Alcohol is a widely used substance among women of reproductive age in Canada [1
]. Researchers have identified sex-specific effects of alcohol on women’s health and national guidelines related to low risk alcohol drinking reflect this evidence, including recommendations for no alcohol use in pregnancy [5
Prenatal alcohol exposure (PAE) is a leading cause of disability and is considered to be a major public health concern in Canada [7
]. Fetal Alcohol Spectrum Disorder (FASD) refers to the constellation of effects associated with PAE. These include a continuum of neurodevelopmental impairments, resulting in problems with learning, memory, attention, executive functioning, affect regulation, adaptive behavior, motor skills and social skills [9
]. A wide range of long-term effects on health have been reported by adults with FASD, including, but not limited to: An increased risk of autoimmune disorders, hypertension, sleep disorders, osteoarthritis, vision and hearing problems, digestive problems, migraines and mental health concerns [11
]. The prevalence of FASD in the Canadian population is estimated to be as high as 4% [12
]. Alcohol use in pregnancy may also lead to preterm birth, defined as a delivery at less than 37 completed weeks of gestation, and the baby being small for gestational age [17
]. Given the considerable negative health effects of heavy alcohol use on women’s health overall, and the risk of FASD when pregnant women consume alcohol, it is critical to understand the level and patterns of women’s drinking before, during and following pregnancy.
There are well-documented barriers for women, and for service providers, related to asking about alcohol use in pregnancy in the clinical context. Service providers report being reluctant to ask women about their alcohol use, due to fears of jeopardizing their clinical relationship, or being perceived as judging and shaming women’s behavior [18
]. While research [20
] and prevention efforts such as the Women Want to Know campaign in Australia (fare.org.au) indicate that women appreciate the opportunity to discuss their alcohol use, and learn about ways to improve their health, persistent barriers exist for women to answering questions about alcohol use during pregnancy. For example, women report not feeling safe to report their alcohol use for fear they will be seen negatively by their service provider for their continued use in pregnancy, out of concern about involvement from the child welfare or justice systems, or out of the perception that accessible and helpful outpatient options are unavailable [19
As such, the confidential research context, in addition to the clinical context, could serve as an important setting for learning about alcohol use before, during and after pregnancy to inform FASD prevention strategies. In the confidential research context, the challenges mentioned above related to the safety of reporting of alcohol use are lowered. But, individual cohort studies generally include a limited number of regionally-distributed participants, and it is often essential to combine data across studies to truly unleash surveillance and research potential. An increasing number of national and international collaborative projects are opting for a co-analysis of individual participant data across studies to obtain sufficient statistical power, perform refined subgroup analysis, increase exposure heterogeneity, and enhance capacity to undertake comparison, cross validation or replication analysis [23
]. The Lifecycle [26
] and Enrieco [27
] projects are examples of such international initiatives. Generally, these initiatives use existing data to support their research activities, but the heterogeneity of the data collected (different questions, categories, etc.) limits the capacity to harmonize data across studies. The use of compatible tools to prospectively collect common measures would facilitate such harmonization [28
The Research Advancement through Cohort Cataloguing and Harmonization (ReACH) initiative (https://www.maelstrom-research.org/mica/network/reach#/
) provides a unique opportunity to leverage the integration of Canadian pregnancy and birth cohort information regarding women’s drinking during pregnancy. At least 12 of the cohorts documented in the ReACH catalogue include information regarding women’s alcohol use during pregnancy. These cohorts are based in Alberta, Ontario, and Quebec, and are studying maternal and pregnancy health and risk factors such as nutrition, weight, smoking exposure, hypertension, mental health, stress, and various environmental exposures. These 12 cohorts combined include data on a total n =
23,856 mothers, and n =
This article will explore how women are being asked information on alcohol use in pregnancy by cohort studies, and examine how a consistent approach that is informed by best practice in screening for alcohol use in clinical contexts can improve this data collection. Harmonization of data could overcome some existing challenges in Canadian FASD prevention research, which is currently limited by factors such as small data sets, and the lack of current population-based surveillance surveys that include information on maternal substance use.
