A Delphi Exercise to Identify Characteristic Features of
Gout — Opinions from Patients and Physicians, the
First Stage in Developing New Classification Criteria
Rebecca L. Prowse, Nicola Dalbeth, Arthur Kavanaugh, Adewale O. Adebajo, Angelo L. Gaffo, Robert Terkeltaub, Brian F. Mandell, Bagus P.P. Suryana, Claudia Goldenstein-Schainberg, Cèsar Diaz-Torne, Dinesh Khanna, Frederic Lioté, Geraldine Mccarthy, Gail S. Kerr,
Hisashi Yamanaka, Hein Janssens, Herbert F. Baraf, Jiunn-Horng Chen, Janitzia Vazquez-Mellado, Leslie R. Harrold, Lisa K. Stamp, Mart A. Van De Laar, Matthijs Janssen, Michael Doherty, Maarten Boers, N. Lawrence Edwards, Peter Gow, Peter Chapman, Puja Khanna,
Philip S. Helliwell, Rebecca Grainger, H. Ralph Schumacher, Tuhina Neogi, Tim L. Jansen, Worawit Louthrenoo, Francisca Sivera, And William J. Taylor
ABSTRACT. Objective. To identify a comprehensive list of features that might discriminate between gout and other rheumatic musculoskeletal conditions, to be used subsequently for a case-control study to
develop and test new classification criteria for gout.
Methods. Two Delphi exercises were conducted using Web-based questionnaires: one with physicians from several countries who had an interest in gout and one with patients from New Zealand
who had gout. Physicians rated a list of potentially discriminating features that were identified by literature review and expert opinion, and patients rated a list of features that they generated themselves. Agreement was defined by the RAND/UCLA disagreement index.
Results. Forty-four experienced physicians and 9 patients responded to all iterations. For physicians, 71 items were identified by literature review and 15 more were suggested by physicians. The physician survey showed agreement for 26 discriminatory features and 15 as not discriminatory. The patients identified 46 features of gout, for which there was agreement on 25 items as being discriminatory and 7 items as not discriminatory.
Conclusion. Patients and physicians agreed upon several key features of gout. Physicians emphasized objective findings, imaging, and patterns of symptoms, whereas patients emphasized severity, functional results, and idiographic perception of symptoms. (First Release Feb 15 2013; J Rheumatol 2013;40:498–505; doi:10.3899/jrheum.121037)
Key Indexing Terms:
GOUT CLASSIFICATION CRITERIA PATIENTS PHYSICIANS
From the University of Otago, Dunedin, New Zealand; University of Auckland, Auckland, New Zealand; Veterans Affairs Healthcare System, University of California, San Diego, California, USA; University of Sheffield, Sheffield, UK; Veterans Affairs Medical Center and University
of Alabama at Birmingham, Birmingham, Alabama, USA; CCF Lerner College of Medicine of Case Western Reserve University, The Cleveland Clinic, Cleveland, Ohio, USA; Brawijaya University and Dr. Saiful Anwar Hospital, Malang, Indonesia; Faculdade de Medicina da Universidade de São Paulo e Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil; Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; University of Michigan, Ann Arbor, Michigan, USA; University Paris Diderot, Sorbonne Paris Cité and Hôpital Lariboisière, Paris, France; University College Dublin, Dublin, Ireland; Veterans Affairs Medical Center, Georgetown and Howard Universities, Washington, DC, USA; Tokyo Women’s Medical University, Tokyo, Japan; Department of Primary and Community Care, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; The Center for Rheumatology and Bone Research, Arthritis and Rheumatism Associates, Wheaton, Maryland, USA; China Medical University, China Medical University Hospital, Taichung, Taiwan; Hospital General de Mexico, Mexico City, Mexico; University of Massachusetts Medical School, Worcester, Massachusetts, USA; University of Otago, Christchurch, New Zealand; Department of Rheumatology and Clinical Immunology, Medisch Spectrum Twente, Enschede, The Netherlands; Department of Rheumatology, Rijnstate, Arnhem, The Netherlands; University of Nottingham, Nottingham, UK; VU University Medical Center, Amsterdam, The Netherlands; University of Florida, Gainesville Florida, USA; Counties-Manukau District Health Board, Auckland, New Zealand; Canterbury District Health Board, Christchurch, New Zealand; University of Leeds, Leeds, UK; University of Otago, Wellington, New Zealand; University of Pennsylvania and Veterans Affairs Medical Center, Philadelphia, Pennsylvania, USA; Boston University School of Medicine, Boston, Massachusetts, USA; Department of Rheumatology, Radboud University Nijmegen Medisch Centrum, Nijmegen, The Netherlands; Chiang Mai University, Chiang Mai, Thailand; and Hospital General de Elda and Universidad Miguel Hernandez, Alicante, Spain.
