Statistical Methods
We used the Kaplan–Meier method to analyze the time elapsing
before the first UTI recurrence and the log-rank test to assess the
differences in cumulative survival functions between the groups.
The occurrence of all episodes of UTI was expressed in terms of
incidence density, calculated by summing the number of UTI
attacks and the time at risk in each group and then calculating
the rate of UTI episodes per person-year at risk (PYR). Each
patient contributed days at risk until the point of dropping out
or receiving antimicrobials for any reason (if for UTI, at least
10 days were subtracted), or until the follow-up ended. For the
children who received antibiotic prophylaxis the intervention
ceased at that point. The differences in UTI incidence density
between the groups were tested on the assumption that the occurrence
of UTI follows a Poisson distribution. The differences
between the groups in the proportion of children with at least
1 UTI recurrence, antimicrobial days per year, antimicrobial
prophylaxis, and dropouts were tested with the binomial SND
(Standardized Normal Deviate) test. The data were analyzed
with PASW Statistics for Windows (version 18.0) and Stats-
Direct software (version 2.7.2).