Skip to main content

National Institute on Alcohol Abuse and Alcoholism (NIAAA)

Substance Abuse Prevention and Other Pregnancy Risk Studies

April 23-24, 1998 • Ramada Inn • Bethesda, Maryland


Substance Abuse in Pregnant Women
Richard S. Schottenfeld, M.D.

From the Department of Psychiatry, Yale University School of Medicine, New Haven, Connecticut

The objective of this program was to evaluate the prevalence, patterns, and correlates of tobacco, alcohol, marijuana, and cocaine use during pregnancy in an urban hospital prenatal clinic and the feasibility of intervening in this setting to facilitate treatment entry for cocaine-dependent pregnant women.

We implemented voluntary substance abuse screening, consisting of a structured interview and urine toxicology testing, in a hospital-based prenatal clinic between July 1, 1991, and June 30, 1994. Women identified as using cocaine during the current pregnancy were offered enrollment in a clinical trial comparing on-site drug abuse counseling and a more intensive, off-site drug treatment program.

Rates of participation in voluntary drug abuse screening averaged 83.8% (2,927 of the 3,491 registrants), and fewer than 1% of women refused to participate. Urine toxicology tests were obtained on 86.9% (n=2,542) of the women interviewed. Nearly half (48.8%; 1,444/2,897) of women with complete data used at least one substance during pregnancy, and most (54.3%; 784/1,444) used more than one substance. Rates of use averaged 35.5% for cigarettes (1/2 pack per day or more), 22.6% for alcohol (two or more standard drinks per occasion), 15.3% for marijuana, 16.1% for cocaine, and 2.0% for heroin. Cigarette smoking was a significant risk factor for use of alcohol (OR=4.4, 95% CI=3.6-5.3), marijuana (OR=3.7, 95% CI=3.0-4.6), and cocaine (OR=10.1, 95% CI=7.9-12.9). Alcohol use was also significantly associated with use of marijuana (OR=3.5, 95% CI=2.8-4.4) and cocaine (OR=5.1, 95% CI=4.1-6.3). While 48.2% of women who reported using alcohol and 29.5% who reported using marijuana during pregnancy had discontinued use by the time of prenatal care registration, only 6.6% of cigarette smokers and 18.1% of cocaine users abstained from use during the month preceding the interview. Brief interventions in the prenatal clinic led to treatment enrollment for 252 of 465 (54.2%) women eligible for cocaine treatment.

Implementation of routine voluntary substance abuse screening in a hospital prenatal clinic is feasible and acceptable to women and can be used to identify substantial numbers of women at risk because of their substance use during pregnancy. Cigarette smoking indicated an increased risk for alcohol use, and use of either cigarettes or alcohol indicated an increased risk for marijuana and cocaine use. Most women continued substance use after learning of the pregnancy, but the high rate of treatment enrollment for cocaine abusers supports the potential efficacy of interventions in the prenatal clinic.

Drug Use Reductions During Pregnancy Among Crack Cocaine-Dependent Women
S.L. Usdan, Med., CHES, J.E. Schumacher, Ph.D., S. Daniel, Ph.D., S. Robinson, Ph.D.
P. Sekar, S. Gilbreath, R.N., and S. Harkless, M.S.

From the University of Alabama at Birmingham, Birmingham, Alabama

Pregnancy has been identified as an important protective factor in the prevention of negative delivery and developmental outcomes as a result of drug use. The purpose of this study was to examine the relationship between pregnancy and drug use among crack cocaine-dependent women. It was hypothesized that drug use would decline during pregnancy. The Substance Use and Pregnancy Interview (SUPI) was administered to 15 drug-dependent women in a CSAT-funded residential drug treatment program for women with children. The SUPI is a retrospective interview of drug use at 6 months prior to, during (first, second, and third trimesters), and 6 months following pregnancy. Participants were 93.3% African American, with an approximate average age of 31 years. Prior to pregnancy, 80% were using crack and alcohol, and 93.3% were using tobacco.

Reductions in the percentage of women using crack (-51.3%) and alcohol (-40%) occurred from the second to the third trimester. Increases 6 months after pregnancy were still below pre-pregnancy levels. Tobacco use did not change across all assessment points. These findings suggest that pregnancy is associated with reductions in drug use from the second to the third trimester. In the next phase of the study, intervening variables such as involvement with treatment and the Department of Human Services will be analyzed and important drug prevention and treatment implications for pregnant women will be discussed.

Intervention Strategies to Decrease the Risk of Malformations Associated with Pre-gestational Diabetes
Lois Jovanovic, M.D.