R. Prowse was supported by a Summer Student Scholarship from Arthritis New Zealand. Dr. Kerr has participated in gout trials by Savient, Nuon, and Ardea. Dr. Doherty is a member of advisory boards for Ardea Biosciences, Ipsen, Menarini, Novartis, and Savient.
R.L. Prowse, BSc, University of Otago; N. Dalbeth, MD, University of Auckland; A. Kavanaugh, MD, Veterans Affairs Healthcare System,
University of California; A.O. Adebajo, MD, University of Sheffield; A.L. Gaffo, MD, MPH, Veterans Affairs Medical Center, University of Alabama at Birmingham; R. Terkeltaub, MD, Veterans Affairs Healthcare System, University of California; B.F. Mandell, MD, CCF Lerner College of Medicine of Case Western Reserve University, The Cleveland Clinic; B.P.P. Suryana, MD, Brawijaya University, Dr. Saiful Anwar Hospital; C. Goldenstein-Schainberg, MD, Faculdade de Medicina da Universidade de São Paulo, Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo; C. Diaz-Torne, MD, Hospital de la Santa Creu i Sant Pau; D. Khanna, MD, MS, University of Michigan; F. Lioté, MD, University Paris Diderot, Sorbonne Paris Cité, Hôpital Lariboisière; G. McCarthy, MD, University College Dublin; G.S. Kerr, MD, Veterans Affairs Medical Center, Howard Universities; H. Yamanaka, MD, Tokyo Women’s Medical University; H. Janssens, MD, Department of Primary and Community Care, Radboud University Nijmegen Medical Centre; H.F. Baraf, MD, The Center for Rheumatology and Bone Research, Arthritis and Rheumatism Associates; J-H. Chen, MD, China Medical University, China Medical University Hospital; J. Vazquez-Mellado, MD, Hospital General de Mexico; L.R. Harrold, MD, MPH, University of Massachusetts Medical School; L.K. Stamp, MD, University of Otago; M.A. van de Laar, MD, Department of Rheumatology and Clinical Immunology, Medisch Spectrum Twente; M. Janssen, MD, Department of Rheumatology, Rijnstate; M. Doherty, MD, University of Nottingham; M. Boers, MD, VU University Medical Center; N.L. Edwards, MD, University of Florida; P. Gow, MBChB, Counties-Manukau District Health Board; P. Chapman, MD, Canterbury District Health Board; P. Khanna, MD, University of Michigan; P.S. Helliwell, MD, PhD, University of Leeds; R. Grainger, MBChB, PhD, University of Otago; H.R. Schumacher, MD, University of Pennsylvania, Veterans Affairs Medical Center; T. Neogi, MD, MPH, Boston University School of Medicine; T.L. Jansen, MD, Department of Rheumatology, Radboud University Nijmegen Medisch Centrum; W. Louthrenoo, MD, Chiang Mai University; F. Sivera, MD, Hospital General de Elda, Universidad Miguel Hernandez; W.J. Taylor, MBChB, PhD, University of Otago.
Address correspondence to Prof. W. Taylor, Department of Medicine, University of Otago Wellington, PO Box 7343, Wellington 6242, New Zealand. E-mail: Will.taylor@otago.ac.nz
Gout is characterized by synovial and tissue deposition of monosodium urate crystals1. The gold standard diagnostic
test for gout is the presence of monosodium urate (MSU) crystals within joint fluid or tissue and this should normally be the preferred approach to diagnosis in clinical practice2.
