Background

Being pregnant and giving birth are described as a transition phase, or an existential threshold that childbearing women have to cross [1]. Childbirth is an experience with many dimensions, multifaceted and unique for each woman, still strongly influenced by her social context [2]. Women’s expectations and experiences of pregnancy and birth are both positive and negative in nature, involving feelings of joy and faith but also worries, anxiety and fears. Despite the fact that maternity care in high income countries is safe, fear of childbirth is a common problem affecting women’s health and wellbeing before and during pregnancy, as well as after childbirth. Fear of childbirth has consequences for women’s relationships with their baby, partner and family [3], and often leads to requests for caesarean section (CS) by women striving for control in an exposed situation [4,5,6,7].

During the last few decades there has been a growing research interest in women’s fear of childbirth. For some women, the fear only relates to childbirth, but for others fear occurs in relation to other types of anxiety also [3, 8]. Fear of parturition is not new, and was described by the French psychiatrist Louis Victor Marcé (1858) [9]. The term ‘fear of childbirth’ (FOC) was characterised in 1981 in a population of Swedish pregnant women, defined as: “a strong anxiety which had impaired their [the women’s] daily functioning and wellbeing” [10, p. 265]. In addition, a more moderate fear was described as a significant anxiety, which did not interfere with the women’s daily life [10]. Later, during the 1990s, studies from Finland defined FOC as a health issue for a pregnant woman related to an anxiety disorder or a phobic fear including physical complications, nightmares and concentration problems, as well as demands for caesarean section [11]. The term ‘clinical FOC’ describes a “disabling fear that interferes with occupational and domestic functioning, as well as social activities and relationships”, and in some cases even reaches the classification for a specific phobia according to the DSM IV [3, p. 141]. The label “tokophobia” is also used [12, 13], characterised as an “unreasoning dread of childbirth” in women, a “specific and harrowing condition” [12, p. 83] including a “pathological dread” and “avoidance of childbirth” [13, p. 506]. Moreover, FOC is strongly related to the increasing caesarean section (CS) rates in Western countries, as being a common cause for women requesting a surgical birth [14, 15].

In early studies, the prevalence of FOC for pregnant women in Scandinavia was reported as 20%, with approximately 5–10% women experiencing intense fear [10]. The prevalence in Europe seems to vary between countries, from 1.9–14% [16], while Australia indicates higher rates of around 30% [17], leaving questions on possible cultural differences in FOC, or differing definitions. However, the variations in prevalence can also depend on the measures used; these can vary from fear being self-defined by women, or self-reported via different questionnaires, or estimated through measurement of physiological indices such as stress hormones in childbirth [3, 18, 19].

For women lacking experience of childbirth (so called primary tokophobia or FOC), their fears may date from adolescence or early adulthood, where experiences of others’ fearful responses to childbirth or a history of anxiety disorders could be important [8, 12, 13]. Secondary tokophobia or FOC is related to the event of birth, and is usually linked to fears developed after a previous negative or traumatic experience of childbirth, sometimes related to posttraumatic stress disorder (PTSD) [3, 8, 11,12,13]. Tokophobia is also described as a symptom of prenatal depression [12, 13]. However, the research on FOC has been criticised for constructing women’s fear as a medical category, having too pathological an approach that searches for errors in women, instead of examining possible causes of women’s fear within maternity care itself [20, 21].

To summarise, the research field is extensive, complex, and difficult to survey without any consensus on definitions. The concept “fear of childbirth” seems to be used as a broad label for all kinds of anxiety and fears that women experience in relation to pregnancy and childbirth. Relevant questions are: How common is FOC? Are there any cross cultural differences? Which measures are pertinent? What is FOC? There is a need for systematic reviews on FOC to be able to direct future research on developing optimal care and effective treatment for women fearing childbirth and to identify factors that reduce, as well as increase, women’s fears. Therefore, as a first step, we conducted a systematic review of all studies demonstrating a prevalence of FOC. The aims of this review were to identify the prevalence of FOC in childbearing women and how it is defined and measured during pregnancy and postpartum, and to search for useful measures of FOC, for research as well as for clinical settings.

Methods

Inclusion criteria

Types of participants

Participants were childbearing women (defined as the period covering pregnancy, labour and birth, and the first year postpartum).

