Despite the significant health impact of gout, there is no consensus on management standards. To guide physician practice, we sought to develop quality of care indicators for gout management.
Despite the significant health impact of gout, there is no consensus on management standards. To guide physician practice, we sought to develop quality of care indicators for gout management.
A systematic literature review of gout therapy was performed using the Medline database. Two abstractors independently reviewed each of the articles for relevance and satisfaction of minimal inclusion criteria. Based on the review of the literature, 11 preliminary quality indicators were developed and then reviewed and refined by an initial feasibility panel of community and academic rheumatologists. A twelfth indicator was added at the request of the first panel. Using a modification of the RAND/University of California at Los Angeles appropriateness method (bridging teleconference and white-board Internet technology were added), a second expert panel rated each of the proposed indicators for validity using a 9-point scale, in which ratings of 1–3, 4–6, and 7–9 were considered “invalid,” “indeterminate,” and “highly valid,” respectively. Indicators were considered valid if the median panel rating was ≥7 and there was no evidence of panel disagreement (defined to occur when 2 of 6 panelists provided a validity rating of 1–3 and 2 panelists provided a validity rating of 7–9).
Ten of the 12 draft indicators were rated to be valid by our second expert panel. Validated indicators pertained to 1) the use of urate-lowering medications in chronic gout, 2) the use of antiinflammatory drugs, and 3) counseling on lifestyle modifications.
Using a combination of evidence and expert opinion, 10 indicators for quality of gout care were developed. These indicators represent an important initial step in quality improvement initiatives for gout care.
Preliminary evidence suggests that medical errors in the treatment of gout are common. An older age at onset of gout and the presence of serious comorbid conditions render patients vulnerable to medication-related errors and substandard quality of care. Indomethacin, a nonsteroidal antiinflammatory drug (NSAID) routinely used in gout treatment, is among the most commonly prescribed inappropriate medications among the elderly (1). In a prospective investigation involving a large cohort of elderly patients who presented to an emergency room (1), gout was among the treatment indications most commonly associated with selection of inappropriate medications. Such medication-related errors have been reported as a complication in more than one-fourth of all orders for intravenous colchicine administered to inpatients (2) and approximately one-half of orders for administration of allopurinol (3), both of which are drugs that are commonly used in the treatment of gout.
Although gout has a substantial health impact and although its treatment appears to be characterized by preventable medical errors, quality of care in gout has not been well studied. A barrier to such research has been a lack of consensus regarding the standards of care in gout therapy. To address this issue, we developed quality indicators for gout management using a comprehensive literature search and an expert panel process.
Prior to formulating the initial draft indicators, we performed a systematic review of the Medline database (from 1966 to 2001) for all articles pertaining to gout therapy. The titles and/or abstracts of all citations were reviewed for relevance. We reviewed the reference lists of all relevant articles for other pertinent reports and a clinical content expert provided citations for older references. Two abstractors (TRM and KGS or JJA) independently reviewed each of the articles for relevance and satisfaction of minimal inclusion criteria. Single case reports, non–English language articles, and reviews, as well as animal and in vitro studies were excluded from the final reference database. Additionally, articles were excluded if the study's focus was not primarily on therapeutics or did not involve subjects with gout. Discordance between the primary reviewers was adjudicated by consensus of all 3 abstractors.
Twelve draft indicators that reflected processes of care were developed, and these could be either directly or indirectly linked to significant clinical outcomes within the reference database. Using methods described by Shekelle et al (4), potential explicit process measures were constructed using an IF—THEN—BECAUSE format, with IF referring to the clinical characteristics that qualify individuals as eligible for the indicator, THEN indicating the actual process that should or should not be performed, and BECAUSE describing the anticipated health impact of the indicator. In contrast to treatment guidelines that are promulgated as best practice strategies, these indicators represent minimal acceptable standards of care in gout management to which health-care providers and health-care organizations could be held accountable. A monograph consisting of the draft process indicators and an evidence review supporting each indicator was circulated to members of our expert panels for their review 7–14 days prior to the panel meetings.