However, in some research settings, examination of
synovial fluid is impractical. For example, in epidemiological studies or in studies of patients recruited from
primary care, there may not be access to synovial fluid microscopy. In such situations, classification criteria that aim to mimic the diagnostic gold standard are needed3.
Classification criteria for gout that do not rely upon MSU crystal identification have previously been developed but may not be sufficiently accurate. Malik, et al4 examined the
validity of the non-crystal-dependent aspects of these criteria in a hospital-based population, using the gold standard of MSU crystal identification as a comparison group; they found imperfect specificity and sensitivity for the Rome, New York5, and American Rheumatism
Association (ARA)6 criteria. Janssens, et al found limited
accuracy of the ARA criteria, with a sensitivity of 80% and specificity of 64% in patients presenting to family practitioners with potential gout symptoms7. Both the Rome and
New York criteria are heavily dependent on verifying the presence of tophi or MSU crystals within a joint, which is not always achievable in research settings. The rising prevalence
of gout8 and its association with the metabolic syndrome9 and cardiovascular disease10 make it
important
to study the disorder accurately. Therefore better classification criteria for gout are required.
A modification of the ARA criteria, termed the Clinical Gout Diagnosis (CGD) criteria set, was shown to have very high sensitivity (97%) and specificity (96%) in a group of rheumatology clinic patients with crystal-proven gout and other rheumatic diseases (osteoarthritis, spondyloarthritis, rheumatoid arthritis)11. However, the non-gout cases in that
rate of tophi (81%) in the cohort limit the general applicability. Another novel approach based in primary care has
been reported, with a positive predictive value of 80%12.
This approach is somewhat limited by the inclusion of items associated with gout such as cardiovascular disease and male sex, rather than items intrinsic to the disease. Traditionally, potential items for classification criteria are identified by physicians on the basis of clinical experience and knowledge of the pathology of the disease. The opinions of patients about the disease in question are rarely sought, yet patients have firsthand knowledge of how a disease is manifest and may be able to identify important clinical diagnostic pointers that could be overlooked by physicians. Patient involvement in outcome measurement13,14, teaching health professionals15, and
self-management16 are well described and so it was thought
to be potentially useful to also include patients’ perceptions regarding classification criteria in this study.
It is important to emphasize that the purpose of the overall project and for classification criteria in general is accurate case ascertainment for clinical research so that populations that are relatively homogeneous (with respect to the disease under study) are recruited. This is distinct from diagnostic criteria, which may be used for the diagnosis of individual patients in clinical practice. Nevertheless, it is usually the case that classification criteria are formed by a restricted set of items that are also used for diagnosis. In our study, we did not wish to restrict the range of items to be elicited, and thus framed questions in terms of diagnosis rather than classification, even though classification criteria are the ultimate aim. Also, in clinical practice, examination of tissue or synovial fluid is the preferred diagnostic approach for gout. In the case of rheumatology care, all rheumatologists should be able to obtain synovial fluid and examine it for the presence of MSU crystals because that is part of the training curriculum17. Classification criteria do
not replace this diagnostic approach. Even in primary care, classification criteria do not necessarily replace the recommended
diagnostic approach but can be useful aides to recalling the key features of the disease.
The objective of our study was to identify a comprehensive list of clinical, laboratory, and imaging features that
could potentially discriminate between gout and other forms
of arthritis or rheumatic musculoskeletal disease in a primary healthcare setting. This study used the Delphi
technique to anonymously obtain opinions from both physicians and patients, and then give them the opportunity to
revise their opinion in light of the group’s average. This information will serve as the basis for a planned multinational case-control study that aims to create and validate
new classification criteria for the identification of gout that is designed for the setting of clinical research independent of patient care. As noted, such criteria should not be used for the diagnosis of individual patients in ordinary clinical care.
MATERIALS AND METHODS
Eighty-one physicians from multiple countries who were interested in gout were identified from an e-mail list accumulated from previous gout studies, and 87 patients with gout were identified from patient registers at 3 New Zealand rheumatology services. Nearly all physicians were rheumatologists. Participants were asked to take part in a series of Web-based
questionnaires to identify features typical of gout to be used to develop new criteria for the classification of gout. Physicians were invited by e-mail and patients were invited by letter.
