Developing Sampling Weights for Statistical Analysis of Parent-Child Pair Data From the National Health Interview Survey.

The National Health Interview Survey (NHIS), conducted by the National Center for Health Statistics since 1957, is the principal source of information on the health of the U.S. civilian noninstitutionalized population. NHIS selects one adult (Sample Adult) and, when applicable, one child (Sample Child) randomly within a family (through 2018) or a household (2019 and forward). Sampling weights for the separate analysis of data from Sample Adults and Sample Children are provided annually by the National Center for Health Statistics. A growing interest in analysis of parent-child pair data using NHIS has been observed, which necessitated the development of appropriate analytic weights. Objective This report explains how dyad weights were created such that data users can analyze NHIS data from both Sample Children and their mothers or fathers, respectively. Methods Using data from the 2019 NHIS, adult-child pair-level sampling weights were developed by combining each pair's conditional selection probability with their household-level sampling weight. The calculated pair weights were then adjusted for pair-level nonresponse, and large sampling weights were trimmed at the 99th percentile of the derived sampling weights. Examples of analyzing parent-child pair data by means of domain estimation methods (that is, statistical analysis for subpopulations or subgroups) are included in this report. Conclusions The National Center for Health Statistics has created dyad or pair weights that can be used for studies using parent-child pairs in NHIS. This method could potentially be adapted to other surveys with similar sampling design and statistical needs.

Calibration Weighting Methods for the National Center for Health Statistics Research and Development Survey.

Objectives The Research and Development Survey (RANDS) is a series of web-based, commercial panel surveys that have been conducted by the National Center for Health Statistics (NCHS) since 2015. RANDS was designed for methodological research purposes,including supplementing NCHS' evaluation of surveys and questionnaires to detect measurement error, and exploring methods to integrate data from commercial survey panels with high-quality data collections to improve survey estimation. The latter goal of improving survey estimation is in response to limitations of web surveys, including coverage and nonresponse bias. To address the potential bias in estimates from RANDS,NCHS has investigated various calibration weighting methods to adjust the RANDS panel weights using one of NCHS' national household surveys, the National Health Interview Survey. This report describes calibration weighting methods and the approaches used to calibrate weights in web-based panel surveys at NCHS.

Propensity-score Adjusted Estimates for Selected Health Outcomes From the Research and Development Survey.

To evaluate the quality of web surveys, the National Center for Health Statistics' Division of Research and Methodology has been conducting a series of studies with survey data from commercially recruited panels,referred to as the Research and Development Survey (RANDS). This report describes the propensity-score adjusted estimates from the second round of RANDS (RANDS 2) using the 2016 National Health Interview Survey (NHIS).

Sample Design and Estimation Structures for the National Health Interview Survey, 2016-2025.

This report presents operating characteristics of the NHIS 2016-2025 sample design. The general sampling structure is presented, along with a discussion of weighting and variance estimation techniques primarily for 2016-2018. This report is organized into four major sections. The first section presents a general overview of NHIS and its sample design. The second section describes the redesign process, updates for 2016-2025, and includes general frame and sample design considerations. The third section provides a more detailed description of the sample design and how the sample was selected. The last two sections present a description of the estimators used in NHIS for analyzing and summarizing survey results. Documentation for subsequent changes to the sampling and weighting procedures is available on the NCHS website as separate reports and through each year's survey description document. This report is intended for general users of NHIS data.

The National Center for Health Statistics' 2015 and 2016 Research and Development Surveys.

Objective: This report provides a general description of the background and operation of the first two rounds of the Research and Development Survey (RANDS), a series of cross-sectional surveys from probability-sampled commercial survey panels. The Division of Research and Methodology of the National Center for Health Statistics (NCHS) conducted the first two rounds of RANDS in 2015 and 2016. RANDS 1 and 2 are being used primarily for question design evaluation and for investigating statistical methodologies for estimation. Methods: NCHS contracted with Gallup, Inc. to conduct RANDS 1 in Fall 2015 and RANDS 2 in Spring 2016. RANDS 1 and 2 were conducted using a web survey mode and included survey questions from the National Health Interview Survey (NHIS) that were specifically chosen to provide comparison and evaluation of the survey methodology properties of web surveys and traditional household surveys. In this report, some demographic and health estimates are provided from both sources to describe the RANDS data. Results: In RANDS 1, 2,304 out of the original 9,809 invited panel members completed the survey, for a completion rate of 23.5%. In RANDS 2, 2,480 of the initial 8,231 invited respondents completed the survey, for a completion rate of 30.1%. RANDS 1 and 2 participants were similar to the quarterly NHIS participants with respect to sex, census region, and whether they had worked for pay in the previous week. Other characteristics varied, including age, race and ethnicity, and income. Most health estimates differed between RANDS and NHIS. Public-use versions of the RANDS data can be found at: https://www.cdc.gov/nchs/rands. Conclusion: RANDS is an ongoing platform for research to understand the properties of probability-sampled recruited panels of primarily web users, investigating and developing statistical methods for using such data in conjunction with large nationally representative health surveys, and for extending question-design evaluations.

