Current Studies on Prevention of Alcohol Use in Pregnancy

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

Abstracts

Activities of the Center for Addiction Research and Education
Michael F. Fleming, M.D., M.P.H.

From the Department of Family Medicine, University of Wisconsin, Madison, Wisconsin

The University of Wisconsin Center for Addiction Research and Education (CARE) has a number of projects that focus on the prevention of alcohol problems in women and children. Project TREAT conducted a randomized clinical trial to test the efficacy of brief physician advice for the treatment of problem drinking in a large sample of men and women (n=774) attending 17 primary care clinics in southern Wisconsin. The results of the 12-month followup interviews were reported in the Journal of the American Medical Association (JAMA) in the spring of 1997 and found a large treatment effect in men and women. The 48-month followup study will be completed in the fall of 1998.

CARE is also preparing to conduct a trial to test the effectiveness of a clinic-based intervention in obstetricians’ offices. The goal of this trial is to increase rates of alcohol screening, brief intervention, and referral in women of childbearing age.

An FAS Clinicians Guide has also been developed, and we are in the process of pilot testing the clinical protocols prepared for the Guide. The goal of the Guide is to help clinicians increase their ability to detect at-risk drinking in women and to identify children who may have been adversely affected by fetal alcohol exposure.

Perinatal Alcohol Abuse: Identification and Intervention
Carole Kenner, D.N.S., R.N.C., F.A.A.H.

From the College of Nursing and Health, University of Cincinnati, Cincinnati, Ohio

The specific aims of this 5-year, CDC-funded intervention demonstration study were to: (1) educate health care professionals about the adverse effect of alcohol on pregnant women and their offspring and the identification of alcohol-abusing pregnant women, and (2) test the effectiveness of the intervention protocol by comparing the experimental group of mothers and infants with the control group using perinatal, maternal, neonatal, and infant outcomes.

The study enrolled pregnant women from three public urban clinics prior to the 28th week of gestation who admitted that alcohol was their primary substance of choice. Women and their infants were followed for 15 months postpartum. The intervention exposed health professionals to an educational workshop that focused on alcohol’s adverse perinatal and neonatal effects and the use of the T-ACE (alcohol abuse screening tool) as part of the routine prenatal intake. The interventions for the experimental group of pregnant women consisted of case management and brief intervention based on the Nora Cox Brief Intervention Model (BIM).

Nurses provided the case management services for the study, which included BIM, weekly client contact, mutually agreed-upon plans of care, and coordination of services such as financial aid, counseling, childbirth education, alcohol support, and other life skills training. The BIM was philosophically compatible with the case management intervention for this project, and its prescribed method of motivational interviewing resulted in more structured contacts. ATransition Questionnaire was used to identify changes in parenting and coping skills. These two scales had a reliability of .92 and .69, respectively.

Following the educational intervention, participating health care professionals increased their rates of reporting patient alcohol use from 2.5% to over 10%. The project developed a Health Care Professional Attitude Scale consisting of three Semantic Differential Scales and one Likert-type scale and administered it at 2, 4, and 6 months postworkshop. The alphas for reliability ranged from .70 to .91. As might be expected, over time attitudes fell back to preworkshop baseline levels.

The project compiled data on initial and eighth-month drinking patterns for 40 pregnant subjects. A paired- t test assessed differences in the amount of absolute alcohol per ounce consumed per week for each variable per group. The variables were beer, wine, liquor, and total. Significant differences emerged between initial and eighth-month drinking patterns in the treatment group for beer and total alcohol with consumption decreasing during pregnancy. A repeated measures ANOVA was performed to include significant differences in drinking over time, which included the 3- to 6-week postpartal drinking measure. There was an overall significant difference for total alcohol (.019 at .05 level). When percent change was calculated, there was an overall decrease in all categories for the treatment group. In both groups, the largest change was for beer. Overall, the percent change for the total variable shows a greater decrease in drinking for the treatment group as compared to the control group. However, five of the eight measures demonstrated an increase in drinking following birth. This trend is supported in the literature for social drinkers. Once case managers built a rapport with the study subjects, we found that some were using other substances.

