Background: Juvenile idiopathic arthritis (JIA) has a generalized impact on physical function, thus functional capacity (FC) is one of the main outcome measures in this population. FC in these children can be evaluated using self-referential questionnaires, being the Childhood Health Assessment Questionnaire (C-HAQ) one of the most widely used, and with observational tests such as the Functional Capacity Scale (CAPFUN). The psychometric properties of this scale have not been tested yet.

Objectives: To evaluate the psychometric properties of the CAPFUN in Argentine children with JIA according to Consesus-based Standards for the selection of health Measurements Instruments (COSMIN).

Methods: Psychometric properties measurement study. Children were admitted between January 2018 and November 2019. Inclusion criteria: age between 6 and 16 years, diagnosis of JIA, complete physical therapy evaluation, including: CAPFUN scale, C-HAQ questionnaire, Visual Numerical Scale of patient´s self-perceived Functional Capacity (VNSFC), Global Rating of Change (GROC) scale and Steinbrocker Scale. Subjects who discontinued follow-up were excluded. The psychometric properties evaluated were reliability, considering internal consistency (Cronbach’s alpha); test-retest, only in stable subjects between T0 (first evaluation) and T1 (15 days later), using the GROC by an ordinal scale with three categories. The family member was asked, “How is your child doing since the last evaluation: better, the same or worse?”. They were considered stable if they answered “The same”; the Intraclass Correlation Coefficient -ICC-, with their respective 95% confidence intervals and the measurement error (the differences of each measurement between T0 and T1 versus the mean of both measurements were expressed in Bland Altman plots with the corresponding 95% agreement limits); construct validity, assessing structural validity (through an exploratory factor analysis -EFA-, including all variables with a correlation coefficient greater than 0.3 and inter-item correlations no greater than 0.65) and hypothesis testing (CAPFUN scores were correlated with the C-HAQ, Steinbrocker and VNSFC. Pearson’s or Spearman’s correlation coefficient were used, as appropriate. Correlation coefficients >0.50, between 0.30 and 0.50 and <0. 30 were considered strong, moderate, and poor, respectively. We also measured responsiveness (by concurrent longitudinal validity correlating changes in CAPFUN scores with the C-HAQ, Steinbrocker and VNSFC, between T0 and T2 at 12 months) and interpretability, assessing the ceiling and floor effect (considered present if more than 15% of participants obtained the minimum or maximum value of the questionnaire at T0) and the minimum clinically important change -MCIC- (we only analyzed those subjects who were stable or improved at T2 according to GROC scale).

Results: 114 children were analyzed. Internal consistency was acceptable with a Cronbach alpha of 0.93. Test-retest reliability showed a CCI of 0.999 (95% CI: 0.998 - 0.999). In the Bland-Altman plot for the measurement error analysis, the observations were uniformly dispersed and no funnel effect was observed. Therefore, no systematic change in variability with respect to the score value was identified. For structural validity, the EFA was acceptable, demonstrating the reliability of the instrument. All correlation coefficients were strong and all established hypotheses were tested. The floor and ceiling effect was not considered present. The MCIC was 0.037 points [AUC 0.90 (95% CI 0.83 to 0.91)].

Conclusion: The CAPFUN demonstrated acceptable values of reliability, construct validity, responsiveness, and interpretability. Key words: juvenile idiopathic arthritis, physiotherapy, validation study, pediatrics, rheumatology, minimal clinically important difference.

REFERENCES: NIL.

Acknowledgements: Special thanks to Mauro Andreu, Soledad Gomez y Mauro Andreu.

Disclosure of Interests: None declared.