Fetal Alcohol Syndrome Warnings: Policing Women’s Behavior Distorts Science

Background

Historically, and cross-culturally, people’s relationship with alcohol differs. In much of European history, unsafe water led people, including children and pregnant women, to consume beer and other alcoholic beverages. ² In the modern United States, after the blip of prohibition ended, alcohol consumption not only increased but also became less of a predominately male-oriented activity. By 1947, 56 percent of American women reported drinking at least occasionally, and by 1953, depending on the type of college they attended, 40 to 89 percent of female students drank (Rotskoff 2002). Indeed, the better educated, the more likely people were (and still are) to drink alcohol and the more often (Armstrong 2003; Jones 2015). Drinking together in the home or out at social events became “a major part of sociability and leisure among middle- and upper-middle-class married couples . . .” (Rotskoff 2002:201). Given the time period and social class, these 1940s and 1950s cocktail-drinking women were likely to have been the mothers of at least some of the doctors who first wrote about FAS in the 1970s and 1980s.

In her 2003 book, sociologist Elizabeth M. Armstrong details how pre-FAS, women were told that alcohol was relaxing and could assist with sleep during pregnancy and was therefore encouraged as long as it did not substantially increase caloric intake putting the mother over her pregnancy weight-gain goal or replace too many food calories; alcohol was even used intravenously by physicians to halt early labor (Armstrong 2003; Golden 2005). One of Armstrong’s obstetrician respondents who graduated from medical school in 1963 noted that his wife informed him that for decades the women of their wealthy town dealt with life’s problems by making sure their martinis were stiff, echoing Rotskoff’s (2002) observation that “[i]n many social circles, if a suburban housewife took a daily drink to spice up her routine (or to overlook her husband’s flaws) such behavior would have seemed unremarkable” (p. 207).

By the end of the twentieth century, however, the views on alcohol consumption had changed radically due to concerns about a new disorder, FAS. In her 2005 book, historian Janet Golden points out that the groundbreaking publication on FAS by physicians Kenneth Lyons Jones, David Smith, Christy Ulleland, and psychologist Ann Streissguth was rushed to publication without undergoing the traditional peer-review process. While the Jones et al. (1973) article and other 1970s FAS articles drew conclusions based on cases of women who drank so much alcohol that they were chronically ill and often were lost to follow-up or dead within a few years of their children’s births (Golden 2005), the warnings issued to pregnant women in the 1980s were not targeted at alcoholic women but rather were aimed at all women and made no distinctions between low or moderate alcohol consumption and heavy drinking. Today, dire warnings to not drink any alcohol while pregnant and recommendations to seek medical advice if any amount of alcohol was consumed before women realized they were pregnant persist despite a host of troubling findings that problematize the direct, unmediated relationship between alcohol and FAS. A 1987 study of lower income and wealthier alcoholic women pointed to mediating effects such as nutrition. Despite controlling for level of alcohol intake, Bingol et al. (1987) found that among children born to poor alcoholic mothers, the rate of FAS diagnosis was 71 percent, but was only 4.5 percent for the children of wealthier alcoholic women. Building on the Bingol et al. study, work by Abel and Hannigan (1995) and others questions whether the pattern of alcohol consumption, in particular binge drinking, may be more likely to have a teratogenic effect on offspring and also points to the cofactors of cigarette smoking, stress, and poverty, as well as the toll on the body from years of alcoholism. Yet despite the gains that would be made if FAS could potentially be reduced by ensuring good nutrition for pregnant women who drink excessively or helping alcoholic women change their patterns of consumption to stabilize their blood alcohol level, surprisingly little work has been done in the past 40 years to elucidate these and other variables that correlate with social class.

Due to a different historical relationship and social norms around alcohol consumption, even after the advent of FAS, Europeans tend to be less Manichean in their advice to pregnant women, and scientists in Europe “take seriously the research question of whether there is in fact ‘a level of consumption which could be considered safe for the fetus which would allow the mother to take part in social gatherings, including meals, without absolute abstention’” (Armstrong 2003:206, citing Florey and Taylor 1992:S6). These studies initially led Europeans and others to be less categorical on the issue of alcohol during pregnancy than Americans who tend to take the puritanical approach of total abstention. More recently, however, there has been a worldwide move to caution women not to drink at all despite a lack of evidence warranting complete abstention. In the United Kingdom, for example, health officials reversed themselves repeatedly over a short period of time (e.g., “Women told glass of wine a day is fine during pregnancy—and too dangerous” Rose 2007 The Times UK). One exception is Singapore, where women are advised:

If you have one or two drinks of alcohol (one or two units), once or twice a week, it is unlikely to harm your unborn baby. However, the amount of alcohol that is safe in pregnancy is not definitely known. Heavy or frequent drinking can seriously harm the baby’s development (Singapore Ministry of Health 2015)