Types of studies

Surveys, cross-sectional studies, experimental and quasi-experimental studies (where the control group could provide data), observational studies, systematic reviews, and meta-analyses, were eligible for inclusion.

Types of outcome

The primary outcome was prevalence of fear of childbirth, where this was defined clearly by study authors. Papers measuring fear during labour were excluded. Many studies used the same population (i.e., a number of PhD students accessed the same population at the same time and conducted different studies, but reported the same prevalence). We only included studies that were the first to report prevalence in a population and at similar time points. Studies reporting the same prevalence, on the same population or on a subsample (less representative, were excluded.

Search and selection strategy

A search strategy was developed and reviewed for accuracy, by one member not involved in its development (CS-L), using the Peer Review of Electronic Search Strategies (PRESS) criteria [22]. No restrictions were applied to years searched, but papers included were limited to English and Swedish publications only. We searched electronic bibliographic databases of The Cochrane Library, PubMed, Scopus, PsycINFO and CINAHL from their inception dates, in March 2015, using the agreed search strategy as described in Additional file 1.

Selection of studies

Studies were selected for inclusion from the papers identified, by team members working in pairs, using the above criteria. Any disagreements were resolved by a third member.

Quality assessment of included studies

The Effective Public Health Practice Project (EPHPP) quality assessment tool [23] was chosen to assess methodological quality of all studies included. Components of study design and methods assessed by this tool include selection and allocation bias, confounding, blinding, methods of data collection, withdrawals/drop-outs and analysis and intervention integrity. As some dimensions of the EPHPP (e.g. the sections on confounding and intervention integrity) are not relevant for reviews of non-intervention studies, the tool did have some shortcomings. For example, following discussion, it was agreed by the team to rate all cross-sectional studies (single cohort) as moderate, to avoid studies of otherwise good quality being excluded. Despite these problems, the tool was useful, especially for identifying ‘Weak’ experimental studies and excluding them. An overall quality rating was assigned to each study following assessment, of Strong (where no weak ratings were assigned), Moderate (one weak rating) or Weak (two or more weak ratings). An a priori decision was made that studies receiving a ‘Weak’ global rating score would be excluded from analysis. Team members in pairs assessed the quality of included studies. Any disagreements were discussed and resolved by consensus, or by a third member of the review team if necessary. In addition, as one of the team-members was co-author in some of the studies, those were assessed by other members of the review team.

Data extraction and analysis

Using a pre-designed data extraction form, data on prevalence, definitions and measurements were extracted independently by each member of the four review teams and checked for accuracy by the other reviewer.

Due to the differing types of studies, it was seldom possible to combine results into a meta-analysis. A narrative synthesis was provided instead.

Results

Results of search and selection strategy

In total, 18,464 citations were identified using the search strategy designed. After removing duplicates, 12,188 unique citations were screened by title and abstract and 11,698 excluded. Full-text papers of the remaining 490 citations were read and 354 of these were subsequently excluded, leaving 136 for inclusion (Fig. 1).

Methodological quality of included studies

The 136 papers were assessed and 76 were rated as “Weak”, and therefore excluded. The main reasons for “Weak” ratings were that the samples were unlikely to be representative of the population and the data collection tools were either not tested for validity, or there was insufficient information on their testing. This resulted in 60 papers meeting the inclusion criteria for this review (Fig. 1). At data extraction stage, however, it was clear that, for nine of the papers, no prevalence data could be identified [24,25,26,27,28,29,30,31,32], for six others, the whole sample of women had FOC [33,34,35,36,37,38], 11 papers were qualitative [39,40,41,42,43,44,45,46,47,48,49], and 10 papers had double reporting or reported the prevalence of a subsample of a population already included in the review [50,51,52,53,54,55,56,57,58,59]; these were excluded leaving a total of 24 papers for inclusion, based on data from 23 study populations (two papers [60, 61] presented FOC during pregnancy and postpartum, based on the same population) (Fig. 1 and Table 1). Thirteen of these had been rated as methodologically “Strong” and 11 were rated as “Moderate” (Table 1).

