Child Health Assessment and Monitoring Program (CHAMP)

For most CHAMP tables, the results are displayed showing characteristics of the child, including sex, race, Hispanic origin, age, school enrollment type (public/private), grade level, health insurance status, and Special Health Care Needs status. In addition, parental education level is also presented. Due to a relatively small sample size, only state-level estimates are presented.

All questions from the survey are not presented in the web tables; however, the full questionnaire for each year may be viewed online. Whenever possible, the exact wording for each question is used as a title, however due to space limitations, the titles may not include some introductory remarks and explanations for some questions. For more technical information about CHAMP, please contact the CHAMP Coordinator.

All of the percentages shown in CHAMP reports are calculated using weighted data and therefore cannot be derived exactly from the numbers in the tables. Weighting adjusts for unequal probabilities of selection due to the disproportionate sampling method and due to people living in households with different numbers of telephones and different numbers of children. The final sample data are also weighted to account for unequal non-response rates among different demographic groups. For example, if parents of children under five constitute 10 percent of the sample respondents, but this group represents 8 percent of the total population of the state, then a factor of 0.80 would be entered into the last weighting process for these respondents to account for this discrepancy. Thus, the weighting procedure makes the CHAMP data more representative of the total population of children in the state.

The main data tables in CHAMP reports show the 95 percent confidence interval associated with each percentage (labeled “C.I.”). All survey results, no matter how large or small the sample, are subject to some degree of sampling error. The 95 percent confidence interval shows the range within which we would expect the true value for the entire population to fall 95 percent of the time. Estimates for smaller sample sizes, such as those by ethnicity or race subgroups, would have a larger margin of error (the confidence intervals will be wider) since the sample size would be smaller in each group.

Given the complex nature of the CHAMP sample (i.e. it is not a simple random sample), the State Center for Health Statistics uses SUDAAN software to calculate the confidence intervals for the estimates. This software takes the complex sampling design into account when computing the errors of the estimates. In general, any percentage with a numerator of less than 20 will have a relatively large degree of sampling error and should be considered cautiously.

Tests of the statistical significance of a difference between two percentages (for example, between the percentages for two age groups) can be performed after calculating the standard error of the difference. The data user should contact CHAMP staff of the Center for assistance with this calculation. Though not exactly technically correct, a rough approximation of the statistical significance of a difference between two percentages can be derived by comparing the confidence intervals. If the confidence intervals of the two percentages being compared do not overlap, then it is likely that the difference between the two percentages is statistically significant at the 95 percent confidence level. Stated another way, one can be 95 percent certain that a difference that large would not be observed just due to random variation in the two percentages.

Respondents who refused to answer or did not know the answer were excluded from any calculations in these tables. However, for a few questions "Don't know/Not sure" responses provide valuable information and were included in the calculations. The denominator for each calculation is the number of parents who responded to the item (exceptions are specified in the footnotes).

** Risk Factors and Calculated Variables Presented in the CHAMP Results **

** Weight for Age – CHAMP Results for 2005 & 2006 **

Due to issues with parental recall of children’s heights, only weight for age data is presented for 2005 and 2006. These data are for children 2 years old or older. These Percentiles are calculated using weight for age percentile growth charts provided by the National Center for Health Statistics. Percentiles rank the position of an individual by indicating what percent of the reference population the individual would equal or exceed. For example, a 5-year-old girl whose weight is at the 25th percentile, weighs the same or more than 25 percent of the reference population of 5-year-old girls, and weighs less than 75 percent of the 5-year-old girls in the reference population.

** Body Mass Index for Age – 2007 & forward **

Improvements in the CHAMP survey design resulted in more accurate height and weight data for children starting in 2007. Beginning with that year, Body Mass Index groupings are calculated for children between the ages of 6 and 17. Weight and height data were available for the majority of these children. The SAS code for calculating BMI for age and gender was obtained from the National Center for Health Statistic's website entitled, "2000 CDC Growth Charts: United States". The calculations use the 2000 CDC growth charts, by age and gender, as the standard for assessing weight status.

BMI is calculated in the same manner for children as it is for adults: weight in kilograms divided by height in centimeters squared. Children are said to be at risk for overweight if their body-mass index is between the 85th and 94th percentiles. They are overweight if their body-mass index is in the 95th percentile or higher -- or greater than at least 95 percent of youngsters the same age and gender in the reference year for the CDC growth chart. For more information on BMI for children, visit the website listed above.

** Strengths and Limitations of the CHAMP Survey Data **

One potential limitation of the CHAMP survey is due to the fact that the data are reported by parents. Research suggests that parental respondents may not accurately report health risk behaviors involving their children, especially those that are illegal or socially undesirable. In addition, in certain cases, parents may not be aware of the health risk behaviors of their children – especially teenagers.

There are some significant advantages of the telephone survey methodology, including better quality control over data collection made possible by a computer-assisted-telephone-interviewing system, relatively low cost, and speed of data collection. The CHAMP methodology is modeled after the BRFSS which has been used and evaluated by the CDC and participating states since 1984. The content of the survey questions, questionnaire design, data collection procedures, interviewing techniques, and editing procedures have been carefully developed to improve data quality and lessen the potential for bias. The data collection is ongoing, and each year new annual results become available.

One limitation of a telephone survey is the lack of coverage of persons who live in households without a telephone. Households without a telephone are, on average, of lower income. Therefore, for many of the health risks measured, the results are likely to understate the true level of risk in the total population of children in North Carolina .

Additionally, the widespread use of cellular phones has impacted both the CHAMP and BRFSS telephone surveys. As more North Carolinians abandon landline phones and become “cell phone only” households, the BRFSS is devising plans to address this change. In 2009, the North Carolina BRFSS will add a cell phone component to its landline survey. It may also be possible to include CHAMP households that are serviced only by cell phones.