Kelly et al. (2009) actually found fewer behavioral problems in the children of light drinkers during pregnancy than in those of abstainers, and McCormack et al. (2018) found higher cognitive scores in 12-month olds of mothers who drank seven or fewer drinks per week during pregnancy than those who abstained, although, as they rightly point out, this could be more about the higher socioeconomic level of the parents, which is correlated with drinking, than about the drinking itself. This type of finding is neither new nor unique, as Florey and Taylor (1992) reported similar results on multiple measures of development. Yet, within the past couple years, FAS has elicited substantial media coverage in the United States reflected in a newly and more strongly issued CDC warning that all sexually active women not using birth control refrain entirely from drinking as well as a NYT article (Belluck 2018) based on the JAMA article (May et al. 2018) claiming that the number of U.S. children detrimentally affected by the alcohol use of their pregnant mothers may be grossly underrepresented. The CDC warning not only echoes the Surgeon General’s original warning to “women who are pregnant (or considering pregnancy) not to drink alcoholic beverages and to be aware of the alcoholic content of food and drugs” (emphasis added), but goes beyond it to women who are not attempting to conceive but simply do not use contraceptives. In a press release on February 2, 2016, Dr. Coleen Boyle, the director of the CDC’s National Center on Birth Defects and Developmental Disabilities, encouraged health care providers to “. . . advise her not to drink at all if she is pregnant, trying to get pregnant or sexually active and not using birth control” (emphasis added; see https://www.cdc.gov/media/releases/2016/p0202-alcohol-exposed-pregnancy.html). The NYT article (Belluck 2018) quoted one of the JAMA study’s authors, Dr. Howard Taras, as saying “People say, ‘Don’t be ridiculous, I went to a wine tasting and my kid came out fine’ . . . But the C.D.C. is saying ‘We don’t know. Maybe you just won the lottery’” (p. A16). This is a revelatory example of experts disparaging women’s lived experiences that match the scientific data about low levels of alcohol consumption not causing birth defects and instead favoring unsubstantiated narratives of heightened risk. In the next sections, we analyze how and why scientists and doctors continue to sound these alarm bells with ever-widening scope despite the fact that no scientific evidence has ever conclusively shown that one drink during pregnancy, or even one drink per week, causes harm to a fetus.

The Studies

While there are numerous articles warning of the dangers of alcohol during pregnancy (Burden, Jacobson, and Jacobson 2005; Charness et al. 2016; Connor and Streissguth 1996; Day and Richardson 1991; DeJong, Olyaei, and Lo 2019; D’Onofrio et al. 2007; Flak et al. 2014; Kesmodel et al. 2019; Lupton, Burd, and Harwood 2004; Zuccolo et al. 2013), they share similar problems. As recent exemplars, here we focus our attention specifically on the scientific studies that lead to the 2016 CDC warning and the JAMA article that became the basis of an NYT article. The 2016 CDC warning is based on a report “Vital Signs: Alcohol-exposed Pregnancies—United States, 2011-2013” by Green et al. (2016). This article states that “[a]lcohol is a known teratogen that can cause adverse reproductive outcomes for women, and serious lifelong problems for a person exposed to it prenatally. These risks occur throughout pregnancy, including the period before a woman knows she is pregnant” (pp. 92–93). This statement, in turn, is based on an article from the American Journal of Preventative Medicine titled, “Alcohol Use Prior to Pregnancy Recognition,” by Floyd, Decouflé, and Hungerford (1999). While the Green et al. report makes it sound as if any amount of alcohol consumption by a pregnant woman can trigger problems, Floyd et al.’s (1999) article begins with a background paragraph: “Frequent alcohol use during the first 8 weeks of pregnancy can result in spontaneous abortion and dysmorphologic changes in the developing organs of the embryo, including the heart, kidneys, and brain” (p. 101, emphasis added). The word “frequent” drops out of the Green et al. article altogether, and Green et al. is the report upon which the CDC recommendation is based. Frequent is later defined by Floyd et al. as six or more drinks per week, but throughout the rest of the article, the goal of which is to determine how many women drink alcohol before realizing they are pregnant, the authors also talk about moderate drinking as seven to 14 drinks per week and heavy drinking as more than 14 drinks per week. They note that while 51 percent of nonpregnant women had consumed alcohol during the month prior to the study, only 3 percent of nonpregnant women reported moderate drinking (7–14 drinks per week) and 1 percent reported heavy drinking (more than 14 drinks per week), with 11 percent reporting binge drinking (five or more alcoholic drinks on at least one occasion). However, these are nonpregnant women. Floyd et al. find that among pregnant women, 45 percent consumed alcohol during the three months before they knew they were pregnant with 25.5 percent imbibing less than one drink per week, an additional 14.8 percent also engaging in infrequent and light (by the authors’ own definition) drinking by consuming one to five drinks per week, and only 4.9 percent consuming six or more drinks per week with 3.6 percent of those frequent drinkers drinking moderately and only 1.3 percent drinking heavily, and this includes the month or months during that three month window when women were not pregnant yet. Approximately 20 percent knew they were pregnant prior to four weeks gestation, more than 50 percent knew by the sixth gestational week, more than 70 percent by the eighth week, and more than 90 percent by the end of the first trimester. Once the 45 percent who drank prior to realizing they were pregnant became aware of the pregnancy, what were predominately low levels of drinking to begin with were typically reduced further: “Prior to pregnancy, 5% of all women drank 6 or more drinks per week; after pregnancy recognition, less than 1% (0.7%) drank at this level” (Floyd et al. 1999:103). Only 0.4 percent engaged in moderate drinking and 0.2 percent engaged in heavy drinking after finding out they were pregnant.

