Michael Fendrich1, Daniel Fuhrmann2, Lisa Berger2, Charles Plate3, Douglas Lewis3, Joseph Jones3. 1. School of Social Work University of Connecticut West Hartford, Connecticut USA. Electronic address: michael.fendrich@uconn.edu. 2. Center for Applied Behavioral Health Research University of Wisconsin-Milwaukee Milwaukee, Wisconsin USA. 3. United States Drug Testing Laboratories, Inc. Des Plaines, Illinois USA.
Abstract
INTRODUCTION: Researchers have increasingly used collateral informants to validate the reports provided by primary research subjects. We assessed the utility of collateral informants for college students in a study that incorporates biomarkers to validate student reports of recent drinking behavior. METHODS: Students from a Midwestern university were randomly selected for a study in which they provided 90-day Timeline Followback data, hair and fingernail specimens for ethylglucuronide (EtG) testing, and information about collateral (friends or peers) informants who were familiar with their drinking behavior. We compared summary measures of recent drinking to collateral informant reports for the subset of 72 students who were selected to participate in the collateral validation process who had complete measures. Kappa, weighted kappa, and McNemar tests were performed to evaluate levels of agreement. We compared levels of use indicated by each informant within the context of EtG findings. We also compared respondent and collateral reports with respect to heavy drinking directly to EtG test results. RESULTS: There was considerable overlap between the reports provided by the student participants and their collateral informants. Within the context of EtG-informed analyses, collaterals rarely provided new information about heavy use beyond that provided by the study subjects. CONCLUSIONS: Collateral informants have limited utility in non-clinical studies of heavy drinking in randomly selected college students.
INTRODUCTION: Researchers have increasingly used collateral informants to validate the reports provided by primary research subjects. We assessed the utility of collateral informants for college students in a study that incorporates biomarkers to validate student reports of recent drinking behavior. METHODS: Students from a Midwestern university were randomly selected for a study in which they provided 90-day Timeline Followback data, hair and fingernail specimens for ethylglucuronide (EtG) testing, and information about collateral (friends or peers) informants who were familiar with their drinking behavior. We compared summary measures of recent drinking to collateral informant reports for the subset of 72 students who were selected to participate in the collateral validation process who had complete measures. Kappa, weighted kappa, and McNemar tests were performed to evaluate levels of agreement. We compared levels of use indicated by each informant within the context of EtG findings. We also compared respondent and collateral reports with respect to heavy drinking directly to EtG test results. RESULTS: There was considerable overlap between the reports provided by the student participants and their collateral informants. Within the context of EtG-informed analyses, collaterals rarely provided new information about heavy use beyond that provided by the study subjects. CONCLUSIONS: Collateral informants have limited utility in non-clinical studies of heavy drinking in randomly selected college students.
Authors: Lisa Berger; Michael Fendrich; Joseph Jones; Daniel Fuhrmann; Charles Plate; Douglas Lewis Journal: Addiction Date: 2013-12-13 Impact factor: 6.526