Stata 12 for mac

Data were described as median and 25–75 % interquartile range (IQR). Metric variables were compared using Mann–Whitney U test and dichotomous variables were compared using Chi-square analysis. Correlation analysis was performed using Spearman’s correlation. Cox regression proportional hazard analysis was performed to assess predictors of 28-day mortality. A forward stepwise procedure was used to identify most potent predictors.
The overall diagnostic test accuracy of ICG-PDR, SOFA, arterial serum lactate, AST levels and INR was assessed by receiver operating characteristics (ROC) expressed as their area under the curve (AUROC). We compared AUROCs at several time points using standard non-parametric methods. Estimates of diagnostic test accuracy (sensitivity, specificity) were calculated using standard methods. One-way repeated/multiple measured analysis of variance (ANOVA) was performed in patients with HH surviving until day 5. A Greenhouse-Geisser correction was used for sphericity. For data management and analyses, we used MS Excel 2008 for Mac, SPSS 21 for Mac (SPSS, Inc. Chicago, IL, USA), and Stata 12 for Mac (Stata Corp., College Station, TX, USA). All p values reported are two sided and p < 0.05 was considered significant.

Since continuous variables were not distributed normally (Shapiro–Wilk test), nonparametric tests (Wilcoxon rank-sum test, Kruskal–Wallis test and Spearman’s correlation) were used for the univariate analysis. Mann–Whitney U test was used to compare groups of thyroid function, considering euthyroid as a control. To assess the relationship between dependent variables (parameters of carotid structure) and thyroid hormones (TSH and FT4), we used a stepwise regression analysis including all traits significantly associated with dependent variables. Although associated, SBD and DBP were not further included in the analyses due to the presence of collinearity. The final model included age, sex, BMI, PP, previous CV events, presence of hypertension, heart rate and FT4. To test the interaction terms, heart rate-by-FT4 was entered in the multilevel model. Tests of significance were two-sided, and a p < 0.05 was assumed as statistically significant in Stata 12 for Mac.