PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Polen MR, Whitlock EP, Wisdom JP, et al. Screening in Primary Care Settings for Illicit Drug Use: Staged Systematic Review for the United States Preventive Services Task Force [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2008 Jan. (Evidence Syntheses, No. 58.1.)


Limitations of the Literature Review

This review was not intended to be a comprehensive, cumulative review of evidence regarding drug misuse screening and treatment. It was designed, rather, to address whether there still is insufficient evidence available to answer critical key questions required for the USPSTF to make a recommendation on this topic as a clinical preventive service in primary care. Our review was limited to the defined scope of work as a staged review to update a previous USPSTF recommendation.

One limitation in this review was our focus on the most prevalently misused substances and those most likely to be associated with abuse or dependence. While the misuse of prescription-type drugs is fairly prevalent (2.5% of persons age 12 and over),2 this category represents at least four different types of medications (pain relievers, tranquilizers, stimulants - including methamphetamine - and sedatives), and multiple individual medications (see Table 1). These different substances represent different misuse profiles, including different average ages of initiation, sources of drug, trends in the number of users, and annual incidence of new users. Misuse of prescription medications is likely to be a growing public health problem and should be considered in future USPSTF updates for this topic.

Drug Misuse Treatment (Key Questions 4/5/5a)

The drug misuse treatment literature is voluminous and heterogeneous with regard to types of drugs, types of drug treatments, and types of study designs. We applied a series of inclusion and exclusion criteria to identify the most relevant and valid research. A clear understanding of these criteria, listed in Appendix B, is necessary to judge the adequacy and applicability of our findings. After reviewing much of this literature, we focused our review on treatment for the four categories of drugs (opiates, cocaine, cannabis, and mixed drugs) that represent the most prevalent and addictive illicit drugs in the US. Also, in order to efficiently examine the evidence regarding the efficacy of drug misuse treatment, we first reviewed existing systematic and authoritative reviews which included evidence from RCTs or from controlled trials comparing drug misuse treatment to placebo or no (minimal) treatment. We created bridge searches (as necessary) to fill the gaps in the literature. Many studies were excluded after careful review, mostly due to design (uncontrolled studies, comparative effectiveness studies, or studies not reporting outcomes designated a priori in our analytic framework). Excluded studies are identified in Appendix G.

Two of the exclusion criteria we applied to the drug misuse treatment literature (key questions 4, 5, and 5a) markedly reduced the volume of included evidence: detoxification/withdrawal studies and studies of comparative treatment effectiveness. We excluded detoxification/withdrawal studies because we conceptualized detoxification as an intermediate step with short-term outcomes designed to stabilize individuals and prepare them for drug misuse treatment, rather than as “treatment” itself. We excluded comparative effectiveness studies (e.g., medication plus counseling versus placebo plus counseling, or medication dosage comparisons) because they did not provide evidence relevant to establishing the efficacy of treatment versus no treatment. This decision has been criticized by some drug misuse treatment researchers, who feel that it is unethical to conduct trials in which treatment-seeking individuals are assigned to no-or-minimal-treatment control conditions, because they believe the efficacy of drug misuse treatment is established. Both detoxification and comparative effectiveness were frequently addressed in systematic reviews and individual trials.

A potential limitation of our review of health outcomes following drug misuse treatment is that by limiting the treatment literature to RCTs and CCTs, which tend to have relatively short follow-up periods, we may have reduced the likelihood of finding studies documenting long-term improvements in morbidity and mortality. We searched explicitly, however, for cohort studies of health effects associated with changes in drug use and believe we would have located most longer-term follow-up trials reviewed for our treatment benefit questions if these trials were available.

The treatments tested in the 17 trials included in our review are relatively heterogeneous. All but one of the treatments for opiate misuse, and all but one of the treatments for cocaine misuse, are medications, whereas all four treatments for cannabis misuse, and the one trial for opiate and cocaine use, are counseling interventions. A common theme is that the studies were conducted among non-screened populations (with one exception, Bernstein 200524). Participants were frequently recruited through advertisements or as they sought treatment at an existing drug treatment agency. Because it is an exception to this norm, the Bernstein trial deserves special comment. Bernstein (2005) was the only trial in which participants were recruited by screening an asymptomatic, outpatient medical clinic population for drug use. The Bernstein population may not have had levels of internal or external motivation to reduce drug use similar to those in the treatment-seeking populations examined in the other studies. The Bermstein trial was also unique in that participants reported sub-diagnostic levels of drug use. These participants may not have met diagnostic criteria for drug misuse (i.e., abuse or dependence criteria from the DSM-IV). It thus differs from the other trials along two dimensions—motivation and addiction severity.

Linking Changes in Drug Misuse to Health Outcomes (Key Question 7)

We identified eleven relevant longitudinal studies that linked reduction in, or cessation of, drug misuse to morbidity or mortality. Results were mixed among studies of young adults or adults, with perhaps the strongest evidence of benefit coming from a Danish study that found the risk of mortality over a 10-year period was reduced by 55% among former opiate addicts who had become abstinent, relative to continuing daily drug users.35 Two other long-term studies of mortality, however, did not find reduced risks among former opiate or injection drug users. Other outcomes at 6–12 months generally support benefit through improvement in compliance with or response to necessary medical care (HAART), improvement in depressive and anxiety symptoms, or improvement in physical health measures, with reduced use or abstinence among injection drug users (of opiates or cocaine) compared to ongoing users. Factors that differentiated those who reduced or stopped drug misuse and those who continued to use may explain some of these differences. Also, these studies frequently examined only two time points, showing cross-sectional correlations between contemporaneous changes in drug misuse and morbidity outcomes, rather than linking reductions in drug misuse with subsequent improvements in health. The two studies of cocaine and cannabis use among pregnant women provided inconsistent results, with one study43 finding no reliable evidence that cocaine or cannabis use during pregnancy was associated with poorer birth outcomes, and one44 finding that stopping cocaine use early in pregnancy was associated with some improvements in birth outcomes relative to continuing users.