We do not know how many of these women reduced their alcohol intake due to pregnancy-related nausea or food/alcohol aversions versus fears about the possibility of harming their children. We do know that Floyd et al. oversampled the mothers of low and moderately low birthweight children, a factor reported among children diagnosed with FAS. Nowhere in the article, however, does it say whether any of the close to 10,000 children, all born in 1988 a decade before the study, had been diagnosed with FAS or related alcohol-induced disorders. In the discussion, the authors jump from stating that many women who do not yet know they are pregnant are consuming alcohol to talking about FAS, but the link to FAS is not based on their data:

Documented cases of fetal alcohol syndrome have been found predominately among women consuming heavy amounts of alcohol, (i.e., 10 drinks or more per day [italics ours]), although more recent studies have found measurable growth and cognitive effects among children whose mothers drank at thresholds comparable to an average of 6 to 7 drinks per week. In addition, increased risks of spontaneous abortion have been reported at this level and lower. (Floyd et al. 1999:105)

The articles they cite for cases of deleterious effects at six to seven drinks run into a common problem, conflating low-level drinking on a regular basis with extreme consumption in one sitting. Given credible questions about negative effects of binge drinking specifically, it could be that having seven drinks in a row every Saturday could be linked to FAS whereas having one drink a day seven days in a row is not, yet both would fall under the category of an average of six to seven drinks per week (Armstrong and Abel 2000). Jacobson and Jacobson (1994), while unwilling to declare any amount of alcohol safe, repeatedly note that there is little evidence of neurobehavioral effects and no functional impairment below seven drinks per week, and although they report a threshold of seven to 28 drinks per week for neurobehavioral effects, their drinkers who consumed seven or more drinks per week drank an average of six drinks per day. In addition, Jacobson and Jacobson (1994) presented a table where only one child out of 61 infants born to “very heavy” drinkers scored in the bottom 10th percentile on mental development, but erroneously report the statistic as 6.7 percent instead of the correct 1.6 percent. The incorrect 6.7 percent is the percentage of children of abstainers (4/60) who scored in the bottom 10th percentile! As for the account of spontaneous abortion, Windham et al. (1997) did find an association between 3.5 or more drinks per week and miscarriage (odds ratio of 1.9 for 0.5 or fewer drinks per week vs. 2.3 for more than three drinks per week), but not only do they cite other work that finds no association, their own results indicate that women who drink one to three drinks per week during pregnancy have the lowest odds ratio for miscarriage (1.0). ³

In a recent CDC report highlighting binge drinking, “Alcohol Use and Binge Drinking Among Women of Childbearing Age—United States, 2011-2013,” by Tan et al. (2015), also cited by the Green et al. CDC report with which it shares two authors (Tan and Denny), the rates of alcohol use by pregnant women were 10.2 percent for consuming at least one alcoholic beverage in the past 30 days and 3.1 percent reporting binge drinking (in comparison to 53.6 percent and 18.2 percent of nonpregnant women). In this study, the definition of binge drinking is lowered from five or more to four or more drinks in one sitting. The authors make much of the finding that pregnant binge drinkers binge more frequently than nonpregnant binge drinkers (4.6 episodes compared with 3.1 episodes) and, while not statistically significant compared with nonpregnant binge drinkers, the highest number of drinks in an episode, 7.5, came from pregnant binge drinkers (Tan et al. 2015). They should make much of this finding. Pregnant binge drinkers are more likely to have problems with alcohol. Tan et al. (2015) suggested that “[o]ne possible explanation for this might be that women who binge drink during pregnancy are more likely to be alcohol-dependent than the average female binge drinker . . .” (p. 1043). Not reducing binge drinking behavior during pregnancy, especially in the era of ubiquitous FAS warnings, is a likely indicator of alcohol dependency, and alcoholic women are at risk of having babies with FAS.

Beyond slights such as turning “frequent drinking” into drinking any amount and redefining categories such as moderate to include less alcohol and consequently increasing the number of women falling into that category, health researchers also mix behaviors that society deems undesirable with behaviors that are actually dangerous. Tan et al. (2015:1042) began their article with the words “excessive alcohol use” and go on to list consequences such as cirrhosis of the liver and violence. However, when they mention FAS disorders they do so without using the word excessive. In fact, “excessive alcohol use” is followed by an asterisk, and that asterisk at the end of the paper defines excessive usage as

binge drinking (≥4 drinks on an occasion for women, ≥5 on an occasion for men), high weekly consumption (≥8 drinks a week for women, ≥15 drinks a week for men), any alcohol consumption by pregnant women, or any alcohol consumption by those under the minimum legal drinking age of 21 years. (Emphasis added)

Clearly, an 18-year-old having one beer at a party is not at risk of cirrhosis of the liver even though she or he may be breaking the law. The authors therefore confuse laws and, in the case of pregnant adult women, norms, with actual detrimental health outcomes. This conflation of nonmedical dangerous social behaviors with health prescriptions in a CDC report is troubling and points toward researchers and health policy experts having more on their minds than science.

