OBJECTIVE Urinary tract infection is one of the most common bacterial infections among children. Difficulty in specimen collection and interpretation of inadequately collected specimens may contribute to misdiagnosis of urinary tract infection. Our objective was to assess the effect of perineal/genital cleaning on bacterial contamination rates of midstream urine collections in toilet-trained children. METHODS We conducted a randomized trial in toilet-trained children who presented to a tertiary care pediatric emergency department between November 1, 2004, and October 1, 2005. All toilet-trained children who were between the ages of 2 and 18 years and had a midstream urine sample requested were eligible. Those whose parents consented were cluster-randomized by week to either cleaning or not cleaning the perineum with soap. The risk for a contaminated urine culture (defined as growth of < 10(8) colony-forming units per liter [< 10(5) colony-forming units per milliliter] of a single organism or a mix of > or = 2 organisms) and the risk for a positive urinalysis (defined as a positive leukocyte esterase and/or nitrites on dipstick or > or = 5 white blood cells per high-powered field on a standard microscopic examination) were analyzed by intention to treat. RESULTS A total of 350 children were enrolled. The rate of contamination in the cleaning group was 14 (7.8%) of 179 vs 41 (23.9%) of 171 in the noncleaning group. Children who were randomly assigned to cleaning were less likely to have a positive urinalysis (37 of 179 [20.6%]) than those in the noncleaning group (63 of 171 [36.8%]). CONCLUSIONS Urine contamination rates are higher in midstream urine that is collected from toilet-trained children when obtained without perineal/genital cleaning. Cleaning may reduce the risk for returning for repeat cultures and for receiving unnecessary antibiotic treatment and investigations.