Depression and anxiety are two of the most common health problems in pregnancy in the UK, affecting 12% and 13% of women, respectively (Heron et al, 2004; Gavin et al, 2005; Vesga-Lopez et al, 2008). The most common treatment is antidepressant medication (McManus et al, 2016). According to Einarson (2013), in 2013, antidepressants were the most researched drugs during pregnancy, and the findings related to risk of adverse effects for the fetus generated more than 30 000 potential outcomes. Outcomes included, spontaneous abortion, low birth weight, preterm birth, congenital malformations, neonatal withdrawal syndrome and persistent pulmonary hypertension of the newborn, in addition to the maternal morbidity incurred as a consequence of poor outcome (Bonari et al, 2004; Chan et al, 2014; UK Teratology Information Service (UKTIS), 2016). However, risks are difficult to quantify due to conflicting findings from teratological research (UKTIS, 2016) and the observational methods used, since pregnant women are not included in randomised controlled trials for antidepressant drugs (Einarson et al, 2015). Ambiguity around risk factors makes decision-making about antidepressant usage during pregnancy a complex process (Bonari et al, 2005; Patel et al, 2011; Stepanuk et al, 2013; Walton et al, 2014). This literature review therefore seeks to explore whether women feel empowered to make complex decisions about antidepressants in pregnancy.
Method and results
This literature review focused on papers using Boolean logic to incorporate the use of the word ‘AND’ within the search strategy. For example, ‘pregnancy AND depression AND decision**’ was used for the majority of the search. To enhance the search, reference lists from identified articles were reviewed for relevance. The search used CINAHL Complete, Intermid, ProQuest and Discover More databases. The process to select relevant articles included screening by title and abstract. To add rigor, an independent review was undertaken. Selection revolved around the exclusion criteria and relevance to the topic.
This paper focused on women's decision-making regarding the use of antidepressants. Articles were therefore excluded if the main focus was on decision-making by clinicians, postpartum depression, the teratogenic effects of antidepressants, or there was no analysis of how women are empowered to make a decision (Figure 1). Countries similar to the UK in demographics were included and the timeline included articles from 2000-2017. The selected articles were analysed using an appraisal tool adapted from the Critical Appraisal Skills Programme (CASP) (2016) and Crombie's (1996) checklists.
Key findings from the literature review are outlined in Table 1, which shows a significant weighting towards quantitative research (Bonari et al, 2005; Patel and Wisner, 2011; Stepanuk et al, 2013) with an additional mixed methods study (Walton et al, 2014). There was a lack of evidence around information provision and decision-making in the antenatal period in a UK setting. The findings suggested a need for clear risk-benefit information for antidepressant usage, and that women wanted collaborative discussions about their antidepressant treatment options. Two of the key themes from this literature review related to information provision and collaborating with clinicians on decision-making, and will be discussed further.
| Article | Design | Sample size and setting | Data collection | Analysis | Key results |
|---|---|---|---|---|---|
| Bonari et al (2005) | Cross-sectional descriptive survey | 100 pregnant women, 400 in total Canada | Telephone survey | Quantitative (n=100) Telephone survey conducted on initial call to help line, using an risk perception analog scale and the Centre for Epidimiologic Studies Depression scale (CES-D). Re-test performed 2 weeks later using risk perception analog scale |
|
| Patel and Wisner (2011) | Cross-sectional descriptive survey | 100 USA | Online survey | Quantitative (n=100) Questionnaires including 3 tools: Decisional Conflict Scale (DCS), Control Preferences Scale (CPS) and Problem Solving Decision Making Scale (PSDMS) |
|
| Stepanuk et al (2013) | Cross-sectional descriptive survey | 128 USA | Online survey | Quantitative (n =143) Questionnaires using 2 tools: Wittmann-Price revised Emancipated Decision-Making scale (EDM) and the Satisfaction With Decision scale (SWD) |
|
| Walton et al (2014) | Cross-sectional, mixed method | 40 Canada | Telephone survey | Qualitative (n=10) Semi-structured interviews to characterise barriers and facilitators to decision making |
|
Themes
Information provision
Information provision was cited as an important aspect of decision-making in all four studies reviewed. A mixed methods, cross-sectional study (n=50), by Walton et al (2014) looked at decisional conflict among pregnant women who were considering antidepressant usage. The quantitative aspect of the study focused on the Decisional Conflict Scale (DCS), a validated tool for studying complex decisions (O'Connor, 1995; Stacey et al, 2012), which provides a score indicating the level of decisional conflict experienced by the user. A DCS score of >37.5 indicated a high level of decisional conflict, a score of 25–37.5 indicated moderate decisional conflict and a score of <25 indicated low decisional conflict. It was found that women who were antidepressant users (n=21) had a much lower state of decisional conflict (DCS mean=17.5) in the ‘Informed’ subscale than women who were non-users (n=19; DCS mean=42.1). This indicated that women who were antidepressant users felt more informed than women who were non-users. The difference in scores was statistically significant (P=0.001); the low P-value indicating there was little chance for error (Crombie, 1996).
