The best magazine
Discussion
This study has sought to investigate alcohol use before and during pregnancy and predictors for drinking during pregnancy in Sweden. Approximately 6% of the women consumed any alcohol during their pregnancy. Older age, living in a large city, using tobacco during pregnancy, lower score for social support, stronger alcohol habit before pregnancy and higher score for social drinking motives were factors found to predict drinking during pregnancy.
We found that 84% of the women consumed alcohol the year preceding pregnancy, with 69% being moderate drinkers, 15% hazardous drinkers and 16% reporting prepregnancy abstinence. Four-fifths of the women with hazardous drinking before the pregnancy continued to drink until they became aware of their pregnancy, suggesting that there is a substantial risk that they consumed alcohol after becoming pregnant but before pregnancy recognition. In a study on pregnancy planning it was reported that only 10% of the women changed their pattern of alcohol consumption during the pregnancy planning period. Since alcohol consumption can harm the fetus in the earliest weeks of pregnancy, even before pregnancy recognition, it is of importance to find ways to prevent drinking in early pregnancy.
Abstinence after pregnancy recognition was achieved by almost all women. Most of the women who did not cease drinking during pregnancy reported drinking small amounts and few drinking occasions. These findings are similar to earlier studies conducted at a single antenatal care center in Sweden, showing prevalence rates of about 6% after pregnancy recognition. However, these studies used a retrospective questionnaire answered at home after giving birth. Other Swedish studies investigating the prevalence of alcohol consumption during pregnancy reported prevalence rates between 12% and 30%. These studies were single-centre studies undertaken in Stockholm, the capital of Sweden, and Uppsala, the fourth largest city in Sweden. These higher prevalence figures are in line with our findings showing that the prevalence of alcohol use during pregnancy is approximately twice as high in the three major cities included in this study compared with smaller cities and rural areas. Population-based data show a similar pattern of higher alcohol consumption in the major cities compared with the rest of Sweden. Since the prevalence rates vary from 0–13.9% in the 30 antenatal care centers included in this study, it is not surprising that the prevalence rates reported in single center studies vary a great deal.
Factors found to predict drinking during pregnancy were higher age, living in a major city, tobacco use during pregnancy, low social support, strong prepregnancy alcohol habits and higher score for social drinking motives. Some of these predictors are consistent with previously identified factors as age and smoking. Since smoking like alcohol is harmful to the foetus it can be of importance to be aware of the association from a preventive point of view. We have not found any international studies that support our finding that a higher share of women in larger cities consumes alcohol during pregnancy compared to women in smaller cities or rural areas. However, this likely echoes the drinking cultures and norms present in the cities and rural areas. The developers of MSSS has found that the scores are correlated with poorer health during pregnancy, contacting antenatal care later and more depressed mood after delivery. Our findings indicate that the instrument could also be useful to identify women with elevated risk for consuming alcohol during pregnancy, although this needs to be further investigated.
Pre-pregnancy habits were also found to predict drinking during pregnancy. Although we have not identified any studies investigating alcohol habits in a pregnant population, pre pregnancy drinking frequency/drinking behavior have been found to be the strongest predictor of drinking during pregnancy in several studies. Repetition of a behavior in a stable context is required for developing a habit, but the association between frequency of enacting a behavior and habit strength is not fully understood. Some behaviors turn into habits quickly, whereas others may require years of repetition. Research in various domains has shown that habits performed in stable contexts are unlikely to be spontaneously reconsidered. Because habits are triggered automatically in response to contextual cues, some sort of contextual change or disruption might be needed to make behavior-relevant information more salient and influential. It seems likely that pregnancy represents such a contextual change, providing a window of opportunity to break habits such as drinking alcohol, as has been suggested in previous research. This preparedness provides an advantage in preventive interventions targeting pregnant women.
The results from this study suggest that the five-question habit instrument SRHI could be used as a screening tool to identify women with risk of drinking after pregnancy recognition although this needs to be further investigated. The instrument has been used with numerous behaviours, but not with pregnant women. In current Swedish antenatal care all women are screened for prepregnancy alcohol use with the 10-item AUDIT questionnaire with the aim of finding women with increased risk of continuing drinking during pregnancy. Three out of four women who drank during pregnancy in our study did not have hazardous consumption in the year preceding pregnancy. Although we did not use the full AUDIT instrument, our results suggest that screening for hazardous prepregnancy alcohol intake might not be optimal to identify the women who will drink during pregnancy. This is in line with Magnusson, who found that most women who drink during pregnancy did not have AUDIT scores indicating likely alcohol dependence why further research to find more effective screening instruments is vital.
Differences in motives for drinking have been found to predict patterns of alcohol consumption and to be a risk factor for drinking in nonpregnant populations. Pregnant women in this study who consumed alcohol during pregnancy were more likely to score higher on social drinking motives for drinking in the year preceding pregnancy. Social motives have previously been associated with moderate alcohol use, coping motives with frequent but not heavier drinking, and enhancement motives with heavy drinking.
This study has some limitations that need to be considered when interpreting the results. We used self-reporting of alcohol consumption, which is a source of uncertainty because the responses might be influenced by social desirability, a bias that tends to be important when the questions deal with socially desirable (or undesirable) attitudes and behaviors. However, when assessment situations are structured to minimize bias self-reports show adequate reliability and validity. To reduce the risk of social desirability bias in the present study, the respondents were guaranteed anonymity, which has been shown to reduce bias in self-reports of sensitive behaviors.
We sought to achieve a sample which was representative of the distribution of pregnant women in Sweden. The invitations went through coordinating midwives who are in charge of a number of antenatal care centers in a region. Some of the coordinating midwives passed the invitation on to all centers in the region while others asked only a few centers. The research team was then contacted by or given contact information to centers willing to participate. Unfortunately, response rates for the invited centers are not available. It was not possible to obtain the precise number and proportion of pregnant women for all location/city size combinations. Thus, the study population does not entirely reflect the pregnant population in Sweden. Since the prevalence of drinking during pregnancy varied between large and smaller cities as well as between the different regions, it is possible that the reported prevalence rate is not representative of the whole population. On the other hand, our study adds information about differences across the country that has not been shown earlier.
This study required active participation by the midwives at the antenatal care centers to give the questionnaires to the pregnant women. We do not have full information on the extent to which midwives neglected to give out the questionnaire to some women, e.g. due to heavy workload. However, based on our informal assessment of several antenatal care centers, this type of omission was uncommon. Furthermore, this factor is unlikely to have biased the results in any specific direction.
Another potential limitation of this study is the risk of selection bias. This might lead to over- or underestimation of the prevalence of drinking during pregnancy differs between responders and non-responders. The drop-out analysis showed that the non-responders were younger than the responders, making overestimation more likely than underestimation. Further, the questionnaire was available only in Swedish, which meant that women who did not understand Swedish were excluded from the study. This might have affected the prevalence rate since drinking cultures and patterns of consumption vary between different cultures and locations.
A strength in our study is that we achieved a low dropout rate, which can be partially attributed to our efforts to establish good relationships with all participating antenatal care centers and to generate interest in the study. The first author visited nearly all centers in person to inform about the study and discuss various issues about the study procedure. To avoid burdening the midwives, data collection was restricted to a 4-week period at each center. All participating centers received a report with their unique data. Compared with previous Swedish studies conducted at single centers, our multicenter approach is an advantage because it increases the generalizability of the findings to the entire population. Another strength is the assessment of several predictors for drinking during pregnancy that have not previously been examined in a Swedish setting.
Source: ...