Several errors occurred in the article by Kosek et al, titled “Epidemiology of Highly Endemic Multiply Antibiotic-Resistant Shigellosis in Children in the Peruvian Amazon” published in the September 2008 issue of Pediatrics (doi: 10.1542/peds.2008-0458).
A data management error occurred that resulted in a small change in the number of surveillance days and age classification of several children, primarily for those categorized as under 6 months of age. The authors have rerun all analysis and ascertained that the overall effects on the results are small. Table 1 on page e543 was most strongly affected and the published values should be replaced by this corrected table:
The only other result changed by this same error was related to the univariate analysis of breastfeeding and risk of shigellosis, which was originally reported as not significant but now does reach statistical significance. In Table 3, page e545, Column 2, lines 20–31, should be replaced with the following values:
This correction does not improve or change the multivariate model as a superior QIC statistic is obtained with the published model (in the absence of breastfeeding) as compared to models that include breastfeeding status or history.
During the copy-editing process several errors were generated in Table 3 on page e545 that were not detected by the authors upon review of the proofs. In Table 3 column 1, lines 10–15, the results for “child <2 y younger than sentinel in household” appeared blank and these results appeared under maternal education. Lines 10–15 in Table 3 should read
In Table 4 line 19 the variable order is reversed. The published version reads “sealed containers vs unsealed” whereas it should read “unsealed vs sealed containers.”
In Figure 2, two errors were generated during copy-editing that were not detected by the authors upon review of the proof. First, the key to Figure 2 puts the age in years (y), while it should be in months (m). Second, the Figure 2 legend in the published version reads “predicted probability (log odds) of shigellosis according to WAS for children at select ages” but should read “Predicted probability (log odds) of shigellosis according to WAZ for children at select ages.”