Supplemental Digital Content 1.
Molecular typing by repetitive sequence-based polymerase chain reaction (rep-PCR)
DNA extraction was performed with the QIAamp DNA Mini Kit (QIAGEN, Inc., Valencia,
IL) according to the manufacturer’s instructions. For rep-PCR, DNA amplification was
performed using random amplified polymorphic DNA Ready-to-go RAPD analysis beads (GE
Life Sciences, Piscataway, NJ) in a final volume of 25 μL, including ~100 ng of total genomic
DNA and 75 pmol of the primer RW3A (Kang HP and Dunne WM, J Clin Microbiol. 2003;41:
2694-2696). Rep-PCR products were resolved using Diversilab DNA chips for the Agilent 2100
(Agilent Technologies, Santa Clara, CA). Diversilab Bacterial Barcodes software was used to
compare banding patterns and determine similarity among isolates (bioMérieux Clinical
Diagnostics, Marcy I’Etoile, France; Frye SR, Healy M, In: Tang YW, Stratton CW, eds.
Advanced Techniques in Diagnostic Microbiology. New York: Springer Science, 2006:444-471).
Isolates with a similarity index ≥95% were considered to represent the same strain.
Statistical methods
Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC) and SPSS
for Windows 15.0 (SPSS, Chicago, IL). All tests for significance were 2-tailed; p-values of ≤0.05
were considered significant. To compare characteristics between cohorts, categorical variables
were analyzed by chi-square or Fisher’s exact test and continuous variables using the KruskalWallis test. Likelihood of nasal colonization (MRSA vs. no MRSA, MSSA vs. no MSSA) during
the follow-up period (at 3, 6, and 12 months) was estimated with mixed models using the SAS
procedure PROC GLIMMIX to model the repeated measures. Logit was used as the link
function to estimate odds ratios (OR) and 95% confidence intervals (CI). Random effects were
included for subject and individual pediatric practice, since observations over time were nested
within subjects and subjects were nested within the pediatric practice. We evaluated
demographic, health, and household variables as potential risk factors for colonization. Time-
varying factors included prior colonization, SSTI in the participant or a household member,
antibiotic use by the participant, and performance of decolonization measures by the subject or
a household member. Risk factors associated with colonization at ≥2 samplings were also
assessed using mixed models, with a random effect for individual pediatric practice only. Each
risk factor was examined separately in univariate analysis. Multivariable models were built in a
forward stepwise fashion; candidate variables were those with p<0.10 in univariate analysis.
Those with p≤0.05 in multivariable analysis were retained in the final models.