The project encountered a major problem when the clinic population shifted due to changes precipitated by the HMO (Medicaid clients were reassigned to a clinic where the study had no access), the prosecution of perinatal substance-abusing women (prenatal clinics were to report women with identified substance abuse problems who then had to attend treatment or lose WIC and TANF support; as a result those women stopped coming to the prenatal clinic), and turnover in case managers. Experimental subjects from the "shifted" clinic population also admitted more problems coping with their new babies than did members of the control group.

The Cincinnati study reinforced the efficacy of brief intervention as a strategy for promoting abstention/reductions in substance abuse during pregnancy among moderate/social drinkers. But the study also demonstrated the impact of provider resistance, time constraints, and sanctions on study implementation. Initially, the study targeted eight clinics, but this proved far too ambitious so the number was reduced to those three sites that had the largest number of drinkers. Because providers were uninformed about the impact of risk drinking on pregnancy and reluctant to accept any intervention that disrupted their routine, the study decided to rely on the short and easily administered T-ACE questionnaire. The advent of sanctions and the number of women with positive screens who were not amenable to attending treatment fueled provider resistance and further undermined study efforts to make identification of risk drinkers an integral part of clinic operations. Echoing lessons learned from other programs, Dr. Kenner observed that the adolescents in the Cincinnati study were confused about alcoholic beverages and did not consider beer, wine, or wine coolers as alcoholic. Likewise, a number of obstetricians encountered in the study thought that drinking was acceptable during the third trimester and clearly demonstrated that education on risk drinking was sorely needed.

Determining Educational Effects and Resources (DEER):
The Influences-During-Pregnancy Project
Lee Ann Kaskutas, Dr.P.H.

From the Alcohol Research Group, Berkeley, California

Abstention from alcohol is recommended during pregnancy to prevent fetal alcohol syndrome. Continued consumption among some subgroups suggests that information about risk either is not reaching some pregnant women, or, if it is reaching them, it is being interpreted skeptically or negatively and thus is ignored. How pregnant women respond to abstention warnings is an understudied area, yet such messages represent a key component of our nation’s FAS prevention strategy. The DEER project is addressing that informational gap by pilot testing recruitment and measurement methodologies among at-risk, ethnically diverse pregnant women in Los Angeles. We are asking culturally appropriate questions of 90 Native American and 150 African-American women (with 30 white women as a comparison group) that assess consumption during pregnancy and interpretation of warnings against drinking while pregnant. Results will inform a larger R01 (NIH-supported investigator-initiated research grant) epidemiological study concerned with how best to influence minority women’s decisions about perinatal drinking.

The primary aims for the pilot are to field-test techniques for recruiting and retaining ethnically diverse pregnant women in a geographically contained urban area so that exposure to health messages from the environment may be accurately assessed early and late in pregnancy; develop and test composite measures of consumption that may be used among ethnically diverse urban gravidas, drawing from instruments originally intended for specific use with Native Americans or African Americans; and gather information about women’s interpretation and application of health messages during pregnancy. A secondary aim is to determine the overall dose of exposure to health messages about drinking during pregnancy from the broad message environment–that includes both impersonal messages such as labels, posters, ads, and magazine articles and personal messages such as conversations with husbands and doctors–to which minority women are exposed during pregnancy and the relationship between message dose and alcohol consumption. In preparation for a future study with more subjects (to enable stable assessment of change), we are using pilot data to measure changes in alcohol consumption over the course of the pregnancy and to determine in a preliminary way the extent of association between changes in consumption and cumulative exposure to health messages.

Assuming an 80% participation rate, over a 6-month period approximately 200 women from the clinic sample (60 Native Americans, 120 African Americans, and 20 whites) and 70 from the purposive sample (30 Native Americans, 30 African Americans, and 10 whites) will be interviewed (n=270). Assuming 65% were first trimester recruits (Campbell et al., 1989), about 139 followup interviews (80% followup response rate) are anticipated (47 Native Americans, 77 African Americans, and 16 whites). Interviewers and subjects will be matched on ethnicity.

