Design

Meta-ethnography is an interpretative methodology for qualitative evidence synthesis²¹. It has become an influential and increasingly popular methodology in health and social care research, drawing on Geertz’s²² concept of ‘thick descriptions’ as well as Turner’s²³ theory of sociological understanding as ‘translation’. Meta-ethnography has the potential to produce new interpretations, models or theories. We conducted the study according to the seven overlapping and iterative phases of meta-ethnography described by Noblit and Hare²⁴ (Table 1) and the eMERGe metaethnography reporting guidance²¹ (Table 2). The guidance consists of 19 reporting criteria, covering all phases of the meta-ethnographic method.

Inclusion criteria and search strategy

We included empirical studies using qualitative methodologies focusing on midwives’ and other HCPs’ experiences of screening pregnant women for alcohol use in pregnancy and counselling them, published between 2010 and 2021. Articles presenting different perspectives were included when it was possible to separate them, and articles using mixed methods were included if they presented discernible findings from the qualitative part. The articles had to be written in English or one of the Nordic languages and published in a peer-reviewed scientific journal. We excluded quantitative studies, theoretical articles, reviews, Master’s theses, and dissertations.

We conducted an initial scoping search in February 2021. This was followed by a discussion with the team and a revision of key terms before new literature searches were conducted in February and August 2021 (Supplementary file Part 1). An updated search was conducted in January 2023 (Supplementary file Part 1). We included four databases: CINAHL, Maternity & Infant Care, MEDLINE, and Scopus. Medical subject headings and text words used for the searches included: ‘alcohol’, ‘alcohol drinking’, ‘alcohol consumption’, ‘pregnant’, ‘gestation’, and ‘maternal’ combined with ‘qualitative’. Boolean operators were used to combine the terms and truncations were used to ensure a sufficiently broad search. We backtracked reference lists in the included articles. The first author was responsible for carrying out the literature searches and experienced librarians assisted in identifying relevant key terms and performing the searches.

Search outcome

We describe the literature search and screening process in detail in a PRISMA flow diagram²⁵ (Figure 1). Article titles were read to remove duplicates and irrelevant publications. The next step involved reviewing abstracts and excluding irrelevant articles or articles employing unsuitable methodologies. The remaining 18 articles were read in full and assessed for eligibility according to the inclusion and exclusion criteria, resulting in the exclusion of four articles due to insufficient data. The reference searches resulted in two articles, but these turned out to be duplicates and were therefore excluded. The updated search did not result in the inclusion of more articles. All authors were involved in the screening process, which began with individual readings and was followed by discussions to reach a consensus on eligibility and inclusion.

Figure 1

The PRISMA 2020 statement: an updated guideline for reporting systematic reviews²⁵

Quality appraisal

To assess the included articles, we used a checklist from the Critical Appraisal Skills Program (CASP)²⁶. CASP is designed as a pedagogical tool. For us it served as a systematic reminder of issues related to study quality, including study aim, methodology, recruitment strategy, data collection, reflexivity, ethical issues, data analysis, statement of findings and study value (assessed by 10 questions given in Table 3). We divided the articles that were eligible for inclusion between the authors and assessed them independently. Afterwards, we discussed minor disagreements to reach a final consensus on inclusion. Two articles answered 8 of 10 questions, lacking information about reflexivity and data analysis. Nine answered 9 of 10 questions, lacking information about reflexivity. The remaining three articles answered all 10 questions. We, therefore, considered all articles to have sufficient quality to be included and provide a detailed list of the quality appraisal in Table 3.

Data extraction, analysis, and synthesis

All authors were involved in the process of repeated reading and data extraction, working in pairs with the first author to discuss possible differences of opinion and overlooked findings. We made use of our experience as qualitative researchers as well as our varied professional (midwife, public health nurse and psychiatric nurse), theoretical and international experience when performing the analysis. We developed a table describing each study’s aim, sample characteristics, research design, data collection, data analysis and key findings to provide context (Table 4) and continued by re-reading the results sections, noting the interpretative metaphors including citations, themes, categories, and concepts in a Microsoft Word table. This enabled us to compare findings across the studies. The next phase, translating the studies into one another, was performed by the first and third authors. This process involved moving back and forth in the data within each study and between the studies without losing sight of the context of the findings. According to Noblit and Hare²⁴, it is important to consider the themes, concepts and metaphors in the studies to determine how they are related. They can be directly comparable (reciprocal), they can stand in opposition to each other (refutational), or they can say something about the whole, based on a set of parts (line of argument).

