[An application of multiple logistic regression analysis to an outcome study on surgery for patients with femoral neck fracture]

Abstract

Sixty five patients with femoral neck fracture were examined to find any risk factor for their postoperative complications. Here, postoperative complications included cardiac catastrophe, pulmonary embolism, central nervous system accident, and respiratory failure requiring mechanical ventilation during the period of one week after operation. Multiple logistic regression analysis was used to predict these factors. Explanatory variables were selected based on the clinical judgement. Those variable are age, sex, preoperative underlying cardiac disease, preoperative underlying respiratory complications, the use of cement for the procedure, and anesthetic technique. The analysis revealed that the sex was the only variable that was considered significant based on the P value less than 0.05. However, multi-colinearity between variables was suspected to exist based on the correlation matrix analysis which may indicate that the factor is not, in fact, significant. In short, further study will be required to find any significant variables or risk factors to predict the outcome of untoward events.