Alcohol use among pregnant African American women: Ecological considerations

Alcohol use during pregnancy causes growth, morphological, and neurobehavioral abnormalities in the offspring (Boone, 1985; Day et al., 1992; Geva, Goldschmidt, Stoffer, & Day, 1993; Larkby & Day, 1997; Oyemade et al., 1994). On confirmation of pregnancy, most women stop or reduce their substance use, for their health and that of the offspring (Cornelius, Goldschmidt, Taylor, & Day, 1999). Unfortunately, not all women reduce their alcohol use, and the continued use of alcohol throughout pregnancy has cumulative effects on the fetus across the trimesters (Day & Richardson, 2000). Although the number of women using alcohol into the third trimester is small, the characteristics, attitudes, and environments of those women warrant consideration.

FACTORS THAT INFLUENCE ALCOHOL USE

The socioecological framework views individuals in the context of interactions with their environments. Factors such as sociodemographic status, race and ethnicity, interpersonal relationships, life events, religiosity, and drug use influence alcohol use. The mental health of the woman, the extent to which the pregnancy is desired, the woman's social support, and the presence of Stressors also, by this model, predict alcohol and other substance use. It is important to understand these factors to help women control their alcohol use during pregnancy. Ethnicity and race have been identified as significant correlates of prenatal alcohol use. Peindl (1992) found that African American women were more likely to continue to drink throughout pregnancy than were white women. Cornelius and colleagues (1999) compared drinking patterns among 413 pregnant teenagers and found that the African American teenagers were more likely to continue to drink after the first trimester compared with white teenagers. Given ethnic and racial differences in the rates and patterns of prenatal and postnatal substance use, it is likely that risk factors that affect substance use during pregnancy may differ by ethnicity and race as well. It is important to identify these factors for different racial groups, because they may affect the design, provision, and acceptability of services to these women.

Alcohol, tobacco, and drug use are not mutually exclusive. Women often use more than one drug concurrently. In a study of 83 African American women in residential treatment facilities, 77 percent reported multiple drug use when they were pregnant, 17.3 percent used alcohol before pregnancy, and 9.7 percent continued use after pregnancy (Bass & Jackson, 1997). Day and Richardson (1991) reported that women who continued drinking into the third trimester were more likely to use tobacco, marijuana, and other illicit drugs. Alcohol is the most frequently used substance, however. In a study by Wiemann, Berenson, and San Miguel (1994), pregnant African American women used alcohol at higher rates than tobacco or illicit drugs.

Other factors also affect substance use during pregnancy. Peindl (1992) reported that women who used alcohol throughout pregnancy were older, with more disruptive life events such as divorce, job loss, and moving. In the clinical sample of Bass and Jackson (1997), the women who used substances during pregnancy realized the danger their behavior posed for their offspring, but indicated that feelings of depression during the pregnancy led to more drug use. Researchers have reported depression, hostility, anxiety, and a lack of self-esteem among pregnant women who used alcohol and other substances prenatally (Clayson, Berkowitz, & Brindis, 1995; Cosden, Peerson, & Elliott, 1997; Day et al., 1991; Hutchins, 1997; Marcenko & Spence, 1995; Oyemade et al., 1994).

Research studies have found that women view relationship problems and stress as situational or environmental causes of drug use (Boyd, Hill, Holmes, & Purnell, 1998; Kauffman, Silver, & Poulin, 1997; Young, Boyd, & Hubbell, 2000). Pregnant substance users need support from their family and friends (Brindis, Berkowitz, & Clayson, 1997; Mumm, Olsen, & Allen, 1998). All too often, these individuals are themselves substance users (Marcenko & Spence, 1995). Drug use by family and friends affects the pregnant user's ability to decrease or abstain; it may reduce the social support available to the pregnant woman and result in economic stress or concerns for physical safety.

In addition to the support from family and friends, religious beliefs and religious activities can have significant nurturing and support roles. Historically, these have been important components of African Americans' lives. Bass and Jackson (1997) found that although 83 percent of drugabusing pregnant African American women acknowledged the importance of religion in their lives, 28 percent reported that they no longer attend church even though they had attended church weekly as adolescents.