Physicians were asked to rate items on the extent to which they believed
that particular feature could distinguish gout from other rheumatic musculoskeletal conditions. Items presented to the physicians in the first iteration
were identified by literature search and expert opinion. Any extra features identified by physicians as being important were also solicited in the first iteration. Features of gout in the patient survey were obtained from the first iteration using the question, “list as many features of gout as possible that help you and your doctor know you have gout and not some other joint condition.” All participants used a 9-point rating scale (1 = not at all discriminatory; 9 = extremely discriminatory). Consensus was defined by the RAND/UCLA disagreement index whereby values > 1 indicated disagreement18. Items that had been suggested by physicians in the first iteration,
a median rating of 4–6 (uncertainty) were re-rated in the second and third
iterations, if needed. In the second iteration of the patient survey, all items from the first round were rated using the 9-point agreement scale. In the third round only
the items for which there was disagreement or those with a median rating of 4–6 were re-rated. Reminders were sent by e-mail to all participants after a week of each iteration and they were given a further week to complete the survey before they were considered a nonrespondent.
According to the principles of the Delphi method19, the participants (patients and physicians) remained anonymous to each other throughout the duration of the study. The responses to the surveys were analyzed after each round and the median and 30th and 70th percentiles were made known to each respondent in subsequent rounds. The surveys were carried out for 3 iterations or until consensus was reached, giving participants the opportunity to change their answers in light of the groups’ average.
The study protocol was approved by the New Zealand Health and Disability Multiregional Ethics Committee (MEC/11/EXP/077).
RESULTS
There were 49 respondents to the first physician survey (60% response rate). The mean age was 52.5 (SD 10.5) years, participants had been in specialist practice for 19.9 (SD 10.8) years, and consulted on a mean of 29.7 (SD 32.9) patients with gout per month. Of these, 44 responded to the second round (90%). There were 71 clinical, laboratory, and imaging features identified by literature review and expert opinion for the first iteration of the physician survey. Of these, 13 features were considered not discriminatory for
gout and 25 were considered discriminatory. All 38 discriminatory and nondiscriminatory features were excluded from
the second iteration. The remaining features with a median rating of 4–6 (30 items) or those for which there was disagreement (2 items) were included in the second iteration, along with 15 additional features nominated by physicians and 8 features from the first iteration for which
respondents had requested clarification. There was agreement on all items during the second iteration so that
a third iteration was not required. The final list of features (Table 1) contained 4 additional discriminatory items and 2 additional nondiscriminatory items. There
were 52 items that were rated as uncertain (median rating 4–6).
There were 14 respondents to the first patient survey (16% response rate). Of these, 13 (93%) responded to the second iteration and 9 (69%) to the third iteration. Patients were a median age of 63 (range 38–89) years and the median duration of disease was 10 (range 4–25) years. In the first round, 46 features were identified by patients. In the second round, it was agreed that 2 of the features were not discriminatory for gout and that 22 of the features were discriminatory.
Patients were uncertain of the diagnostic importance of 19 of the features or were in disagreement concerning 3 items and these were re-rated in the final iteration. After the final iteration of the patient survey (Table 2) there was agreement that 7 items were not discriminatory for gout, 25 items were discriminatory for gout, and 14 items were rated with uncertainty or disagreement.
Comparison of the patient and physician data showed consensus on the following general characteristics thought to be specific for gout: the suddenness of onset, redness and swelling of the affected joint, the marked tenderness of the joint, elevated serum urate levels, presence of tophi, the presence of MSU crystals in synovial fluid, and involvement of the first metatarsophalangeal joint (Figure 1).
DISCUSSION
This Delphi exercise identified 26 features of gout that expert physicians believed were potentially appropriate to distinguish gout from other rheumatic musculoskeletal diseases. Patients with chronic gout further supported these findings by identifying many of the same features as physicians.
