Heather G. Miller

Audio-computer interviewing to measure risk behaviour for HIV among injecting drug users: a quasi-randomised trial.

BACKGROUND We aimed to assess audio-computer-assisted self-interviewing (audio-CASI) as a method of reducing under-reporting of HIV risk behaviour among injecting drug users. METHODS Injecting drug users were interviewed at syringe-exchange programmes in four US cities. Potential respondents were randomly selected from participants in the syringe exchanges, with weekly alternate assignment to either traditional face-to-face interviews or audio-CASI. The questionnaire included items on sociodemographic characteristics, drug use, and HIV risk behaviours for 30 days preceding the interview. We calculated odds ratios for the difference in reporting of HIV risk behaviours between interview methods. FINDINGS 757 respondents were interviewed face-to-face, and 724 were interviewed by audio-CASI. More respondents reported HIV risk behaviours and other sensitive behaviours in audio-CASI than in face-to-face interviews (odds ratios for reporting of rented or bought used injection equipment in audio-CASI vs face-to-face interview 2.1 [95% CI 1.4-3.3] p=0.001; for injection with borrowed used injection equipment 1.5 [1.1-2.2] p=0.02; for renting or selling used equipment 2.3 [1.3-4.0] p=0.003). INTERPRETATION Although validation of these self-reported behaviours was not possible, we propose that audio-CASI enables substantially more complete reporting of HIV risk behaviour. More complete reporting might increase understanding of the dynamics of HIV transmission and make the assessment of HIV-prevention efforts easier.

AIDS: Sexual Behavior and Intravenous Drug Use

Since the acquired immunodeficiency syndrome (AIDS) epidemic is partly a social phenomenon changes in the social behaviors that spread the disease are as necessary as the development of vaccines and other therapies. To help in the design implementation and evaluation of programs to curb the spread of AIDS there is a need for greater understanding of the human behaviors that transmit human immunodeficiency virus (HIV) infection as well as the social contexts in which these behaviors occur. Toward this end the US Committee on AIDS Research and the Behavioral Social and Statistical Sciences was asked to: 1) describe what is known about the spread of HIV and AIDS in the US with special attention to the quality of information at hand and the kind of additional information that is needed; 2) identify critical populations and indicate objectives and tasks related to them; 3) describe existing research findings in the behavioral and social sciences that should be useful in planning and choosing among interventions designed to control the spread of HIV infection; 4) describe existing research on interventions intended to facilitate behavior changes and ways to evaluate their effectiveness; and 5) identify new research that should be undertaken in these areas. The 7 chapters of the resultant report are divided into 3 parts: Understanding the Spread of HIV Infection; Intervening to Limit the Spread of HIV Infection; and Impediments to Research and Intervention. Also included are 6 background papers. Among the recommendations of the Committee are: vigorous programs of basic social and behavioral research on human sexual behavior; availability of condoms and treatment for sexually transmitted diseases through local public health authorities; drug treatment upon request for intravenous drug users; trials of sterile needle programs; and anonymous HIV antibody testing with appropriate counseling on a voluntary basis.

Interview Mode and Measurement of Sexual Behaviors: Methodological Issues.

Studies of sexual and other sensitive behaviors are often fraught with a variety of reporting biases. When IAQs are used to collect data, respondents may underreport certain sensitive behaviors and overreport normative behaviors. SAQs can also pose problems: requiring that respondents be literate and able to follow skip patterns. In recent years, the development of computerized technologies--audio-CASI and T-ACASI--have begun to overcome some of the limitations of IAQs and SAQs. By providing a more private mode for data collection and standardized delivery of all questions, as well as automated skip patterns and range checks, audio-CASI and T-ACASI have been tested in a number of studies and found to be an effective way of reducing response bias, and thus, contributing to a better understanding of the prevalence and patterns of sexual and other sensitive behaviors.

Correlates of sexually transmitted bacterial infections among U.S. women in 1995.