The social work person-in-environment model seeks to identify interactions in the individual, family, and environment and the factors that affect the ability of these systems to adapt to changing conditions (Brueggemann, 1996). Thus, identifying and better understanding the factors that affect drinking and patterns of drinking during pregnancy among African American women is necessary to design programs for intervention. This article is an analysis of data from an ongoing study of alcohol use during pregnancy. The relations between prenatal drinking and risk factors from multiple domains, including emotional, social, demographic, and religious characteristics, were explored, focusing on identifying characteristics that correlated with drinking during the first and third trimesters pregnancy and predicted drinking throughout pregnancy. We hypothesized that variables in each of the domains would be associated with alcohol use in the first and third trimester and would predict the use of alcohol throughout pregnancy.

METHOD

Sample Selection and Interview Schedule

The Maternal Health Practices and Child Development project is a prospective study of the effects of alcohol and marijuana use on childhood outcomes. Initial recruitment occurred between 1982 and 1985. Women attending a hospital-based prenatal clinic who were at least 18 years of age were recruited into the study. A cohort of 1,360 successive clinic attendees was interviewed during the fourth or fifth prenatal month. Two samples were then selected. The first cohort included all women who drank three or more drinks per week and the next woman interviewed who drank less than that amount. The second cohort included all women who smoked two or more joints of marijuana per month and the next woman interviewed who smoked less. The two cohorts were combined for this analysis. Women in the study were reinterviewed at their seventh prenatal month and again at delivery. Substance use was assessed at each contact period. The Institutional Review Boards of the Magee Women's Hospital and the University of Pittsburgh approved the study. There were 763 live single births in the final study cohort; 393 were to African American women, and 370 were to white women. Our analyses used only the African American women.

At the first trimester interview, 17 percent (n = 67) of the African American women were married. They had a mean age of 23 (range 18 to 42) and a mean of 12 years of education (range nine to 18 years). Eighty-seven percent of the women had a monthly income of $599 or less; 5 percent had incomes over $1,000. Thirty percent were primiparous.

Measures

Alcohol Use. Alcohol use was assessed at each trimester of pregnancy using established assessment measures (Day & Robles, 1989). Women were asked about their minimum, maximum, and usual quantity and frequency of beer, wine, liquor, and wine coolers. For each trimester, a summary measure of the total amount of alcohol used was calculated and expressed as the average daily volume (ADV) of alcohol. For this report alcohol use during the first trimester and the third trimester of pregnancy were used in the analyses.

Other Substance Use. Marijuana, tobacco, hashish, and sinsemilla were assessed for each trimester, using the same measures as for alcohol. A summary scale, average number of joints per day (ADJ), was based on the combined use of these substances weighted by the amount of tetrahydrocannabinol in each. Tobacco was expressed as the average number of cigarettes smoked per day. Illicit drugs, including cocaine, amphetamines, and barbiturates, were combined and dichotomized as use/no use.

Maternal Psychological Characteristics. Maternal depression was assessed using the Center for Epidemiological Studies Depression Scale (CESD) (Radloff, 1977); hostility and anxiety were measured using the Spielberger State-Trait Anxiety Inventory (Spielberger, Gorsuch, & Lushene, 1970). Self-esteem was assessed with the Rosenberg Self-Esteem Scale (Rosenberg, 1965).

Religious Membership and Attendance. Two variables were used to represent church membership, involvement, and attendance. A dichotomous variable indicated whether the participant was a member of a church. The second variable assessed how often the participant went to church, ranging from 0 = never to 5 = more than once a week.

Social Support. The social support measure was adapted from the Human Population Laboratory studies (Berkman & Syme, 1979). Three questions were selected for this analysis to represent social support: "How many contacts (relatives and friends) (1) are you close to, (2) do you talk to, or (3) do you borrow money from?" These three questions were factored together to represent the number of social contacts.

Feelings about Pregnancy. Feelings about the pregnancy were assessed by asking, "How do you feel about your pregnancy now?" Two variables were rated to measure this area (happy versus all others and sad versus all others).

Effects of the Pregnancy. A dichotomous variable was defined to indicate whether the pregnancy did or did not change any plans. Some of the plans identified as being interrupted by the pregnancy included work, going to school, sports, vacation, getting married, wearing certain clothes, partying, active duty in military, buying a house, remodeling home, riding a motorcycle, drinking alcohol, and sex.

Household Structure. Household structure was assessed using four dichotomous variables: (1) lives alone, (2) lives with husband or boyfriend; (3) lives with parents; or (4) lives as a single adult with children under 18 years.

Care for Baby. The question used to assess the comfort level with the baby was, "Do you feel comfortable caring for your baby?" The response choices ranged from 1 = very to 4 = not at all.

