PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of rheumatologyLink to Publisher's site
 
Rheumatology (Oxford). 2011 July; 50(7): 1331–1336.
Published online 2011 March 3. doi:  10.1093/rheumatology/ker023
PMCID: PMC3307519

Minimally important differences of the gout impact scale in a randomized controlled trial

Abstract

Objective. The Gout Impact Scale (GIS) is a gout-specific quality of life instrument that assesses impact of gout during an attack and impact of overall gout. The GIS has five scales and each is scored from 0 to 100 (worse health). Our objective was to assess minimally important differences (MIDs) for the GIS administered in a randomized controlled trial (RCT) assessing rilonacept vs placebo for prevention of gout flares during initiation of allopurinol therapy.

Methods. Trial subjects ( n = 83) included those with two or more gout flares (self-reported) in the past year. Of these, 73 had data for Weeks 8 vs 4 and formed the MID analysis group and were analysed irrespective of the treatment assignment. Subjects completed the GIS and seven patient-reported anchors. Subjects with a one-step change (e.g. from very poor to poor) were considered as the MID group for each anchor. The mean change in GIS scores and effect size (ES) was calculated for each anchor’s MID group. The average of these created the overall summary MID statistics for each GIS. An ES of 0.2–0.5 was considered to represent MID estimates.

Results. Trial subjects (n = 73) were males (96.0%), White (90.4%), with mean age of 50.5 years and serum uric acid of 9.0 mg/dl. The mean change score for the MID improvement group for scales ranged from −5.24 to −7.61 (0–100 scale). The ES for the MID improvement group for the four scales ranged from 0.22 to 0.38.

Conclusion. The MID estimates for GIS scales are between 5 and 8 points (0–100 scale). This information can aid in interpreting the GIS results in future gout RCTs.

Trial Registration. Clinicaltrials.gov, www.clinicaltrials.gov, NCT00610363.

Keywords: Gout assessment questionnaire, Gout impact scale, Minimally important difference, Minimal clinically important differences, Rilonacept, Clinical trial design, Health-related quality of life, Health status

Introduction

Gout is a common chronic disorder of uric acid metabolism punctuated by acute, painful arthritis attacks [1]. Both acute and chronic gout have a powerful impact on patients’ health-related quality of life (HRQOL). Historically, impact of gout on HRQOL was under recognized or trivialized. However, recently observations from clinical trials and observational studies have shown a detrimental impact of acute and chronic gout on HRQOL [2–5]. HRQOL measures are often used in monitoring outcomes in clinical encounters, assessing population health and as end points in clinical trials [6]. Disease-specific HRQOL measures are increasingly being developed to evaluate medical treatments as they provide better face validity for the population under study and at the same time are more sensitive to smaller differences and smaller changes that occur over time [6].

In this study, we examine the recently developed gout-specific HRQOL instrument, called the Gout Impact Scale (GIS). This scale was originally the Gout Impact Section of the Gout Assessment Questionnaire 2.0 (GAQ2.0) [7, 8]. The GIS allows patients to describe the impact of gout on their HRQOL, while the remaining GAQ2.0 questions allow patients to describe their gout overall (e.g. recent gout attacks, treatment, gout history and demographics). Although the GIS was found to be feasible, reliable and valid in a large cross-sectional cohort [7, 8], the minimally important differences (MIDs) of the GIS are yet to be established. MID score is defined as the smallest difference in score in the domain of interest that patients perceive as beneficial (or worse) and that would mandate, in the absence of troublesome side effects and excessive costs, a change in the patient’s management [3, 6].

It is expected that future clinical trials and observational studies in gout will incorporate the GIS. Therefore, it is important to estimate the MIDs of the GIS. The MID can help clinicians to understand whether score differences in the GIS between two treatment groups are considered clinically meaningful and also if the changes within one group over time are clinically meaningful and relevant [9].

MID estimates can be assessed using anchor- and distribution-based approaches [10]. An anchor is a clinically relevant indicator or pointer to which a patient-reported outcome change can be tied [6, 10, 11]. These measures are of clinical relevance and can be subjective, such as self-reports of change, or objective, such as clinical indicators of response to treatment (disease severity). Subjective anchors rely on an individual’s (subject’s or his/her physician’s) assessment of the patient’s condition. The distribution-based approach utilizes the statistical properties of the data generated by an instrument and generally complements the anchor-based approach in estimating an MID [12]. Our objective was to estimate the MID estimates for GIS scales using both approaches: prospective patient-reported anchors and distribution-based measure within a randomized controlled trial (RCT).