The CDC argument is that the United States has a very high rate of unintended pregnancies for a developed nation, only 55 percent of pregnancies are intended (Guttmacher Institute 2016). Of the 45 percent that are not, 27 percent are mistimed and 18 percent are unwanted (Guttmacher Institute 2016). As many women get pregnant without realizing it, public health officials therefore believe all potentially fertile women must be cautioned not to drink. While a woman is indeed unlikely to know that she is pregnant until a few weeks after conception, not only will 15 percent of pregnancies end in miscarriage during the first trimester (Weeks and Danielsson 2006), but of those that do not, 42 percent of unintended pregnancies will end in abortion (Guttmacher Institute 2016). Consequently, the rate of babies born from unintended pregnancies drops sharply. Coupled with the statistic that 41 percent of American women do not drink at all, the number of alcohol-affected births plummets (Armstrong 2003; Jones 2015; Tan et al. 2015). Even fewer women abuse alcohol by drinking excessively, and while self-reports of excessive drinking and binge drinking may indeed suffer from underreporting, and granting that women who binge drink may be among those more likely to get pregnant unintentionally, the rate of women who do not know they are pregnant and are engaging in dangerous levels of drinking is much smaller than the hyperbole of these studies would lead us to believe. Reassuringly, recent research by McCormack et al. (2018) not only found that 12-month olds whose mothers drank seven drinks or fewer per week during pregnancy had better cognitive scores than the children of abstainers, but, addressing fears about women who drink before they know they are pregnant, they also found that “[a]lcohol exposure at any level [italics ours] in the first 6 weeks of Trimester 1 was not associated with any difference in Bayley cognitive score in unadjusted analyses nor following any level of statistical adjustment” (p. 334). ⁴ Finally, even among women with alcoholism, not all babies have FAS. Sokol, Ager, and Martier (1988) set a threshold for FAS at 42 or more drinks per week, but that does not mean that all babies of mothers who drink that heavily exhibit FAS (Sokol et al. 1988). Positive diagnosis is strongly linked to the length of time the mother has been an alcoholic and her class status (Abel and Hannigan 1995; Armstrong and Abel 2000; Bingol et al. 1987).

The 2018 JAMA article that is the foundation for the NYT article also provokes questions. The authors include three levels of disorders: diagnoses of FAS, the less severe but more broadly defined (no growth deficiency required) partial FAS, and also alcohol-related neurodevelopmental disorder that diagnostically does not include facial dysmorphia or other physical characteristics but is categorized by markedly poor performance (normally at least two standard deviations from the mean) in multiple areas such as executive functioning, visuospatial perception/skills, social skills, attention, and learning ability/retentiveness. For a classification of alcohol-related neurodevelopmental disorder, there has to be evidence of maternal drinking during the pregnancy. Such evidence is not required for diagnoses of FAS or partial FAS. In addition to including a large range of FAS disorders, ⁵ for their criteria, the researchers (May et al. 2018:476) “applied a less stringent cutoff of more than one standard deviation below the mean for specific learning impairments,” which will necessarily increase the number of children included. They admit that their decision to include functioning with classical FAS behaviors could be problematic but argue that it was done to make it easier to test first graders across schools. Ultimately, the impact is broadening the criteria for counting children as alcohol-damaged.

Pregnancies were counted as alcohol-exposed in this study if mothers or others close to them reported that she drank six or more drinks per week for two weeks or more of the pregnancy or had three drinks or more on an occasion twice or more during the pregnancy, had trouble with the law due to drinking, or was treated for an alcohol-related problem. The authors, then, were unwilling to diagnose alcohol-related neurodevelopmental disorder in mothers who had only two drinks in a row or who consumed fewer than six drinks per week. Researchers were not expecting to see even the mildest of ostensible FAS disorders, neurodevelopmental, in children whose mothers drank some but less than their cutoff criteria, and this again makes us critical of health agencies’ recommendations that bar any amount of alcohol at all during or before pregnancy. Yet, May et al. (2018) cannot answer the question of whether drinking more than this amount actually led to problems with spelling, math, attention, or mood regulation or whether some or all of these children would have had such problems even in the absence of their mother’s drinking. Combining their four research sites and three sampling methods, May et al. found 222 cases of FAS disorders in 6,639 children for a total of 3 percent. However, they note that full blown FAS accounted for less than 20 percent of FAS disorders in all of their samples. Only 0.4 percent of children (27) were diagnosed with FAS; 1.6 percent (104) were given a partial FAS diagnosis, and 1.4 percent (91) were classified as having alcohol-related neurodevelopmental disorder. In their limitations section, the authors discuss that the number of cases was small and admit that “in the absence of a definitive biomarker for fetal alcohol spectrum disorders, it is impossible to know what proportion of these deficits were caused by fetal alcohol exposure. Therefore, prevalence estimates, particularly for alcohol-related neurodevelopmental disorder, could be overestimated” (p. 481). Yet the NYT article title is dramatic precisely because it warns that more children may have fetal alcohol disorders than originally thought. If we do not know that neurodevelopmental problems are alcohol-induced, and neurodevelopmental disorders made up 41 percent of the 3 percent of FAS disorders cases being claimed in the JAMA study, with partial FAS, which is also diagnosed subjectively, making up most of the rest, then this conclusion is a stretch.