Statistical Analysis

Initial descriptive analyses examined the relations between the independent and dependent variables. The dependent variables were alcohol use in the first trimester, alcohol use in the third trimester, and alcohol use throughout the pregnancy. Alcohol use was dichotomized as abstain or light (ADV < or = .40) and moderate to heavy use (ADV > .40). An ADV of .40 is approximately three drinks per week. The independent variables included age, income, education, gravidity, depression, self-esteem, hostility, anxiety, marijuana, cigarettes, drugs, member of a church, involvement in religious activities (attends church), attitude toward pregnancy, and contacts with family and friends. To examine differences between covariates and alcohol use groups, we used Student's t test for normally distributed variables, the nonparametric Mann-Whitney ranksum test for skewed variables, and chi-square tests for dichotomous variables.

To compare drinking patterns across pregnancy, three groups were created: (1) none to light throughout pregnancy, (2) reduced alcohol use between the first and third trimesters, and (3) continued moderate to heavy throughout pregnancy. Continued alcohol use throughout pregnancy was defined as moderate to heavy use in both the first and third trimesters (ADV > .40). Reduction in alcohol use included women who drank moderately to heavily during the first trimester and none to lightly in the third trimester. Analysis of variance was used to identify the characteristics that differentiated the three patterns of use. The few nondrinkers during the first trimester who became moderate to heavy drinkers during the third trimester (n = 8) were excluded from this analysis.

After testing each independent variable separately, a stepwise polychotomous logistic regression was used to find the most parsimonious set of variables that best distinguished between none or light, reduced, and continued use among the women. In this statistical model, at each step the variable that improves the likelihood the most is entered into the model, and the contribution of the variable to the goodness of fit is evaluated. This methodology allows us to identify a core set of variables that independently differentiate between the groups.

RESULTS

Characteristics of Drinkers in the Fust Trimester

The 118 women who were moderate to heavy drinkers in the first trimester were less educated, had lower self-esteem, had more hostility, and were more depressed than the none or light drinkers (Table 1). Moderate to heavy drinkers were more likely to use more marijuana, tobacco, and other illicit drugs than were the none or light drinkers. Marijuana was used by 58 percent of the moderate to heavy drinkers, compared with 44 percent of none to light drinkers. The number of cigarettes reported by moderate to heavy drinkers was more than double that for none or light drinkers, 8.6 and 3.8 per day, respectively. Moderate to heavy drinkers were also less happy about their pregnancy. Age (older), gravidity (greater), and anxiety (more) differed between the groups, with marginal significance (p < .1). Marital status, income, work status, church attendance, number of social contacts, and household structure were not significantly associated with drinking at the first trimester.

Characteristics of Drinkers in the Third Trimester

There were 38 moderate to heavy drinkers in the third trimester. These heavier drinkers were less educated, smoked more cigarettes, and attended religious services less often compared with none to light drinkers in the third trimester. Fifty-three percent of the women who drank at the moderate to heavy level also used marijuana, compared with 19 percent of women who were in the none or light category. Moderate to heavy drinkers' ADJ use was .8 joints compared with .17 in the none to light drinkers (Table 2). Moderate to heavy drinkers were marginally more hostile, more anxious, and less likely to belong to a church. No relations were observed between alcohol use at third trimester and age, income, work, marital status, gravidity, depression, self-esteem, drugs, happiness or sadness about pregnancy, number of social contacts, pregnancy plans, care for baby, or household structure.

Change in Patterns of Drinking during Pregnancy

Nearly all (97 percent) of the 267 women who were none or light drinkers in the first trimester continued to abstain or remained light drinkers in the third trimester (Table 3). Among the women who were moderate to heavy drinkers in the first trimester, 88 (74 percent) reduced their alcohol consumption by the third trimester, whereas 30 (25 percent) continued to drink at a moderate to heavy level in the third trimester. Four percent (n = 8) of the women who were none to light drinkers became moderate to heavy drinkers at third trimester.

First Trimester Factors Associated with Change in Drinking Patterns during Pregnancy. Women who were abstainers or light drinkers at both the first and third trimesters had, on average, 12 years of education and one pregnancy, compared with those who continued to be moderate to heavy drinkers who had 11 years of education and two pregnancies (Table 4).

