To the Editor.
—In their article regarding single-dose antibiotic therapy for uncomplicated urinary tract infections, Drs Philbrick and Bracikowski1 emphasized that the studies they examined did not have sample sizes large enough to guard against a type II error. Therefore, the finding of "no statistically significant difference" between single-dose and multiple-dose therapy may be erroneous. The importance of the type II error and sample size in the interpretation of "negative" trials has been examined by other authors as well.2-4When Philbrick and Bracikowski reanalyzed the results using pooled data, they found a statistically significant difference in the cure rates from studies comparing single- and multiple-dose therapy with ampicillin or amoxicillin but no significant difference in the trimethoprim-sulfamethoxazole studies. Even with pooled data, however, the sample sizes for trimethoprim-sulfamethoxazole therapy were not large enough to guard against a type II error.After this meticulous analysis, the authors make a
Jordan NS. Single-Dose Antibiotic Therapy for Urinary Tract Infections and Type II Error. Arch Intern Med. 1986;146(2):413. doi:10.1001/archinte.1986.00360140269043