Data from the primary studies in the matrix, along with the translations and our notes from the individual and team discussions constituted the basis for the final synthesis. Meta-ethnography has the potential to create metaphors to enable deeper understanding²⁴. The synthesis process allowed themes to emerge and resulted in an overarching lines-of-argument model²⁴. The translations of the findings of each primary study into each other encompassed all the translations into an overarching metaphor in an interpretative yet systematic process. We began with an assumption of reciprocity between the findings, i.e. incorporating all the findings in an analogous interpretation, moving on to check possible contradictions that emerged, and finally constructing a lines-of-argument synthesis²⁴ in which we identified different aspects of our topic, moving beyond the original primary studies and probing the metaphor. This type of interpretative synthesis, including the use of metaphors, is a characteristic feature of meta-ethnography. The synthesis should accurately portray both the shared and unique findings of the included studies. GRADE CERQual²⁷ (Supplementary file Part 2) was used to assess confidence in the evidence from the study findings. The tool comprises four components: methodological limitations, relevance, coherence, and adequacy of data. The CERQual assessment revealed moderate confidence in the themes.

The synthesis

In the synthesis, we use the metaphor of Pandora’s box (Figure 2) to deepen our understanding of the topic. In Greek mythology, Pandora’s Box is related to the myth of Pandora, a woman whose curiosity led her to open a box that she was forbidden to open. When she removed the lid, all the troubles of the world swarmed out, never to be recaptured. Pandora was scared when she saw all the evil coming out and tried to close the box, without success. Only Hope was left in the box, stuck under the lid. In modern times, Pandora’s box has become an idiom referring to any source of great and unexpected trouble. Anything that looks ordinary but may produce unpredictable, harmful results can thus be called a Pandora’s box²⁸.

Figure 2

Pandora’s box

Our findings show that many HCPs ‘tiptoe around the box’, not wanting to face the consequences and responsibilities of asking women about their alcohol use. These HCPs argue that women do not drink during pregnancy or have stopped drinking. They perceive young partying women to be alcohol users and Muslim women to be abstinent. Some consider alcohol to be a problem only in early pregnancy, and question present recommendations to completely avoid AIP. Altogether, their practice reflects a silent unspoken culture based on stereotypical assumptions rather than a practice based on evidence.

Healthcare providers sometimes ‘refuse or are reluctant to open Pandora’s box’ and prioritize screening and counselling in other areas, seemingly due to lack of time. Further, routines for asking about AIP are absent or are only part of the booking visit. They believe that women have no interest in talking about AIP and try to avoid making women who have used alcohol feel guilty. Moreover, they seem to lack knowledge about AIP, display underlying doubts about whether alcohol use in pregnancy exists as a problem, and disregard the follow-up of alcohol use in antenatal care. This may result in the marginalization of health-promoting antenatal care.

The midwives and HCPs that eventually ‘open Pandora’s box’ understand the importance of establishing a trusting relationship with their patients to address the topic of alcohol use. They see the need for knowledge and screening tools, and they keep information at hand for women. In these cases, Hope is left in the box and may help create trust in the presence of the midwife, allowing the woman to share her fear, shame, worries, and health problems related to alcohol. Rather than closing Pandora’s box, the midwife listens to the woman and helps her plan how to stop or reduce her alcohol intake, creating hope for better health and a better future for her unborn baby and herself. An overview of themes and sub-themes is presented in Table 5.

Tiptoeing around Pandora’s box

Doubts regarding the risk of alcohol use in pregnancy

Healthcare providers considered small amounts of AIP harmless^29-31. Some were skeptical that moderate use would pose a risk to the foetus^29,31 and associated FASD with heavy, sustained drinking³¹. Uncertainty regarding the threshold at which alcohol could harm the baby prompted midwives to take a cautious stance in relation to advising abstinence³¹. Some argued that alcohol consumption in late pregnancy was less harmful than drinking in the first trimester³¹ while others supported the precautionary principle in the guidelines but acknowledged the uncertainty of risk of alcohol-related harm to the fetus at low levels of drinking³². Thus, the motivation to dissuade pregnant patients from occasionally drinking, if they were not considered at-risk, was low²⁹. They believed that the recommendations to avoid alcohol consumption was introduced as a risk mitigation strategy and that evidence that small quantities of alcohol are harmful is lacking³⁰.