Semi-structured interviews with women (n=10) reported a lack of high-quality information (n=5) about risks to their baby (Walton et al, 2014). All cited a requirement to weigh their own health against the health of their baby as a barrier to decision-making. This is in contrast to the previous quantitative finding from this same study that women felt less decisional conflict when they were informed. The number of participants involved (n=10) could be seen as limited, although Guest et al (2006) suggested that overarching themes can emerge from as few as six interviews. The limited number does impact on the generalisability of the findings, which was acknowledged by the authors.
Information provision was also tested by Bonari et al (2005) in their structured telephone survey (n=100). They found a positive correlation (Pearson's r=0.391, P=<0.01) between information received and continued antidepressant usage. However, there was no rationale given for the chosen sample size (n=100); therefore the correlation should be considered as weak because it may be underpowered (Bowling and Ebrahaim, 2005).
In a further study (Stepanuk et al, 2013), information provision by health professionals was found to impact on the level of emancipation in the decision-making process, which was linked to satisfaction about the decision made. Satisfaction was not significantly different between women who chose to take or not take antidepressant medication. Stepanuk et al (2013) also explored empowerment in decision-making using a cross-sectional, descriptive, web-based survey based on Wittman-Price's (2004) theory of emancipated decision-making (n=143). A statistically significant relationship between personal knowledge and satisfaction with decision (P=<0.001) was found. Personal knowledge was defined as the information that women needed to weigh up the risks and benefits of taking antidepressant medication during pregnancy (Stepanuk et al, 2013). This finding can be treated as robust since the power calculation to define the sample size was set at 0.8, which is accepted as a good level in order to be able to detect statistically significant relationships (Sedgwick, 2011).
Collaborative decision-making
It was apparent that women wanted to be actively involved in collaborative decision-making about anti-depressant usage (Patel and Wisner, 2011; Stepanuk et al, 2013; Walton et al, 2014). Collaborative decision-making was tested in Patel and Wisner's (2013) study, in which pregnant women (n=100) completed a structured online survey using three validated tools: the DCS, the Control Preferences Scale (CPS) and the Problem Solving Decision Making Scale (PSDMS) (O'Connor, 1995; Deber et al, 1996; Degner et al, 1997). This survey split the decision-making process into two parts: problem-solving and decision-making. The study found that 72% of women wanted to defer; 24% wanted to share, and only 4% wanted to keep sole responsibility for solving the problem. In contrast, 41% wanted to share and 55% wanted sole responsibility for decision-making, while only 4% wanted to defer the process. One interpretation suggests that women see the clinician as playing a crucial role in helping to understand how to weigh risks and benefits and would like to take a passive role in this process. However, when making a decision, they preferred that clinicians took the role of facilitating a conversation, rather than taking the lead in the decision. Regardless of the interpretation, the findings demonstrated a clear desire from women surveyed to play an active role in decision-making around usage of antidepressant medication.
Strengths and limitations
There was an imbalance of available evidence to critique, since three studies were quantitative and one study used mixed methods. The balance within the mixed methods study reflected a weighting towards 40 quantitative responses and only 10 qualitative responses. Therefore, the perspective in this literature review reflects that imbalance, commenting on quantitative findings mostly and adding context and depth from the small amount of qualitative information available. All studies demonstrated a structured approach in their analysis and highlighted information about a little-known topic, namely, how women make decisions about the use of antidepressants in pregnancy. To the authors' knowledge, there are no comparable studies related to how women decide to use antidepressants in pregnancy in the UK, with all four of the studies reviewed being set in the USA or Canada. It was apparent that bias was identified in all four of the studies, as demographics suggested that the majority of participating women were described as white, married or living with a partner, or educated to a high school diploma level or higher. A comparison of the demographic details can be seen in Table 2.
| Source | n | Ethnicity | Education | Living arrangement | Age | |||
|---|---|---|---|---|---|---|---|---|
| White/Caucasian | Black, Asian, Pacific Islander, Other | Above high school level | To high school level | Married/living with partner | Single | Mean age | ||
| Patel and Wisner (2011) | 100 | 93 (93%) | 7 (7%) | 100 (100%) | 0 (0%) | 98 (98%) | 2 (2%) | 31 |
| Stapanuk et al (2013) | 143 | 136 (95%) | 7 (5%) | 127 (89%) | 16 (11%) | 130 (91%) | 12 (9%) | (100) 70% aged 25–34 |
| Walton et al (2014) | 40 | 31 (77.5%) born in Canada | 33 (82.5%) | 7 (17.5%) | 32 (80%) | 8 (20%) | 31 | |
| Bonari et al (2005) | 100 | No demographic information provided other than to highlight that there were no statistical differences between groups in maternal demographics, which included socioeconomic and educational status | ||||||
Discussion
This literature review has provided robust information showing that women are not empowered to make decisions about the use of antidepressants in pregnancy. Walton et al's results (2014) showed that women wanted clearer information from a trusted source, and identified a proven relationship between feeling informed and having less conflict when deciding to take antidepressants. However, Walton et al's (2014) finding that there was a significant difference in decisional conflict between antidepressant users and non-users differs from Stepanuk at al (2013), who found no statistically significant difference. Stepanuk et al (2013) did not provide any qualitative data, so it is not possible to look at the reason for this difference. As a result, there is a need to explore in depth how women decide to take antidepressant medication in pregnancy, especially in the UK.