The Challenges of Protecting the Next Pregnancy
Janet R. Hankin, Ph.D.

From the Department of Sociology, Wayne State University, Detroit, Michigan

Protecting the Next Pregnancy is an indicated prevention intervention that targets women who drank at risk levels during an index pregnancy. The goal is to prevent alcohol use during the next pregnancy.

The study is a randomized clinical trial. The intervention program is based on a brief intervention model that uses cognitive behavior techniques. Women who drank at least 0.3 ounces of absolute alcohol per day at the time of conceiving the index pregnancy were invited to participate in the program after they delivered their index infant at Hutzel Hospital, an inner-city hospital located in Detroit, which is affiliated with Wayne State University and the Detroit Medical Center. Four weeks postpartum, they were randomly assigned to the Control or Experimental Group using a ratio of 1:3. The Control Group women received the standard clinic warning that they could have a healthier baby if they cut way down or stopped drinking. Women in the Experimental Group received a series of brief interventions designed to teach them about a standard drink, setting drinking goals, and methods to slow down their drinking. Women were followed for up to 5 years after delivering the index infant, with periodic booster sessions encouraging them to decrease their alcohol consumption. If they conceived a subsequent fetus, they were contacted three times during the course of pregnancy and encouraged to reduce their alcohol consumption.

Alcohol histories were obtained each time the women were contacted. Index infants were tested using the Bayley II Developmental Index at 13 months of age. Currently, we are in the process of testing subsequent infants with the same developmental assessment tools and collecting data on a variety of outcome measures. In addition to comparing index and subsequent siblings on the Bayley Index, we will compare subsequent Control and Experimental infants. Important outcome measures will include drinking during the subsequent pregnancy, birth weight of the subsequent infants, and number of prenatal visits. We will compare the Control and Experimental groups on interpregnancy intervals, use of birth control methods, and sterilization rates.

Data on 78 women who have delivered subsequent infants indicate that compared to the Control Group (n=21), experimental subsequent infants (n=57) are less likely to have birth weights less than 2,500 grams and experimental mothers drink less during the subsequent pregnancy and have more prenatal visits. Given the small numbers, however, these results need to be interpreted with caution.

This program has presented numerous challenges during the past 5 years. The first problem was tracking the women. Given their chaotic lives, we spent hours trying to contact them by mail and phone (if they had one). Women in our study also disappeared into the streets of Detroit, and we lost a number of them to followup (during the analysis phase, we will be examining the selection bias that results). When women were reluctant to return to the prenatal clinic to participate in our study, we relocated our program to the university campus. We also found that providing gifts, money, and transportation helped facilitate retention. Because many in our sample were low literacy, we learned to tailor our messages to their level of comprehension, using pictures as well as words to educate them. We also discovered that telephone interviews were more likely to be completed than in-person interviews.

Another issue affecting the study was the lower than expected pregnancy rate. Since the goal of the intervention was to separate alcohol from pregnancy, this lower rate of conception was not all bad. The smaller than expected number of infants was compounded, however, when some of the index infants were placed for adoption; receiving permission for developmental testing of the adopted infants has been difficult. Despite the difficulties and challenges of implementing this study, Protecting the Next Pregnancy is a program that seems to have a positive impact on infant birth outcomes among a particular high-risk population.

Preventing FAS Among American Indians
Philip A. May, Ph.D.

From CASAA, University of New Mexico, Albuquerque, New Mexico

My colleagues and I have just completed one FAS Prevention Research Project and are in the middle of another. From 1992 through 1998, the Centers for Disease Control and Prevention (CDC), Division of Birth Defects, funded a Fetal Alcohol Syndrome Statewide Epidemiology and Prevention research study. Our goal was to establish the baseline prevalence of FAS in New Mexico’s three major ethnic groups (non-Hispanic white, Hispanic, and American Indian) and use the resulting data to set the stage for prevention programs. From our work on this project, we have learned a number of lessons. We found it much more difficult to conduct active case ascertainment in FAS prevention research and practice in non-Indian than Indian communities because of the lack of centralized health care services, lack of coordination between agencies, and reduced commitment to alcohol abuse prevention in general, and FAS in particular, as a public health problem. We also learned that training private health practitioners for cooperative research/prevention programs is not always successful.