A method for calculating BMI z-scores and percentiles above the 95th percentile of the CDC growth charts.

BACKGROUND: The 2000 CDC growth charts are based on national data collected between 1963 and 1994 and include a set of selected percentiles between the 3rd and 97th and LMS parameters that can be used to obtain other percentiles and associated z-scores. Obesity is defined as a sex- and age-specific body mass index (BMI) at or above the 95th percentile. Extrapolating beyond the 97th percentile is not recommended and leads to compressed z-score values. AIM: This study attempts to overcome this limitation by constructing a new method for calculating BMI distributions above the 95th percentile using an extended reference population. SUBJECTS AND METHODS: Data from youth at or above the 95th percentile of BMI-for-age in national surveys between 1963 and 2016 were modelled as half-normal distributions. Scale parameters for these distributions were estimated at each sex-specific 6-month age-interval, from 24 to 239 months, and then smoothed as a function of age using regression procedures. RESULTS: The modelled distributions above the 95th percentile can be used to calculate percentiles and non-compressed z-scores for extreme BMI values among youth. CONCLUSION: This method can be used, in conjunction with the current CDC BMI-for-age growth charts, to track extreme values of BMI among youth.

Comparison of Quarterly and Yearly Calibration Data for Propensity Score Adjusted Web Survey Estimates.

While web surveys have become increasingly popular as a method of data collection, there is concern that estimates obtained from web surveys may not reflect the target population of interest. Web survey estimates can be calibrated to existing national surveys using a propensity score adjustment, although requirements for the size and collection timeline of the reference data set have not been investigated. We evaluate health outcomes estimates from the National Center for Health Statistics' Research and Development web survey. In our study, the 2016 National Health Interview Survey as well as its quarterly subsets are considered as reference datasets for the web data. It is demonstrated that the calibrated health estimates overall vary little when using the quarterly or yearly data, suggesting that there is flexibility in selecting the reference dataset. This finding has many practical implications for constructing reference data, including the reduced cost and burden of a smaller sample size and a more flexible timeline.

Overview and Initial Results of the National Center for Health Statistics' Research and Development Survey.

The National Center for Health Statistics is assessing the usefulness of recruited web panels in multiple research areas. One research area examines the use of close-ended probe questions and split-panel experiments for evaluating question-response patterns. Another research area is the development of statistical methodology to leverage the strength of national survey data to evaluate, and possibly improve, health estimates from recruited panels. Recruited web panels, with their lower cost and faster production cycle, in combination with established population health surveys, may be useful for some purposes for statistical agencies. Our initial results indicate that web survey data from a recruited panel can be used for question evaluation studies without affecting other survey content. However, the success of these data to provide estimates that align with those from large national surveys will depend on many factors, including further understanding of design features of the recruited panel (e.g. coverage and mode effects), the statistical methods and covariates used to obtain the original and adjusted weights, and the health outcomes of interest.

Small Area Estimation of Cancer Risk Factors and Screening Behaviors in US Counties by Combining Two Large National Health Surveys.

BACKGROUND: National health surveys, such as the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS), collect data on cancer screening and smoking-related measures in the US noninstitutionalized population. These surveys are designed to produce reliable estimates at the national and state levels. However, county-level data are often needed for cancer surveillance and related research. METHODS: To use the large sample sizes of BRFSS and the high response rates and better coverage of NHIS, we applied multilevel models that combined information from both surveys. We also used relevant sources such as census and administrative records. By using these methods, we generated estimates for several cancer risk factors and screening behaviors that are more precise than design-based estimates. RESULTS: We produced reliable, modeled estimates for 11 outcomes related to smoking and to screening for female breast cancer, cervical cancer, and colorectal cancer. The estimates were produced for 3,112 counties in the United States for the data period from 2008 through 2010. CONCLUSION: The modeled estimates corrected for potential noncoverage bias and nonresponse bias in the BRFSS and reduced the variability in NHIS estimates that is attributable to small sample size. The small area estimates produced in this study can serve as a useful resource to the cancer surveillance community.

