These factors were potentially important confounding variables. The one year time frame between study periods also allowed for stabilization of the new FTA and acted as a “wash-out” period. Data collection methods Data was retrospectively extracted by the researcher and data analyst from the routine hospital information system for each patient. The data analyst who had earlier captured the original data was blinded

to the hypothesis since this was a retrospective study. The computerized system was built on a Microsoft sequel server with the capability to access ordered interventions and results. A standardized data collection spreadsheet was used. There was no change Inhibitors,research,lifescience,medical in the health information system during both study periods. The key times were hand written and entered

at the time of discharge onto a Microsoft Excel spreadsheet. Inhibitors,research,lifescience,medical Data was collected retrospectively from the electronic hospital system for all patients registered at the ED before and after the opening of the FTA (i.e. January 2005 and January 2006 respectively). Data validation consisted of checking Inhibitors,research,lifescience,medical for incomplete or missing data and correlating data items. Range checks were done to identify outliers in the data. The accuracy of all fields in the data was cross checked to selleck compound ensure that all transfers, recodes and calculations were correct. Double checking against paper charts was performed by the data analyst with invalid or excessive WTs and randomly with 1% of patient records. The data entered for each study patient comprised of the following information: Inhibitors,research,lifescience,medical date of arrival to the ED, arrival time to the ED, WT, LOS, LWBS, discharge time, died or survived, the triage category and hospital disposition. Statistics Data analyses were performed using MedCalc for Windows, version 9.20 (MedCalc Software, Mariakerke, Inhibitors,research,lifescience,medical Belgium). Data screening and a check for the plausibility and distribution of data were conducted before performing descriptive statistics to ensure that the data met the statistical assumptions necessary

for data analysis. The outcome measures of the study were divided into effectiveness measures (WTs and LOS) and quality measures (LWBS and mortality rate). Univariate descriptive analysis was computed for the effectiveness measures and expressed as the mean and standard deviation. Bivariate analyses were used to determine differences MRIP in the effectiveness measures of WTs and LOS between the control and intervention groups. The independent sample t-test was used to calculate the differences in the mean WTs and mean LOS between the two study groups and the differences were expressed as 95% confidence intervals. With a large sample size (as in our study), the independent sample t-test is robust and the P value will be nearly correct even if a population is far from Gaussian [25]. Quality measures (mortality and LWBS rates) were analyzed using frequencies and proportions.