Commentary on Ondersma et al. (2012): Beyond the quest for the perfect test—drug use screening in pregnancy

Abstract

Unchanged since its articulation in 1968 by the World Health Organization, the object of screening remains ‘to discover among the apparently well who are in fact suffering from disease’ 1. Guidelines for a good screening test remain the following: the availability of treatment, latent disease state, patient acceptability and a balance of benefits to harms on the population level. Ondersma et al. 2 propose a novel screening instrument for drug use in the prenatal period. The authors are to be commended in approaching the subject area with an understanding of the shame and stigma that often accompanies drug use during pregnancy. To overcome what they term ‘hesitancy in disclosure’, the authors propose and validate an indirect screening method (Wayne Indirect Drug Use Screener: WIDUS). Using both urine and hair toxicology as the gold standard, the indirect screening instrument performed similarly to other published screening instruments and better than a direct screener, the Drug Abuse Screening Test (DAST), in their sample. As an obstetrician gynecologist who provides care to pregnant women with drug problems, I understand the importance of this topic. However, I am uncertain as to the benefit of a new screening instrument in general, and have reservations about the clinical utility of indirect methods in particular. From the 4Ps Plus to the Zuckerman–Kuhlman Personality Questionnaire (ZKPQ), the Substance Use Screening and Assessment Instruments Database managed by the Alcohol and Drug Abuse Institute at the University of Washington currently catalogues 904 unique instruments, of which 15 have been developed specifically for or validated in pregnant women 3. To my knowledge, there is a single systematic review of substance screening during pregnancy. Bradley et al. investigated seven alcohol screening instruments and failed to demonstrate any large differences in test characteristics between them 4. To date, no single method is recommended by professional societies in the United States 5, 6 or Canada 7. The overall absence of any evidence privileging one screening method over another begs the question as to whether additional instruments need to be developed. One of the problems with screening for drug use during pregnancy is the inconsistent uptake of screening among providers. Although universal screening is recommended 8, it is far from achieved 9. More common in the United States is selective screening based on race, poverty and poor pregnancy outcomes 10. This selective screening is highly discretionary 11, and betrays a disturbing bias among providers of women's health care towards the legal coercion of pregnant women with drug problems 12. A new screening instrument is unlikely to change these realities. Unlike direct screening methods which rely on self-report, indirect methods attempt to capture use without using a tool that refers to drug use. The concept and appeal of indirect screening has existed for several decades 13. However, the evidence to support its superiority is lacking. For example, a recent review of one of the most used indirect screeners (Substance Abuse Subtle Screening Inventory: SASSI) showed no benefit in detecting substance use correctly compared with direct methods 14. Screening only makes public health sense if linked to evidence-based interventions. Clinically, a positive screen invites the opportunity to discuss drug problems and provide brief interventions based upon the individual's readiness to change. Ondersma et al. note that the primary motivation for indirect screening is that under-reporting leads to missed opportunities for interventions. However, non-disclosure reflects a lack of readiness to change (a context in which brief interventions are not indicated) or a mistrust of the provider/clinical encounter (a context in which brief interventions may be harmful), or both. Therefore, it is difficult to see how indirect screening would lead to brief interventions without confrontation, an approach that Ondersma et al. describe correctly as ‘inaccurate, unethical and counterproductive’ 2. In this regard, the clinical utility of indirect screening appears more similar to toxicology testing than to direct screening and therefore raises similar ethical questions 15, specifically the potential harms of false positive results. Using the data from Ondersma et al.'s Table 3, I calculate a false positive rate of at least 7% (cut score 4) to as high as 34% (cut score 2) in the WIDUS screen, which is higher than the DAST-10 in their sample (range 2–10%). This is concerning, as one of the consequences of a positive screen may be a report to Child Protective Services. In conclusion, I believe that we need to move beyond the quest for the perfect test and focus instead on educating providers and empowering mothers such that the complex issues of drug use in pregnancy can be addressed safely and with compassion. None.