Methods and results for small area estimation using smoking data from the 2008 National Health Interview Survey.

The National Health Interview Survey, conducted by the National Center for Health Statistics, is designed to provide reliable design-based estimates for a wide range of health-related variables for national and four major geographical regions of the USA. However, state-level or substate-level estimates are likely to be unreliable because they are based on small sample sizes. In this paper, we compare the efficiency of different area-level models in estimating smoking prevalence for the 50 US states and the District of Columbia. Our study is based on survey data from the 2008 National Health Interview Survey in conjunction with a number of potentially related auxiliary variables obtained from the American Community Survey, an ongoing large complex survey conducted by the US Census. A major portion of this study is devoted to the investigation of several methods for estimating survey sampling variances needed to implement an area-level hierarchical model. Based on our findings, a hierarchical Bayesian method that uses a survey-adjusted random sampling variance model to capture the complex survey sampling variability appears to be somewhat superior to the other considered area-level models in accounting for small sample behavior of estimated survey sampling variances. Several diagnostic procedures are presented to compare the proposed methods.

Design and estimation for the national health interview survey, 2006-2015.

OBJECTIVES: This report presents an overview, a detailed description of the sample design features, and estimation structures for the 2006-2015 National Health Interview Survey NHIS). It fulfills the same role for the current 2006-2015 NHIS design as NCHS Series 2, No. 130, "Design and Estimation for the National Health Interview Survey, 1995-2004" provided for the previous design, which was extended through 2005. METHODS: The 2006-2015 NHIS sample design uses cost-effective complex sampling techniques including stratification, clustering, and differential sampling rates to achieve several objectives, among them improved reliability of racial, ethnic, and geographical domains. This report describes these methods. RESULTS: This report presents operating characteristics of NHIS 2006-2015. The general sampling structure is presented, along with a discussion of weighting and variance estimation techniques. This report is intended for general users of NHIS data systems.

Estimating standard errors for life expectancies based on complex survey data with mortality follow-up: A case study using the National Health Interview Survey Linked Mortality Files.

Life expectancy is an important measure for health research and policymaking. Linking individual survey records to mortality data can overcome limitations in vital statistics data used to examine differential mortality by permitting the construction of death rates based on information collected from respondents at the time of interview and facilitating estimation of life expectancies for subgroups of interest. However, use of complex survey data linked to mortality data can complicate the estimation of standard errors. This paper presents a case study of approaches to variance estimation for life expectancies based on life tables, using the National Health Interview Survey Linked Mortality Files. The approaches considered include application of Chiang's traditional method, which is straightforward but does not account for the complex design features of the data; balanced repeated replication (BRR), which is more complicated but accounts more fully for the design features; and compromise, 'hybrid' approaches, which can be less difficult to implement than BRR but still account partially for the design features. Two tentative conclusions are drawn. First, it is important to account for the effects of the complex sample design, at least within life-table age intervals. Second, accounting for the effects within age intervals but not across age intervals, as is done by the hybrid methods, can yield reasonably accurate estimates of standard errors, especially for subgroups of interest with more homogeneous characteristics among their members.

State-based estimates of mammography screening rates based on information from two health surveys.

OBJECTIVES: We compared national and state-based estimates for the prevalence of mammography screening from the National Health Interview Survey (NHIS), the Behavioral Risk Factor Surveillance System (BRFSS), and a model-based approach that combines information from the two surveys. METHODS: At the state and national levels, we compared the three estimates of prevalence for two time periods (1997-1999 and 2000-2003) and the estimated difference between the periods. We included state-level covariates in the model-based approach through principal components. RESULTS: The national mammography screening prevalence estimate based on the BRFSS was substantially larger than the NHIS estimate for both time periods. This difference may have been due to nonresponse and noncoverage biases, response mode (telephone vs. in-person) differences, or other factors. However, the estimated change between the two periods was similar for the two surveys. Consistent with the model assumptions, the model-based estimates were more similar to the NHIS estimates than to the BRFSS prevalence estimates. The state-level covariates (through the principal components) were shown to be related to the mammography prevalence with the expected positive relationship for socioeconomic status and urbanicity. In addition, several principal components were significantly related to the difference between NHIS and BRFSS telephone prevalence estimates. CONCLUSIONS: Model-based estimates, based on information from the two surveys, are useful tools in representing combined information about mammography prevalence estimates from the two surveys. The model-based approach adjusts for the possible nonresponse and noncoverage biases of the telephone survey while using the large BRFSS state sample size to increase precision.