Objectives: To reach consensus with recommendations made by an OMERACT Special Interest Group (SIG).
Methods: Rheumatologists and industry representatives interested in gout rated and clarified, in three iterations, the importance of domains proposed by the OMERACT SIG for use in acute and chronic gout intervention studies. Consensus was defined as a value of less than 1 of the UCLA/RAND disagreement index.
Results: There were 33 respondents (61% response rate); all agreed the initial items were necessary, except “total body urate pool”. Additional domains were suggested and clarification sought for defining “joint inflammation” and “musculoskeletal function”. Items that demonstrated no clear decision were re-rated in the final iteration. There were six highly rated items (rating 1–2) with four slightly lower rating items (rating 3) for acute gout; and 11 highly rated items with eight slightly lower ratings for chronic gout.
Conclusions: Consensus is that the following domains be considered mandatory for acute gout studies: pain, joint swelling, joint tenderness, patient global, physician global, functional disability; and for chronic gout studies: serum urate, gout flares, tophus regression, health-related quality of life, functional disability, pain, patient global, physician global, work disability and joint inflammation. Several additional domains were considered discretionary.
Statistics from Altmetric.com
Gout is a systemic metabolic disease manifested by hyperuricaemia, acute and chronic arthritis, monosodium urate crystal deposits in connective tissue producing tophi, and uric acid nephrolithiasis. It is the most common inflammatory arthritis in men. It may be associated with the metabolic syndrome and with an increased risk of cardiovascular disease.
There has been renewed interest in the treatment of gout with recent reported intervention studies of new agents, including etoricoxib,1 lumiracoxib2 in acute gout or attacks of gout and febuxostat,3 pegylated-uricase4 in chronic gout. These studies have highlighted the relative paucity of validated outcome measures with which to judge efficacy. Gout has been discussed at OMERACT 75 and OMERACT 86 and a preliminary list of relevant domains has been developed by a Special Interest Group (SIG). These domains include: pain, inflammation, function, patient global, safety for acute gout studies; serum urate, gout flare recurrence, tophus regression, joint damage imaging, health-related quality of life (HRQOL), musculoskeletal function, patient global assessment, participation, safety and tolerability for chronic gout studies. A key aim of treatment for chronic gout is likely to include elimination of the deposited urate crystals.
The purpose of this study was to formally determine the extent of consensus with these recommendations using a Delphi approach. This technique iteratively and anonymously solicits opinions from participants, who have the opportunity to revise their opinion in the light of feedback on the opinion of the group as a whole.7
Fifty-four rheumatologists and industry representatives interested in gout were identified from the OMERACT Gout mailing list and a previous Delphi exercise that examined the question of gout flare.8 Respondents were asked to rate the importance of measuring domains in acute gout and chronic gout intervention studies on a seven-point scale (1 = definitely necessary to 7 = definitely not necessary), using a web-based questionnaire. Studies for acute gout refer to interventions that aim to limit the severity or duration of an acute gout flare. The precise definition of acute gout or gout flare is the subject of an ongoing study under the OMERACT umbrella.9 Studies for chronic gout refer to interventions that aim to prevent recurrent episodes of acute gout or limit the overall impact of persistent symptoms, disease activity or functional consequences of persistent disease over an extended time period. The domains suggested by the OMERACT SIG were used for the first iteration, supplemented by domains of “physician global assessment” and “work disability” for acute gout and “work disability”, “physician global assessment”, “joint inflammation”, “pain” and “total body urate pool” for chronic gout studies. These extra domains were selected on the basis of literature review and expert opinion. Additional domains felt to be of importance were also solicited from Delphi respondents. Consensus was defined by the UCLA/RAND disagreement index, whereby values of less than 1 indicated agreement.10 This index is essentially calculated from the 30th and 70th percentile of the respondents’ ratings, adjusted for symmetry between the central point of the interpercentile range and the mid-point of the rating scale. The adjustment factor was derived from experimental work that compared different definitions of what constituted “disagreement” among panels of various sizes.
New items, re-worded items, and items for which there was disagreement and/or median rating of 4 (neither agreement nor disagreement) were re-rated in the second iteration. In the final (third) iteration, no new items were introduced and only items for which there was disagreement and/or median rating of 4 were re-rated. Reminders were emailed at 2 weeks following the start of each round and potential respondents were given a further week to respond before being declared a non-respondent.
According to the principles of the Delphi technique,7 respondents were not known to each other during the survey and the group response (median and interpercentile range) was made known to respondents at the time any item was rated for a second or third time.
The study protocol was reviewed by the New Zealand Health and Disability Central Region Ethics Committee.
There were 33 respondents (61% response rate) to the first survey; 54% were from North America, 18% from Europe and 27% from Asia-Pacific. Of these, 29 (88%) responded to the second and third rounds of the survey. There were three participants from industry, all of whom had first-hand knowledge of design and conduct of pharmaceutical trials in gout.