Despite these problems, we also had some positive results. We were able to improve our maternal risk factor screening methodology by including better questions on topics such as maternal history, siblings, and the quantity, frequency, and timing of prenatal drinking. We found that it aids data collection greatly to employ a maternal locator for the maternal interviews and to offer a $20 respondent fee to help attract birth mothers. We also learned that attracting FAS cases and maternal subjects in non-Indian communities is best done via emergency rooms, child protection agencies, social welfare agencies, and in some cases, through school-based special education programs.

Currently, NIAAA is funding phase one of a proposed three-phase FAS epidemiology research project that builds on previous CDC and Indian Health Service (IHS) efforts to obtain the data needed to define a specific, comprehensive (universal, selected, and indicated) prevention program for Indian communities. Phase one consists of baseline research in four American Indian communities in the Northern Plains of the United States (one in Montana, one in North Dakota, and two in South Dakota). Specific activities include a prevalence survey of adult drinking (e.g., norms, knowledge, attitudes, and beliefs about alcohol and drinking in general), an FAS prevalence and epidemiology study using active case ascertainment methods, and an extensive series of maternal risk factor surveys.

The currently unfunded part of the project (phases two and three) would require an additional 5 years of activity. In phase two, the four communities currently under study would become prevention sites with 2.5 prevention specialists coordinating their efforts with existing agencies to carry out those FAS prevention activities recommended in the Institute of Medicine’s Report on Fetal Alcohol Syndrome (e.g., case management, referral to inpatient alcohol treatment and birth control services, brief motivational therapy, and the use of the community reinforcement approach to outpatient treatment). A menu of other selected and universal prevention techniques also would be used in an attempt to lower the rate of FAS established in phase one of the research.

Phase three would repeat the phase one surveys of adult drinking and maternal risk factors to determine whether risk and outcome variables had been reduced. These findings would be compared with those obtained from two control communities that would have been subjected to the same surveys but would not have received prevention services.

Issues inhibiting study progress center on the fact that alcohol does not always lead to damage, not all children have birth defects that are caused by FAS, and the timing of exposure affects the degree of alcohol-related birth defects (ARBD) and alcohol-related neurological disorders (ARND). In conjunction with Sharon Wilsnack, project staff are in the process of developing a better consumption instrument. We are also working on a more sensitive neurological/psychological test for young children.

Dr. May concluded his presentation with descriptions of two compelling case studies–one of a Plains Indian patient and one of a woman in South Africa–that illustrate some clinical observations about the impact of age/parity on FAS, the role of spirituality in supporting behavior change, and the window of opportunity for preventing FAS. Both the American Indian and South African patient began drinking regularly at an early age (a factor Dr. May views as a major risk for FAS) and bore at least two children during their 20s who did not have FAS. Both then had FAS children in their 30s (in the case of the Plains Indian, her fifth child had severe FAS). Both women then experienced a religious conversion, became abstainers, and subsequently delivered normal children. For these women, heavy-drinking peer groups played an instrumental role in encouraging the women’s hazardous drinking behavior. The combination of bearing FAS children and a spiritual reawakening ultimately precipitated positive change (it should be noted, however, that the Plains Indian bore three FAS children prior to her conversion so the experience of producing an FAS child did not account solely for her behavioral transformation). In the future, Dr. May believes that more attention should be paid to the potential impact of spirituality in preventing FAS. Another striking feature of the two case studies is that both women engaged in abusive drinking for more than 15 years, but the health care community was unable to reach them with an effective intervention. In Dr. May’s opinion, we need to look at the reasons for this failure.

The Seattle Advocacy Model for High-Risk Mothers
Abusing Alcohol and Drugs During Pregnancy
Ann Streissguth, Ph.D., Therese Grant, Ph.D., Cara Ernst, M.A., Pam Phipps, B.S.

