pregnant women continue to drink at hazardous levels despite factual knowledge about fetal risks, it is vital to understand the personal and social-environmental risk factors that support maternal drinking (May, in press; Waterson and Murray-Lyon, 1990; Weiner et al., 1989). Related research on women's drinking more generally may provide some answers for understanding the determinants of pregnant women's drinking behavior.
The following sections discuss (1) methodological considerations, (2) definitions and patterns of drinking among women, and (3) needed research on pregnant women's drinking.
Conceptualizing and measuring factors that increase the likelihood of drinking among pregnant women represent a complex task. These factors are likely to differ (1) for drinking in early pregnancy versus continued drinking throughout pregnancy; (2) for any alcohol use during pregnancy versus heavier consumption during pregnancy; and (3) for alcohol use during pregnancy versus alcohol abuse or dependence. In addition, learning more about factors that influence women of childbearing age to drink and to abuse alcohol could tell us which women are most likely to be drinking at high-risk levels before pregnancy is recognized.
In addition to personal and environmental risk factors that affect maternal drinking behavior, other factors may combine with alcohol either to reduce or to exacerbate fetal risks. For example, the incidence of FAS among "heavy" drinkers, variously defined (see discussion below), ranges widely but has never been found to be more than 40 percent in any study (Abel, in press). FAS is not unlike other teratogens in this regard. Very few, if any, teratogens have "attack rates" of 100%. A better understanding of biological and life-style characteristics associated with variations in fetal risk at comparable levels of maternal alcohol consumption is needed to understand the diversity in fetal outcome and might suggest prevention strategies that could strengthen naturally occurring protection against adverse fetal effects (Faden and Hanna, 1994).
During the past two decades, research on alcohol use in pregnancy has become increasingly international, including studies from the United Kingdom, France, Norway, Sweden, New Zealand, a multinational European collaborative project, and others (e.g., Plant et al., 1993; Tolo and Little, in press; Waterson and Murray-Lyon, 1989, 1990). However, studies abroad have uncertain utility for detecting social influences on drinking during pregnancy in the United States. For example, frequent alcohol consumption during pregnancy was not related to smoking behavior in a recent national study of women in New Zealand (Counsell et al., 1994). This finding diverges from those in most U.S. studies, which usually find drinking in pregnancy to be associated with smoking (Serdula et al., 1991). Because of the questionable relevance of some international data to social