OBJECTIVE: To identify a core set of outcome variables for the assessment of children with juvenile arthritis (JA), to use the core set to develop a definition of improvement to determine whether individual patients demonstrate clinically important improvement, and to promote this definition as a single efficacy measure in JA clinical trials. METHODS: A core set of outcome variables was established using a combination of statistical and consensus formation techniques. Variables in the core set consisted of 1) physician global assessment of disease activity; 2) parent/patient assessment of overall well-being; 3) functional ability; 4) number of joints with active arthritis; 5) number of joints with limited range of motion; and 6) erythrocyte sedimentation rate. To establish a definition of improvement using this core set, 21 pediatric rheumatologists from 14 countries met, and, using consensus formation techniques, scored each of 72 patient profiles as improved or not improved. Using the physicians' consensus as the gold standard, the chi-square, sensitivity, and specificity were calculated for each of 240 possible definitions of improvement. Definitions with sensitivity or specificity of <80% were eliminated. The ability of the remaining definitions to discriminate between the effects of active agent and those of placebo, using actual trial data, was then observed. Each definition was also ranked for face validity, and the sum of the ranks was then multiplied by the kappa statistic. RESULTS: The definition of improvement with the highest final score was as follows: at least 30% improvement from baseline in 3 of any 6 variables in the core set, with no more than 1 of the remaining variables worsening by >30%. The second highest scoring definition was closely related to the first; the third highest was similar to the Paulus criteria used in adult rheumatoid arthritis trials, except with different variables. This indicates convergent validity of the process used. CONCLUSION: We propose a definition of improvement for JA. Use of a uniform definition will help standardize the conduct and reporting of clinical trials, and should help practitioners decide if a child with JA has responded adequately to therapy. We are in the process of prospectively validating this definition and several others that scored highly.
AimThe purpose of this project was to develop and promulgate a core set of endpoints that can be used in future clinical trials in children with JA, to describe the amount of change in each variable that is considered clinically important, and to use the entire core set to develop a definition of improvement to aid in the classification of individual patients as either improved or not improved.
ContributorsGiannini, E. H. Ruperto, N. Ravelli, A. Lovell, D. J. Felson, D. T. Martini, A.
Disease Category: Child health, Rheumatology
Disease Name: Juvenile arthritis
Age Range: 0 - 18
Sex: Either
Nature of Intervention: Any
- Academic research representatives
- Clinical experts
- Conference participants
- Individuals with a known interest
- Methodologists
- Prioritising
- COS for clinical trials or clinical research
- COS for practice
- Consensus meeting
- Nominal group technique (NGT)
- Survey
1. Selection of the preliminary core set of response variables. In 1993, a 16 member Advisory Council was formed. Prior to convening a meeting of this committee, a brief questionnaire was mailed to each member asking about response variables used when assessing clinical response in patients with JA. The questionnaire listed 25 variables that had been used in the reporting of JA clinical trials and asked the physicians to rank-order their top 6 choices. An ‘other’ category was provided to add variables not included in the list. Variables were ranked in order of priority votes received and a total of 16 variables received votes. These became known as the candidate variables for inclusion in the core set.
2. The performance characteristics (validity, reliability, sensitivity to change, redundancy) of the candidate variables were reviewed using existing literature.
3. In 1994, a conference of the advisory council was convened in Florida. Using NGT, attendees developed a preliminary set of response variables.
4. Ascertainment of international consensus on the core set of variables was achieved by sending out a survey questionnaire to a larger, more international sample of practitioners. When variables were ranked according to their priority score, the same 6 chosen in Florida were scored highest in the larger survey.
5. Development and selection of a definition of improvement, a second conference in Italy in May 1996. The overall goal of the meeting was to decide upon a preliminary definition of improvement based on the core set of endpoints, using a combination of statistical and consensus techniques.