From the Sansum Medical Research Institute, Santa Barbara, California

The discovery of insulin in 1922 dramatically decreased the mortality rate of infants born of diabetic women from 100% to 25%. In the 1980s, the advent of programs to create normoglycemia further decreased the rate to 8% to 10%, with malformations accounting for the residual mortality of these infants. The next objective was to determine whether the malformation rate could be reduced by normalizing maternal glucose levels prior to pregnancy. Since the diabetes-specific malformations of caudal agenesis, cardiac cushion defects, and renal agenesis occur before the eighth week of gestation, however, it is too late to begin normalization programs once pregnancy is documented.

Five university centers participated in the Diabetes in Early Pregnancy Trial, an effort sponsored by the National Institute of Child Health and Human Development (NICHD) to document the relationship between glucose levels at the time of conception and the risk of malformations and spontaneous abortions. As recruitment strategies we used pharmacies, churches, schools, and physician offices to advertise our project. Once women were recruited, we needed to interact on a daily basis to motivate them and ensure continued attention to medications, diet, and a balanced exercise program. The outcome of the study showed that improved glucose control prior to pregnancy decreases the risk for spontaneous abortions and malformation.

We also learned that a subset of women do not come for care prior to pregnancy despite a clear understanding of the decreased risk. Women forgoing preconception care tended to be in the worst control, and did not have a consistent health care provider. In addition, we learned that accidental pregnancy in unwed women was universally associated with the highest glycemic levels. Thus attention to safe methods of birth control became a paramount concern. Results from the multicenter trials showed that strategies to decrease malformations in offspring of diabetic mothers must include public awareness and access to health care systems to provide needed services.


Dr. Schottenfeld’s program had high participation rates (84% of pregnant women were screened) and reasonable rates of disclosure (3/4 of the women using cocaine admitted it; 1/4 who were using, denied it), possibly because there is no reporting requirement in the State of Connecticut for prenatal substance abuse. Although patients used multiple drugs, spontaneous quit rates varied substantially by drug type. Close to half of women reporting alcohol use and one-fourth reporting marijuana use stopped on their own during pregnancy. A much smaller percentage of cocaine and tobacco smokers reported stopping, possibly indicating higher levels of dependence on those substances.

Screening revealed high rates of use for tobacco, alcohol, marijuana, and cocaine, with 53% reporting use of more than one substance. Analysis of results further revealed that cigarettes were a risk factor for drinking and that women who drank were more likely to use marijuana and cocaine.

The New Haven program demonstrated that substance abuse screening was feasible and acceptable to patients in a health care context. Screening also proved to be an effective intervention. Fifty-five percent of women screening positive for cocaine enrolled in treatment, with close to two-thirds agreeing to enrollment at the time the screening results were discussed. Although several follow-on contacts during prenatal visits were required to convince the remaining one-third to enroll, repeated referral ultimately proved successful. It should be noted that none of the women who denied using cocaine (although urine tests indicated that they were) enrolled in treatment.

New Haven’s experience shows that interventions in primary care settings are effective in persuading pregnant substance abusers to enter treatment. Despite their success, however, Dr. Schottenfeld and his colleagues were unable to sustain the intervention because they could not identify incentives compelling enough to convince the hospital to institutionalize the program.

One interesting finding from the University of Alabama at Birmingham’s work with pregnant cocaine abusers was that study women believed that the harm associated with cocaine use during pregnancy occurred later in pregnancy (the study showed that with the exception of cigarettes, the use of alcohol and cocaine decreased during the third trimester). Since these women also tended to delay prenatal care until the third trimester and many had lost custody of their children, the links between perception of harmfulness and changed behavior are not clear. Nevertheless, the lack of understanding about the impact of alcohol and other drugs on pregnancy supports the need for continued educational efforts targeted to at-risk women.

Workgroup members were interested in probing who the Alabama and New Haven program dropouts were and whether those women who were able to abstain during pregnancy were dependent users or may have replaced cocaine with methamphetamine. Both Dr. Schottenfeld and Dr. Schumacher agreed that it is relatively easy for nondependent users to stop during pregnancy and since methamphetamine is not yet a problem in either Alabama or New Haven, a switch in drugs did not explain the decrease in use. Both also acknowledged that dependent users comprised the majority of the dropout group.

In addition to exploring the role of pregnancy in preventing/reducing substance abuse among cocaine-dependent women, Dr. Schumacher and colleagues also examined the role of community-based obstetricians in preventing/reducing substance abuse. They found that most obstetricians believed they lacked the training to intervene in this area. Other barriers to playing a more active prevention role included biases ("my mother drank, and I’m okay," "substance abuse treatment doesn’t work"), fear ("I don’t know how to respond to admissions of use so I don’t ask"), ignorance ("I know nothing about AA"), and stereotypical thinking about the class, race, and ethnicity of patients likely to have substance abuse problems.

The Alabama training program for obstetricians was office-based and used incentives (e.g., conducting sessions during lunch and providing the meal) to encourage participation, a tactic that worked with this target group. Practitioners could select among several topics (e.g., assessment, referral, risk management–a popular choice) that were offered once per month over a 6-month period. They also were given a video to reinforce learning. As a result of the training, obstetricians significantly improved their knowledge of substance abuse and their feelings of self-efficacy. Changes in practice have yet to be evaluated.