Table 1 Characteristics of included studies on Fear of Childbirth
Full size table

Description of studies

The studies had been conducted in twelve countries, with Sweden emerging as the country with the most research into FOC (Table 1). Ten papers (reporting on nine studies) were from Sweden (one of which also included a cohort of women from Australia), two from Norway, four from Finland, two from Denmark, two from the United States of America, and one each from Canada, Australia, Switzerland, and Croatia. One study had included data from six countries: Belgium, Norway, Iceland, Denmark, Estonia, and Sweden [62]. The majority (n = 20) had collected data using either postal or self-completed and personally returned surveys, one in the post-natal period, 16 in the antenatal period and six at both time periods. Two studies used telephone interviews, one in the third trimester of pregnancy and 1 month after birth [63] and the second in early and late pregnancy [64]. One study used a randomised controlled trial methodology and we used the data from the population screened by W-DEQ to identify FOC in early pregnancy, prior to trial entry [65]. One study [66] used a retrospective analysis of a national database to gather data on FOC, defined according to ICD code 099.80 (Table 1). Sample sizes ranged from 200 to 788,317.

Eight papers were published based on studies using the same population, with surveys administered at four time-points, and care has been taken to ensure that rates have not been double-counted. Of these eight papers we only included three [60, 61, 67]; the others were excluded due to repeated prevalence reporting or because they reported on a subsample of a previously reported population [50,51,52,53,54]. Hildingsson et al. 2010 [60] included 1212 women and presented the results for FOC in mid-pregnancy, and Nilsson et al. [61] used the same sample, but presented the results for FOC in women 1 year after birth (n = 763). Haines et al. [67] took a sub-set of 386 women from the full population, who had attended one regional hospital, and recruited 122 Australian women also. As a VAS was used for the response format, these results are presented separately and are not merged with any other.

Fig. 1
figure 1

Prisma flow chart Single file. Prisma flow-diagram of the search and selection process

Full size image

Definitions and measurements of FOC used by included studies

Eleven papers used the W-DEQ questionnaire, an instrument specifically designed to measure fear of labour and birth (Table 2). It consists of a 33-item questionnaire, and can be scored from 0 to 165 [68]. Various different cut-off points are used to define ‘severe fear of birth,’ from scores of ‘66 or greater’ to ‘greater than 100’, which makes comparison of prevalence difficult.

Table 2 Measurement tools used, prevalence, and parity in Fear of Childbirth studies
Full size table

The authors of two papers [64, 69] used a 3-point scale to measure fear/anxiety about birth, in answer to similar questions relating to FOC (Table 2). The three response options were variously expressed as: “no, I am not afraid/not at all;” “yes, I am a bit afraid/yes, a little;” “yes, I am very afraid/yes, a lot,” respectively, which were deemed suitable to merge (Table 2). Their definition of severe FOC was thus “yes, I am very afraid/yes, a lot.” Laursen et al. [64] measured FOC in both early and late pregnancy, but as the other authors, Geissbuehler and Eberhard [69], measured FOC in late pregnancy only, a combination of the late pregnancy rates was made.

A question with a four-point response scale was used in a study of a Swedish population, reported by Hildingsson et al. [60], and Nilsson et al. [61], with all their data coming from the same cohort of women and using similar questions relating to FOC; “to what extent do you experience worries and fear?”, With the addition of “when thinking of coming births” in the questionnaire 1 year after birth (Table 2). The authors dichotomised the scale into ‘no fear’ and ‘fear’. ‘No fear’ was described variously as ‘not at all + very little’ and ‘not at all + somewhat’, and ‘fear’ was described as ‘a lot + very much’ and ‘a great deal + very much’ (Table 2).

The remaining papers used a heterogeneous mix of various scales for measurement, none of which could be combined. Elvander et al. [63] graded fear on a 5-point scale, where women answered 6 questions on feeling: ‘nervous,’ ‘worried,’ ‘fearful,’ ‘relaxed,’ ‘terrified,’ and ‘calm’, in relation to their impending birth. Scores were divided into 3 categories: 6–13 (low fear), 14–20 (intermediate fear) and 21–30 (high fear, their definition of FOC). Waldenström et al. [70] also assessed FOC on a 5-point rating scale using the single question: “How do you feel when thinking about labour and birth?” Women ticking the “very negative” response alternative were defined as having childbirth related fear. Eriksson et al. [71] graded fear on a 6-point scale, from “no fear at all” to “very high fear.” Intense fear was defined as four or above and agreement with the statement that “childbirth-related fear influences your daily life in a negative sense” (Table 2).