Prejudices related to women using AIP

While some midwives believed that the risk of prenatal alcohol consumption crossed social strata³³, others argued that AIP was related to socio-economic status^30,34. They considered AIP to be a concern for vulnerable groups^29,34 and argued that there were other underlying issues present, such as mental health problems³⁰. Muslim women were considered non-drinkers³⁴, and some HCPs found it difficult not to fall for the stereotype that women who drink during pregnancy come from lower income families²⁹. It was important to understand the role of alcohol in a woman’s life, as some women might be affected by their own mothers’ substance use or by trauma or the previous removal of a child³⁵. They considered alcohol-related harm to be linked to the problem of addiction or binge drinking²⁹, and FASD was considered an issue particularly for families in which there were other issues associated with alcohol use³³.

Alcohol use as an accepted part of social life

The attitude that it is acceptable to drink a little during pregnancy is widespread in society, and it informs women’s decisions about drinking³³. The media, including social media, played an important role in women’s perception of AIP, and midwives believed that women felt social pressure to drink during pregnancy³⁰ or felt that drinking was unproblematic due to the legal and social acceptance of alcohol³⁵. Australia is considered a drinking culture and midwives’ advice was considered secondary to contrasting advice given by family and friends³⁶. Midwives considered nine months a long time to abstain given the centrality of alcohol in many social situations³⁷. It was challenging to discuss AIP in social settings³³, and partners rarely considered changing their own behaviors to support future mothers³⁷. Thus, spouses should also be encouraged to limit their alcohol consumption²⁹.

Alcohol in pregnancy is a health policy challenge

Healthcare providers questioned the reluctance to highlight health risks related to AIP³³. They believed that alcohol-related harm was minimized in public discourse because of politicians’ personal views on and experiences with alcohol, as well as political and economic pressure on the government by the alcohol industry³³. Drinking in general, and women’s drinking in particular, is a public health concern, and alcohol consumption is a massive issue³⁷. Communication problems around alcohol use in pregnancy could be related to broader social patterns of alcohol use³³.

Refusing to open Pandora’s box

Controversy regarding the responsibility to inform about alcohol in pregnancy

Screening for substance use in pregnancy was part of midwives’ role, and some felt confident discussing AIP with their clients^31,37-40. Midwives who did not drink alcohol were strongly against AIP, while midwives who did drink alcohol were skeptical of advising abstinence³¹. Some argued that advising women about AIP was ‘none of their business’³⁶ while others felt disinclined to discuss alcohol use because of personal beliefs³⁶. Often, they advised women to abstain from alcohol during pregnancy^{31,32,37,39,41}. However, some did not address moderate consumption³⁴ or recommended not to drink alcohol except on special occasions³⁹. Midwives should provide evidence-based information³², and one midwife called for clarity from the government to explicitly state that abstinence is the safest choice and for all HCPs to give the same advice³⁷.

Informing about alcohol use in pregnancy is challenging

Some midwives and other HCPs found it difficult to ask questions about AIP^32,33. Others considered the topic sensitive but not difficult to raise with clients³⁷. Midwives argued that recommending abstinence can lead women to develop feelings of guilt or anxiety if they have unwittingly consumed AIP^34,39. Thus, alleviating anxiety about unintended exposure to alcohol before the woman knew she was pregnant was important³². According to one midwife, if a client were to disclose that she had had a drink on a special occasion, she would alleviate her fears but continue to advise no alcohol⁴¹. However, it was important to have good rapport with the women^34,38 as asking about AIP could appear stigmatising²⁹ or ‘prying’⁴¹. Thus, midwives had to be careful not to alienate the women from antenatal care when discussing sensitive alcohol issues^31,34, as stigma created a threat of disengagement³⁸. In this regard, sometimes harm reduction was a more beneficial approach^29,40.

