To begin to resolve conflicts among current competing taxonomies of child and adolescent psychopathology, the authors developed an interview covering the symptoms of anxiety, depression, inattention, and disruptive behavior used in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994), the International Statistical Classification of Diseases and Related Health Problems (ICD-10; World Health Organization, 1992), and several implicit taxonomies. This interview will be used in the future to compare the internal and external validity of alternative taxonomies. To provide an informative framework for future hypothesis-testing studies, the authors used principal factor analysis to induce new testable hypotheses regarding the structure of this item pool in a representative sample of 1,358 children and adolescents ranging in age from 4 to 17 years. The resulting hypotheses differed from the DSM-IV, particularly in suggesting that some anxiety symptoms are part of the same syndrome as depression, whereas separation anxiety, fears, and compulsions constitute a separate anxiety dimension.

Trena M. Ezzati-Ricea, Martin R. Frankelb, David C. Hoaglinb, John D. Loftc, Victor G. Coronadod and Robert A. Wrighta aCenters for Disease Control and Prevention, National Center for Health Statistics, 6525 Belcrest Road, Room 915, Hyattsville, MD 20782, USA bAbt Associates Inc., 55 Wheeler Street, Cambridge, MA 02138, USA cResearch Triangle Institute, 203 North Wabash, Suite 1220, Chicago, IL 60601, USA dCenters for Disease Control and Prevention, National Immunization Program, Corporate Square Boulevard, Atlanta, GA 30329, USA

Background The National Immunization Survey (NIS) provides coverage estimates for children aged 19 to 35 months in each of 78 state and urban areas designated as Immunization Action Plan (IAP) areas by the National Immunization Program in the Centers for Disease Control and Prevention (CDC). Coverage-level estimates for the 78 areas and for the nation as a whole are published in Morbidity and Mortality Weekly Report (MMWR), U. S. Department of Health and Human Services/Public Health Service (Ezzati-Rice et al., 1995). The NIS uses a list-assisted random-digit-dial (RDD) sample design. Households with eligible children are interviewed by telephone and invited to participate in the survey by providing immunization data about each eligible child in the household. Respondents are encouraged to rely on shot cards given to them by physicians and other immunization providers. Although more than half of the respondents do use these records as they respond to the survey, a significant proportion do not. Even when respondents do refer to shot records during the interview, they may have incomplete records. Parents can lose a copy of the shot card and be given a new one that omits earlier vaccinations, or they may have changed physicians and kept only the shot card from the newer provider. The NIS Provider Record Check Study was designed to address these shortcomings in the household data by obtaining vaccination records from providers themselves. During the telephone survey with eligible households, interviewers request permission from the respondent to contact immunization providers and obtain mailing addresses for those providers. Obtaining correct mailing addresses from household respondents is an essential component of contacting providers to request the necessary immunization records. This paper describes an enhancement to the NIS CATI system questionnaire that is designed to increase the accuracy of provider address information; reduce the time required for interviewers to gather address information; and reduce the time required to edit and locate provider addresses before mailing requests to obtain or complete children's immunization information.

Nonresponse causes bias in survey results when nonrespondents differ in important ways from respondents. This paper focuses on potential nonresponse bias in surveys that first screen a sample to determine eligibility for the study and then complete an interview with eligible respondents. Nonresponse to the screening interview may be correlated with eligibility criteria in such a way that typical response rate calculations fail to accurately represent the proportion of eligible sample members that participate in the survey. An estimate of the eligibility rate among sample members that cannot be contacted and screened during the field data collection period is an important component of the formula used to calculate overall response rates. As suggested in recommendations published recently by the American Association for Public Opinion Research (AAPOR, 1998), if there is reason to suspect that screening respondents and nonrespondents have different eligibility rates, then it is important to use distinct final dispositions (e.g., telephone dialing outcomes or in-person contact outcomes) for screened and unscreened cases and develop separate estimates of eligibility rates. This paper extends an approach suggested by Massey (1995) for estimating the response rate in telephone surveys with screening by using external data to estimate the percentage of eligible households. In his paper Massey used as an example the telephone call outcomes from the first three calendar quarters (April 1994-December 1994) of the National Immunization Survey (NIS), a large random-digit-dialing (RDD) telephone survey . However, the conceptual approach is applicable in any survey, regardless of data collection mode, where screening for

This evaluation, sponsored by the Centers for Disease Control and Prevention (CDC) and conducted by RTI International, seeks to estimate hospital screening rates for HIV and other perinatal infectious diseases among new mothers and their infants as documented in hospital medical charts. The evaluation is underway in selected geographic areas (counties and states) in 7 states and in selected hospitals within these areas. We will discuss the impact of efforts to anticipate concerns about the applicability of HIPAA to this project's protocol. We are sampling births using state birth certificates from calendar year 2003 as the sample frame and abstracting data from the corresponding charts on screening for HIV and other infectious diseases. A letter outlining the status of the project with respect to HIPAA was included in the project materials mailed to health departments and to hospitals selected for the assessment. We believe this letter was significant in allaying potential HIPAA concerns. The project has been approved by state health departments and HIPAA concerns have not been a barrier in any hospital thus far.