New Findings from the 2007 National Survey of Children’s Health (NSCH)
June 02, 2009, 2:00 – 3:00 p.m. ET (1:00 – 2:00 p.m. CT, 12:00 – 1:00 p.m. MT, 11:00 – 12:00 p.m. PT)
Q: Is there a charge for requesting confidential data?
A: Access to data that have not been publicly released is available through an NCHS Research Data Center (http://www.cdc.gov/nchs/r&d/rdc.htm). The minimum charge is approximately $1000, which covers set-up of the dataset and 2 days on site or 30 days remote access.
Q: how reliable are height and weight data for children, particularly adolescents?
A: A recently published article by Lara Akinbami and Cynthia Ogden in the journal Obesity addresses that question: http://www.nature.com/oby/journal/vaop/ncurrent/abs/oby20091a.html
Q: Is there a way to access county level data? What about data for cities or area codes?
A: ZIP Codes and telephone area codes for NSCH respondents are available through the NCHS Research Data Center (http://www.cdc.gov/nchs/r&d/rdc.htm).
Q: Are there limitations to the NSCH survey given it is a telephone survey, e.g. by socio-economic status?
A: All sample surveys are subject to error due to nonresponse and noncoverage. Telephone surveys have lower response rates than face-to-face surveys, and therefore they have a greater potential for bias than face-to-face surveys. In addition, the 2007 NSCH was a landline-only telephone survey, and therefore it could not include the roughly 13-14% of children living in cell-phone-only households and the 2% of children living in households with no phone coverage. Sampling weights were adjusted to account for potential nonresponse and noncoverage bias, to the extent possible.
Q: How were communities/neighborhoods defined? If a child lives in a community with a park, does that mean it is close (within walking distance) or just within their city?
A: The survey asked respondents about their “neighborhood and community.” The terms were not defined.
Q: What were the 4 Asian languages that were used to conduct the survey?
A: Mandarin, Cantonese, Vietnamese, and Korean.
Q: How does health insurance coverage rates compare with CPS estimates?
A: To our knowledge, this analysis has not yet been done with the 2007 NSCH data.
Q: If we were to filter the responses to try to get at respondents whose children are disabled or have special health care needs, which items (questions) on the survey would you use to screen their answers?
A: The survey used the CSHCN Screener to identify children with special health care needs. In the survey instrument, these are questions K2Q10-K2Q23. On the DRC, this population has already been defined and identified and can be used to stratify the other indicators.
Q: Does the National Survey of Adoptive Parents include any data on foster children and their families?
A: The NSAP permits data analysts to identify children who were adopted from the US foster care system, but it does not include data for children still living in foster care.
Q: Are data on Medicaid coverage collected?
A: Children with public insurance (which may include Medicaid, CHIP, or state programs) can be identified in the survey.
Q: Can SLAITS be used for a state-only survey?
A: States can add questions to the SLAITS surveys if they cover the cost of asking the additional questions. States may also increase the sample size of SLAITS surveys if they cover the cost of the additional interviews. Finally, states may sponsor their own SLAITS surveys if they cover all of the costs of data collection.
Q: Can one tell from the data sets if the child is a military child?
A: No.
Q: Did the survey determine what percentage of asthmatic children with inadequate insurance live in households with smokers?
A: The population of children who meet those three criteria can be identified. On the state level, however, be aware that the sample may be too small for an estimate to be reliable.
Q: Do we have any variables to recognize children with Down syndrome or spina bifida? If it is, how many children will be in the sample?
A: Down syndrome and spina bifida were not among the 16 conditions that the survey asked about specifically.
Q: Does the special health care needs data set include specific mental health data?
A: Yes. A number of different questions addressed mental health needs and issues. Among the 16 specific chronic conditions addressed in the survey were depression, anxiety, and behavioral or conduct problems, as well as cognitive disorders such as autism spectrum disorder, learning disabilities, and ADD/ADHD. The CSHCN Screener explicitly includes emotional, developmental, or behavioral conditions that are expected to last 12 months or longer. In addition, parents of children from birth to age 5 were asked if they had any concerns about their child’s learning, development or behavior.
Q: Any idea why preventive care visit rates in the 2007 NSCH are 3 times as high as suggested by the NAMCS?
A: If true (and we have not confirmed the premise), this may reflect the limitations of parent report. A visit defined by a parent as “preventive medical care” may be classified by the doctor for record keeping, accounting, or billing as having an alternative purpose. The NAMCS is based on medical record review, not parent report.
Q: Do you collect data on chronic fatigue syndrome in children and adolescents?
A: No, chronic fatigue syndrome was not among the 16 conditions asked about in the survey.
Q: It was reported that only 40% of children did not receive mental health services, can you elaborate on this? Was the lack of mental health service utilization due to inadequacy of insurance coverage? Cultural beliefs?