Methods

Treatment-blinded data from a 16-week RCT assessing rilonacept vs placebo for prevention of gout flares during the initiation of allopurinol therapy were used for this study. The study population was comprised of male and female patients >18 years of age with inter-critical gout (experienced no gout flare within 2 weeks of baseline). Subjects had a self-reported history of two or more gout flares in the past year, were not taking either colchicine or steroids within 1 month of enrolment and had a serum uric acid (sUA) level of >7.5 mg/dl [13, 14]. The study received ethical approval at each study site [Goodwyn (Central IRB); Biomedical Research Alliance of New York, LLC (BRANY) IRB and Duke University Health System IRB] and every patient signed an approved voluntary consent form.

The GIS contains five scales—three of these assess the impact of gout overall (Gout Concern Overall, Gout Medication Side Effects and Unmet Gout Treatment Need) and two assess the impact of gout during an attack (Gout Well-being during Attack and Gout Concern during Attack). Response options for GIS items are on a five-point ordinal scale (e.g. strongly agree to strongly disagree or all of the time to none of the time). Each GIS scale is scored from 0 to 100, with higher scores on each scale indicating worse condition or greater gout impact.

The patients completed the GIS and seven anchor questions every 4 weeks. Of seven anchor questions, Questions 1–4 asked: Because of your gout, how would you rate your physical health, quality of life, mental health and pain in the past 4 weeks? Very poor, poor, fair, good, very good or excellent. Anchor question #5 asked: Considering all the ways gout affects you, circle a number on the scale for how well you have been doing for the past 4 weeks. Answer choices ranged from 1 (no disease activity) to 10 (severe disease activity) on a 1–10 Likert scale. Question #6 asked: Circle a number on the scale indicating the severity of pain within the past 4 weeks. Answer choices ranged from 1 (no pain) to 10 (severe pain) on a 1–10 Likert scale. The last anchor question was the number of flares in the past 4 weeks, which was collected from patient diaries.

Analysis

For this analysis, the data were pooled and analysed without the knowledge of the treatment group to avoid bias. We chose mean change scores from Weeks 8 vs 4 to estimate MID, since we hypothesized that the anticipated treatment effect early in the study may lead to larger than usual improvements and may not be reflective of MID estimates. In addition, we did not choose 4 weeks vs baseline comparison, since any baseline assessments would be influenced by trial requirements such as the 2-week screening period, exclusion of all patients with an acute gout flare in the past 4 weeks and restriction of the medication use. MID was not estimated for one of the GIS scales, the Gout Medication Side Effects, because of lack of an appropriate anchor question in the trial and the low rate of side effects experienced in the clinical trial. For all anchor questions, participants with a one-step change in the positive direction (e.g. from very poor to poor or 1 point on a 10-point scale or one flare difference) were considered as the MID group for each anchor, respectively [12, 15]. The number of patients who worsened was too small to assess estimates for worsening. The mean change in the GIS scores across the seven anchor questions and effect size (ES) was calculated for each anchor’s MID group as well as the improved (greater than minimal group change), worsened and no change groups. ES is the ratio of observed change to a measure of variance (also known as signal to noise) [16]. For ES, the numerator is the mean change in the GIS scales from the Weeks 8 to 4 and the denominator is the s.d. of GIS scales at Week 4. An ES of 0.2–0.5 [17] is considered to represent MID estimates.

We also assessed MID estimates using the distribution approach [12, 18] by calculating the standard error of measurement (SEM). SEM is a measure of the precision of a test instrument and SEM was computed as SEM = σx (1_relx)1/2, where σx is the s.d. of the scale and relx is the reliability (internal consistency) of the scale or aggregate score. Cronbach alpha coefficients were calculated to determine internal consistency reliability for the four scales, and were 0.82 for Gout Concern Overall, 0.59 for Unmet Gout Treatment Need, 0.94 for Gout Well-being during Attack and 0.79 for Gout Concern during Attack.