CONTEXT Sexually transmitted diseases (STDs) of bacterial origin such as gonorrhea and chlamydial infection can lead to pelvic inflammatory disease (PID) and infertility. Identifying behaviors and characteristics associated with infection may assist in preventing these often asymptomatic diseases and their sequelae. METHODS Data from 9,882 sexually active women who participated in the 1995 National Survey of Family Growth describe the characteristics of women who report a history of infection with a bacterial STD or of treatment for PID. Multivariate analysis is used to determine which demographic characteristics and sexual and health-related behaviors affect the likelihood of infection or the occurrence of complications. RESULTS Overall, 6% of sexually active women reported a history of a bacterial STD, and 8% reported a history of PID. Women who first had sexual intercourse before age 15 were nearly four times as likely to report a bacterial STD, and more than twice as likely to report PID, as were women who first had sex after age 18. Having more than five lifetime sexual partners also was associated with both having an STD and having PID. PID was more common among women reporting a history of a bacterial STD (23%) than among women who reported no such history (7%). In multivariate analyses, age, race, age at first intercourse and lifetime number of sexual partners had a significant effect on the risk of a bacterial STD. Education, age, a history of IUD use, douching and a history of a bacterial STD had a significant impact on the risk of PID, but early onset of intercourse did not, and lifetime number of partners had only a marginal effect. CONCLUSIONS The pattern of characteristics and behaviors that place women at risk of infection with bacterial STDs is not uniform among groups of women. Further, the level of self-reported PID would suggest higher rates of gonorrhea and chlamydial infection than reported.

Zenilman's anomaly reconsidered : Fallible reports, ceteris paribus, and other hypotheses

Background and Objectives: In the January‐February, 1995 issue of Sexually Transmitted Diseases, Zenilman and colleagues reported a null association between incident sexually transmitted diseases (STDs) and self‐reported condom use. That anomalous finding generated a flurry of letters to the editor, some of which were quite heated. This article reconsiders the Zenilman teams results. Study Design: New statistical analyses were conducted to test two hypotheses that sought to account for the null association: (1) deviation from study protocol, and (2) differential risks of acquiring an incident STD among segments of the study population that varied by reported level of condom use. Results: No support was found for hypotheses concerning deviation from study protocol and differential risk of acquiring an incident STD by level of condom use. Indeed, for respondents who reported multiple sexual partners, the analyses found increased rates of infection among those who reported more consistent condom use. Conclusions: Two of the most promising hypotheses for explaining Zenilmans anomalous findings are unsupported by reanalysis of the available empirical evidence. It is still possible that respondents who reported that they used condoms consistently differed from self‐reported nonusers or inconsistent users in some way that altered their risk of acquiring an STD and thus obscured the protective effects of properly used condoms. Nonetheless, as Zenilman and others suggest, fallibility in self‐reports of condom use remains the primary suspect as the cause of these anomalous results. Such fallibility may be particularly pronounced when self‐reported behavioral data are collected in contexts that include strong educational campaigns or other norm‐setting interventions.

Effects of interview mode on bias in survey measurements of drug use: do respondent characteristics make a difference?

Three recent empirical studies have provided strong evidence that self-administered questionnaires (SAQs), compared with interviewer questioning, substantially improve the reporting of drug use in population surveys. Specifically, SAQs appear to diminish underreporting bias. Two of these studies previously reported that this effect of interview mode varied significantly across gender, race/ethnicity, and age. Data from a randomized experiment embedded in the 1990 National Household Survey of Drug Abuse (NHSDA) field test were reanalyzed to test for those interaction effects. To better replicate prior studies, the NHSDA field test sample was restricted to people ages 18 to 45 (N = 1,877). The results of our statistical analyses generally replicated the finding of a main effect of SAQs on the reporting of drug use. However, only weak evidence was found to support the hypothesis that the advantage of SAQs varies substantially by the gender, race/ethnicity, or age of the respondent.

NAAT-identified and self-reported gonorrhea and chlamydial infections: Different at-risk population subgroups?

