Objective: To compare the risks of contaminated culture results and consequent adverse clinical outcomes in urine specimens obtained by “clean-voided” bag method versus catheterization. Study design: Hospital-based cohort study of all children ≤24 months with outpatient urine cultures (n = 7584) obtained from January 1993 to December 1995. Medical records were followed up for all children with contaminated culture results who had 1 or more additional cultures within 7 days of the original culture. Contamination rates of bag urine cultures from the emergency department and a pediatric test center were compared. Results: Contamination rates were 62.8% and 9.1% (P < .001) in bag versus catheter specimens, respectively. Contamination rates of bag urine specimens collected in the emergency department and pediatric test center were 56.4% versus 69.25%, respectively. Of the 3440 contaminated urines, 132 (1.7%) resulted in 1 or more adverse clinical outcomes. Adjusted odds ratios (and 95% CI) for these outcomes in bag versus catheter specimens were as follows: 4.9 (2.3 to 10.5) for unnecessary recall, infinite for delayed diagnosis and treatment, 4.8 (1.8 to 12.4) for unnecessary treatment, 15.6 (2.1 to 116.8) for unnecessary prolonged treatment, 4.1 (1.4 to 12.1) for unnecessary radiologic investigation, and 12.4 (1.6 to 95.5) for unnecessary hospital admission. Conclusions: The risks of the “noninvasive” bag urine culture appear to exceed its benefits. (J Pediatr 2000;137:221-6)