By Emily Lawler, Associate Professor, Department of Public Administration and Policy at University of Georgia and NBER

Question 1: Tell us about a project or projects you’ve used this data for. What stage are these in (published, working paper, work-in-progress)?

The National Immunization Survey (NIS) consists of two main surveys: the NIS-Child, which interviews households with children aged 19-35 months, and the NIS-Teen, which surveys households with teens aged 13-17 years.  To date, I’ve used these data in five published articles, and I am currently using them in one working paper and one early-stage project. As their name suggests, these surveys collect information about child and teen immunizations, and so most of the work I’ve used them for has focused on the impacts of vaccine policies. For example, in this paper, published in American Economic Journal: Economic Policy (2019), my co-author and I examined the effects of state policies requiring middle school students to receive a Tdap booster shot. Using the NIS-Teen data, we show that the policies increased uptake not only of the Tdap vaccine, but also other recommended preventive care, including the probability of having a well-child visit and receiving the HPV vaccine.

More recently, I’ve used the NIS-Child data in a series of projects examining the impacts of state hospital regulations that are intended to increase breastfeeding. Although you wouldn’t know it from the name, the NIS-Child household survey asks questions about breastfeeding initiation and duration for the focal child, and these data are one of the main sources used to monitor breastfeeding rates in the United States. Using this information, my co-authors and I have shown that the hospital regulations worked as intended and significantly increased breastfeeding initiation and duration. This finding then served as a jumping off point for examining the effects of the breastfeeding support policies on a range of other outcomes in additional datasets. For example, in one paper, my co-author and I show that the breastfeeding support policies resulted in increased maternal time spent on childcare and decreased maternal employment (American Economic Journal: Applied Economics, 2023). In a current working paper, another co-author and I explore the impacts of the breastfeeding support policies on infant health and find significant reductions in infant mortality following policy adoption.

Question 2: What was the application process like? How long did it take to write up the application? Did it require any revisions? How long was the waiting period (application to approval, approval to receiving the data)?

A key advantage of the NIS-Child and NIS-Teen data is that much of it is publicly available and can be downloaded directly from the CDC website. In the public-use data, you will have information on state of residence at time of survey, year of survey, and age of the focal child at time of survey. For nearly all of my projects that have used these data, the public-use version has had all the information I needed. If you need more precise geographic information, such as state of birth or county of residence, or more precise date information, such as birth date, then that requires accessing the restricted-use version of the data.

The restricted-use NIS data must be accessed through a Research Data Center (RDC). These applications can be pretty lengthy, because they require information about the project background and analytic approach, as well as shells for all results tables and a listing of all variables from the restricted and any secondary data sources that need to be merged on. My most recent application was nearly 90 pages long. From submission to approval, it took only a couple of months, and that included having to make one small set of revisions to the proposal. From there, it took a couple more months to take care of paperwork, payments, etc., before finally being able to touch the data.

Question 3: How is the data accessed (sent and stored locally, via VPN)? What kind of security measures are required?

For the public-use data, you can download it directly from the CDC website and work with it on your own computer. For the restricted-use files, you do have to be on site at an RDC. If you access the data at a Federal Statistical RDC, you also have to have Special Sworn Status. To ensure security of the restricted data, you are not able to directly bring any information in or out of the RDC. Instead, once your project is approved you will be assigned to an NCHS analyst. Your analyst then reviews and transfers any files that you want to take into the RDC, such as program code or reference materials, or any results that you want disclosed.

Question 4: What aspects of the data are particularly useful (variables, sample size, timeframe, etc.), in general and for you specifically?

The NIS-Child is a long-running survey – it has been fielded annually each year from 1995 to the present. That, combined with the fact that it has publicly available state of residence information, makes it well suited for examining a broad range of questions. It is also particularly useful that the data consist of not just a household survey with self-reported vaccination information, but also a health-care-provider survey, where providers are contacted directly for the child’s immunization records. Having provider-verified vaccination information increases confidence in the quality of the data and helps reduce concerns that observed policy effects are being driven by changes in vaccination recall.

Question 5: Are there limitations to keep in mind, or that researchers might find surprising?

As I think is typical with most long-running surveys, there have been a number of changes to some survey questions and to the sampling frame across the years, and this can make it challenging to make certain comparisons over time. I have found, though, that the data documentation is fairly detailed and clear, making it much easier to navigate those changes than with some of the other data I work with.

Question 6: Is there something interesting/cool/useful about this data that is likely unknown to most researchers? Anything else you want to share?

When I first started working with the data, I was surprised to find the questions about breastfeeding in the “immunization” survey! The NIS-Teen also has some interesting non-vaccination-related measures of health and health care use, including information about timing of most recent check-up, school days missed due to illness or injury, and diagnoses of certain conditions, such as asthma. I think adolescent health and health care use is fairly understudied in our field, and these data could be a useful starting point for answering some important questions.