STATISTICAL ANALYSIS
All analyzes of the study were performed using SPSS 20.0 package
program. Nominal scale (categorical) variables were presented using
frequency and percentage; proportional scale (numerical) variables with
normal distribution using mean±SS and with non normal distribution using
median (interquartile range) in tables and graphics. All the
proportional scale variables were checked for whether they were normally
distributed or not using Kolmogorov-Smirnov and Shapiro-Wilk analyzes.
Age, bALP, Bone US BMD and Mg values were found to be normally
distributed in the groups (p>0,05), and the remaining
values were found to be non-normally distributed with a high skewness
value. Therefore, Student t-test was used in independent two-group
comparison for parameters of Age, bALP, Bone US BMD and Mg since they
were parametric variables and Mann-Whitney U test was preferred for
nonparametric variables. Monte Carlo corrected chi-square analysis was
used to determine the significance of the association between
categorical variables. Pearson correlation coefficient was used when
both parameters were normally distributed to establish the correlation
between the numerical variables. Spearman’s Rho correlation coefficient
was used when any of the parameters was non normally distributed. Type-1
error level was accepted as 5 % in all the analyzes and
p<0,05 was accepted as statistically significant.