To assess the usefulness of an anchor, previous research has recommended reporting the correlation between the anchor and the changed score: for example, a correlation of zero will make the anchor useless and a correlation coefficient of 0.30–0.35 has been suggested to be clinically meaningful [12, 19]. We assessed the association between the anchors and the changed scores for GIS scales using the Spearman’s correlation coefficient (as the anchor is an ordinal variable). The data were analysed using SPSS software. P < 0.05 was deemed to be indicative of statistical significance.

Role of the funding source

One funding source provided funds to conduct the analyses of GIS data that were collected within a clinical trial conducted by the other funding source. The funding sources participated fully with the investigators in the design and interpretation of MID analyses; however, they did not influence the decision to submit this manuscript, or in any way contribute or influence the content of the manuscript.

Results

Eighty-three patients were enrolled in the RCT. Of these, 73 had data available for Weeks 4 and 8 and formed our final cohort for this analysis. The mean (s.d.) age of participants was 50.5 (10.6) years; 90.4% White and 96.0% of patients were men (Table 1). The mean (s.d.) baseline age at first attack was 40.5 (12.0) years, the average number of attacks per year was 4.6 (3.6) and sUA was 9.0 (1.3) mg/dl. For the four GIS scales, mean (s.d.) baseline scores were 79.9 (17.2) for Gout Concern Overall, 50.1 (17.0) for Unmet Treatment Need, 52.8 (24.0) for Well-being during Attack and 53.1 (23.7) for Gout Concern during Attack (all scales on a 0–100 scale with high scores associated with worse HRQOL). The Spearman’s correlation between the anchors and changed scores in the GIS scales ranged from 0.12 to 0.26 for Gout Concern Overall, 0.06 to 0.27 for Unmet Treatment Need, 0.05 to 0.20 for Well-being during Attack and 0.04 to 0.24 for Concern during Gout Attack.

Table 1
Demographic characteristics of study participants (n = 73)

The mean change score for the MID improvement group for Gout Concern Overall was −7.16, Unmet Gout Treatment Need was −6.88, Gout Well-Being during Attack was −5.24 and Gout Concern during Attack was −7.61 (Table 2). The ES for the MID improvement group for the four scales ranged from 0.22 to 0.38. Using ES, the MID estimates were generally larger in magnitude than the no change group (ES = 0.18–0.33; Table 2). There was overlap in the estimates between the MID group and markedly improved group for each scale likely related to very small sample size in the latter group (n = 2–13). Using the SEM (distribution-based anchor), the SEM estimates ranged from 5.9 (Well-Being Impact) to 11.0 (Unmet Treatment Need; Table 2).

Table 2
MIDs of the four GIS scales

Discussion

GIS is a validated HRQOL instrument for gout [7, 8]. With a renewed interest in the development of new therapeutics in gout, assessment of HRQOL using validated measures will become increasingly important. We show that a change of 5–8 points (0–100 scale) for the four different GIS scales is the MID for improvement in an RCT of patients with gout. This estimate is similar to other HRQOL measures scored on a 0–100 scale. For example, Kosinski et al. [20] found that 5–10 points constituted MID estimates for short form (SF)-36 scales in patients with active RA. Also, a 10-point change in patient-reported outcome on a 0–100 mm visual analogue scale is considered an MID estimate for improvement [11, 21].

MID estimates provide a benchmark for the future design of gout clinical trials by helping researchers and clinicians understand whether HRQOL-score differences between two treatment groups are meaningful, or if changes within one group over time are meaningful [10]. For example, an average change of 3 points on the Gout Concern Overall scale (0–100 scale) may be statistically significant for a new treatment in a gout clinic trial, but may not be perceived as beneficial by the subjects. Thus, differences in scores smaller than the MID are considered unimportant, regardless of whether statistical significance is reached. MID can also be useful for determining sample size for future studies [22]. It is important to note that MID estimates are applicable at the group level and not at the individual level. Other statistical tests have been recommended to assess statistical significance at an individual level [11, 23].