We convened 2 separate panels. Panel members and their affiliations are shown in Table 1. Two rounds of indicator ratings were performed with each panel, the first by mail (or facsimile) and the second at the time the panel was formally convened. The first group, consisting of 2 academic and 2 private practice rheumatologists, met on March 1, 2002. The group reviewed and refined 11 initial draft indicators. At the request of the initial panel, an additional process indicator was added (process indicator 7 in Table 2) for review by the second panel.
|Panel member||Institutional affiliation|
|John S. Cowdery, MD||University of Iowa|
|Richard E. Jones, MD, PhD||Private practice, Tuscaloosa, AL|
|Gerald F. Moore, MD||University of Nebraska|
|Paul Utrie, MD||Private practice, Appleton, WI|
|Gene V. Ball, MD||University of Alabama at Birmingham|
|S. Louis Bridges, Jr., MD, PhD||University of Alabama at Birmingham|
|John R. Kirwan, MD||Bristol Royal Infirmary, UK|
|Brian F. Mandell, MD, PhD||Cleveland Clinic Foundation|
|H. Ralph Schumacher, Jr., MD||University of Pennsylvania|
|Brian L. Strom, MD, PhD||University of Pennsylvania|
|Topic area, process indicators||Median panel rating (range)||Supporting references|
|Use of uric acid–lowering therapy|
|1.IF a gout patient is receiving an initial prescription for allopurinol and has significant renal impairment (defined as a serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute), THEN the initial daily allopurinol dose should be <300 mg per day, BECAUSE the risk of allopurinol-related toxicity is increased in the presence of significant renal impairment in gout patients given a daily allopurinol dose equal to or exceeding 300 mg.||8.5 (3–9)||32–43|
|2.IF a gout patient is given a prescription for xanthine oxidase inhibitor in the setting of required therapy with either 1) azathioprine (Imuran) or 2) 6-MP, THEN the dose of azathioprine/6-MP should be reduced by a minimum of 50%, BECAUSE concurrent use of a xanthine oxidase inhibitor leads to a substantial increase in serum levels of azathioprine (and 6-MP) and increases the risk for severe drug-related myelosuppression.||8.5 (5–9)||44–48|
|3.IF a patient with tophaceous gout is given an initial prescription for a urate-lowering medication (xanthine oxidase inhibitor, probenecid, or sulfinpyrazone) and lacks both 1) significant renal impairment (a serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute) and 2) peptic ulcer disease, THEN a prophylactic antiinflammatory agent (colchicine or NSAID) should be given concomitantly, BECAUSE prophylactic antiinflammatory therapy reduces the risk of rebound gout attacks, which frequently follow the initiation of urate-lowering therapy.||7.5 (5–9)||17,49–61|
|4.IF a patient has asymptomatic hyperuricemia characterized by 1) no prior history of gouty arthritis or tophaceous deposits and 2) no prior history of nephrolithiasis or hyperuricosuria and 3) no ongoing treatment of malignancy, THEN urate-lowering therapies should not be initiated, BECAUSE there is currently no widely accepted indication for the treatment of asymptomatic hyperuricemia.||8 (7–9)||18, 32, 35, 38, 39, 42, 62, 63|
|5.IF a gout patient is started on urate-lowering therapy and has either 1) a history of nephrolithiasis or 2) significant renal insufficiency (serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute), THEN a xanthine oxidase inhibitor should be started as the initial urate-lowering medication rather than a uricosuric agent (probenecid or sulfinpyrazone), BECAUSE in contrast to xanthine oxidase inhibitors, uricosuric agents increase the renal excretion of urate, enhancing the risk of nephrolithiasis, and may have diminished efficacy in the context of significant renal insufficiency.||8.5 (5–9)||18, 19, 23, 33, 34, 49, 50, 52, 55, 56, 58, 59, 64–68|
|6.IF a patient has hyperuricemia and gouty arthritis characterized by any of the following clinical characteristics, 1) tophaceous deposits, 2) gouty erosive changes on radiographs, or 3) gout attack frequency of ≥2 attacks per year, THEN the patient should be offered treatment with a urate-lowering drug, BECAUSE urate-lowering drugs have been well-tolerated and effective in decreasing the attack frequency and disease severity for those with severe gout.||8 (7–9)||22–24,49–52,54–60,64–77|
|7.IF a gout patient is given a prescription for a xanthine oxidase inhibitor, THEN a serum urate level should be checked at least once during the first 6 months of continued use, BECAUSE periodic serum urate measurements are required for appropriate dose adjustments of xanthine oxidase inhibitors (escalations or reductions).||9 (6–9)||23, 32, 51, 52, 54, 58, 72, 77–79|