One difference between patients and physicians was the different emphasis on functional disability. Patients believed that the inability to carry out everyday tasks such as walking was an important diagnostic feature and rated it highly
whereas physicians believed that it was not at all discriminatory. There was more emphasis by patients on the severity
tender joints that prevent sleep and normal everyday functioning. The response to treatment and the triggers for gout attacks were also seen by patients to be more important than physicians. In contrast, physicians tended to emphasize imaging, the pattern of joint involvement, and its behavior over time. Overall, physicians were more focused on diagnostic criteria and patients on disease severity criteria. There was greater disagreement among patients
regarding the specificity of features they suggested, compared to among physicians. This is consistent with substantial interindividual variation in how diseases manifest and how symptoms are interpreted by patients. Physicians are trained to recognize nomothetic commonalities, patterns, symptom clusters, and pathology, rather
than idiographic variations of symptoms. An obvious key difference between patients and physicians that is relevant here is that physicians have experience in distinguishing between different rheumatic diseases, whereas patients have experience only in distinguishing between having and not having gout, and may not be able to easily determine when symptoms are due to gout and not some other rheumatic disease.
Many of the items for which there was agreement between patients and physicians already appear within existing classification criteria. This is not surprising, since such features are likely to be highly typical or characteristic of the disease. An improvement upon existing criteria may still be achievable with different criteria formats (for example, weighting of different features) and inclusion of new items (for example, modern imaging techniques). Unfortunately, the patient response rate in our study was much lower than expected. Five patients did not complete all iterations and thus were considered nonrespondents, we received 8 “return to sender” letters due to incorrect addresses, and we received at least 1 letter and some telephone messages from patients who wanted to participate but had no access to a computer. But the reason for
light of the low response rate, the patient results cannot be considered representative of the gout patient population. In addition, the patients reported features such as tophi that may occur only in more severely affected patients. Also, it should be noted that all patient participants were from New Zealand whereas the physicians were from several
countries. It would be of interest to obtain opinions from a larger number of patients from different countries. Finally, patients and physicians were hospital-based rather than recruited from primary care settings, which may tend to bias opinion toward more severe gout. Overall, it should not be considered that the patients in our study were representative of the gout population. Nonetheless, their opinions are of value.
This Delphi consensus methodology has provided some direction toward features that could be tested for possible new gout classification criteria. The next phase of this project is to conduct a case-control study to establish the most accurate combinations of these features for classifying gout when compared to the gold standard diagnostic procedure of MSU identification in tissue or synovial fluid.
ACKNOWLEDGMENT
List of contributors: Participants in the physician Delphi panel: Prof. Rieke Alten, Assoc. Prof. William Taylor, Assist. Prof. Angelo L. Gaffo, Dr. Puja Khanna, Prof. Ralph Schumacher, Dr. Mart van De Laar, Prof. Geraldine McCarthy, Prof. Ric Day, Dr. Janitzia Vazquez-Mellado, Dr. Fernando Perez-Ruiz, Prof. Maarten Boers, Dr. Francisca Sivera, Dr. Matthijs Janssen, Assoc. Prof. Lisa Stamp, Dr. Fred Lioté, Dr. Herbert F. Baraf, Dr. Hein Janssens, Prof. Worawit Louthrenoo, Prof. Michael Doherty, Dr. Tim L. Jansen, Assoc. Prof. Tuhina Neogi, Assoc. Prof. Peter Gow, Prof. Michael Becker, Dr. Leslie Harrold, Prof. Cláudia Schainberg, Dr. Gail Kerr, Dr. Peter Chapman, Dr. Bagus P.P. Suryana, Dr. Hisashi Yamanaka, Assoc. Prof. Nicola Dalbeth, Prof. Arthur Kavanaugh, Dr. Brian Mandel, Prof. Naomi Schlesinger, Prof. Dinesh Khanna, Dr. Philip Helliwell, Prof. Larry Edwards, Prof. Ade Adebajo, Dr. Rebecca Grainger, Dr. Cèsar Diaz-Torne, Prof. Ruben Burgos-Vargas, Prof. Bob Terkeltaub, Dr. Ted Mikuls, and Dr. Jiunn-Horng Chen.