Although we show that a change of 5–8 points in GIS scales is the MID, this should not be interpreted that a change of <5 points is not clinically important as there is an inherent uncertainty around MID estimates. Previous studies have reported this uncertainty around the MID estimates [12, 24]; hence, experts recommend using several anchors. In addition, they suggest gathering data from both observational and clinical trials to gather confidence in MID estimates [12], as it is unlikely that a single MID estimate is applicable to all patient populations: future studies should address MID estimates of GIS in other gout cohorts.

We corroborated our data by calculating ES and SEM. ES provides a uniform platform for different HRQOL instruments and explores the extent to which MID estimates are similar or vary across instruments [19]. Our study is in alignment with other studies that have shown that an ES 0.20–0.50 corresponds to the MID for a patient-reported outcome measure [12, 24]. SEM is a distribution method and complements the anchor-based MID estimates. Previously, researchers have found a close correspondence between the anchor-based approach and a criterion of one SEM [25]. In our case, we found similar MID estimates for Gout Concern Overall and Well-Being Impact scales between anchor- and distribution-based methods. For Unmet Gout Treatment Need and Gout Concern during Attack, MID estimates using anchor-based were 2–4 points lower than the SEM method indicated. Since anchor-based methods were our primary methodology [12, 26], we present our MID estimates calculated using anchor-based methods acknowledging the variability around it (as presented in the previous paragraph).

Our study has several strengths. First, our MID estimates are based on a sample of patients participating in an RCT in which gout symptoms, and thus impact on HRQOL, changed. Secondly, we prospectively incorporated anchors with an a priori aim to calculate MID estimates. Thirdly, our estimates were similar using the anchor- and distribution-based approaches, giving confidence in our estimates.

Our study had some limitations as well. Our MID estimates are based on a small number of patients; specifically, those experiencing only minimal improvement between Weeks 4 and 8 of a clinical trial with only total 73 patients. For the current analysis, we found lower correlations than recommended by experts partly caused by a relatively small sample size. This may have led to increased variability observed in MID estimates [12]. The low correlation coefficients found in our study do not invalidate the MID estimates per se, but should be considered preliminary and further validated in the future. In conclusion, the MID estimates for GIS scales are between 5 and 8 points using anchor- and distribution-based methods. This information can facilitate interpretation of GIS in future gout trials and day-to-day clinical practice and care.

An external file that holds a picture, illustration, etc.
Object name is ker023b1.jpg

Acknowledgements

P.P.K. was supported by Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grant NIAMS 1 T32 AR053463 and ACR Research and Education Foundation Clinical Investigator Fellowship Award 2009–11.

Funding: This study was supported by Takeda Pharmaceutical International, Inc. and Regeneron Pharmaceuticals.

Disclosure statement: J.A.S. has received speaker honoraria from Abbott; research and travel grants from Allergan, Takeda, Savient, Wyeth and Amgen; and consultant fees from Savient, URL pharmaceuticals and Novartis. D.K. was supported by a National Institutes of Health Award (NIAMS K23 AR053858-04). D.K. is a consultant, member of speakers’ bureau or received grants from Novartis, Takeda, Savient and URL. P.P.K. is a member of a speakers’ bureau for Takeda Pharmaceuticals and has received grant support from the ACR Research and Education Foundation Clinical Investigator Fellowship Award 2009–11. R.A.T. is a consultant for Takeda, URL and Savient. All other authors have declared no conflicts of interest.

