Commentary on Ondersma et al. (2019): Will better self-report screening instruments be enough to detect drug use during pregnancy?

Abstract

Better screening instruments to detect drug use during pregnancy are needed, but under-reporting of use in this population will probably continue as long as many pregnant women who use drugs fear social and/or legal repercussions..

An estimated 8.5% of pregnant women in the United States have used an illicit drug in the past month [1]. However, prevalence is probably much higher, as survey studies including biological testing have confirmed extensive under-reporting of use in this population [2,3]. For example, in a recent study of 279 457 pregnant women in the state of California, the estimated prevalence of prenatal cannabis use increased from 4.2% when estimated via self-report to 7.1% when correcting for positive detection via urine testing [2]. As the prevalence of illicit drug use continues to increase among pregnant women in the United States [1,4], accurate estimates of use are needed more than ever to help guide the public health response.

Ondersma et al. [5] conducted a study examining the accuracy of five self-report instruments to detect drug use during pregnancy. None of the measures examined had both high sensitivity and high specificity in detecting reported and/or biologically verified drug use. This lack of accuracy could have been due, in part, to various scale-related factors including: (1) assessment of problems related to drug use rather than assessment of drug use itself, (2) non-specific drug use time-frames and (3) specific time-frames too wide to accurately detect use while pregnant (e.g. past-year use).

Results of this study indicate that validated self-report measures are needed to detect prenatal drug use. Measures specifically querying past-month use or use while pregnant would be optimal. However, we must not forget that under-reporting is especially common in this population and will probably continue regardless of the self-report measures used. Thus, perhaps the elephant in the room is that even valid self-report measures will not accurately detect a substantial portion of drug use among pregnant women.

Drug use is often stigmatized, especially among pregnant women, so we should not expect under-reporting of drug use to stop among this population any time soon. Anonymity, as was promised to women in Ondersma et al.’s [5] study, may decrease under-reporting in research settings, but anonymity or confidentiality will probably not solve this problem. Many pregnant women who use drugs—perhaps rightfully—fear repercussions from disclosing use. Research suggests that some pregnant women who use drugs even delay or avoid prenatal care appointments altogether because they fear stigma and/or child protection services [6,7]. As long as pregnant women who use drugs have to fear such social and legal repercussions, we cannot expect all pregnant women to disclose drug use—in research or in health-care settings.

As under-reporting in this population is prevalent and of specific concern, biological testing as an adjunct to self-report measures in research settings would help to provide better prevalence estimates. However, urine testing is the most common biological measure used in such settings, and this method has some limitations. Although urine testing is often considered a ‘gold standard’ for toxicological detection of drug use, it can usually only detect use that occurred in the past few days [8]. Hair testing, although more expensive, can detect wider window periods of drug use [8-10], and hair segments can be measured to detect approximately when drug use occurred [11]. Mass-produced urine tests, which are most commonly used, are also limited because they often can only detect a handful of common drugs. There are many prevalent drugs, and hundreds of new psychoactive substances have emerged in recent years [12]; therefore, it is likely to be impossible to detect every drug. However, identifying effective means of laboratory testing for a wider variety of drugs would allow for more accurate detection of use.

More accurate self-report measures are needed to detect drug use during pregnancy, but we need to remain cognizant that a large portion of drug use among pregnant women will remain undetected as long as fear of social and legal repercussions exists.