A significant benefit of the cohort data available in the ReACH catalogue is that it includes birth outcomes, and longitudinal data collected about infants and children, some up to 12 years post-partum. Previous research has indicated that women who consume alcohol during pregnancy may not represent a homogeneous group [30
]. The larger sample size and increased statistical power afforded by harmonization will allow for a more nuanced analysis, such as identifying differing trajectories of alcohol use before, during and after pregnancy, as well as the effects of these trajectories, and the relationship of social demographic characteristics on infant and child outcomes. Such data on women’s alcohol use will provide information which is difficult to obtain through clinical interventions, and will be highly useful to inform FASD prevention strategies. Identifying social demographic correlates and trajectories of alcohol use among Canadian women will allow for the development of more targeted FASD prevention efforts. However, leveraging the development of a Canadian standard to identify alcohol consumption by women is essential to facilitate the harmonization and co-analysis of data across cohorts.
2. Materials and Methods
In this paper we identify: (1) The data that can be collected from formal validated tools; (2) the approach to data collection from Canadian provincial/territorial perinatal surveillance efforts; and (3) the data collected in a research context from the pregnancy cohorts included in the ReACH Catalogue as of November 2018. We use these three sources to make recommendations on guidelines for data collection about women’s alcohol use by future pregnancy cohorts. As there is no gold standard for the best practice related to screening and brief intervention for alcohol use and alcohol use problems with pregnant women, we describe below the lessons from each of these sources, and make recommendations for a consistent approach to asking about alcohol in the research context.
2.1. Identification of Current Practices in Screening for Alcohol Use in Clinical and Surveillance Contexts
We present data on current approaches to asking about alcohol use in pregnancy from structured screening tools used to identify risky drinking developed by organizations such as the World Health Organization (WHO). The screening tools and approaches discussed were identified during a literature review conducted by the Centre of Excellence for Women’s Health in 2017 with the goal of summarizing existing evidence on brief interventions with girls and women, including strategies for engaging their partners and support networks [33
]. The literature review included academic literature published between 2004 and 2017, identified using EBSCOHost Research Databases and grey literature (e.g., unpublished reports, practice guidelines) identified though targeted web searches [33
We also present the screening questions used and documented on antenatal records by provinces and territories in Canada. In April 2018, an online meeting of Canadian perinatal data collection (surveillance) experts was sponsored by the Centre of Excellence for Women’s Health for the purpose of discussing data collection methods in use across Canada, and how these practices align with what is known about effective brief interventions with women who use substances [33
]. We discuss these in the context of the evidence underlying Canada’s Low Risk Alcohol Drinking Guidelines [6
2.2. Identification of Data Collected in the Research Context
In November 2018, there were twelve pregnancy cohorts in the ReACH Catalogue that included at least one variable on maternal alcohol use during pregnancy. The data labels and response categories for all alcohol use variables in these cohorts were extracted, and variables that were collected during seven data collection periods (during each of the three trimesters, birth, 1–6 months, 7–18 months and 1.5 to 3 years postpartum) were charted in an Excel™ (version 16, Microsoft, Redmond, WA, USA) spreadsheet. Three studies also include data on alcohol use at 3–6 years and one at 6–12 years postpartum, but this information was not included in the present analysis. Questions with different leading statements that collected the same data, using the same response categories, were considered the same question.
For example, questions such as “Since you were pregnant, how often did you drink?” asked during the first trimester were combined with “In the last three months, how often did you drink?” However, if similar questions had different response categories (e.g., the same question with different categories during pregnancy compared with postpartum), or included different periods of recall, they were counted as separate questions (e.g., “In the 12 months prior to your pregnancy, how often did you drink?” was considered separate from “Currently how often do you drink?”). The data labels and response categories were compared across the cohorts to describe the information collected by each study, and to identify specific questions used to gather information regarding the frequency, quantity and heavy alcohol use before, during and after pregnancy. Information about the actual questions used during data collection were available for seven of the 12 cohorts extracted from the ReACH Catalogue, while the remaining five cohorts provided only data labels with information about the variable content (e.g., “Frequency of red wine (1 glass = 4 oz.)”).
There are challenges to gathering information about women’s alcohol use in the clinical context, related to barriers to reporting by women and to asking by health care providers. There are also barriers in establishing common Canadian standards to collect information across cohort studies. It is however important to know about alcohol use in pregnancy, in order to develop FASD prevention strategies based on accurate information about the prevalence, patterns and influences on women’s alcohol consumption. Accurate information about alcohol use in pregnancy also informs FASD diagnostic processes.
