34 Data and statistical analysis In Table 1, the data of both control groups 5,11 have been combined and compared with the recruited patients with OI. Only the first control group11 was used for the data presented in the Supplementary Appendix. For each age category, a comparison was made with the age-matched control patients to test if the null-hypothesis (no differences between OI and controls) could be rejected. Then, the OI types were reciprocally compared. To calculate a Δ score, the score of the youngest patient group was subtracted from the eldest patient group. As the oldest OI type 3 group consisted only of three individuals, the Δ score was not calculated. Given that the questionnaire score cannot be reliably estimated for participants with extreme scores, floor and ceiling effects were examined (Table 1). Variables were tested for normal distribution with the Kolmogorov–Smirnov test, Shapiro–Wilk test, and q-q plots. Means and SDs were given for normally distributed continuous variables. Non-normally distributed continuous variables were presented as median, interquartile range (IQR). Differences in means comparing patients with OI with the controls were in normally distributed data tested using the summary independent sample t tests and in not normally distributed data tested with the one-sample Wilcoxon signed-rank test. Comparisons between OI types of different ages were done using ANOVA in normally distributed data, and with independent-samples Kruskal-Wallis tests with Dunn’s comparison for post hoc testing in not normally distributed data. A two-sided p value of 0.05 was considered significant. Significance values for comparison between OI types have been adjusted by the Bonferroni correction for multiple testing. Significance values for comparing patients with OI with controls are presented with three decimals for adequate interpretation. Analyses were performed using SPSS 25 (SPSS, Inc., Chicago, IL, USA) for Windows. We did not assess separately modifiers of QoL such as fracture history, scoliosis, and pulmonary function. Results Clinical characteristics A total of 372 patients with OI were identified for participation in the current study. Fifty patients were excluded as their SF-36 questionnaires were unavailable. Therefore, 322 patients were available for analysis.
RkJQdWJsaXNoZXIy MTk4NDMw