Hypothesis testing in neurosurgical trials.

Controlled clinical trials represent the most scientific methods of evaluating a new form of treatment. In designing such a trial, one must avoid committing two kinds of errors. The Type I error is defined as falsely concluding that a difference between two treatments exists, when they are equal. The Type II error is committed when one concludes that two treatments are the same, when a real difference exists. To reduce the probability of committing these errors, large sample sizes are required. A survey of neurosurgical trials showed that the majority of these trials have an unacceptably high probability of committing a Type II error because of inadequate sample size.