The central goal of the staged review process is to establish the sufficiency of evidence for answering critical key questions about drug misuse screening as a clinical preventive service in primary care. The following details our provisional conclusions about this evidence, organized by critical key question (Table 5).

Table 5. Summary of Evidence Quality by Key Question and Population.

Table 5

Summary of Evidence Quality by Key Question and Population.

Key Question 1

We found no studies addressing whether drug misuse screening programs in primary care reduce morbidity or mortality in any of the four population subgroups we examined, and therefore provisionally conclude there is insufficient evidence for this key question.

Key Questions 4/5/5a

Among screened individuals

We found one trial by Bernstein 200524 providing evidence that drug misuse treatment decreases drug misuse in screened, asymptomatic individuals (key question 5). No studies in screened individuals addressed morbidity or mortality (key question 4), or intermediate social or legal outcomes (key question 5a), in any of the four populations.

We, therefore, provisionally conclude that there is there is some evidence that drug misuse treatment reduces drug misuse in screened, asymptomatic individuals. There is insufficient evidence, however, that drug misuse treatment in such individuals improves morbidity or mortality, or intermediate social and legal outcomes.

Among treatment-seeking individuals

All but one of the treatment studies we examined reported on treatment-seeking individuals who may have presented for treatment as a result of internal motivators, external motivators, or a combination of both. Because our inclusion criteria (Appendix B) set a high threshold for study design and quality, we expect that our results represent the strongest evidence available for the health, drug, and intermediate outcomes we considered. This evidence is very limited for health outcomes, since most studies did not report health outcomes. Among those that did, only three reported significant treatment effects on symptoms of depression or anxiety, and two of these reported multiple non-significant effects on other psychiatric measures.

The evidence supporting the efficacy of drug misuse treatment on drug use intermediate outcomes was more robust, with at least one trial showing significant improvements in drug use behaviors in each of the drug categories of opioids, cocaine, and cannabis. Social and legal intermediate outcomes, however, were not frequently reported in the evidence. One opiate treatment trial reported significant treatment effects on legal outcomes, and one cannabis trial reported mixed results on an employment measure (although three found improvements in cannabis-related problems).

We provisionally conclude: a) there is insufficient evidence that drug misuse treatment in treatment-seeking individuals improves morbidity or mortality (key question 4); b) there is good evidence that drug misuse treatment in treatment-seeking individuals reliably reduces drug misuse (key question 5); and c) that there is insufficient evidence that drug misuse treatment in treatment-seeking individuals improves intermediate social and legal outcomes.

Key Question 7

Given the dearth of evidence from the drug misuse treatment studies on outcomes other than drug misuse behaviors, the evidence link between these intermediate outcomes and health outcomes becomes quite important. While the evidence we identified from eleven studies is mixed, there is evidence that stopping heroin addiction is associated with reduced mortality risk, and that stopping injection drug use is associated with better adherence and response to medical treatment (among individuals with HIV) and with better mental and physical health functioning. We provisionally conclude that there is fair evidence that reducing or stopping drug misuse is associated with some health outcomes, in some populations. The generalizability of these studies to general primary care populations may be limited.


Our provisional conclusions for each of the critical key questions reviewed suggest the state of the evidence regarding drug misuse screening in primary care essentially has not changed since the previous USPSTF review of drug abuse screening.1 Although many advances in drug misuse treatment have occurred during the past decade, the vast majority of studies are conducted in treatment-seeking populations, and thus the relevance of outcomes from such studies is of uncertain applicability to asymptomatic primary care populations that could be screened for drug misuse. The Bernstein trial of a brief, motivational counseling intervention to reduce opiate and cocaine use in a screened, outpatient clinic population may herald a new generation of drug misuse treatment research that will provide evidence more applicable to primary care populations.

Our finding of continuing evidence insufficiency is also consistent with the perspective described in recently initiated research by the National Quality Forum.45 The project, “Evidence-based Practices to Treat Substance Use Disorders” funded by the Robert Wood Johnson Foundation, is attempting to achieve national consensus on effective practices for treating substance use disorders. Seven practice categories were defined in an expert workshop and the project is seeking input about specific practices within each area. The workshop panel concluded that the evidence on opportunistic screening for drugs in health care settings was not strong enough or general enough to warrant inclusion as a general best practice. In contrast, opportunistic screening for alcohol use disorders in health care settings was included.

Cover of Screening in Primary Care Settings for Illicit Drug Use: Staged Systematic Review for the United States Preventive Services Task Force
Screening in Primary Care Settings for Illicit Drug Use: Staged Systematic Review for the United States Preventive Services Task Force [Internet].
Evidence Syntheses, No. 58.1.
Polen MR, Whitlock EP, Wisdom JP, et al.


PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...