Substance use and disorders associated with it occur simultaneously with PTSD. Whether it's a clinician attempting to gain a full clinical picture from a single patient or a researcher planning a nationwide survey, obtaining self-report data is the first step to examining an individual's substance use. Validation of self-report information with additional data such as biological tests may be necessary because self-reports, even with the most honest individuals, are subject to bias.

When urine toxicology screening is used to verify self-report data in studies of adult populations, approximately 92 percent to 99 percent who self-report current substance use are biologically validated, while the remaining 1 percent to 8 percent who report substance use receive negative toxicology screen results. The latter disparity may be a product of the limitations of biological testing. However, researchers and clinicians have the most concern with the approximately 5 percent to 13 percent of individuals who fail to endorse self-report measures but demonstrate evidence of substance use on biological measures.

In studies examining PTSD samples, Roger Weiss and colleagues at Harvard University reported that 4.7 percent of civilian women in treatment for concurrent PTSD and substance abuse gave false negative self-reports, while Patrick Calhoun and colleagues at Duke University found an 8 percent rate among male veterans at a PTSD outpatient clinic.

So how is it possible to obtain an accurate self-report of substance use? In general, the more specific the information requested, the more likely the individual will respond in a valid fashion. To offer a familiar analogy, when obtaining trauma histories, inquiries using loaded and more general terms such as "physical abuse" result in underreporting, while inquiries about specific acts, such as whether one was hit or slapped, improve reporting.

Similarly, in obtaining histories of substance use, it is preferable to ask substance-specific questions about quantity, frequency and recent use as a part of a comprehensive assessment, rather than the more global, "Do you use alcohol or drugs?" Improved phrasing might start with, "When did you last use [a given substance]?" and "How much did you use at that time?" These questions usually are followed by inquiries concerning the number of days in the past week or month the respondent used a given substance and how long the respondent used the substance over time. Questions about each substance should be asked separately in a systematic manner. Self-administered questionnaires, including those administered on computers, result in an increased rate of valid reporting when compared with reports obtained using interviewer-administered questionnaires.

When examining recall of substance use over an extended period of time, an effective approach is the Time-line Follow-back Method, as developed by Linda and Mark Sobell at the University of Toronto Addictions Research Foundation and further validated by William Fals-Stewart and colleagues at Old Dominion University. This calendar-based approach uses anchoring time points, such as holidays and birthdays, to stimulate recall of substance use in relationship to these events, and the preceding and following days and weeks. In addition to acting as a stimulus to memory, this method implicitly acknowledges that individuals often change their usual pattern of substance use in relation to potentially positive or negative emotionally charged events, therefore permitting increased accuracy in reporting.

Biological tests often serve to confirm or disconfirm self-reports. The ability of biological tests to detect parent compounds and metabolites over time varies in a substance-specific fashion, commensurate with quantity ingested and metabolic and excretion patterns. Complicating matters further, within a given methodology such as urine or hair testing, the duration over which parent compounds and relevant metabolites may be detected also varies with the sample collection and detection methods used. For example, alcohol generally is detectable by urine testing for up to four to six hours after use, while measurable by the Breathalyzer up to eight to ten hours after use. Conversely, the cocaine metabolite, benzylecognine, is measurable in urine specimens for up to three to seven days after use. Hair testing is the biological method currently providing reliable measures for use over the longest duration of time. This method detects cocaine or heroin as far back as three months after use.

It is crucial to remember that all biological tests have the potential of false negatives or positives and, in hair toxicology testing, for racial biases. Therefore, a systematic means of resolving contradictions between self-reports and biological tests must be used. In most research protocols, a positive self-report is accepted as sufficient evidence of substance use. An exception is when a research protocol places substance users at excessive risk; here, for safety and ethical reasons, multiple urine toxicology screens are obtained over time as primary evidence, with self-report as corroborative.

For more information about substance use disorders, see the textbook edited by Mark Galanter and Herb Kleber (The American Psychiatric Press Textbook of Substance Abuse Treatment, Second Edition, American Psychiatric Press Inc., 1999). The 1998 article by Weiss and colleagues in volume 155 of the American Journal of Psychiatry and the 2000 publication by Calhoun and associates in volume 68 of the Journal of Consulting and Clinical Psychology also provide further research-related insights into the issues surrounding measuring substance use in patients with PTSD.

This brief report is sponsored by ISTSS's Research Methodology Special Interest Group. For more information or to become a member of this Special Interest Group, contact chairs Daniel and Lynda King, National Center for PTSD; e-mail: king.daniel@boston.va.gov or lking@world.std.com.