohort (table 3).

Table 3

Adherence to quality indicators (QI) in subgroups of patients

In a separate analysis according to the function of care, treatment-related QIs were met in significantly more eligible patients (84.3%) followed by monitoring (50.5%) and diagnostic (44.6%) QIs (p=0.03) (table 4).

Table 4

Adherence to quality indicators (QIs) grouped according to function of care


Disease-related outcomes are summarised in online supplemental table 5. Patients were followed up for a median of 2 years (IQR 2–4). SDI progression was observed in 22.3% of patients incidence rate (IR)=13/100 patient-years (pys). A total of 310 flares were captured over the follow-up corresponding to 0.58 per py. The IR of hospitalisations was 15.4/100 py, attributed mainly to major flares (7.8/100 py), serious infections (6.1/100 py) and cardiovascular events (1.5/100 py).

Overall, QI adherence did not differ among patients experiencing CAO and patients without CAO throughout the observation period (54.0% vs 54.7%, p=0.71). However, patients with CAO during the last year of follow-up had lower adherence rates in monitoring QIs when compared with patients without a CAO (47.6% vs 53.9%, p=0.02) (online supplemental table 6). We also explored possible associations between adherence to specific QIs and outcomes. Patients who achieved sustained remission or Lupus Low Disease Activity State (LLDAS) (QI13), patients who fulfilled QI16 regarding vaccination and patients who received low-dose GC (QI7) had lower odds of experiencing a flare during the observation period (OR 0.15, 95% CI 0.07 to 0.31 OR 0.46, 95% CI 0.21 to 0.98 and OR 0.23, 95% CI 0.05 to 0.94, respectively). A lower risk of CAO during the last year of follow-up was also found in patients who met QI13 on remission/LLDAS and QI16 on vaccination (OR 0.09, 95% CI 0.04 to 0.18 and OR 0.52, 95% CI 0.28 to 0.99 respectively). As expected, patients who achieved sustained remission or LLDAS (QI13) had lower odds of damage accrual during the observation period (OR 0.35, 95% CI 0.14 to 0.84). Patients assessed for SDI accrual (QI12) and CVD risk stratification (QI4) had higher probability to exhibit any CAO (OR 2.62, 95% CI 1.18 to 5.71 and OR 1.77, 95% CI 1.01 to 3.12, respectively) (table 5).

Table 5

Risk of adverse events associated with the delivered care in an SLE cohort of 220 patients


SLE is notorious for its clinical heterogeneity, which may in turn increase the risk of inconsistency and variations in the care received by patients. To ensure improved and more homogeneous care, EULAR has developed evidence-based and expert opinion-based recommendations for the management of various aspects of the disease.9–12 Nonetheless, since management recommendations are often followed incompletely in real-life settings, efforts have been made to create tools which can transform them into easily applicable, ‘user friendly’ instructions for daily practice. In this regard, QIs can be useful instruments for the quantification of gaps and shortcomings in medical care. Herein, we created a set of QIs based on the EULAR recommendations for SLE, using a validated, systematic methodology supported by expert opinion. In addition, we examined the adherence to the proposed QIs in 220 patients of the ‘Attikon’ lupus cohort, a readily available patient cohort, to take an initial ‘glimpse’ on potential gaps of care in daily practice and assess their impact on disease outcomes.

QIs have been previously proposed for SLE,5 23 but for the most part they were not based on a comprehensive SLR. This is the first set of QIs based on such a comprehensive SLR of the various aspects of SLE (ie, diagnosis, monitoring and therapy), which was performed in the context of the updated EULAR recommendations. The credibility of the proposed QI set is reinforced by the robust methodology of the procedure (ie, the RAND/UCLA modified Delphi method), which involved assessment of a large number of initial candidate QIs for validity and feasibility, followed by two rounds of voting, all performed by a panel of experts with expertise in SLE.

Our initial findings suggest moderate adherence (54%) with great variability in certain types of QIs. The low rates of CVD protection and reproductive health counselling are consistent with data from previous studies;24 25 rates for sunscreen protection and individual components for osteoporosis and vaccination (influenza, pneumococcal) QIs are also consistent with published data.25 Looking for potential explanations, in the case of CVD-related QIs, the complexity of prescribing statins by rheumatologists in some countries and, in case of osteoporosis prophylaxis the plethora of recommendations by various scientific societies, may account at least in part for these low adherence rates. In our view, this reality highlights the need to actively involve nurse specialists in the care of SLE patients, especially in the settings of expert SLE referral centres. Such nurse practitioners could monitor the assessment and fulfilment of these QIs, which may not be a priority in a busy physician outpatient clinic.