1. Harrold LR, Saag KG, Yood RA, Mikuls TR, Andrade SE, Fouayzi H, et al. Validity of gout diagnoses in administrative data. Arthritis Care Res 2007;57:103-8.
2. Zhang W, Doherty M, Pascual E, Bardin T, Barskova V, Conaghan P, et al. EULAR evidence based recommendations for gout. Part I: Diagnosis. Report of a task force of the Standing Committee for International Clinical Studies Including Therapeutics (ESCISIT). Ann Rheum Dis 2006;65:1301-11.
3. Taylor WJ. Diagnosis of gout: Considering clinical and research settings. Curr Rheumatol Rev 2011;7:97-105.
4. Malik A, Schumacher R, Dinnella JE, Clayburne GM. Clinical diagnostic criteria for gout — Comparison with the gold standard of synovial fluid crystal analysis. J Clin Rheumatol 2009;15:22–4. 5. Decker JL. Report from the Subcommittee on Diagnostic Criteria for Gout. In: Bennett PH, Wood PHN, editors. Population studies of the rheumatic diseases — Proceedings of the Third International Symposium New York, June 5-10, 1966. Amsterdam: Excerpta Medica Foundation; 1968:385-7.
6. Wallace SL, Robinson H, Masi AT, Decker JL, McCarty DJ, Yu TF. Preliminary criteria for the classification of the acute arthritis of primary gout. Arthritis Rheum 1977;20:895-900.
7. Janssens H, Janssen M, van de Lisdonk EH, Fransen J, van Riel P. Limited validity of the American College of Rheumatology criteria for classifying patients with gout in primary care. Ann Rheum Dis 2010;69:1255–6.
8. Winnard D, Wright C, Taylor WJ, Jackson G, Te Karu L, Gow P, et al. National prevalence of gout derived from administrative health data in Aotearoa New Zealand. Rheumatology 2012;51:901-9. 9. Choi HK, Ford ES, Li C, Curhan G. Prevalence of the metabolic syndrome in patients with gout: The Third National Health and Nutrition Examination Survey. Arthritis Rheum 2007;57:109-15.
10. Gaffo AL, Edwards NL, Saag KG. Gout. Hyperuricemia and cardiovascular disease: How strong is the evidence for a causal link? Arthritis Res Ther 2009;11:240.
11. Vazquez-Mellado J, Hernandez-Cuevas CB, Alvarez Hernandez E, Ventura-Rios L, Pelaez-Ballestas I, Casasola-Vargas J, et al. The diagnostic value of the proposal for clinical gout diagnosis (CGD). Clin Rheumatol 2012;31:429-34.
C, Janssen M. A diagnostic rule for acute gouty arthritis in primary care without joint fluid analysis. Arch Intern Med 2010;170:1120-6. 13. Kirwan JR, Tugwell PS. Overview of the patient perspective at OMERACT 10 — Conceptualizing methods for developing patient-reported outcomes. J Rheumatol 2011;38:1699-701.
14. Kirwan JR, Newman S, Tugwell PS, Wells GA. Patient perspective on outcomes in rheumatology — A position paper for OMERACT 9. J Rheumatol 2009;36:2067-70.
15. Ottewill R, Demain S, Ellis-Hill C, Greenyer CH, Kileff J. An expert patient-led approach to learning and teaching: The case of physiotherapy. Med Teach 2006;28:e120-6.
16. Abraham C, Gardner B. What psychological and behaviour changes are initiated by ‘expert patient’ training and what training
techniques are most helpful? Psychol Health 2009;24:1153-65. 17. Schumacher HR Jr, Chen LX, Mandell BF. The time has come to incorporate more teaching and formalised assessment of skills in synovial fluid analysis into rheumatology training programs. Arthritis Care Res 2012;64:1271-3.
18. Fitch K, Bernstein SJ, Aguilar MS, Burnand B, LaCalle JR, Lazaro P, et al. The RAND/UCLA appropriateness method user’s manual. Santa Monica, CA: RAND; 2001.
19. Graham B, Regher G, Wright J. Delphi as a method to establish consensus for diagnostic criteria. J Clin Epidemiol 2003;56:1150-6.