From the Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, Washington

The Seattle Advocacy Model, also known as the Seattle Birth to 3 Program, was developed to enable communities to respond, through outreach and advocacy, to the problems of mothers who abused alcohol and other drugs during pregnancy and to the needs of their children. The program focused on the highest-risk mothers: those who received little or no prenatal care, were unconnected with other community resources, and abused alcohol and/or drugs during pregnancy. Women were recruited within a month of delivery at the hospital or through community referral. The Seattle program found that screening for drugs and alcohol at delivery was an effective method for identifying high-risk mothers and, as in the New Haven project, encountered little resistance from women. Participating mothers were offered a personal advocate to work with them and their families on an individualized basis for a 3-year period.

In the initial 5-year demonstration project, five Birth to 3 advocates worked with 65 women for 3 years after the child’s birth, addressing complex social, behavioral, medical, legal, and financial issues. Paraprofessional advocates were given intensive training and individual weekly supervision, and were engaged in weekly group meetings. Advocates also worked with community agencies to reduce service barriers.

Of the 65 mothers enrolled, 49% were homeless or living in a temporary shelter at the time of the delivery. Prior to delivery of the target infant, these women had 131 children; 62% of the children did not live with them, 86% of their new babies were unplanned, and 49% had considered pregnancy termination. After their first year in the project, 43% of the mothers had been alcohol/drug free for 90 days or more; 98% of the infants were receiving well-child care, and 74% had received three sets of immunizations. The use of birth control on a regular basis had increased from 3% to 45%. Results of the 36-month exit interview indicated a statistically significant positive impact among clients compared with similarly recruited controls in terms of alcohol/drug treatment, abstinence from alcohol/drugs, family planning, child well-being, and connection to services.

A post-program followup study was carried out to determine whether the successful outcomes were maintained. The followup study occurred 19 to 40 months after program completion and consisted of a short (10- to 15-minute) version of the exit interview, administered either over the phone or in person. Data on 43 clients from the followup interviews and the exit interviews were compared. The proportion of clients abstinent from alcohol/drugs for a period of 1 and 2 years had increased from the time that the clients completed the program (from 36% to 48% and from 17% to 29%, respectively). Overall use of birth control remained fairly stable, changing only from 76% to 74%, and use of the most reliable birth control methods (Norplant, Depo Provera, and tubal ligation) remained at approximately 46%. Subsequent pregnancies and births decreased (from 54% to 33% for pregnancies and from 27% to 12% for births). Results to date suggest that the advocacy program set in motion some processes through which ongoing improvement was possible for high-risk women, even 2.4 years (on average) after the conclusion of their access to advocacy services.

After completion of the 5-year demonstration in 1996, the Seattle Birth to 3 Program was funded by a local philanthropist and a new Birth to 3 Program in Tacoma was funded by the Governor’s Emergency Fund. In 1997, both the Seattle and the Tacoma Birth to 3 Programs were funded by the Washington State Legislature for the biennium (the $1.52 million budget includes funding for administration, training, and ongoing program evaluation). Data analyses indicate that we are continuing to recruit high-risk women and similarly impact their lives and the lives of their children. We estimate that there are at least 130 mothers per year in the Seattle/King County area who meet criteria for Birth to 3 services. The cost of the Birth to 3 Program is $3,800 per mother per year in direct services.

Alcohol Use in Pregnancy:
Results From a Randomized Trial of a Brief Intervention
Grace Chang, M.D., M.P.H.

From the Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts

The purpose of this randomized trial was to assess the impact of a brief intervention on antepartum alcohol consumption. The trial randomized 250 eligible women initiating prenatal care at the obstetric practices of the Brigham and Women’s Hospital who scored positive on the T-ACE alcohol screen to receive either a comprehensive assessment of their alcohol use or the same comprehensive assessment with a brief intervention. *

After delivery, 99% (n=247) of the 250 subjects provided antepartum drinking histories. Both the assessment only (AO) and brief intervention groups (BI) had reductions in antepartum alcohol consumption. Risk of antepartum drinking after the AO or BI was increased nearly threefold if the subject had any prenatal alcohol consumption before assessment (p=.0001) or was breastfeeding (p=.003). Of the 143 subjects who were abstinent at assessment, however, those who received the BI maintained higher rates of abstinence (86% versus 72%, p=.04).

