Background.Cerebrospinal fluid shunts transfer cerebrospinal fluid (CSF) from the lateral ventricles in the brain to the peritoneum (ventriculoperitoneal shunt) or the right atrium (ventriculoatrial shunt) via subcutaneous Silastic tubing. As with any implanted foreign body, infection is a serious complication. Although there are several therapeutic modalities currently used for the treatment of shunt infections, controversy remains as to which is best given that there has been only one randomized trial comparing their effectiveness.Objective.To determine which treatment modality is most effective by using decision analysis to compare three approaches with regard to cure rate, morbidity and mortality.Methods.We constructed a decision tree to map out the different treatment modalities and assigned probability values obtained from previously published studies. A utility value was assigned to each treatment outcome, ranging from 0 to 1.0, with a higher score indicating a more favorable outcome. Calculations were performed using Decision Analysis TreeAge computer software.Results.The removal of an infected shunt with establishment of external ventricular drainage or ventricular taps and administration of antibiotics leads to the highest expected value, 0.86. Removal of an infected shunt followed by immediate replacement and administration of antibiotics is less effective, with an expected value of 0.76. The use of antibiotics alone results in the lowest expected value, 0.61. Sensitivity analysis showed the above findings to be robust with respect to clinically relevant changes for the baseline probabilities and utility values.Conclusion.A protocol of shunt removal, external ventricular drainage placement or ventricular taps and antibiotics, followed by creation of a new shunt when CSF sterility is achieved, is the most effective method of treatment for CSF shunt infection.