Walton et al's (2014) research also suggests that many women believed information from a specialist mental health clinic to be more helpful to their decision-making process. Walton et al (2014) did not provide the number of women that reported this and so it should be treated with a degree of caution. However, in the UK there are gaps in mental health services, associated with lack of funding, shortage of appropriately qualified staff and poor resources (Hogg, 2013).
Corroboration that women wanted information provided by specially trained mental health professionals is evidenced by the British Psychological Society and the Royal College of Psychiatrists (2015) in their systematic review of qualitative research of UK women's experience of antenatal and postnatal care. The report suggested that there was a lack of information, meaning that many women stopped taking their medication after finding out that they were pregnant. This evidence formed the basis of the recommendation from the National Institute for Health and Care Excellence (NICE) (2014) that women should be provided with information on the risks and benefits of medication, the consequences of no treatment, and what might happen if medication is stopped. The guideline was UK-based, but focused on postnatal mental health only and was therefore excluded from this review. This finding highlights the lack of research situated in the antenatal period around how women are empowered to make decisions about using antidepressants. It is therefore easy to understand why women in the antenatal period find decision-making difficult (Walton et al, 2014).
Both Bonari et al (2005) and Walton et al (2014) found a relationship between information provided and continuation with antidepressant use. Despite the weakness of correlation and small sample sizes involved in the studies, the link between information received and the decision to take antidepressant medication should be acknowledged. This is particularly important in light of the known risks of untreated depression, such as reduced ability to form attachment to infants (Hayes et al, 2013), increased rates of premature birth, increased frequency of admission to neonatal intensive care (Engelstad et al, 2014) and suicide, one of the largest contributors to maternal death (Knight et al, 2016).
Providing clear information about antidepressant use in pregnancy is difficult for health professionals due to the complexity of evidence surrounding the issue (Einarson et al, 2015). Health professionals should therefore consider how information is communicated, as Stepanuk et al (2013) found that women were more satisfied with their decision when they felt more informed. Therefore, providing information in an accessible way that empowers women during the decision-making process makes a difference to their decisions and to their satisfaction; even if the woman opts for no medication.
The link that Patel and Wisner (2011) found between complex information and a desire to defer the problem-solving aspect of the decision-making process must also be acknowledged. Only 4% of women wanted sole responsibility for this aspect, which included weighing risk and considering treatment options. This indicates that women did not feel empowered at this stage of the process but did not want to relinquish the decision-making since 96% of women wanted shared (41%) or sole (55%) responsibility in actually making the decision. Active decision-making can empower women and influence positive treatment outcomes, from satisfaction with the decision made (Bruera et al, 2002; Kunneman et al, 2014) to positive health outcomes up to 3 years after treatment (Hack et al, 2006). Patel and Wisner (2011) recommended exploring women's expectations of the decision-making process, to improve communication and satisfaction.
A key finding within the studies reviewed focused on the limited demographics (Table 2). Patel and Wisner's study (2011) found that younger women surveyed had a higher score in the ‘Informed’ subscale of the Decisional Conflict Scale (P=0.02), and experienced more conflict about antidepressant use when they feel uninformed. This highlights the need for further research encompassing a much wider demographic. Younger women were also identified by de Jonge (2001) as needing specific and relevant information about mental health. There are also important considerations for those who identify as part of minority groups related to ethnicity, sexuality, learning disability or those who are less educated.
It is important to note that all women participating in the evidence presented in this review had actively sought further advice or information about their depression diagnosis (Bonari et al, 2005; Patel and Wisner, 2011; Stepanuk et al, 2013; Walton et al, 2014, suggesting that this self-help, coupled with the narrow demographic, might make the participants more motivated to gather information and share or take control of their situation. The general high level of education could also be said to contribute to a stronger need for detailed information. Further research with a wider demographic is needed to better understand this potential source of bias before a firm conclusions can be made.
Conclusion and recommendations
It was clear from this analysis that women in the USA and Canada were not empowered to make decisions about the use of antidepressants in pregnancy. Four studies identified information provision and a collaborative decision-making process as the two most important factors for women. Women requested clarity of information, which was undermined by complexity when interpreting research about risks and benefits of taking antidepressants during pregnancy. Contradictory information confused women and health professionals. Midwives should consider that, when women feel informed, there is a clear association with feeling satisfied with the decision made (Stepanuk et al, 2013); therefore the impact of unclear information about using antidepressants needs more in-depth qualitative research in the UK and globally to identify accessible information that women will find useful. There is a need to understand more about how women problem-solve around the uptake of antidepressants, as the topic is crucial to the health and wellbeing of women and their families when depression in pregnancy is experienced.
It was found that women wanted to actively participate in their treatment decisions (Patel and Wisner, 2011). Midwives need to have clear instructions about holding collaborative discussions with women in a sensitive and caring way, which could be taught in undergraduate and postgraduate health education. Gaps in providing appropriately trained health professionals in perinatal mental health services also need to be addressed to improve standards of care.