Statistical analysis

Data was organized, categorized, and typed into Excel 2016 version 16.0, Microsoft Corporation, Redmond, USA. Statistical analysis was performed using the program Calc from LibreOffice version 5.3.7.2 (x64), The Document Foundation, Berlin, Germany, and the statistics program R version 3.4.2 (2017), The R Foundation, Vienna, Austria, using packages IswR 2.0–7, bootLR 1.0, epiR 0.9–93 and epibasix 1.3. The results were initially evaluated using a descriptive analysis, expressed by absolute numbers and percentages in categorical variables such as gender or presence of chronic diseases and by means and standard deviation (SD) for continuous variables such as age and Body Mass Index (BMI). The accuracy, sensitivity, specificity, true prevalence, estimated prevalence, positive predictive value and negative predictive value were calculated, as well as the 95% confidence interval (CI) of the results were obtained. The agreement between the methods, venous percussion and US with venous Doppler, were evaluated by the index kappa and LR (likelihood ratio) with 95% CIs. The agreement between observers measuring the signal venous percussion was evaluated by the kappa index with 95% CI’s. A Z test was performed to confirm if the p value <0.05 for Z test where H0 is kappa = or <0.21 (at least Fair or better agreement as measured by kappa value was statistically significant) for kappa measurements of agreement between methods (venous percussion and ultrasound) and between observers (medical student and vascular surgeon. Kappa index was interpreted accordingly to Viera & Garrett (2005) [23]as slight agreement, fair agreement, moderate agreement, substantial agreement and almost perfect agreement.