Concurrent validity of a brief self-report Drug Use Frequency measure
Introduction
Reliable and valid self-report information on the use of alcohol and illicit drugs is needed to evaluate outcomes of substance abuse treatment and prevention programs and to estimate the extent of use and abuse of these substances in epidemiological studies. An extensive literature supports the reliability and validity of self-reported alcohol use data that are gathered with carefully developed, well-tested methods under appropriate conditions Brown et al., 1992, O'Farrell & Maisto, 1987, but self-reports of illicit drug use have received considerably less investigation. The Timeline Followback interview method (TLFB; Sobell & Sobell, 1996), a widely used and well-validated measure of self-reported alcohol use, has also been used recently to measure illicit drug use.
Fals-Stewart, O'Farrell, Rutigliano, Freitas, and McFarlin (2000) recently conducted a comprehensive analysis of the psychometric properties of the TLFB for measuring illicit drug use. TLFB interviews were conducted with adult drug-abusing patients seeking treatment for substance abuse (N=113) at baseline, posttreatment, and quarterly thereafter for 12 months. It was found that patients' reports about their drug consumption using this method generally had high (a) test–retest reliability, (b) convergent and discriminant validity with other measures, (c) agreement with collateral informants' reports of patients' substance use, and (d) agreement with results from patients' urine assays.
Results of the Fals-Stewart et al. (2000) investigation, when combined with other studies supporting the reliability and validity of TLFB self-reports of illicit drug use (e.g., Ehrman & Robins, 1994, Hersh et al., 1999), suggest the TLFB is a psychometrically sound measure of illicit drug use. However, given the relatively large amount of time required to conduct a TLFB interview, there are a number of situations for which the TLFB may not be a practical assessment method. If a study or other assessment situation (e.g., assessment as part of treatment planning) requires a brief measure of drug use, then the extensive TLFB interview method would not be appropriate.
We faced this dilemma when designing a study of alcohol and drug use (a) in a sample of married or cohabiting male patients entering substance abuse treatment and their female partners and (b) in a community comparison sample. We chose the TLFB for the patients in the clinical sample because study aims required a detailed, in-depth measurement of patients' substance use. However, study aims and practical constraints required a briefer, less in-depth measure of substance use for the other study participants (i.e., the treatment sample patients' female partners and couples in the community comparison sample). To measure alcohol use, we used the Quantity–Frequency (QF) items from the Health and Daily Living Form (HDL; Moos, Cronkite, Billings, & Finney, 1984). To measure illicit drug use, we developed a brief Drug Use Frequency (DUF) measure that was adapted from the HDL section on medication use to which we added items for the same drug classes we assessed with the TLFB. Use of each drug was rated on the same frequency scale used in the QF items to gauge frequency of drinking.
This report provides initial validity information on the DUF measure. We examined two sources of concurrent validity information for the DUF — the well-validated TLFB measure and knowledgeable collateral informants (i.e., spouses). Data from 106 male alcoholic patients and their female partners (treatment sample couples) and 130 male and female participants without current alcohol and drug problems (community sample couples) provided information on the extent of self-reported drug use in the prior 6 months and the concurrent validity of the DUF measure in these treatment and community sample couples.
Section snippets
Participants
Participants gave written informed consent to take part in the Harvard Couples Project, a study comparing individual and relationship functioning in two groups of couples.
Characteristics of study participants
Table 1 has demographic characteristics and current and lifetime substance problem diagnoses from the SCID for the treatment and community sample couples. On average these couples had been together over 10 years (mean±S.D. years married or living together=11.6±10.1 for treatment sample, and 14.2±9.2 for community sample).
Treatment sample male alcoholic patients' self-reported drug use
The left side of Table 2 presents for the 6 months prior to the assessment (1) number of patients and percent of the sample using six types of drugs and (2) frequency of use
Discussion
This initial psychometric evaluation of a brief DUF measure showed generally good concurrent validity of self-reported frequency of illicit drug use on the DUF. Concurrent validity was high for the DUF in relation to the TLFB for all drug classes examined. Concurrent validity, as measured by agreement between self- and collateral reports on the DUF, was high for the most frequently used drugs, namely cannabis, heroin, and cocaine. Collateral agreement was lower for less frequently used drugs
Acknowledgements
This research was supported in part by the Department of Veterans Affairs, and by grants from the National Institute on Alcohol Abuse and Alcoholism (R01AA10356 and K02AA00234) and the National Institute on Drug Abuse (R01DA12189).
References (9)
- et al.
The utility of self-report and biological measures of alcohol consumption in alcoholism treatment outcome studies
Advances in Behavior Research and Therapy
(1987) Diagnostic and statistical manual of mental disorders
(1994)- et al.
Self-reports by alcohol and drug abuse inpatients: factors affecting reliability and validity
British Journal of Addictions
(1992) - et al.
Reliability and validity of 6 month timeline reports of cocaine and heroin use in a methadone population
Journal of Consulting and Clinical Psychology
(1994)