Two studies [67, 72] used the Fear of Birth (FOBS) visual analogue scale but with different cut-off points (>50 and ≥60), so results could not be merged. The FOBS scale consists of one question – “How do you feel right now about the approaching birth?”- graded by marking two 100 mm visual analogue scales, which are anchored with the words “calm/worried” and “no fear/strong fear” [67].

Fabian et al. [73] asked women to answer one question on whether they had attended, or felt they needed to attend, a clinic for counseling in relation to FOC, to which they answered ‘yes’ or ‘no’. Lowe et al. [74] used a 15 item Childbirth Attitudes Questionnaire (high fear = 1 Standard Deviation above mean) and Poikkeus et al. [75] used a revised version of the Fear-of-Childbirth questionnaire, developed by Saisto et al. [76], and pregnancy anxiety scale. Total scores equal to or higher than the 90th percentile in the revised Fear-of-Childbirth questionnaire (total scores ≥6 and pregnancy anxiety scale, total scores ≥30) were considered “severe fear” and “severe pregnancy related anxiety” (Table 2). None of these results could be combined.

The majority of studies took measurements at only one time point or at varying times throughout pregnancy without comparing rates from different times. Laursen et al. [64] found rates in early pregnancy to be 2308 out of 30,480 (7.6%), similar to rates in late pregnancy, of 2245 out of 30,480 (7.4%). Postpartum rates of FOC were measured in only two studies [61, 77]. Due to differing measurement times and tools, no comparisons could be made but postpartum rates did not seem to differ greatly from those in the antenatal period (Table 2).

Prevalence

The prevalence of FOC varied, which may be due, in part, to the differing measurement scales. The most common scale was the W-DEQ, used in 11 studies, with four different cut-off points (some papers used more than one point). Three studies [65, 78, 79] used a cut-off point of greater than or equal to 100, with rates varying from 5.5 to 8.1%, giving an average ‘very severe’ FOC rate of 7.1% (536 out of 7531). A cut-off of greater than or equal to 85 in seven studies [62, 79,80,81,82,83,84] two cut-off levels gave FOC rates of 7.5% (165 out of 2206) to 15.5% (254 out of 1635), giving an average ‘severe’ rate of FOC of 11.1% (1545 out of 14,163) (Table 2). The final two studies using W-DEQ had cut-off points of greater than or equal to 66 [85] or 71 [77], which gave rates of ‘moderate’ FOC of 24.9 - 26.2% (Table 2).

The two studies using a 3-point scale [64, 69] had merged “extreme fear” FOC rates of 7.0% (2707 out of 38,801). The data collection using the dichotomised 4-point scale had rates (based on the same population) of 14% [60], in mid pregnancy, and 15.1% 1 year after birth [61] (Table 2).

Rates for the studies using 5- and 6-point scales [70, 71] varied considerably from 3.6 to 22.9%, and the 5-point scale was based on the question “How do you feel when thinking about labour and birth?” with the response “very negative” deemed to equate to FOC [70], which may not be accurate. Rates for all other prospective studies varied from 11 to 31.1%; the register study based on data from medical records [66] showed a rate of 28,960 out of 788,317 (3.7%) (Table 2).

Comparison of country rates

Rates of severe FOC in each country (as measured by seven studies using W-DEQ with ≥85 cut-off point) were combined, the average taken for each country, and then compared. The average rates varied from 6.3% in Belgium to 14.8% in Estonia, a significant difference in the seven countries (chi-square = 104.44, d.f. = 6, p < 0.0001) (Table 3).

Table 3 Prevalence of FOC in different countries, measured by W-DEQ (cut-off ≥85)
Full size table

Discussion

W-DEQ gave an average ‘very severe’ (greater than or equal to 100) FOC rate of 7.1%, severe (greater than or equal to 85) rate of 11.1%, and a ‘moderate’ rate of 25.3 - 26.2% (greater than or equal to 66 or 71). Merged rates from two studies [64, 69] using a question with a 3-point response scale and similar questions on fear (Table 2) had an average “extreme fear” rate in early pregnancy of 7%, very similar to the “very severe” rates measured by W-DEQ. The study using a similar question with a dichotomised 4-point Likert response scale (Table 2) had a rate in early pregnancy of 14.0% for fear or strong fear of childbirth [60].