A: The survey did not ask about the reasons why mental health services were delayed or not received.
Q: Any plans for a future Early Childhood national survey? MCHB funds almost all states to develop early childhood comprehensive systems which often rely on national surveys on young children (0-5 yrs) for data & accountability.
A: No, there are no plans for MCHB to do this. The NSCH has about 6,000 children between 6 months and 5 years old, and questions specific to that age group.
Q: Have you considered analyzing whether the increase in preventive health visits between 2003 and 2007 may be associated with the decrease in private insurance coverage?
A: We have not looked into that yet. It’s a good idea to do so.
Q: How would you suggest identifying children whose special health care need is more severe than are identified by the screener? e.g. medically complex, technologically dependent, severely disabled, etc.
A: Terms such as “medically complex” or “technology dependent” were not used in the survey. However, for each of the 16 specific conditions in the survey, parents were asked if their child’s condition was mild, moderate, or severe. In addition, parents of children with special health care needs were asked if their children’s conditions limited their ability to participate in a number of specific activities, like playing with other children, making friends, attending school, or participating in sports or clubs. Finally, recently published work by Matt Bramlett and his colleagues in the Maternal and Child Health Journal (March 2009, vol 13, pp 151-163) suggest that the complexity of care needs can be inferred by the number or types of health care consequences that the child experiences.
Q: Does this survey ask any questions about childhood immunizations?
A: Yes, there is a section on adolescent immunizations that includes tetanus booster, meningitis, and HPV. There are no NSCH-specific immunization questions for younger children, though linked data from the National Immunization Survey is available through an NCHS Research Data Center for NSCH children 19 to 35 months of age.
Q: Did the survey ask about whether children had been told they had autism? If so, what were the results?
A: The question about autism asked parents whether they had been told by a doctor that their child had the condition and whether the child still has the condition. (Whether the children themselves were told was not asked.). The results will be forthcoming in a chartbook in the near future.
Q: What did you use as the definition of "exclusively breastfed?"
A: In this analysis, “exclusively breastfed” was defined as receiving only breastmilk and no formula, water, or solid foods for the first six months of the child’s life. This indicator was derived from a series of questions: whether the child was ever fed breastmilk; at what age the child was first given formula (if ever); at what age the child was first fed anything else, including juice, cow’s milk, water, or baby food. An indicator defining exclusive breastfeeding differently could be constructed as well (for example, allowing for solid food or water at an earlier age but not formula).
Q: Mental health service usage was measured for children under five years of age, or was that a typo in the slide?
A: The questions about mental health service use were asked for children 2-17 years of age.
Q: Was Dental Disease/caries one of the choices for parents when selecting conditions? We often hear that it is five times more common than asthma but it doesn't appear on this list.
A: The survey includes a series of questions about the condition of children’s teeth. This series includes questions about decayed teeth or cavities, toothaches, broken teeth, and bleeding gums.
Q: Is the 'medical home' specific to CSHCN, or is it for all children?
A: Medical home can be analyzed for all children.
Q: For those interested in investigating the influence of "social determinants of health," it would be great to allow subgroup analysis by the neighborhood characteristics variables.
A: To the extent that neighborhood characteristics correlate with income level, starting with income level would be a good place to begin. The idea of creating a neighborhood characteristics variable for use as a subgroup is being considered, but more work needs to be conducted to ensure that the variable is a meaningful and accurate measure.
Q: Would analysis of a sub-set of children aged 10-14 at State level appropriate given the sample size and sampling strategy? (WA state)
A: The NSCH randomly selects telephone numbers and then randomly selects a single child within each household with children. Yes, one could look just at the 10-14 age group in a specific state and it would be appropriate. There were approximately 1,700 children sampled from Washington so there would be approximately 450-500 children in the age group you are asking about. Keep in mind that analysts working with the raw data should not subset the dataset to just a specific age group. Use SUBPOPN (in SUDAAN) or another similar subpopulation statement in the analysis in order to give the statistical package access to the complete design structure.
Q: Are NSCH data available by family structure? In particular, I'm wondering if it was possible to break out data for children with lesbian and gay parents.
A: Data were collected on family structure, and family structure is reported on the publicly available datasets in four categories: two-parent (biological or adoptive) families, single-mother families, stepfamilies, and “other family structures”. Information about the sexual orientation of the parents was not collected.
Q: Can we compare 2005-2006 NS-CSHCN with 2007 NSCH?
A: The surveys are conducted in a different manner, so comparisons should be approached cautiously. The NS-CSHCN samples based on a child in the household meeting the criteria for having a special health care need. In the NSCH, children are also screened for having a special health care need. On the DRC website it is possible to look at variables from the 2007 NSCH by whether or not the child has a special health care need or not. Keep in mind, however, that the surveys each ask a unique set of questions. On content that does overlap (such as medical home), the results can be compared.