Analysis

Is this much ado about very little? Based on the genealogy of the articles leading to the CDC warning and overblown press coverage (“Far More Children . . .”), we concur with previous assessments that argue that what we are seeing is a classic example of a moral panic (Armstrong and Abel 2000; Bell, McNaughton, and Salmon 2009; Cohen 1972). In the book Culture of Fear: Why Americans Are Afraid of the Wrong Things, sociologist Barry Glassner (1999) explained how the construction of social problems follows set patterns such as extending the scope of the problem beyond the original parameters, relying on misleading statistics, and depending on experts who make a name for themselves, in careers and/or in the media, through their claims about the issue. The tracing of FAS articles above provides numerous examples of the definition of FAS being extended to partial FAS, Fetal Alcohol Effect, and alcohol-related neurodevelopmental disorder, collectively fetal alcohol spectrum disorders (FASD), and claims made with problematic numbers such as how many nonpregnant women drink, whether an average number of drinks per week is one a day or all at once, changing criteria for abnormal from two standard deviations to 1.5, and so on. The authors of the JAMA article in 2018 include the original researcher who discovered FAS in 1973, Kenneth Lyons Jones and one of the lead authors on the study is a colleague of his, Christina Chambers, with whom he has been publishing about FAS for more than 20 years. The research was funded by two grants from the National Institute on Alcohol Abuse and Alcoholism (NIAAA). While “[a] group that raises money for research into a particular disease is not likely to negate concerns about that disease” (Glassner 1999:xxiii), the continuing insistence on FAS as an epidemic warranting media attention and federal dollars despite evidence to the contrary is bigger than the careers of individual scientists and advocacy organizations that stand to gain from increased interest in their specialty. As pointed out by Gray et al. (2009) in an article about the dearth of convincing data that low-level drinking during pregnancy is dangerous or whether it is a significant factor in neurodisability, “evidence alone is rarely sufficient to guide policy, which is shaped by culture and politics as well” (p. 1270).

Glassner (1999) made the case that the advent of a social problem based on pseudodangers does not occur in a vacuum but rather emerges at a given point in time because it resonates with specific cultural anxieties. Others have argued that FAS arose in the 1970s precisely because it was a moment in which environmental toxins reached public awareness, concerns about protecting precious children were heightened, and perhaps most importantly, women’s roles in society were changing leading more and more women to work outside the home and lobby for greater rights, including the right to legally terminate a pregnancy (Armstrong 2003; Best 1990; Golden 2005). Abortion raised the issue of a pregnant woman and a fetus potentially having an antagonistic relationship, and a doubling down on messages about what it means to be a good mother can be seen in that moment as a form of backlash to women’s increased independence. Yet if those were the cultural anxieties of the 1970s, why the resurgence of a FAS scare in the second decade of the twenty-first century?

As it was 40+ years ago, the misplaced emphasis on FAS risk prevention aimed at all women is not simply a natural outgrowth of doctors’ desire to protect patients and policymakers’ desire to protect the public, but rather is situated in a larger social context. The 1970s and 1980s saw a conservative political swing, in part focused on personal responsibility, that sought to counteract the social progress and rights for minority groups of the 1960s (Armstrong and Abel 2000). In the 1990s, the fear that mothers were destroying society by giving birth to alcohol-exposed infants switched to a fear that cocaine use during pregnancy would ruin the future of the country (Glassner 1999; Roberts 1997). Not only did evidence mount that these fears about cocaine were unfounded, but the U.S. progressively moved away from a no-tolerance war on drugs to legalizing marijuana in some states and calls to stop prosecuting low-level drug offenses in other states a few years into the 2000s. Yet a return to conservative ideology resulting in, as well as fanned by, the election of President Trump in 2016 has led to attacks on women’s health and rights that mimic arguments made in the 1970s. While a spate of recent challenges to abortion rights entail restrictions on timing of the procedure or the type of facility where it can be performed, the very question of Roe v. Wade, long thought to be settled law, could again reach the Supreme Court and this time with a possibly different outcome.