Commenting on the effectiveness of physician training, Barbara Morse said that Henry Rosett and Lynn Weiner trained providers via grand rounds and a training-of-trainers approach that was institutionalized in hospitals throughout the State of Massachusetts. Not only did screening and referral increase, but clinical behavior changed. In her opinion, educational programs do influence physicians to change their practices. However, she believes that to redress current misperceptions about treating at-risk pregnant women, it is important to focus on improving training in medical schools and during internships and residencies. Marc Usatin agreed that physicians have serious misconceptions about substance abuse treatment. For example, many are unaware that compliance with substance abuse treatment regimens is on a par with those for hypertension and diabetes. Susan Rich suggested that one way of improving knowledge and changing practice would be to post a substance abuse curriculum on the Web that would provide continuing medical education units to those clinicians who completed it. Barbara Wesley, an obstetrician at Howard University, said that the health care system compounds the problem by its failure to support or reward clinicians for their involvement in substance abuse intervention programs. The physician typically generates the billing for the entire staff, and under the current system, there is no reimbursement for these services. Dr. Coles and Dr. Fleming endorsed Dr. Wesley’s observation and said they had confronted the same dilemma when trying to implement interventions in primary care settings.

Dr. Jovanovic’s description of the problems encountered with pregnant diabetic women is strikingly similar to those posed by risk drinkers. Much like FAS, malformations in children of diabetic mothers occur during the third or fourth week of gestation, frequently before a woman knows she is pregnant. For diabetic women, preconceptional programs are essential to normalize glucose and ensure that malformations do not occur. Unfortunately, women with Type II diabetes who have had normal babies prior to developing diabetes often do not take precautions despite knowledge of the possible impact of their decision and access to good health care. They are, in Dr. Jovanovic’s words, "in denial."

Having a healthy pregnancy requires a substantial commitment from the diabetic mother-to-be. Dr. Jovanovic and colleagues have found that if the health care team puts forth an intensive effort, the pregnant woman will match it, but if the health care team loses interest, the woman will let go. Dr. Jovanovic observed that many Type I diabetic teens and young women have an underdeveloped locus of control, because their mothers have been so protective due to their disease. In her program, the health care team assumes the role of the "nagging mother" during a young diabetic woman’s pregnancy. In cases where young women and their mothers have had good relationships, this approach works effectively; when young women are in conflict with their mothers, compliance with the treatment plan is difficult to achieve. Another lesson learned relates to postpartum behavior. After delivery, Dr. Jovanovic’s patients were moved to another service in the hospital and immediately reverted to unhealthy dietary practices. Psychiatrists consulted to understand what precipitated the change in their behavior said that the new mothers felt abandoned by the health care team. To remedy that situation, the women were transferred back to endocrinology. To Dr. Jovanovic, their reaction and the success of the remedy applied underscored the paramount importance of the therapeutic relationship in facilitating compliance.

In response to workgroup questions about contraception, Dr. Jovanovic said her program routinely paid diabetic teenage patients $1 per day for proving they were not pregnant. They discovered that cash incentives rather than coupons or vouchers for service appealed to this age group. Since the incentive program was initiated, teenage pregnancy rates have declined. However, the incentives program has not worked well with older women, many of whom believe that motherhood is their reason for being or who have a fatalistic outlook on life ("what will be, will be"). Dr. Jovanovic believes that another reason the program has been so successful in achieving compliance relates to its policy of partnering with patients. The women are not only full members of the health care team, but they have a tool for self-care so that they participate in their own treatment. The fact that they can prove compliance to family members (who have been frequently distrustful of their efforts in the past) as well as to their health care providers has also contributed to their willingness to adhere to treatment requirements. In this regard, Dr. Jovanovic wondered whether negative urine and other tests could be employed as positive rewards for substance-abusing women.

Dr. Schottenfeld observed that experience with pregnant cocaine-abusing and diabetic women raised key questions about intervention strategies for increasing compliance. Future research efforts might explore, for example, the role of persistent telephone contacts, provision of meaningful incentives, and interaction with health care providers in improving adherence to program requirements.


  • Screening programs in primary care settings are acceptable to high-risk pregnant women (when there is no substance abuse reporting requirement) and effective in moving nondependent women to treatment.
  • Dependent women remain the most nonresponsive to intervention efforts.
  • Tailored incentives, sensitive to a woman’s age, outlook, and circumstances, promote treatment compliance.
  • Clinician training positively affects knowledge and feelings of self-efficacy and appears to influence practice.
  • Involving women as full partners in their own care, maximizing relationships with the health care team, and providing opportunities to prove compliance foster positive outcomes.
Looking for U.S. government information and services?