Background Information on the characteristics and behaviors of persons at high risk for gonorrhea and chlamydial infection has typically been derived from studies of sexually transmitted disease (STD) clinic populations. The Baltimore STD and Behavior Survey (BSBS) used urine-based nucleic acid amplification testing (NAAT) to assess the prevalence and behavioral correlates of gonorrhea and chlamydial infection in a population-based cross-sectional survey of adults in Baltimore, Maryland. Goal The goal of this study was to examine the demographic characteristics and behavioral markers of gonorrhea and chlamydial infection as reported by adults with a self-reported history of gonorrhea and chlamydial infection and to compare these to the characteristics and behaviors of individuals with current NAAT-identified gonorrhea and/or chlamydial infection. Study Design A probability sample of adults aged 18 to 35 years residing in Baltimore was evaluated with collection of urine specimens and administration of a health and behavior survey. Data and specimens were collected between January 1997 and September 1998. Results Respondents with NAAT-detected gonorrhea and/or chlamydial infection (7.9%) did not report a history of high-risk behaviors or more recent occurrences of those behaviors, and the majority were asymptomatic. However, adults in our study who self-reported a history of infection (26.0%) were more likely than those with no history of infection to report multiple partners, paid sex, partners with prior STDs, and STD symptoms—a pattern consistent with findings described in previous clinic-based reports. Conclusion The risk profile generated from studies of clinic populations, with a focus on symptomatic disease, may not characterize the broader population with current, untreated, largely asymptomatic infection.

Entretiens autoadministrés sur ordinateur et mesure des comportements sensibles

Rogers Susan M., Gribble James N., Turner Charles F., Miller Heather G.- Encuestas auto-administradas a traves de ordenador y medicion de cuestiones delicadas Las encuestas que confian en las respuestas de los individuos sobre comportamientos controvertidos, estigmatizados о ilegales pueden contener sesgos. Investigadores del Research Triangle Institute, de Estados Unidos, han desarrollado una tecnica audio de auto- encuesta asistida por ordenador (audio-CASI) para superar las limitaciones de los cuestionarios auto-administrados de forma tradicional, «con lapiz y papel » (SAQs) o reali- zados por un entrevistador (IAQs). El desarrollo de la tecnica audio-CASI ha alterado de forma significativa el contexto de las entrevistas para medir comportamientos sexuales y otras cuestiones delicadas, ya que respeta la privacidad pero no requiere que el individuo sepa leer y escribir. Este articulo analiza el desarrollo cientifico de audio-CASI y describe los resultados de experimentos metodologicos que se han realizado con el objetivo de comparar este meto- do a otras formas de entrevista. En 1995 se probo audio-CASI en dos encuestas norteameri- canas de gran envergadura: la Encuesta Nacional de Hombres Adolescentes (NSAM) y la Encuesta Nacional de Crecimiento de la Familia (NSFG). Las encuestas hallaron niveles si- gnificativamente mas elevados en la declaracion del uso de drogas, contactos sexuales con individuos del mismo sexo y recurso al aborto provocado cuando se utilizo audio-CASI que a traves de IAQs y SAQs.

Measuring AIDS-Related behaviors in older populations - Methodological issues

Because of a dearth of research on reporting biases in the measurement of HIV-relatedsexual and drug use behaviors in older populations, it is frequently assumed thatmethodological findings of research conducted with younger populations will generalize to older respondents. In this study, estimates of the effect of the experimentalmanipulation of interview mode (interviewer administered vs. self-administered)were derived separately for three age strata: 12 to 49, 50 to 64, and 65+. Results ofthese analyses indicate that there were a number of noteworthy reversals in whichinterviewer-administered questioning in the older age strata produced higher estimates of the prevalence of substance use or alcohol-related problem behaviors. Theseresults suggest that caution should be exercised in making generalizations fromstudies of reporting bias for HIV-related behaviors to older populations.

Strategies for Understanding Drug Use in the 21st Century: Preface

0 More frequent use of survey technologies-such as audio computerassisted self-interviewing (audio-CAS1)-to assure complete privacy for respondents and thus encourage more accurate reporting Routine collection and analysis of data-such as medical treatment records and biological specimens-which are not subject to the same constellation of biases that afflict survey measurements Systematic attempts to model drug use trends over time using both survey data and other indicators of drug use