Where do scientists and physicians fit in? Why, despite years of research that disputes risks at low levels of drinking, do they continue to generate alarmist warnings? While certainly not all, or even most, of public health researchers, ob/gyns, pediatricians, and developmental biologists are conservative or hostile to women’s rights, the inherent sexism of modern U.S. society affects perceptions and behaviors of everyone raised here at this time in history, male or female, conservative or liberal. As Allan Johnson (1997) wrote, “[s]ince the [social system] we’re participating in is patriarchal, we tend to behave in ways that create a patriarchal world from one moment to the next” (p. 41). This sexism can slip out in a variety of ways either in direct hostility or much more benevolently (Glick and Fiske 1996; Hebl et al. 2007), like a well-meaning doctor telling a healthy, pregnant woman who wants to have a glass of wine on her birthday not to—after all, what kind of mother would risk her baby’s health?—perhaps without realizing that he or she recommends a glass of wine to pregnant patients following amniocentesis. The difference, of course, is about the proper role of a mother and (often subconscious) beliefs about morality. Glassner (1999:xxvi) discussed anthropologist Mary Douglas’ (1966; Douglas and Wildavsky 1982) work on risk: “Dangers get selected for special emphasis, Douglas showed, either because they offend the basic moral principles of the society or because they enable criticism of disliked groups and institutions.” Choosing to have a drink to celebrate a social occasion while pregnant can be read as selfish, whereas having a drink because a doctor said it might prevent dangerous contractions after a medical procedure is seen as being a good girl and following doctor’s orders as any mother (even if she as yet has no children) should do. Selfishness is deemed antithetical to motherhood (Hays 1996). While seeping across social categories, this is especially ubiquitous among middle and upper class women who perform “intensive mothering” and somewhat less characteristic of working class women from backgrounds where children’s needs can be viewed as second to the parents’ or family’s needs (Hays 1996), suggesting an interesting class lens for the treatment of alcohol use by women of different backgrounds. Golden (2005) traced how middle class, white mothers were initially portrayed as unwitting victims who hurt their children unknowingly through their alcohol consumption compared with later media reports highlighting FAS among less advantaged women, predominately of color, who were portrayed as choosing their desire for alcohol over their love of their children. ⁶ Hays’ assertion that the overinvestment in the mother-child relationship is so extreme precisely because society demands a bulwark against selfishness in an era of family disintegration and capitalism run amuck provides an explanation for the vehemence of reactions to mothers who appear to reject self-sacrifice, even for nonconsequential actions.

Bell et al. (2009)

suggest that the current assessment of the state of the evidence regarding the health “risks” attached to foetal [sic] alcohol exposure, childhood exposure to tobacco smoke and childhood overnutrition should be viewed as an ideological project as much as an empirically driven one. (P. 158)

They go on to point out that “each issue has stimulated the creation of large, well-funded and politically powerful movements that have sought to denormalise [sic] these behaviors and even enact legal sanctions against those who place their children ‘at risk’” (p. 160). The societal consternation and its attendant need to control “bad mothers” (Bell et al. 2009; Glassner 1999; Hunting and Browne 2012; Paltrow and Flavin 2013) thus permeates the “medico-moral discourse” reaching into grant-funding institutions, doctors’ practices, and the media (Bell et al. 2009). It also affects the penal code. Writing on the intersection of health guidelines about alcohol consumption during pregnancy and the law, Seiler (2016) stated that “[a]ll of these laws may be grounded in concepts of health protection, and one might presume that the focus on reporting infant outcomes reflects a nonpunitive approach to mothers” (p. 624). She goes on to point out the numerous punishments levied on women who use substances, including alcohol, during pregnancy and how these sanctions disproportionately target women of color despite equal levels of drug use. ⁷ In only one-third of these cases was an actual adverse pregnancy outcome reported, and even in these cases, “causal evidence linking the woman’s behavior to the outcome was often weak or nonexistent” (Seiler 2016:625). So women are prosecuted despite any evidence of actual harm to their offspring.

Roberts et al. (2017) studied states by counting the number of punitive, supportive, or mixed alcohol and pregnancy policies they have implemented between 1970 and 2013; they find that alcohol policies concerning pregnant women are becoming increasingly punitive. Punitive polices attempt to control women’s behavior by reporting their “abusive behaviors” (typically after the child is already born) and removing their children or terminating their parental rights. Supportive polices seek to provide women with information (in their definition this includes warning signs in restaurants and bars) and substance abuse treatment programs that prioritize pregnant women. Mixed policies include both punitive and supportive policies. Evidence suggests that punitive policies cause women to be more cautious and delay prenatal care and substance abuse treatment, thus undermining rather than aiding efforts to improve fetal and neonatal health (see also Jessup et al. 2003; Roberts and Pies 2011). After showing that punitive policies are positively associated with state policies that “restrict women’s reproductive autonomy,” Roberts et al. (2017) concluded that “a primary goal of pursuing such policies appears to be restricting women’s reproductive rights rather than improving public health” (p. 6).