The level of marijuana, tobacco, and other illicit drug use also differed significantly among the three drinking groups. The average number of marijuana joints per day for none to light drinkers was .47, for reducers was .75, and for continued moderate to heavy drinkers was .81. Forty-three percent of none to light drinkers and 56 percent of reducers used marijuana, compared with 63 percent of the women who continued to be moderate to heavy drinkers. The average number of cigarettes smoked per day among none to light, reduced, and continued to drink were 4, 7, and 14, respectively. Continued moderate to heavy drinkers also used more illicit drugs (13 percent) than none to light drinkers (4 percent) or reducers (11 percent) at the first trimester. Self-esteem, hostility, and happiness about pregnancy were marginally significantly different (p <. 1). The women who continued to be moderate to heavy drinkers had less self-esteem, had more hostility, and were less happy about their pregnancy.

Third Trimester Factors Associated with Continued Alcohol Use. The variables that were significant predictors of continued drinking at first trimester were also significant covariates at third trimester (Table 5). In addition, depression and hostility were significantly higher in the third trimester among the women who continued to be moderate to heavy drinkers. Fewer of the continued moderate to heavy drinkers were members of a church (53 percent), compared with 90 percent for those in the none to light category, and the moderate to heavy drinkers were significantly less likely to attend religious activities. Third trimester marijuana and tobacco use were also significantly associated with the drinking pattern across pregnancy. Among the continued moderate to heavy users, the average number of cigarettes per day was 13.7, compared with 6.9 and 4.1 among those who reduced their use and the continued none to light drinkers, respectively.

Women who were continuous moderate to heavy drinkers were more anxious and less happy about their pregnancy at the third trimester, although the significance level was marginal. Third trimester measures of age, income, work, marital status, self-esteem, anxiety, attitude toward pregnancy, social contacts, and household structure were not significantly associated with the drinking pattern across pregnancy.

The covariates that were significant in the bivariate analyses mentioned earlier were entered into a polychotomous stepwise logistic regression to find the most parsimonious model. Significant covariates in the bivariate analyses included education, gravidity, social contacts, church membership, church attendance, depression, hostility, marijuana, illicit drugs, and cigarettes. When all significant covariates were considered together, education and measures of church attendance, depression, and cigarette use remained significant predictors of the drinking pattern across pregnancy (Table 6). Thus, women with more education were less likely to drink across pregnancy, as were those who attended church. The relative risks of drinking across pregnancy were .75 (confidence interval [CI]: .60 to .92) and .74 (CI: .58 to .94) for these two variables, respectively. Depression and cigarette smoking, on the other hand, predicted the continuation of drinking across pregnancy. The relative risks were 1.04 (CI: 1.02 to 1.07) and 1.06 (CI: 1.03 to 1.08), respectively. Gravidity, social contacts, hostility, marijuana, and illicit drugs were not significant after the effects of those variables were considered.

DISCUSSION

These results are from a longitudinal study of the effects of prenatal substance use. This report assessed correlates and predictors of alcohol use among African American women during pregnancy. Environmental, personal, and interpersonal characteristics of the women predicted alcohol use in the first and third trimesters as well as the pattern of use across pregnancy. The delineation of these factors is important in helping us identify women at risk and areas where interventions may be successful.

In the first trimester, women who reported moderate to heavy use of alcohol were significantly different from women who had a pattern of none to light use. They had less education; were more likely to be depressed, to have low self-esteem, to have higher rates of hostility; and were less happy about their pregnancy. They also reported greater use of marijuana, tobacco, and other illicit drugs. In the third trimester, women who used alcohol at a moderate to heavy level were less educated, more hostile, more likely to use marijuana and tobacco, and they attended church less often, compared with women who used less alcohol.

Patterns of drinking across pregnancy defined women who reduced their drinking, maintained continuous moderate to heavy drinking throughout pregnancy, and all other women. In the first trimester, the factors that significantly predicted continued drinking moderate to heavy amounts throughout pregnancy included less education, higher gravidity, and more use of tobacco, marijuana, and other illicit drugs. Third trimester characteristics of these women included less education; higher gravidity; more depression, hostility, and marijuana and tobacco use; and less frequent church attendance. Four factors remained significant predictors: education, depression, tobacco use, and church attendance. Each of these factors has been shown to be a correlate of alcohol use during pregnancy (Bass & Jackson, 1997; Clayson et al., 1995; Cosden et al., 1997; Day et al., 1991; Hutchins, 1997; Marcenko & Spence, 1995; Oyemade et al., 1994). This analysis demonstrates that these factors also predict patterns of alcohol use across pregnancy.

