Search

Non-response is a common problem in household sample surveys. The Medical Expenditure Panel Survey (MEPS), sponsored by the Agency for Healthcare Research and Quality (AHRQ), is a complex national probability sample survey. The survey is designed to produce annual national and regional estimates of health-care use, expenditures, sources of payment, and insurance coverage for the U.S. civilian non-institutionalized population. The MEPS sample is a sub-sample of respondents to the prior year's National Health Interview Survey (NHIS) conducted by the National Center for Health Statistics (NCHS). The MEPS, like most sample surveys, experiences unit, or total, non-response despite intensive efforts to maximize response rates. This paper summarizes research on comparing alternative approaches for modelling response propensity to compensate for dwelling unit (DU), i.e. household level non-response in the MEPS.Non-response in sample surveys is usually compensated for by some form of weighting adjustment to reduce the bias in survey estimates. To compensate for potential bias in survey estimates in the MEPS, two separate non-response adjustments are carried out. The first is an adjustment for DU level non-response at the round one interview to account for non-response among those households subsampled from NHIS for the MEPS. The second non-response adjustment is a person level adjustment to compensate for attrition across the five rounds of data collection. This paper deals only with the DU level non-response adjustment. Currently, the categorical search tree algorithm method, the chi-squared automatic interaction detector (CHAID), is used to model the response probability at the DU level and to create the non-response adjustment cells. In this study, we investigate an alternative approach, i.e. logistic regression to model the response probability. Main effects models and models with interaction terms are both evaluated. We further examine inclusion of the base weights as a covariate in the logistic models. We compare variability of weights of the two alternative response propensity approaches as well as direct use of propensity scores. The logistic regression approaches produce results similar to CHAID; however, using propensity scores from logistic models with interaction terms to form five classification groups for weight adjustment appears to perform best in terms of limiting variability and bias. Published in 2007 by John Wiley & Sons, Ltd.

Objectives: The National Immunization Survey (NIS) uses two phases of data collection to obtain vaccination information from a sample of young children: a random-digit-dialing (RDD) survey for identifying households with children 19-35 months of age, followed by a mail survey for obtaining provider-reported vaccination histories about these children. Provider-reported vaccination histories are used to estimate vaccination coverage rates. In 1998, provider-reported vaccination histories were not obtained for 32.9% of children with a completed RDD interview. This report describes the statistical methods adopted in 1998 to reduce the bias in vaccination coverage estimates that could result from "vaccination history nonresponse," that is, differences between children for whom provider data was obtained and those for whom it was not obtained., Methods: In the methods adopted in 1998, children with completed NIS RDD interviews are grouped into adjustment cells defined by their propensity to have adequate provider data. Sampling weights of children with adequate provider data are divided by the cell-specific weighted response rate to allow these children to represent all children in the cell., Results: Using an "optimal" number of cells, the overall extent of bias reduction was 0.5%, suggesting that provider nonresponse bias was small. Authoritative literature suggests using five cells. No statistically significant differences were observed in vaccination coverage estimates when comparing estimates based on the "optimal" number of cells with five cells. Thus, five adjustment cells are used to reduce provider nonresponse bias in the NIS vaccination coverage estimates. No substantively important differences were observed between estimates based on the methodology used prior to 1998 and the methodology adopted in 1998.

Background: The National Immunization Survey (NIS) and the National Health Interview Survey (NHIS) produce national coverage estimates for children aged 19 months to 35 months. The NIS is a cost-effective, random-digit-dialing telephone survey that produces national and state-level vaccination coverage estimates. The National Immunization Provider Record Check Study (NIPRCS) is conducted in conjunction with the annual NHIS, which is a face-to-face household survey. As the NIS is a telephone survey, potential coverage bias exists as the survey excludes children living in nontelephone households., Methods: To assess the validity of estimates of vaccine coverage from the NIS, we compared 1995 and 1996 NIS national estimates with results from the NHIS/NIPRCS for the same years., Results: Both the NIS and the NHIS/NIPRCS produce similar results., Conclusion: The NHIS/NIPRCS supports the findings of the NIS.

Abstract: The National Immunization Survey (NIS) is a large federally funded survey designed to estimate vaccination coverage rates for children residing in the United States aged 19 to 35 months. In 1999, over 8 million telephone call attempts were made to obtain provider-reported vaccination histories on 22,521 children in the age range of interest.

Objective: This article reviews four surveys methodologies that have been used over the past 40 years to assess immunization rates in young children in the United States. These methods include three national surveys: (1) United States Immunization Survey (1959-1985), which was first a household and then a telephone survey; (2) National Health Interview Survey (1991-present), which interviews people in their homes; and (3) National Immunization Survey (1994-present), a random-digit-dialing telephone survey. In addition, a series of retrospective school record surveys that used standard sampling and assessment methodologies were conducted nationally during 4 school years September 1990-May 1991., Methods: Federal publications, National Immunization Conference proceedings, and Centers for Disease Control and Prevention (CDC) internal reports regarding national immunization surveys were reviewed. The methodology used in each survey is presented, and selected examples of previously tabulated results are presented., Conclusions: The assessment of immunization coverage in American preschool children requires ongoing commitment and survey expertise. Over the past 40 years the CDC's efforts to determine vaccination coverage in young children has evolved from the comparatively simple United States Immunization Survey to the current National Immunization Survey that utilizes sophisticated statistical and survey techniques to obtain the most-accurate results yet available.

The National Immunization Survey (NIS) was designed to measure vaccination coverage estimates for the US, the 50 states, and selected urban areas for children ages 19-35 months. The NIS includes a random-digit-dialed telephone survey and a provider record check study. Data are weighted to account for the sample design and to reduce nonresponse and non-coverage biases in order to improve vaccination coverage estimates. Adjustments are made for biases resulting from nonresponse and nontelephone households, and estimation procedures are used to reduce measurement bias. The NIS coverage estimates represent all US children, not just children living in households with telephones. NIS estimates are highly comparable to vaccination estimates derived from the National Health Interview Survey. The NIS allows comparisons between states and urban areas over time and is used to evaluate current and new vaccination strategies.