In addition to the concrete instances of punitive actions toward women (e.g., losing custody of their children and/or going to jail), it is crucial to point out that in a study of perceptions of women’s behavior during pregnancy, those respondents most likely to express support for punitive actions toward pregnant women who engage in strenuous exercise, eat seafood, or drink alcohol reported that they were motivated by concerns other than a belief that the mother’s actions would hurt the fetus (Murphy et al. 2011). ⁸ This is striking. While there was an association with beliefs that the behavior was dangerous to the mother, but not the fetus, the strongest association was with negative feelings about pregnant women who drink. The authors report that respondents who believed these behaviors were indeed detrimental to the fetus were less rather than more likely to endorse punitive actions directed at pregnant women engaging in these actions (Murphy et al. 2011). Taken together the evidence above, from a psychological study, courtrooms, and state policies, suggests that the desire to punish women who ignore admonitions not to drink while pregnant is more about controlling women’s nonnormative behavior than it is about a genuine desire to ensure healthy babies. The fact that studies of alcohol and other possible teratogens carried by men’s sperm were slow to gain scientific acceptance and research support (Daniels 2006) and continued widespread ignorance in mainstream society of men’s contributions to the health of their offspring lend further credence to the real impetus being making women behave rather than actual health concerns.

As Foucault (2003) argued, “discipline tries to rule a multiplicity of men to the extent that their multiplicity can and must be dissolved into individual bodies that can be kept under surveillance, trained, used, and, if need be, punished” (p. 242). This applies to a further degree to women, who, through their bodies, produce the nation, and therefore who must be subject to more intensive purity policing (Peterson 1998). In her work on Alabama’s Chemical Endangerment to a Minor law, used to target drug using pregnant women, Howard (2014) made this connection explicitly. She asserts that “the disproportionately white, poor, and methamphetamine related arrest demographics are reflective of anxiety about perceived white degeneracy-that this population of deviant whites are perceived as polluting the white race and violating the norms of supposed white moral superiority” (p. 375). The Surgeon General’s warning to women who are not yet pregnant but thinking about conceiving, and more recently, the CDC’s warning to women who are not intending to conceive demonstrate that for government agencies, all women are potential mothers and are treated according to that role: producers of citizens rather than as citizens themselves. The benevolent (and sometimes hostile) sexism of denying adult women choices over what to consume thus belies structural violence in the name of the state/Foucault’s “biopolitic.” Reproduction is not a private act, but, rather, is so intensely politicized because “the coherence and continuity of the group—and the gender hierarchy it imposes—is ‘maintained and secured only by limiting the autonomy, freedom of choice and social adulthood of the group’s physical and social reproducers’ (Vickers 1990:482)” (Peterson 1998:42). The extension of warnings about alcohol and monitoring of all pregnant women’s behavior, not just those abusing alcohol, as well as the tendency to punish rather than help those women who are indeed suffering from an addiction, is an indication of the patriarchal underpinnings at play. In the paternalistic, masculine state, false interpretation of problematic drinking for pregnant women has permeated American’s two problem solving systems, the health care system and the criminal justice system. The misapplication of specious logic with respect to FAS in both systems is mutually reinforcing, perpetuating, in Merton’s terms, a “reign of error” (Merton 1948:195).

Conclusion

Locating social problems within individuals, who are then portrayed as either foolishly making bad choices or lacking in morality, rather than as a manifestation of larger societal issues, has consequences. When poverty appears to play a significant role in FAS among pregnant alcoholics, but we focus only on alcohol, and when we ignore the conditions that lead to alcoholism and/or correlate with it, such a polysubstance abuse, underemployment, domestic violence, and compromised living environments (high stress, increased exposure to lead and other toxins, failure to meet basic health and nutritional needs, etc.), we take the responsibility for a solution off of society and put it on the backs of individual women who are already struggling to cope. Indeed, the women who most need help, not only for the fetuses they may carry but also for their own health and lives, face increased challenges (e.g., an inability to seek medical help for fear of punishment; losing a job and the health insurance that goes with it after being jailed) precisely because of health workers’ insistence that their behavior is to blame. This is ironic because the NIAAA worked hard to change society’s view of alcoholism from antisocial behavior resulting from a failure of character in the early 1900s to a disease model in which the alcoholic was to be treated for addiction as a medical problem in the second half of the 1900s (Rotskoff 2002). Yet, after the advent of FAS, and the research on it funded by millions of dollars of NIAAA grant money, pregnant women are ultimately excluded from being ill alcoholics and instead thrust back into the camp of moral failing. ⁹ In admonishing the mothers of FAS children, all women are warned by their example and cautioned to control their behavior to an extreme and nonsensical degree so as not to be liable. At the same time, very little research has been conducted in decades to study the interaction between nutrition, alcohol consumption, and FAS or other factors that are related to social class (Abel and Hannigan 1995; Armstrong and Abel 2000; Bingol et al. 1987), either due to lack of interest by researchers or an inability to secure support from funders. If the goal is truly the health of babies, then interactive effects (poverty, nutrition, length of alcoholism, pattern of drinking, and stress) all need to be systematically studied.