Since subjects in both the BI and AO groups reduced their antepartum alcohol consumption, it is unclear whether brief interventions contribute to reductions in drinking after a comprehensive assessment of alcohol use. The importance of screening for prenatal alcohol use is underscored by the findings that any prenatal alcohol consumption increases the risk of continued antepartum drinking, and that breastfeeding does not protect against prenatal alcohol use. A potentially important and as yet unresolved question arising from the study is whether a comprehensive assessment, in and of itself, functions as a form of intervention since both control and experimental groups exhibited declines in drinking.

* The T-ACE was embedded in a brief "Health and Habits Survey," which subjects completed while waiting for their first prenatal visit.

Perinatal Substance Abuse Prevention Program - "Early Start"
Marc S. Usatin, M.D..

From Kaiser Permanente, Northern California

In the early 1990s, several prevalence studies confirmed that approximately 10% of the 30,000 pregnant women served each year by Kaiser Permanente, Northern California used alcohol and other drugs during pregnancy. The Early Start Model was proposed to address concerns prompted by these numbers as well as by the lack of early intervention to prevent problems among drug- and alcohol-exposed newborns.

The Early Start Model consisted of four major components:

1. In-service training for obstetricians and other health professionals on early identification of pregnant women who are abusing drugs or alcohol.
2. A questionnaire (modified CAGE) and universal toxicology screening at entry to prenatal care to assist the physician in the identification of pregnant substance abusers.
3. A licensed clinical social worker with expertise in chemical dependency assigned to and present in the Ob/Gyn department to provide immediate assessment and intervention, as well as ongoing counseling and case management.
4. Routine urine toxicology screens and specialized obstetrical care throughout pregnancy.

The cornerstone of the Early Start Model is the colocation of substance abuse services with traditional Ob/Gyn care. Once physicians have identified patients with a substance abuse problem, they can escort them to the Early Start specialist for immediate intervention. Most patients are willing to accept a short initial meeting with the substance abuse specialist when it is introduced this way. Prior experience with referring these same patients to another department or program, even when it was located in the same facility, met with very little success.

The Early Start Model was originally piloted at Kaiser Permanente’s Oakland Medical Center from 1991 to 1993. During the 15-month pilot project, 81 pregnant substance abusers were identified, 92% of whom agreed to see the Early Start counselor for intervention. Approximately 69% of the study group who delivered at Kaiser (n=51) stopped using drugs and alcohol before the 32nd week of gestation. This group was compared to study women who continued to use throughout pregnancy and with a group of pregnant women with no history of drug or alcohol abuse. The study found that babies born to the Early Start women who had stopped their substance abuse before the third trimester had statistically significant better birth outcomes and shorter hospital stays than their counterparts who continued use. Shorter hospital stays also translated into substantial cost savings.

In 1993, the Permanente Medical Group Board of Directors authorized the expansion of Early Start. Today Early Start is available at 15 Ob/Gyn sites in northern California, which serve over 20,000 pregnant women each year.

Discussion

In response to Dr. May’s presentation, several workgroup members commented on spirituality. Joseph Schumacher observed that in Alabama if a respected preacher advises a congregation to stop smoking, many will; however, Barbara Wesley questioned whether the birth of the FAS child was the factor provoking the spiritual change. According to Dr. May, the woman in the South African case study acknowledged that the birth of her FAS child did spark a change, while the Native American mother admitted that the birth of her third obviously damaged FAS child made it impossible for her to deny the impact of her alcoholism any longer. Dr. May also said that research in Native American communities exploring the role of intercessory prayer in changing risk-drinking behavior did not confirm an effect. Nevertheless, in Dr. May’s view, while spirituality is probably not "the single variable" that prompts a quantum change, it may be an important contributing factor and should not be overlooked when designing intervention programs.

Several questions arose about control groups. Workgroup members asked Dr. Kaskutas why a white control group was used. She said that her colleagues strongly suggested using a white control group because more data were available on the impact of white women’s exposure to alcohol during pregnancy. Dr. Kaskutas’ study also illustrates the impact of external environmental forces on study design and execution. When the target Native American clinic lost its Indian Health Service funding (a circumstance duplicated in many IHS-supported programs at the time), study clients were dispersed to services throughout greater Los Angeles and the study had to readjust its focus to include other clinics and referral sources for subject recruitment.