In this paper we have briefly outlined approaches to asking questions about maternal alcohol use employed by: (1) Evidence-based standardized alcohol screening tools, (2) the Canadian provinces and territories for the documentation of alcohol consumption for surveillance purposes, and (3) Canadian pregnancy cohort studies. Based on these three sources, we offer analysis and recommendations for a harmonization of data collection in the research context, which could contribute significantly to a more adequate understanding of alcohol use by women in the perinatal period to inform FASD prevention strategies.
Key elements about effective approaches asking about alcohol use that could be incorporated in pregnancy cohort research contexts, where confidentiality is ensured, are:
Asking about frequency of use using standardized categories, such as those in the AUDIT-C ((0) Never (1) Monthly or less (2) 2 to 4 times a month (3) 2 to 3 times a week (4) 4 or more times a week). This allows for a comparison across cohorts, and with other studies that employ the commonly used AUDIT-C, and corrects the problem that some cohorts had the lowest response category of “one or less” or “less than once a month”; i.e., it did not allow for the separation of no drinking, and low-level drinking.
Asking about the quantity of alcohol consumption using standardized categories such as those in the AUDIT-C which are (0) 1 or 2 (1) 3 or 4 (2) 5 or 6 (3) 7 to 9 (4) more than 10, or a continuous number that can be re-categorized as necessary. The use of visual tools that picture what comprises a standard drink has been found to be helpful when asking about quantity, as many women do not know about standard drink sizes [46
]. These tools show pictures of standard amounts of various types of alcohol, which can help women see that any type of alcohol may affect the fetus, given that misconceptions prevail about the benefits of some types of alcohol (e.g., red wine). Using both low and high examples of the number of drinks, such as is in the AUDIT categories, makes room for those who do drink at risky levels to see that researchers are interested in, and can handle, hearing that more than 10 is a possibility.
Asking about the number of drinks consumed on an occasion, as this has a bearing on the risk of adverse effects on the fetus, as well as on women’s health. It is recommended that “more than 3 drinks on an occasion” be used (or “4 or more”) to align with what is considered risky for all women in Canada’s Low Risk Alcohol Drinking Guidelines (LRDGs). In this way women will be able to see the consistency of any questions about alcohol that are being asked when they are accessing health care services generally, to those asked about consumption before, during and after pregnancy in the context of the pregnancy cohort studies. Learning about these LRDGs and reporting on alcohol consumption, referring to LRDGs guidance, is being recommended in alcohol screening with all men and women.
Asking about alcohol consumption before pregnancy (using the same standardized response categories for frequency and quantity described above), changes in use since pregnancy recognition, and quit date are also important information that could be gathered uniformly in the pregnancy cohorts. This information is important to understand for several reasons. When linked to birth outcomes, it can provide critical information to understanding risk, for example, the risk associated with any heavy preconception use.
It is also recommended that, no matter in what context that questions are asked about alcohol use, that this be done in a trauma informed and culturally sensitive way, and in a way that makes it safe for women to report harm reduction efforts, and not only abstinence. Women who have experienced trauma are more likely to drink during pregnancy [47
], and experience more mental health issues [48
]. Trauma-informed approaches are based in safety and collaboration [49
], and culturally safe, strengths-based and wellness -oriented approaches are foundational to any discussions of alcohol with Indigenous girls and women [50
]. Clinical guidelines on the topic of pregnancy and substance use are now making specific recommendations about using trauma informed, equity informed and culturally safe approaches, and such approaches are relevant for the research context as well [52
Harmonizing and integrating data from multiple studies across Canada would provide a unique opportunity to document information regarding women’s drinking during pregnancy. The co-analysis of harmonized data would also facilitate any investigation of the factors and characteristics that are associated with women drinking during pregnancy, such as age, education, household income, ethnicity, drug use, and smoking (We are currently undergoing this analysis, and the results will be described in a separate paper).
The creation of a larger sample could also be used to investigate the outcomes and correlates of more rare behaviors, such as heavy binge drinking during pregnancy, and to identify women’s trajectories of alcohol use before, during and after pregnancy [30
]. Harmonizing data sets that include birth outcomes and longitudinal data collection with infants and children will also allow for greater opportunities to measure the effects of different trajectories of alcohol use. However, the potential to harmonize cohort data is limited by the similarity of any information collected, and how it is collected.