Women neither want nor need information about alcohol use during pregnancy

There was a widespread belief among midwives that very few women drink during pregnancy^34,37,41 or that they stop drinking once their pregnancy is confirmed^29-31. Pregnancy worked as a motivator to stop drinking³⁵, and women knew not to drink or to drink minimally³⁴. Consequently, some did not ask directly about AIP as they were convinced that their clients already abstained from alcohol^34,39. Further, the use of AIP raised the issue of women’s autonomy over their own bodies and the rights of fetuses versus the rights of women³³. Some argued that women who are ‘well-off’ are aware of risks and handle their consumption²⁹, but at the same time, midwives reported that they were aware that some alcohol or specific drinks were considered acceptable among women of higher socioeconomic status³⁷. Thus, recommendations and prenatal care would be different for ‘well-off’ women and ‘at-risk’ women²⁹.

Women who drink during pregnancy do not disclose their alcohol consumption

According to midwives and other HCPs, some women were not truthful when they claimed to abstain from AIP^29,39,41. Women knew that they should not drink³¹, but midwives were uncertain if this really stopped them from drinking³⁰. Thus, women could be recovering alcoholics or binge drinkers without HCPs being aware²⁹. They believed that the underreported alcohol intake was related to the stigma around AIP³⁷, arguing that women did not disclose if they were drinking³². Some were not ready to disclose³⁸ while others were unwilling to disclose³⁵, or they would ‘bargain’ about alcohol use³⁶. Non-disclosure or partial disclosure was associated with a lack of knowledge about risks and feelings of guilt and shame about use³⁸. Other reasons not to disclose were a lack of continuity of care and insufficient rapport with the care provider³⁸. Thus, lack of honesty and willingness to disclose were barriers to getting help⁴².

Providing information about alcohol in pregnancy is low priority owing to time constraints and a lack of routines

Screening for AIP added to midwives’ workload^{31,34,38,40,42}, resulting in alcohol consumption only being discussed at the booking visit⁴¹. Alcohol intervention was given low priority^31,34,40 as there was ‘so much other stuff’ to handle³⁰. Also, midwives needed to prioritize other issues that were more pressing³². The short duration of pregnancy restricted midwives’ ability to establish adequate support to enable women to reduce or cease alcohol consumption³⁴. Nevertheless, midwives’ knowledge and understanding of the risks of AIP influenced the level of priority, and practices and procedures played an important role in providing information³¹.

Opening Pandora’s box

Healthcare providers’ knowledge about alcohol in pregnancy is fragmentary

Most midwives had a good understanding of problems related to AIP³⁰ and were aware that alcohol use could cause birth defects^31,37. Some were able to discuss the consequences of AIP for the fetus and for reproductive health⁴², but misconceptions were also reported³⁹. Some were aware of FAS but could not explicitly describe the condition^30,39 or believed it was caused only by high alcohol consumption³⁶. Lack of knowledge about the risks of alcohol³³ or the effects of alcohol on infants⁴¹, including knowledge about FASD, was also described⁴⁰. HCPs stated that they had sufficient knowledge to deal with AIP appropriately^30,42 but lacked the knowledge to fully support their clients⁴² and wondered whether telling women they could have one or two drinks would encourage them to drink more²⁹. They were reluctant to engage too deeply with women’s substance-use problems if they did not know how to handle the information⁴², and some had no plans to look for additional information, although they were aware that additional knowledge would allow them to give better advice³⁹.

Guidelines and the media present conflicting messages

Some midwives were unaware of the guidelines regarding alcohol consumption, and many described a need for adequate materials to give to their clients⁴² to educate them about AIP⁴⁰. In addition to verbal advice, they used different sources of information including written^39,41 and digital⁴¹. They described challenges related to the use of recommendations as information and guidance were sometimes conflicting³² and expressed a lack of clarity about recommendations³⁶. Furthermore, changing guidelines and conflicting information in the media caused confusion about safe levels of AIP³⁷ and not all midwives were aware of the current official recommendations^32,39. Information about AIP was lacking in the media³⁰, and some midwives found that the media presented mixed messages about AIP³³. In addition, their clients would receive conflicting information and advice from relatives and friends, causing confusion about safe levels of AIP³⁷.