References

1. Terkeltaub R. Update on gout: new therapeutic strategies and options. Nat Rev Rheumatol. 2010;6:30–8. [PubMed]
2. Kim SY, Choi HK. Gout and quality of life. J Rheumatol. 2009;36:865–8. [PubMed]
3. Becker MA, Schumacher HR, Benjamin KL, et al. Quality of life and disability in patients with treatment-failure gout. J Rheumatol. 2009;36:1041–8. [PubMed]
4. Khanna D, Ahmed M, Yontz D, et al. The disutility of chronic gout. Qual Life Res. 2008;17:815–22. [PubMed]
5. Singh JA, Strand V. Gout is associated with more comorbidities, poorer health related quality of life and higher health care utilization in US veterans. Ann Rheum Dis. 2008;67:1310–16. [PubMed]
6. Khanna D, Tsevat J. Health-related quality of life–an introduction. Am J Manag Care. 2007;13(Suppl 9):S218–23. [PubMed]
7. Hirsch JD, Lee SJ, Terkeltaub R, et al. Evaluation of an instrument assessing influence of gout on health-related quality of life. J Rheumatol. 2008;35:2406–14. [PubMed]
8. Lee SJ, Hirsch JD, Terkeltaub R, et al. Perceptions of disease and health-related quality of life among patients with gout. Rheumatology. 2009;48:582–6. [PMC free article] [PubMed]
9. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol. 2003;56:395–407. [PubMed]
10. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol. 2003;56:395–407. [PubMed]
11. Khanna D, Pope JE, Khanna PP, et al. The minimally important difference for the fatigue visual analog scale in patients with rheumatoid arthritis followed in an academic clinical practice. J Rheumatol. 2008;35:2339–43. [PMC free article] [PubMed]
12. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol. 2008;61:102–9. [PubMed]
13. Schumacher HR, Sundy JS, Terkeltaub R, et al. Placebo-controlled study of rilonacept for prevention of gout flares during initiation of urate lowering therapy [abstract] Ann Rheum Dis. 2009;68(Suppl 3):680. [PubMed]
14. Stahl N, Radin A, Mellis S. Rilonacept–CAPS and beyond. Ann NY Acad Sci. 2009;1182:124–34. [PubMed]
15. Khanna PP, Maranian P, Gregory J, Khanna D. The minimally important difference and patient acceptable symptom state for the Raynaud’s condition score in patients with Raynaud’s phenomenon in a large randomized controlled clinical trial. Ann Rheum Dis. 2009;69:588–81. [PMC free article] [PubMed]
16. Hays RD. Reliability and validity (including responsiveness) In: Fayers P, Hays RD, editors. Assessing quality of life in clinical trials. 2nd edition. New York: Oxford University Press; 2005. pp. 25–39.
17. Sloan JA, Cella D, Hays RD. Clinical significance of patient-reported questionnaire data: another step toward consensus. J Clin Epidemiol. 2005;58:1217–9. [PubMed]
18. Crosby RD, Kolotkin RL, Williams GR. An integrated method to determine meaningful changes in health-related quality of life. J Clin Epidemiol. 2004;57:1153–60. [PubMed]
19. Hays RD, Farivar S, Liu H. Approaches and recommendations for estimating minimally important differences for health-related quality of life measures. COPD. 2005;2:63–7. [PubMed]
20. Kosinski M, Zhao SZ, Dedhiya S, Osterhaus JT, Ware JE., Jr Determining minimally important changes in generic and disease-specific health-related quality of life questionnaires in clinical trials of rheumatoid arthritis. Arthritis Rheum. 2000;43:1478–87. [PubMed]
21. Wells G, Li T, Maxwell L, Maclean R, Tugwell P. Determining the minimal clinically important differences in activity, fatigue, and sleep quality in patients with rheumatoid arthritis. J Rheumatol. 2007;34:280–9. [PubMed]
22. Wyrwich KW, Bullinger M, Aaronson N, Hays RD, Patrick DL, Symonds T. Estimating clinically significant differences in quality of life outcomes. Qual Life Res. 2005;14:285–95. [PubMed]
23. Hays RD, Brodsky M, Johnston MF, Spritzer KL, Hui KK. Evaluating the statistical significance of health-related quality-of-life change in individual patients. Eval Health Prof. 2005;28:160–71. [PubMed]
24. Khanna D, Furst DE, Hays RD, et al. Minimally important difference in diffuse systemic sclerosis: results from the D-penicillamine study. Ann Rheum Dis. 2006;65:1325–9. [PMC free article] [PubMed]
25. Wyrwich KW, Tierney WM, Wolinsky FD. Further evidence supporting an SEM-based criterion for identifying meaningful intra-individual changes in health-related quality of life. J Clin Epidemiol. 1999;52:861–73. [PubMed]
26. Revicki DA, Cella D, Hays RD, Sloan JA, Lenderking WR, Aaronson NK. Responsiveness and minimal important differences for patient reported outcomes. Health Qual Life Outcomes. 2006;4:70. [PMC free article] [PubMed]

Articles from Rheumatology (Oxford, England) are provided here courtesy of Oxford University Press