Grasping the pernicious links between the health care system and the criminal justice system, Seiler (2016) asked public health officials to consider the legal consequences of their health recommendations before making them. She details the price that some women have paid, including incarceration and loss of custody of their children, even when harm to their children is not proven, due to laws written or interpreted a particular way because of public health recommendations. Yet, Seiler herself misses a crucial point. Although she warns that the CDC materials do not caution women that they may face legal penalties for admitting their drinking during pregnancy to a doctor or nurse, she concedes that a recommendation to seek care (if they drank alcohol before knowing they were pregnant or if they believe they have a drinking problem) is “appropriate” (p. 625). While few would disagree that it is generally a good idea for people of any sex with a drinking problem to seek help, how is it appropriate or understandable for trained scientists and health policymakers to scare and shame women who drank before they knew they were pregnant, particularly when the health care provider has nothing to offer them after their admission? Short of abortion, there is nothing a pregnant woman can do about having consumed alcohol while pregnant, and, as all the evidence indicates, an abortion is an entirely inappropriate suggestion for a light drinker who wishes to continue her pregnancy, yet the public health warnings lead some women to unnecessarily consider it (Armstrong and Abel 2000). When members of the public hear a CDC recommendation, they usually assume that it is based on data, yet there is absolutely no evidence that a single drink during pregnancy causes an adverse birth outcome. In fact, evidence is debatable that drinking during pregnancy will result in FAS unless the woman is, in effect, drinking to the level that she likely should be concerned about a possible drinking problem and possibly her own health (e.g., binge drinking), and results for neurodevelopmental/cognitive deficiencies at lower levels of consumption are contradictory and hard to tie directly to alcohol rather than class level of the parents, genetic factors, and so on.

Multiple studies over the course of the last 30 years find that the optimal rate of drinking for most adults is somewhere between three drinks per week and a drink a day; those who consume at this rate have lower rates of cancer and heart disease and better longevity than those who drink more and those who abstain from alcohol completely (Kunzmann et al. 2018; Moore and Pearson 1986). ¹⁰ Given our critiques of FAS studies and their coverage in the press, we recommend scrutiny of all research (medical or other) and question whether a third factor such as socioeconomic status is influencing both light drinking behavior and longevity. Nevertheless, given the CDC’s misguided warning on FAS, it would make sense for them to issue more appropriate warnings in line with the fact that FAS researchers are not looking for effects, even neurodevelopmental ones, below six or seven drinks per week or two drinks in a row. Thus, instead of total abstention, the CDC and the Surgeon General could follow Singapore’s example and use existing research to advise people in general, including pregnant women, that they can safely drink light amounts of alcohol but that more frequent drinking and larger amounts of alcohol in one sitting should be avoided. Binge drinking is not healthy and presents risks, such as impaired decision making, to anyone of any age, class, sex, or pregnancy status, so there is no need to focus warnings about binge drinking on pregnant women. Even if public health officials issue a conservative recommendation well below thresholds where researchers start to look for the mildest forms of FASD, public health sanctioning of light drinking during pregnancy would profoundly affect how Americans view pregnant women who drink and how pregnant women think about their own behaviors. It would also encourage women to ask questions and do their own research into drinking while pregnant. It may even open the door to women discussing their drinking behavior while pregnant more honestly with their medical care providers.

The Surgeon General and CDC’s targeting of pregnant women for alarms about alcohol not only neglects the higher rates of harm caused by men who have imbibed (see Armstrong 2003 on gendered rates of assault and vehicular accidents while under the influence of alcohol), but puts all women, not just those at risk of having babies with FAS, under increased surveillance—both by the state and the benevolently sexist bartender, waiter, or neighboring restaurant patron. While this is worse and leads to more consequential penalties for poor women and for women of color (Bell et al. 2009; Howard 2014; Paltrow and Flavin 2013; Roberts 1997; Seiler 2016), it leads to costs, both societal and psychic, for all women. Striving to convince women that even one drink during pregnancy is selfish and disqualifies them from the good mother category perpetuates sexism and the unequal treatment of women. In an era of increased restrictions on women’s bodies and behavior, attacks on their autonomy, and continued harassment, it is as important as ever to challenge messages to women that come from a place of social control, perhaps especially when they are coming from a trusted source of medical knowledge.

Related Resources

The American College of Obstetricians and Gynecologists (ACOG) issued a Committee Opinion entitled “Substance Abuse Reporting and Pregnancy: The Role of the Obstetrician-Gynecologist,” reminding providers that incarceration and threats of incarceration do not effectively curb alcohol and drug abuse and urging OBGYNs to work to prevent and repeal state laws that harm women and their pregnancies: https://www.acog.org/Clinical-Guidance-and-Publications/Committee-Opinions/Committee-on-Health-Care-for-Underserved-Women/Substance-Abuse-Reporting-and-Pregnancy-The-Role-of-the-Obstetrician-Gynecologist

National Advocates for Pregnant Women (NAPW) is a nonprofit organization that provides legal counsel to women facing legal consequences and/or loss of their children for substance abuse during pregnancy: http://advocatesforpregnantwomen.org/main/publications/

Sister Song Women of Color Reproductive Justice Collaborative works to ensure reproductive health and rights for marginalized communities: https://www.sistersong.net/reproductive-justice