|8.IF a patient is diagnosed with gout and has either 1) obesity (defined as a body mass index ≥28 kg/m2) or 2) frequent alcohol use (≥1 alcoholic beverage per day), THEN as part of their overall therapy, patients should be advised on the importance of weight loss and/or decreased alcohol use, BECAUSE weight loss and reduction of alcohol intake may be beneficial components of gout therapy.||8.5 (4–9)||80, 81|
|Use of antiinflammatory agents|
|9.IF a patient has acute gouty arthritis and lacks both of the relative contraindications to gout treatment, 1) significant renal impairment (a serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute) and 2) peptic ulcer disease, THEN the patient should be treated with an antiinflammatory agent to include one of the following: 1) NSAID, 2) ACTH or glucocorticoid (either systemic or intraarticular administration), or 3) colchicine, BECAUSE antiinflammatory agents have been shown to be both effective and well-tolerated for the short-term treatment of acute gout. Patients with renal impairment and a history of peptic ulcer disease may be at a higher risk for gout medication toxicity.||8.5 (5–9)||16, 82–110|
|10.IF a gout patient receives long-term prophylactic oral colchicine (defined as a minimum daily dose of 0.5 mg for a duration of 6 months or longer) and has significant renal insufficiency (a serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute), THEN a complete blood cell count and creatine kinase should be evaluated a minimum of one time for every 6 months of continued use, BECAUSE the risk of colchicine-related myopathy and myelosuppression appears to be substantially increased in the context of reduced renal function.||7.5 (2–9)||31, 111|
For the second panel, we selected international experts in gout (all of whom had at least one sentinel review or original article in the area), an international authority in pharmacoepidemiology and quality research, and a British rheumatologist with expertise in quantitative research. At their meeting on March 18, 2002, members of the second panel further reviewed, refined, and voted on the validity of the 12 proposed indicators using a modified version of the RAND/University of California at Los Angeles (UCLA) appropriateness method. This formal group method uses 2 rounds of ratings by experts to determine the validity of proposed indicators. Typically, the first round of voting is done by mail in advance of a face-to-face meeting at which time all indicators are discussed by panelists and re-rated. All voting is performed confidentially and all votes are weighted equally. The RAND/UCLA appropriateness method has been shown to be highly reproducible (5) and to have content, construct, and predictive validity (6–9).
For this study, the panel meetings were conducted with the use of a bridging teleconference call aided by white-board Internet technology (CE City, Inc., Pittsburgh, PA) that allowed for direct text visualization, real-time editing of the process indicators, and instant anonymous voting that could be viewed by each of the panelists.
After review, discussion, and refinement, the panel rated the validity of each indicator using a 9-point ordinal scale, in which ratings of 1–3, 4–6, and 7–9 were considered “invalid,” “intermediate,” and “highly valid,” respectively. After the final ratings, median scores and agreement among the panelists was determined. Agreement was defined to occur when all ratings fell into the same validity grouping (1–3 or 4–6 or 7–9), whereas disagreement was observed when at least 2 ratings fell in the lowest grouping (1–3) and at least 2 ratings fell in the highest grouping (7–9). Agreement for all other ratings was considered to be indeterminate. The process indicators were considered valid if the median rating by the expert panel was ≥7 and there was either consensus or indeterminate agreement (4).
The Medline search resulted in 310 potential articles for abstraction. After scanning the titles, 182 articles were excluded (144 articles were deemed not to be relevant and 38 reports were in a foreign language). Review of the articles' reference lists resulted in an additional 159 potentially relevant citations. Our clinical content expert provided 5 additional references, all of which predated the Medline database (publication dates prior to 1966). The search resulted in 292 potentially relevant articles relating to gout therapy. There was 88% concordance between the primary article abstractors (kappa coefficient of 0.76 for interobserver reliability). After reaching consensus, a total of 120 articles were included in the final reference database. Of the 120 articles in the final literature database, only 23 (19%) were randomized controlled trials (RCTs); the remainder consisted predominantly (n = 80 [67%]) of open-label designs or case series. The 23 RCTs included 11 studies of NSAIDs in the treatment of acute gout, 2 studies of colchicine (one for the treatment of acute gout, one for prophylaxis), and 10 studies of urate-lowering medicines in long-term management (7 involving allopurinol).
