Analyses of child-level health outcomes included caregiver surveys administered to 330 caregivers
across all CALM sites. All survey variables were examined for consistency and completeness prior to
analysis. The outcome variables include days and nights of symptoms, days of activity limitation and
days of rescue inhaler use in response to asthma symptoms over the past 4 weeks, health care utilization
and school absenteeism over the past 12 months, and presence of an asthma action plan. The distribution
of the variables with continuous values (days and nights of symptoms, days of rescue inhaler use, days
of activity limitation, and days absent from school) were assessed for normality and found to be skewed to the left, for this reason Wilcoxon signed rank tests for non-parametric, paired data were used to test
significant changes over time. Asthma control was calculated based on the Expert Panel Report 3 (EPR-3)
guidelines15 using symptom days and nights and rescue inhaler use to produce 3 categories of control:
very poorly controlled, not well controlled, and well controlled asthma. Change over time in asthma
control as well as the other categorical variables—any urgent health care use in the past 12 months and
presence of an asthma action plan—were tested using a measure of symmetry, McNemar’s test for variables
with only 2 categories and Bowker’s test for more than 2 categories. All analyses were conducted using
SAS 9.3 (SAS Institute, Inc., Cary, NC).