All initial items were agreed as important (median rating 1–3), except for “total body urate pool” (median rating 4). Additional domains were suggested and clarification sought for the meaning of “joint inflammation” and “musculoskeletal function”. Subsequently the new items and “total body urate pool” were (re)rated in the second and third (final) iteration.
For acute gout (table 1) the final list contained six items (median rating 1–2) and four items (median rating 3); and for chronic gout 11 items (median rating 1–2) and eight items (median rating 4) (table 2). Additional domains that this group of respondents felt were important for studies of chronic gout were health care utilisation, costs, patient utilities and comorbidities. Functional status should be assessed in terms of joint function (range of motion), activity limitation (disability) and participation restriction. It remained unclear whether an acute phase marker should be a core domain for studies of chronic gout.
This Delphi exercise has confirmed that the domains identified by the OMERACT SIG are important for studies of acute and chronic gout. However, additional domains were also seen as important, creating a list of outcome areas that might be too unwieldy for clinical trials. In addition, this exercise has clarified the meaning of “function” and “inflammation” listed in the OMERACT SIG recommendation, creating further domains. It is important to emphasise that identification of monosodium urate crystals was not considered sufficiently important by participants to be included in the final list, probably as it was a necessary prerequisite that the recommendations only applied to patients who had a confirmed diagnosis of gout.
One approach to resolving the tension between practical feasibility and desire to measure everything that is relevant, is to consider listing some items as mandatory for clinical trials and some items as discretionary (while still being seen as important). This approach has been successful for psoriatic arthritis,11 lupus,12 osteoporosis13 and osteoarthritis.14 Based on the results from this Delphi exercise, it is proposed that outcome domains for gout studies be organised into the groups shown in fig 1, and that this proposal be formally ratified at the OMERACT 9 meeting during 2008. Domains that scored a median of 1 or 2 were placed in the inner ellipse, domains that scored a median of 3 into the next ellipse and those that scored a median of 4 into the outermost ellipse. For all studies, it is mandated that safety and tolerability of investigational products also be assessed. Once core domains are agreed, a subsequent task is to identify or develop appropriate tools to measure each domain.
This exercise could be combined with the development of core sets for the International Classification of Functioning, Disability and Health (ICF), where core sets projects aim to identify those factors important for specific health states, including personal perspectives on activities and participation to complement the views of health professionals.15
It is important to recognise that the method or instrument to measure these domains was not addressed in this Delphi exercise. In particular, the rating of the importance of the domain by respondents reflected their own expert judgement, rather than the extent to which the domain could be measured in a way that satisfies the OMERACT filter (truth, discrimination, feasibility).16 Specific outcome measures used in clinical trials of gout treatment were reviewed recently. This review describes currently used measures in terms of the OMERACT filter.9 Other studies that address the measurement properties of instruments for some of these domains are currently in progress. It may be necessary to modify fig 1 in light of currently available tools for outcome measurement in gout, and this is likely to be a key area of discussion at the OMERACT 9 meeting. Even so, this Delphi exercise does point the way to where researchers should focus attention on producing validated outcome tools for gout studies.
A further limitation to this study is represented by the selected group of Delphi participants and non-response rate. This might limit the validity of the final recommendations; in particular, patients with gout were not invited to participate in this exercise. None the less, the initial response rate was actually very satisfactory in comparison with other similar Delphi exercises and the very broad geographical coverage of the panel does suggest adequate international representation.
The Delphi panel participants were: Alexander So, Lausanne; Arthur Kavanaugh, San Diego; Brian Mandel, Cleveland; Cesar Diaz-Torne, Barcelona; Connie Chen, Pfizer, New York; Daniel Clegg, Salt Lake City; Daniel Furst, Los Angeles; Dinesh Khanna, Los Angeles; Fiona McQueen, Auckland; H. Ralph Schumacher, Philadelphia; Herb Baraf, Washington; Hisushi Yamanaka, Tokyo; Hyon Choi, Vancouver; Janitzia Vazquez-Mellado, Mexico City; Jasvinder Singh, Minneapolis; Lee Simon, Boston; Lisa Stamp, Christchurch; Maarten Boers, Amsterdam; Michael Becker, Chicago; Michael Doherty, Nottingham; Nancy Joseph-Ridge, TAP Pharmaceuticals, Lake Forest; Naomi Schlesinger, New Jersey; Nicola Dalbeth, Auckland; Patricia Macdonald, TAP Pharmaceuticals, Lake Forest; Peter Chapman, Christchurch; Peter Gow, Auckland; Rebecca Grainger, Wellington; Ted Mikuls, Omaha; Thomas Bardin, Paris; Vibeke Strand, Pala Alto; Victoria Barskova, Moscow; Worawit Louthrenoo, Chiang Mai; Xuejun Zeng, Beijing.
For numbered affiliations see end of article
Competing interests: None.
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.