NIS Survey Methods
Each year, the CDC National Immunization Program, in partnership with CDC’s National Center for Health Statistics, conducts the National Immunization Survey (NIS). The NIS uses random-digit dialing to survey households with children aged 19–35 months. The telephone survey asks questions about childhood immunization. Beginning July 2001 and continuing through December 2002, a sample of respondents was asked about breastfeeding using the original questions listed in Table A1. Starting January 2003, all respondents to the telephone survey were asked the breastfeeding questions. The Council of American Survey and Research Organizations response rates for landline sample of NIS years 2001–2013 were consistently above 60% (range 61.6%–76.1%). Response rates for the cellular telephone sample of NIS years 2011-2013 ranged from 25.2% to 30.6%. A more detailed description of the methods can be found at the National Immunization Survey Web site.
In 2011, a cellular telephone sample of respondents was included in the NIS along with the landline telephone sampling frame used for all previous survey years (referred to as dual-frame sampling). To evaluate the impact on breastfeeding rates when the NIS added a cellular phone sample of respondents, CDC compared the differences in breastfeeding rates between landline and dual-frame sample using data from the same birth cohort.
Key findings of study comparing breastfeeding rates from landline and dual-frame sample
- Adding a cell phone sample to the NIS survey had a minimal impact on national estimates of the percent of infants ever breastfed and the percent exclusively breastfed at 3 and 6 months. Dual-frame prevalence estimates for these indicators differed from landline frame estimates by less than one percentage point for infants born in 2010, ranging from 0.5 percentage points lower for exclusively breastfed at 3 months to 0.2 percentage points higher for ever breastfeeding and 0.8 percentage points higher for exclusively breastfed at 6 months.
- Adding a cell phone sample had a slightly larger impact on national estimates of the duration of any breastfeeding. Among infants born in 2010, the dual-frame estimates were lower than the landline frame estimates by 1.5 and 1.6 percentage points for breastfed at 6 months and at 12 months, respectively.
- Adding a cell phone sample also had an effect on estimates of breastfeeding at the state level, with some states showing higher rates and some showing lower rates. State breastfeeding rates estimated from the National Immunization Survey do fluctuate from year to year due to the small size of the state samples.
- Based on our findings, we caution against making comparisons of current dual-frame birth year estimates of breastfeeding to more recent landline frame estimates (e.g. 2007 or 2008 birth years) when there was an increasing number of U.S. children living in households with only cell phone service and the National Immunization Survey did not yet include a cell phone sample of respondents. We have less concern about comparison of current dual-frame birth year estimates to landline frame estimates from birth years earlier in the decade when few households were cell phone only.
A full study report will be posted on this website in fall, 2014.
Wording of the Breastfeeding Questions
The breastfeeding questions used in the NIS were modified in 2004 and 2006 (Table A1). In January 2004, the third question was modified slightly to remove “or water” from the first part of the question and add it to the list of items other than breast milk. Additionally, the interviewers were instructed to always read the clarification of foods and liquids to consider. Our examination of the influence of this question change on the rate of exclusive breastfeeding at 3 and 6 months of age revealed a minimal effect.
In 2006, additional changes were made to two of the NIS breastfeeding questions (Table A1). These changes were on the basis of cognitive testing results for the questions used prior to 2006. Question 2 was revised to read “How old was [child’s name] when [child’s name] completely stopped breastfeeding or being fed breast milk?” The wording change on question 2 resulted in a minimal effect on our reported duration of breastfeeding. Question 3, “How old was [child] when s/he was first fed something other than breast milk?” was asked as two separate questions in 2006, which yielded significantly lower estimates of exclusive breastfeeding. Because of this large effect, we only show trends of exclusive breastfeeding and formula supplementation of breast milk for children whose caregivers were interviewed after December 31, 2005.
The 2006 breastfeeding questions continue to be used in the NIS survey.
Table A1. Changes in breastfeeding questions, National Immunization Survey 2001–2013
|Questions 2001–2003||Questions 2004–2005||Questions 2006-Present|
|1. Was [child] ever breastfed or fed breast milk?||1. Was [child] ever breastfed or fed breast milk?||1. Was [child] ever breastfed or fed breast milk?|
|2. How long was [child] breastfed or fed breast milk?||2. How long was [child] breastfed or fed breast milk?||2. How old was [child's name] when [child's name] completely stopped breastfeeding or being fed breast milk?|
|3. How old was [child] when s/he was first fed something other than breast milk or water? [If respondent hesitates, add: This includes formula, juice, cow's milk, sugar water, solid foods or anything else.]||3. How old was [child] when s/he was first fed something other than breast milk? This includes formula, juice, cow's milk, water, sugar water, solid foods or anything else.||3. How old was [child's name] when (he/she) was first fed formula?|
|4. This next question is about the first thing that [child] was given other than breast milk or formula. Please include juice, cow's milk, sugar water, baby food, or anything else that [child] may have been given, even water. How old was [child's name] when (he/she) was first fed anything other than breast milk or formula?|
Analysis of Breastfeeding Practice Data by Year of Child Birth
We combine survey years to calculate breastfeeding indicators by year of child birth (Smith et al., 2006). Because children were 19–35 months of age at the time of the parent interview through NIS survey year 2010 and 19-35 months of age any time during the survey year quarter starting with the 2011 NIS survey, each survey year represents children born over three years (Table A2). Breastfeeding rates among children in a birth year are released when approximately 2/3rds of the children born in that year have been surveyed. In the past, the rates were labeled provisional until they were replaced the following year with final rates based on all children in the birth year. Starting with 2010 births, CDC releases as final breastfeeding rates that are based on two survey years. Adding a third year of survey data to a birth cohort has little impact on the breastfeeding rate and results in a margin of error that is only about 20% smaller than with two survey years. It was determined that this small benefit did not compensate for the confusion created by issuing new breastfeeding rates one year after the Breastfeeding Report Card is published each year. Therefore the breastfeeding rates published in the 2013 Breastfeeding Report Card and onwards are considered as the final rates. CDC is also issuing an update to the 2009 rates based on two survey years (NIS 2011 and 2012) to convert what were final landline rates to final dual-frame rates. All analyses are conducted using statistical software that accounts for complex sample design.
Table A2. Years of breastfeeding data from National Immunization Survey used to calculate breastfeeding rates for each birth cohorta
aStarting with the 2010 birth cohort, CDC includes only the first two survey years of data for annual estimates of breastfeeding rates. For 2009 birth cohort, the second two years were used to estimate dual-frame rates.
Smith PJ, Zhao Z, Wolter KM, Singleton JA, Nuorti JP. Age-period-cohort analyses of public health data collected from independent serial cross-sectional complex probability sample surveys. Paper presented at the Joint Statistical Meeting: 2006; Seattle, WA. http://www.amstat.org/Sections/Srms/Proceedings/y2006/Files/JSM2006-000537.pdf - [PDF-125k]