Screening for alcohol in pregnancy seems to be haphazard, focusing on high-risk groups and uncertainties related to use

Midwives’ sense of responsibility and willingness to help and reduce harm was a strong facilitator to screening³⁸, but the way AIP was assessed varied³². Routine screening was important⁴² and generally took place in early pregnancy^{30,37,38,40,41}. If nothing came up at the first visit, women were not asked again^30,37. Further, the quantity and frequency of consumption was not always investigated³⁰. Asking all women about AIP rather than focusing on highrisk groups normalized the topic^34,37, and when questions were part of routine practice, they could easily be followed up³⁴. For some HCPs, the decision to screen was based on personal feelings about and perceptions of women in their care²⁹. Midwives argued that they would talk with their clients about AIP if the clients requested the information or if there were obvious signs of alcohol abuse³⁰, when clients admitted alcohol use³⁹, or if they perceived the woman to be at risk of using AIP^36,41. They were generally able to identify women who drank high levels of alcohol³⁴. However, supporting heavy users was difficult³⁴. Some referred women with alcohol problems to specialist clinics⁴¹ but described a lack of multidisciplinary care and specialized treatment³⁸. Having an appropriate and validated screening tool was important for midwives to enhance the effectiveness of screening³⁸ and some used screening tools as part of their assessment for sensitive topics⁴⁰, for example, the AUDIT screening questionnaire^32,37. The questionnaire was sometimes used in person during appointments, or it was sent to women ahead of their appointments³². One study reported involving partners in screening³⁷. Some providers did not use tools at all³⁸ while others used a directive style of counselling⁴² or a strategy to involve women in a discussion about their current drinking habits³². The importance of a non-judgmental and supportive approach, as well as a causal and conversational assessment style was underlined⁴⁰. Discussion about AIP should be non-stigmatising³⁵ and non-confrontational, making disclosing comfortable³². Some midwives mentioned the value of having a checklist with questions about AIP³⁴ as well as knowledge about evidence-based strategies to support women⁴⁰. Experience and training on how to initiate a conversation about alcohol in a non-judgmental manner were considered facilitators to screening for AIP³⁸. However, midwives described having limited³⁶ or no formal training or education about substance use screening^38,40,42. Some had received training in motivational interviewing^32,37, but learning was also selfdirected through Internet searches, reviewing of articles and informal information sharing³⁶. Some had learned about AIP during their midwifery training³⁹, through voluntary extra courses, or from the media³⁹, but many expressed a strong desire to learn more about effective ways of responding to the problem⁴², arguing that a lack of skills and resources prevented them from raising the subject³⁴.

A supportive, non-judgmental and sustained relationship makes it easier to discuss alcohol in pregnancy with women

Pregnancy was believed to motivate changes in health behaviour³¹. Thus, it was important for providers to build a strong relationship with their clients based on an empathic, non-judgmental approach^29,35,37. Establishing and maintaining relationships facilitated trust, and women were more likely to disclose sensitive information, seek support, and be receptive to information and advice from a midwife with whom they had a trusting relationship^32,40. Screening for AIP should take place later in pregnancy, when the trusting relationship is established, to make it easier for the woman to disclose³². Screening at the first appointment may not be appropriate because of its potential impact on the provider–client relationship³¹. Women should be allowed to feel supported rather than judged²⁹, and advice and guidance should be tailored to their individual situation³⁷. Thus, attitudes towards women who used AIP were important as women were only likely to disclose if they had a good rapport with their midwife and encountered persons with a non-judgmental attitude⁴².

Strength and limitations

This meta-ethnography synthesizes midwives’ and other HCPs’ experiences with screening and counselling for AIP. A strength of the study is the rigorous methodology of systematic literature searches employed in 2021 and 2023. Altogether, more than 6000 articles published between 2010 and 2023 were screened. However, only 14 articles satisfied the inclusion criteria, demonstrating that qualitative research about AIP is scarce. We were also surprised to find that research on AIP in the Nordic countries was mainly quantitative. Our sample included 311 midwives and HCPs from six different countries, who pointed to challenges related to screening and counselling on different levels, in different cultures and policies and healthcare systems. Thus, it provided empirical variation and sufficient data to achieve the study aim. We are aware that we may have left out studies, that our keywords may not have been comprehensive, and that the inclusion criteria may have resulted in the loss of relevant information. Still, the use of e-MERGe meta-ethnography reporting guidance provided transparency and reflexivity to the process, and the use of the CERQual Qualitative Evidence Profile contributed to transparency and trustworthiness. Another strength is that our research team consisted of three people with different healthcare backgrounds (public health nurse, psychiatric nurse, midwife) and expertise in the subject under study (alcohol use, antenatal care, public health nursing) as well as expertise in methodology (qualitative research, metasynthesis).