In reference to potential causes related to better performance in certain indicators, we found that QI adherence rates were higher in patients with disease duration shorter than 2 years and in patients with severe disease. These observations may reflect the fact that physicians are more likely to adhere early after diagnosis to ensure better disease control, and in patients who are more likely to develop irreversible organ damage, respectively. Despite variable rates of adherence, we did not find strong associations between non-adherence to QIs and adverse outcomes, except that patients who were in a low disease activity state had lower rates of flares and damage progression. A possible explanation is that the adherence to a single QI may not suffice to provide a clinically favourable outcome, if not combined with consistent and adequate care. Thus, in a study by Yazdany et al, patients who met ≥85% of the eligible QIs had lower odds of damage accrual, however, the difference was not significant for any of the individual QIs alone.26 The modest associations between quality of care and outcomes in our study may also reflect the relatively short follow-up, especially because many SLE outcomes develop within years, and longer observation time is needed to detect any association. To further address this issue, prospective long-term follow-up studies evaluating a combination set instead of single indicators with varied settings and outcomes are needed.

Our study has several limitations. The duration of follow- up was modest and data represent the experience of a single academic centre. Consequently, our results may not be representative of other clinical settings and in non-academic centres, gaps in patient care may be even greater. Conversely, a tertiary care hospital that serves as a referral centre may follow patients with a higher burden of the disease and higher risk of progression.27 28 Risk-adjusted and casemix models would help to account for differences in patient-level and hospital-level risk, however, the relatively small study sample, limited access to administrative data and the absence of electronic health record systems prevented us from performing this methodology. Yet, development of the current QIs was based on an extensive systematic review and panels of experts working for over one decade to develop recommendations for SLE. To this end, longitudinal and nationwide population-based studies are warranted to validate these QIs in various time and clinical settings.

In summary, we have developed a set of EULAR recommendations-based QIs for SLE patient care, following a comprehensive SLR and supported by expert opinion. Initial real-life data suggest a variable degree of adherence and areas for further improvement. Nevertheless, these QIs may be used as a ‘checklist’ to be fulfilled in an outpatient setting, in order to improve SLE patient care by facilitating the implementation of the EULAR recommendations.

Data availability statement

All data relevant to the study are included in the article or uploaded as online supplemental information.

Ethics statements

Ethics approval

The study was approved by the Ethics Committee of ’Attikon’ University Hospital of Athens.


We are thankful to the staff physicians and nurses of the Rheumatology and Clinical Immunology Unit of ’Attikon’ University Hospital for providing care to the patients with SLE and other rheumatologic diseases.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

  • Lay summary

    Disclaimer : This is a summary of a scientific article written by a medical professional (“the Original Article”). The Summary is written to assist non medically trained readers to understand general points of the Original Article. It is supplied “as is” without any warranty. You should note that the Original Article (and Summary) may not be fully relevant nor accurate as medical science is constantly changing and errors can occur. It is therefore very important that readers not rely on the content in the Summary and consult their medical professionals for all aspects of their health care and only rely on the Summary if directed to do so by their medical professional. Please view our full Website Terms and Conditions.
    Copyright © 2021 BMJ Publishing Group Ltd & European League Against Rheumatism. Medical professionals may print copies for their and their patients and students non commercial use. Other individuals may print a single copy for their personal, non commercial use. For other uses please contact our Rights and Licensing Team.


  • GB, AF and DTB are joint senior authors.

  • Handling editor Josef S Smolen

  • Twitter @none

  • KC, MK, GB, AF and DTB contributed equally.

  • Contributors KC and OG collected data from patient medical charts and KC drafted the manuscript. MK performed statistical analyses and edited the manuscript. DN edited the manuscript. KT assessed patient medical charts for eligibility in the study. LA, MA, JB, AD, FAH, DJ, MM, ES and AT evaluated the quality indicators, provided voting and critically reviewed the manuscript. GB and AF supervised the study and edited the manuscript. DTB conceived and supervised the study and edited the manuscript.

  • Funding DTB was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 742390). DJ was supported by the NIHR Cambridge Biomedical Research Centre (BRC-1215 20014).

  • Competing interests DTB is an Editorial Board member in the Annals of the Rheumatic Diseases. The remaining authors declare no competing interests relevant to this work.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.