With respect to Dr. Chang’s experience, Dr. Howard asked whether the control group could have been contaminated. Dr. Chang indicated that those conducting the assessments were blind to patient status; however, she reiterated her opinion that conducting an assessment in and of itself may be an intervention. Workgroup members identified case management as an important feature of current research. Dr. Streissguth said that in Birth to 3, advocates (the term used in Seattle for advocacy-based case management) provide the "hooks" that sustain participation by responding to clients’ immediate needs for diapers, cribs, and housing. Advocates use a system of cards called the "difference game" to assist program women in identifying unmet needs. Another advantage of the Birth to 3 Program design is that it gives the advocates three years to work with their clients, a period of time sufficient for fostering behavioral change.

Since Birth to 3 advocates provide home-based services, workgroup members inquired about safety issues. According to Dr. Streissguth, there have been no security problems. However, each advocate carries a cell phone, and if there is any question about possible danger, advocates accompany one another to the client’s home, although this is not the norm.

Dr. Streissguth pointed out that while her program uses the extended advocacy approach after delivery to reach women at highest risk, the program could also be used at prenatal clinics to enhance compliance with nondrinking/reduced drinking goals. For example, the social worker in Kaiser’s Early Start program could link at-risk women with a community-based advocate who functions as part of the chemical dependency program. Dr. Usatin noted that the "Born Free" program in Contra Costa County, California, uses lay advocates in much the same way. In a related observation, Migs Woodside reported that a recent study of pediatricians commissioned by the Children of Alcoholics Foundation revealed that pediatricians spend on average 8 minutes with their patients. Since they do not have time to probe for parental drinking, nor are they trained to do so, she wondered whether the Early Start program could also be adapted for use in pediatric settings. Dr. Usatin agreed that pediatric clinics represented an important screening opportunity for mothers that could be readily justified on the grounds that a mother’s risk-drinking status affects her child’s health and well-being. Dr. Usatin noted further that Early Start has resolved several problems over the course of its implementation. It learned that while physicians were happy to take patients to the social worker, they were uncomfortable raising the subject of substance abuse with patients themselves. Further, clinicians had to be trained to view substance abuse as an illness to be treated as opposed to a condition to be punished (e.g., grounds for initiating an immediate call to Child Protective Services). Dr. Usatin is still not certain whether Early Start is as effective with chemically dependent pregnant women as it is with those who abuse substances, but are not dependent.

Since case management appears to offer a promising intervention for risk-drinking women prior to and during pregnancy, workgroup members discussed strategies for funding case managers. Drs. Russell and Usatin urged programs to enlist the services of health economists in order to justify the costs involved in funding case management on an ongoing basis. Economic arguments, supported by cost data, convinced Kaiser to provide such services. Dr. Coles warned that the problem of funding needed services extended beyond the insurance companies and HMOs. With costs steadily shifting to the public sector, policymakers must also be educated about the value of such services to patients and the cost offsets that can be achieved by putting such programs in place. Dr. Fleming concurred and noted that the National Women’s Resource Center released a report on currently available research that discusses the costs of treating substance abuse. The report was intended for distribution to legislators and may provide information useful to program developers when seeking funding and ongoing support for services such as case management.

Commenting on the choice of screening instruments used in the programs just discussed, Marcia Russell drew attention to the availability of the TWEAK. This instrument is more sensitive than the CAGE in detecting risk drinking in pregnancy, is easy to use, and is nonthreatening.

Since the presentations on "Current Studies on Prevention of Alcohol Use in Pregnancy" covered such a wide range of topics, Migs Woodside, a workgroup member and founder of the Children of Alcoholics Foundation, was asked to summarize the discussion and relate it to the key issues raised earlier in the meeting. The following integrates Ms. Woodside’s synthesis with Dr. Jan Howard’s concluding remarks.

