Epidat 4.1 and Review Manager Software 5.3 statistical proyect for meta-analysis (10-12 articles).

This is a meta-analysis with data on the cost and time of hospitalization of two groups of patients with and without preoperative pathology, based on information from several works. It is also interesting to know the mean hospitalization cost and the average hospitalization time for each of the two groups and if there is Statistic Significance based on these data. The meta-analysis of both cost and time (if applicable) would include the studies for which it was possible to obtain the mean, standard deviation and the sample size of each group. The standardized mean difference was considered as a measure of the effect of the sample size for cost, as the same currency was not used in all the articles included, and mean difference was used for time (days). A confidence interval of 95% (CI) of the sample size was estimated in both cases. First, an analysis of the statistical heterogeneity between the studies using the Higgins and Thompson test (I2), and Galbraith’sgraph was performed to show the dispersion between the studies, and the [1]2 test with p<0.005 was carried out to determine heterogeneity with statistical significance. If we found sufficient data from the articles selected, a meta-analysis was subsequently conducted using the Z test to analyze the statistical significance. If high clinical heterogeneity(I2>50%) was observed, the estimate of the effect was calculated using the random effects model proposed by the DerSimonian and Laird method. If no significant heterogeneity was observed, the fixed-effects model was applied using the inverse variance method. An analysis of the sensitivity of the results (influence graphics) and publication bias (tests of Begg and Egger and the Funnel chart) was finally carried out. All the statistical analysis would performed with the Epidat 4.1 and Review Manager Software 5.3 statistical packages developed by the Cochrane Collaboration institution.