Postoperative wound infection: application of multiple regression analysis to patients parameters

The effect of patient parameters (age, sex, wound drainage, and wound description) on postoperative wound infection is evaluated with the aid of multiple regression analysis. The resulting multiple correlation matrix confirmed that the four parameters are not correlated. Hence, the possibility that parameter interactions introduce a bias into the observations was excluded. In the multiple regression model developed, the regression coefficients of all four parameters were statistically significant. When the relative significance of each parameter was considered, the most significant one was wound contamination, followed by the presence of wound drains. The least significant were the patient's age and sex. Male sex was a relatively more significant parameter when compared with age up to 42 years. However, beyond that critical point patient age appear to be relatively more significant than male sex. The usefulness of the model developed and possible application areas are fully discussed.