Summary and Conclusions

• Advocacy-based case management offers a promising strategy for intervening with, engaging, and retaining those women at "highest risk" for bearing FAS, ARBD, ARND, and substance abuse-affected children, both prior to and during pregnancy.
• Specially trained social workers assigned to hospital-based Ob/Gyn units are effective in identifying at-risk women and moving them to treatment. This strategy appears to have broader application for other primary care settings, including pediatric clinics.
• Spirituality may be an important factor in promoting positive behavioral change among at-risk women.
• Funders are likely to be more responsive to requests for support when programs demonstrate cost offsets from such services as screening, referral, and case management.
• Program image (e.g., readily remembered name, easily identified logo, use of attractive and relevant recruitment materials), facility (e.g., convenience, location), and staff (e.g., knowledgeable, empathetic, capable of making women feel connected to the program) have a major impact on recruiting and retaining risk-drinking women in programs.
• Externally imposed sanctions (e.g., requirements to report substance abuse) negatively affect at-risk pregnant women’s participation in interventions.
• Negative associations with sobriety (e.g., becoming the permanent designated driver or babysitter, being excluded from social functions because of nondrinking status) undermine women’s motivation to abstain from drinking during pregnancy.
• Ongoing education about alcoholic beverages and the impact of alcohol throughout pregnancy is essential for both preconceptional and pregnant women and health care providers (e.g., some women do not consider beer, wine, and wine coolers as alcoholic; some practitioners believe that drinking is acceptable during the third trimester).
• Incentives for patients and providers increase participation in studies and should be prominently displayed/discussed during the recruitment phase.
• Interventions must compensate for low levels of literacy and comprehension when preparing written materials and developing questionnaires (even when questionnaires are administered by staff). Examples of varying containers and products help promote complete responses to questions about type and frequency of alcoholic beverages imbibed.
• FAS programs would benefit from improved consumption instruments and more sensitive neurological/psychological tests for young children.
• Assessment in and of itself may constitute an intervention.

Sharing experiences in implementing research interventions aimed at preventing risk drinking and drug use during pregnancy surfaced an array of common concerns. From the workgroups’ discussions, it appears that women at highest risk are the most difficult to reach. Given the paucity of proven strategies for affecting chemically dependent women in denial about their problem, it may be more feasible to target efforts one tier down and focus interventions on those whose involvement in social or religious organizations provides some leverage for implementing change.

Research to date indicates that interventions, including screening, brief intervention, referral, and advocacy-based case management, are effective with moderate and social drinkers, nondependent cocaine abusers, and non-addicted smokers. As might be expected, the more intensive the intervention, the better the outcome. Interventions that are women-centered, culturally sensitive, and delivered by empathic and knowledgeable care givers are also more effective. Experience in several programs further demonstrates the important role of advocacy-based case management in encouraging program participation. When immediate needs are met and women feel a connection to a program and its staff, they are more likely to modify their behavior.

Although pregnancy presents an ideal opportunity to intervene with women at risk, more insight is needed into the sociodemographic factors that encourage/discourage drinking during pregnancy; other vulnerabilities affecting the capacity to abstain or reduce consumption; the role of fathers and extended families in promoting needed behavioral changes; and the potential efficacy of mandated substance abuse treatment, contraception, and related services. To prevent FAS, ARBD, and ARND, researchers and program developers also need a better understanding of why some women continue to drink during pregnancy despite adverse consequences. The impact of confusing messages about the health benefits of alcohol (e.g., advice to drink one glass of red wine per day for cardiovascular health) also requires investigation. In addition, more information is needed about:

• Access to reimbursement for health care providers;
• Impact of deception and the need for a "dipstick" technology for alcohol;
• Relapse prevention after pregnancy;
• Strategies for raising treatment adherence (e.g., who should be on the health care team? What incentives work best with which target groups?);
• Naturally occurring environmental opportunities for research; and
• Prevalence of FAS, ARBD, and ARND so programs can clearly measure progress and demonstrate positive outcomes.

Finally, the research and service delivery community must clarify what it is trying to prevent: heavy drinking, any drinking, pregnancy among some women who drink, or a combination?