Statistic in Orthopaedic Surgery and Traumatology

Objetive: Epidemiologic research is based on statistical methods and usually studies representative samples to attain knowledge about population. These studies are based on statistical inference. There are two ways to make inference. One is the parameter estimation with confidence intervals, another is the hypothesis test. A confidence interval allow us to control the precision (a range of the estimated variable) and the confidence (the probability of the population parameter to be included in that range) of the results. A hypothesis test quantifies the chance that a result could be explained by the random nature of the sample measuring type I and type II errors. Every research question has a matching hypothesis test according to the types of the variables, their distribution, the number of sampled units and the groups to be compared. The use of p-values is subject to controversy. Other methods like Bayesian statistics or confidence intervals offer alternatives. Survival studies measure events along time intervals. There are different methods regarding the type of intervals or the data to be analyzed. Multivariate statistics allow us to analyze two or more variables at a time. The exploratory methods can simplify the data, both the number of subjects and the variables. Confirmatory methods like logistic regression enable the analysis and importance of confounding and interacting variables.