Aim: The reliability evaluation of the Italian version of the Scoliosis Research Society-22 Patient Questionnaire (SRS-22-I), administered to adolescents with mild vertebral deformities, is evaluated.
Study design: forward-backward translation, pretest, final version according to a focus-group evaluation, and 1 week test/retest.
Population: pretest: 35 subjects (22 females), age range 8.5-19 years, 28 idiopathic scoliosis (17 degrees +/-7 degrees ), 7 hyperkyphosis (54 degrees +/-4 degrees ); test/retest: 20 subjects (11 females), age range 12-17.5 years, 15 idiopathic scoliosis (16 degrees +/-8 degrees ), 5 hyperkyphosis (55 degrees +/-5 degrees ).
Statistical analysis: Spearman rank test, percent of agreement. Statistical significance: 0.05. Software: Statgraphics 3.0. We calculated the results of the questionnaire and performed the statistical analysis using non parametric test because of the charcteristics of the data: this differ from previously published
Results: Response rate was 100%. Total score: median 4 (range 3-5); results for different domains (median, range): function/activity 4 (3-5), pain 5 (3-5), self image/ appearance 3 (2-5), mental health 4 (3-5), satisfaction with management 4 (2-5). Time required to answer the questionnaire ranged from 5 to 20'; for 12 to 17.5 years old subjects from 5 to 10'. pretest showed difficulties with questions on pain; the questionnaire was changed accordingly. Spearman's rho ranged from 0.42 to 1, not significantly different from null value for questions 12 and 21. Percent of agreement (evaluated only for question 11a) was 100%. The ceiling and floor effects have been found high.
Conclusions: The SRS-22-I was found to be reliable for young patients with mild vertebral deformities of different type. Lack of reliability for questions 12 and 21 should be better understood. Until now the psychometric properties of SRS-22 questionnaire have been evaluated only in USA populations, and in idiopathic scoliosis. Moreover, only one study has been proposed including patients with mild idiopathic scoliosis. This, together with the non-parametric statistical analysis used, could explain the high ceiling and floor effects we found in our data.