Sample size and statistical analysis
The primary objective was to estimate the prevalence of in- adequate vitamin concentrations in morbidly obese patients
seeking weight-loss treatment. Reference intervals for each vi- tamin were calculated separately for women and men based on the mean 2 SD in the control group. (Reference intervals for vitamins with a log normal distribution were obtained by calcu- lating the mean 2 SD of log-transformed values and back- transforming the result). Vitamin inadequacy was defined as a vitamin concentration below the reference interval. Assuming vitamin inadequacy rates of 2.5% in the control group and 10% in the obese patients, we used an exact binomial test with a 0.05 significance level to estimate that a sample size of 110 patients would give 93% power to detect a difference between the groups. Data are presented as means SDs unless otherwise noted. Skewed data were log transformed when appropriate for further statistical analysis (ie, for the comparison of means or calculation of reference intervals). A Fisher’s exact test or chi-square test was used to compare categorical data between groups, a Stu- dent’s t test or Mann-Whitney U test was used to compare con- tinuous data between groups, and Spearman’s rank correlation was calculated to explore relations between continuous vari- ables. The degree of association between vitamin B-6 (log
mol/L) and various clinical variables was determined by linear regression models that included dichotomous and continuous predictor variables. Covariates were variables known to influ- ence vitamin B-6 concentrations and significant predictors in univariate analyses. The final model was obtained using a back- ward stepwise method. No significant interactions between the remaining variables were identified. The level of significance was P 0.05, and all P values are 2-tailed. No adjustment was made for multiple testing. Statistical analyses were done in SPSS
14.0 (SPSS Inc, Chicago, IL) for WINDOWS
Sample size and statistical analysis
The primary objective was to estimate the prevalence of in- adequate vitamin concentrations in morbidly obese patients
seeking weight-loss treatment. Reference intervals for each vi- tamin were calculated separately for women and men based on the mean 2 SD in the control group. (Reference intervals for vitamins with a log normal distribution were obtained by calcu- lating the mean 2 SD of log-transformed values and back- transforming the result). Vitamin inadequacy was defined as a vitamin concentration below the reference interval. Assuming vitamin inadequacy rates of 2.5% in the control group and 10% in the obese patients, we used an exact binomial test with a 0.05 significance level to estimate that a sample size of 110 patients would give 93% power to detect a difference between the groups. Data are presented as means SDs unless otherwise noted. Skewed data were log transformed when appropriate for further statistical analysis (ie, for the comparison of means or calculation of reference intervals). A Fisher’s exact test or chi-square test was used to compare categorical data between groups, a Stu- dent’s t test or Mann-Whitney U test was used to compare con- tinuous data between groups, and Spearman’s rank correlation was calculated to explore relations between continuous vari- ables. The degree of association between vitamin B-6 (log
mol/L) and various clinical variables was determined by linear regression models that included dichotomous and continuous predictor variables. Covariates were variables known to influ- ence vitamin B-6 concentrations and significant predictors in univariate analyses. The final model was obtained using a back- ward stepwise method. No significant interactions between the remaining variables were identified. The level of significance was P 0.05, and all P values are 2-tailed. No adjustment was made for multiple testing. Statistical analyses were done in SPSS
14.0 (SPSS Inc, Chicago, IL) for WINDOWS
การแปล กรุณารอสักครู่..