These findings must be considered in the context of this study. The participants were selected from an urban prenatal clinic that serves a lower socioeconomic population. Consequently, the findings cannot be generalized to the larger population. Women were not interviewed if they did not receive prenatal care by the fifth month. Although this ensured that the study population received prenatal care, it excluded a small group of women that represented the most problematic population in terms of prenatal substance use. These data were based on the participants' report of use during pregnancy. Laboratory tests of alcohol and other substance use cannot measure exposure across the trimester. Careful attention was paid to the questions that elicited the alcohol information and to the training of the interviewers to ensure that the data were of the highest quality.

These data allow identification of characteristics at the first and third trimesters of women who are at increased risk of using alcohol. In both trimesters, women who had less education; higher levels of hostility, depression, and anxiety; higher rates of tobacco, marijuana, and illicit drug use; who were less happy about their pregnancy, and who were less likely to attend or be a member of a church were more likely to be moderate to heavy drinkers. These characteristics identify women who need intervention. It is critical to provide counseling about their alcohol use during pregnancy. It is also important to note that these predictors affect the fetus. The combination of alcohol and other factors associated with alcohol use during pregnancy create a high-risk situation for both the child and the mother.

It is important to note which factors do not correlate with alcohol use during pregnancy. In this cohort there was little variation in income; therefore this variable was unlikely to enter into the analyses. However, other factors that have been hypothesized to be important, including marital status, work status, social support, and household constellation, did not significantly differentiate lighter from heavier drinkers in this African American population. By contrast, although social support, marital status, and household constellation did not correlate with drinking, church membership and attendance did. Other researchers have reported these findings (Bass & Jackson, 1997). As noted earlier, religious beliefs and religious activities have significant nurturing and support roles and historically have been important components of the lives of African Americans. The effect of religiosity on alcohol use during pregnancy is not understood, but is worthy of further consideration.

The data allow us to identify first trimester characteristics of women who continued to drink at a moderate to heavy rate throughout pregnancy. This pattern of drinking places the fetus at considerable risk of life-long disabilities. It would be important to explore, through an intervention trial, whether treatment of depression, social engagement of these women, perhaps focused on church activities and connections, as well as counseling about decreasing alcohol and tobacco use during pregnancy, would lead to a reduction in alcohol use.

SOCIAL WELFARE IMPLICATIONS

The results of this study have programmatic and social welfare policy implications for developing effective educational programs for alcohol-using pregnant women. Educational approaches that emphasize the health and developmental consequences of alcohol use must also explore psychological and socioenvironmental influences that increase a woman's vulnerability for continued alcohol use during pregnancy. The characteristics of women in this study who continued to drink across their pregnancy indicate that a constellation of factors influenced their drinking behavior. Understanding these influences is key to the development of effective intervention strategies.

Immediate attention should be given women who use drugs during pregnancy. Referral to appropriate drug treatment programs and close monitoring by health care and social service professionals need to become routine on disclosure of any drug use. Given the environmental and personal characteristics of these women at risk, postreferral follow-up offers additional needed support to seek treatment. Active participation in a treatment program should influence drug use after the first trimester.

Early interventions (that is, substance abuse education, family counseling, and support groups) can alter drug-using behaviors that can have serious consequences for the health of both mother and offspring (Mumm et al., 1998). The provision of social support services that would address depression and hostility, explore the reasons that these women are unhappy about their pregnancies, and increase family and community (including church) connections, would help motivate these women to reduce their drinking during pregnancy and to accept treatment for their alcohol use. These interventions would also increase the ability of the woman to parent and could decrease problems in subsequent pregnancies.

The role of church membership and attendance also deserves consideration. In this study, women who drank more heavily were less likely to attend or be a member of a church. Although the relation between alcohol use and church membership is not understood, this may suggest to churches the effectiveness of nontraditional outreach approaches as they respond with social and health services for their membership and in the community.

Adequate government and private funding of programs to address prenatal drug use are necessary. The developmental needs of children who have been prenatally exposed to alcohol and illicit substances place continuous demands on existing health and mental health systems. The parallel needs of the mothers of these children for substance abuse counseling, mental and physical health care, as well as parenting education, also create a large, and often unanswered, need for care. Proactive actions that target pregnant women and emphasize healthy lifestyles could offset costly services for children prenatally exposed to alcohol and drugs. This would reduce medical contacts and minimize deficiencies in physical and social development. Action now is critical as we experience massive changes in health care delivery systems, particularly decreases in mental health and substance abuse treatment services.