The process indicators approved by the expert panel (shown in Table 2) cover topic areas including the use of urate-lowering medicines (7 indicators), lifestyle modifications (1 indicator), and the use of antiinflammatory therapies (2 indicators). The median panel ratings and the literature supporting each of the indicators are also shown in Table 2; each of the indicators was supported to a varying degree by citations included in the final literature database.
Median validity scores were ≥7 for 10 of the 12 preliminary process indicators. There was panel agreement for 2 of the 10 final indicators (indicators 4 and 6); agreement was indeterminate for the remaining validated indicators. For the 2 draft indicators that were not rated to be valid by the expert panel, the median validity scores were <7 and there was disagreement among the panelists (Table 3).
|Topic area||Process indicators||Median panel rating (range)|
|Use of uric acid–lowering therapy||1. IF a gout patient is given an initial prescription for allopurinol, THEN the medication should be administered as a single daily dose rather than in divided doses, BECAUSE a single daily dose is equally effective as the same amount given in divided doses and, given its relative convenience of administration compared with a divided dose regimen, a single daily dose likely leads to improved adherence.||6 (1–9)|
|Use of antiinflammatory agents||2. IF a gout patient is treated with intravenous colchicine and has significant renal impairment (a serum creatinine level ≥2 mg/dl or measured/estimated creatinine clearance ≤50 ml/minute), THEN the cumulative intravenous colchicine dose for a single attack should not exceed 4 mg within a 7-day period, BECAUSE the risk–benefit ratio of intravenous colchicine appears to be unfavorable with cumulative doses >4 mg.||5.5 (1–9)|
We present the results of an explicit method for the development of process quality indicators for gout management based on the results of a systematic literature review and expert consensus. The 10 process indicators developed and rated as valid by the expert panel cover 3 domains of gout therapy: the use of antiinflammatory agents (in prophylaxis and treatment), the use of urate-lowering therapies, and counseling for lifestyle modification. These process measures are based on expert opinion and the best available evidence and may prove to be a valuable resource for researchers interested in examining the quality of health-care delivery to patients with gout.
In the absence of unequivocal literature-based support, expert review and consensus have been used in the development of treatment guidelines and process indicators for a number of other chronic conditions (10–15). Although there is debate about whether quality measures should focus on processes or clinical outcomes, there is strong rationale favoring the use of process measures in assessing the quality of gout management. Since numerous factors influence patient level outcomes in gout (e.g., comorbidity, concurrent therapies, patient compliance), our process indicators circumvent some of these confounders and identify health care provider behaviors that should affect outcomes.
The results of our systematic literature review corroborate gout as an understudied disease among musculoskeletal disorders. Despite its moderate prevalence in older adults and a significant impact on quality of life, there are relatively few published clinical trials in gout. Our search identified only 2 RCTs that examined the efficacy and tolerability of colchicine (16, 17) and 7 RCTs that examined the use of allopurinol in the management of gout (18–24).
Although there is no standard technique for the development of process indicators, several methods have been used (25–28). Our method follows the RAND experience of combining evidence with expert judgment, an approach that is increasingly being utilized in health services research (29). Nevertheless, it has been suggested that this approach may lead to a collective error in judgment or inference and can be influenced by panel composition (30). This may be particularly true with panels relatively small in size, such as those convened for the purposes of this study. Our use of an initial expert feasibility panel, convened to improve the content and wording of the quality indicators and evidence report, adds further face validity to this process.
In contrast to previously published gout-specific treatment recommendations (31, 32), the process measures rated to be valid by our expert panel were not designed to represent best practice guidelines, but rather minimal acceptable standards of care in gout management. Thus, as with quality indicators developed for other conditions, we would anticipate that adherence to these indicators would lead to a higher, albeit not necessarily optimal, quality of gout care.
Our indicators represent a necessary step in the development of quality improvement initiatives in gout care. In addition to using the indicators as a measurement tool for quality improvement efforts, the aims of future investigations will be to evaluate the degree to which health care providers and organizations adhere to these performance measures, to define the patient and provider characteristics that determine indicator adherence, and to explore associations of indicator adherence with patient level outcomes and satisfaction with gout care.
The authors wish to thank Mr. Jorge Nunez for his technical assistance in the development of these quality indicators.