Once W-DEQ cut-off points went below 85, or other scales had more than three (or four) points, the rates increased and also varied considerably from the average seen at the cut-off ≥85. This might indicate that using the W-DEQ tool with a cut-off point of ≥85, or a measurement of FOC using a single question such as ‘Are you afraid about the birth?’, with three responses such as ‘no, I am not afraid; yes, I am a bit afraid; yes, I am very afraid’ [69] would measure FOC equally but that needs further testing.

Rouhe et al. [86] have previously compared FOC levels in 1348 women using the 33-item W-DEQ scale (with a cut-off point of ≥85) with a one-item VAS scale, and a cut-off point of 5. Although there was some correlation they found it not to be as accurate as the W-DEQ, but pronounced it suitable for initial screening. Haines et al. [87] also compared the W-DEQ scale (using a cut-off point of ≥85) with the Fear of Birth scale (FOBS), a two-item VAS scale, involving 1410 women. Results showed a strong correlation. However, the FOBS’s cut-off point was 54, approximately equivalent to the cut-off of 5 used in their work examined here [67], which resulted in quite high FOC rates.

Using the question with a four-point Likert response scale appeared to over-estimate FOC rates, especially when results were dichotomised so that ‘extreme fear’ was merged with more moderate fear. Dichotomising results before analysis appears to be common, but ultimately does not produce useful results if a true measure of severe FOC is the main aim.

Rates of severe FOC, measured in the same way, varied in different countries from 6.3 to 14.8%. These results are new as, although one study in this review involved measuring FOC in six countries, no paper has analysed data from all seven countries where FOC has been measured. Reasons why FOC might differ in different countries are unknown, and further research in this field is required. One reason may be poor translation, or insufficient testing of the translated version of W-DEQ, both problems highlighted by previous authors [77, 88]. Another possible explanation is that some factors may remain unidentified when measuring with W-DEQ. For instance most scales measuring fear of childbirth do not consider important dimensions such as fear of abandonment by staff during birth [89], fear of medical interventions, loss of autonomy and control, as well as fear of mistreatment and obstetrical violence [90]. In addition, the W-DEQ scale assesses a range of emotions about labour and birth, where fear is only one emotion among many others. However, despite its shortcomings, the W-DEQ has been highlighted in a recent systematic review on validated instruments used for measuring women’s childbirth experiences as, currently, the best, most used and validated tool to measure FOC [91]. Culture-specific aspects in relation to fear of childbirth have been recognised in medicalised birth cultures, where young adults prefer CS over vaginal birth, and negative impressions of birth through visual media can be an important factor for generating fear [92, 93]. Moreover, traditions surrounding birth, women’s rights, how antenatal and maternity care is organised, CS rates, and which professions (midwives, GPs, obstetricians) are involved in pregnant and childbearing women’s care, could all influence women’s fear of childbirth.

As FOC has been shown in a large systematic review and meta-analysis to be strongly associated with post-traumatic stress disorder [94], simple and early diagnosis and intervention for women with severe FOC is recommended. Antenatal education, a relatively cheap and cost-effective intervention has been shown to decrease fear of childbirth [95, 96], as has cognitive behavioural therapy, although the sample size was small [97]. We therefore also examined the published tools to see which might be most useful in the clinical area. As shorter tools have greater clinical utility than the W-DEQ, a comparison between W-DEQ and a measurement of FOC using a single question with three or four responses (that are not dichotomised before analysis) would be useful in making decisions as to how best to measure FOC swiftly and accurately in the clinical location. A comparison between W-DEQ and the FOBS scale, possibly after further testing using a higher cut-off point, would also be very useful to the research and maternity care communities.

Conclusions

Rates of severe FOC, measured in the same way, varied in different countries. Reasons why FOC might differ are unknown, and further research is necessary. Using the W-DEQ tool with a cut-off point of ≥85, or a measurement tool using a single question with three responses, are consistent in measuring levels of severe FOC. Research comparing W-DEQ and a measurement tool using a single question with three responses, or four responses that are not dichotomised before analysis, would be useful to both the research and clinical communities. Similarly, continued and further testing of the FOBS scale, especially using a higher cut-off point to separate out “severe” FOC from more moderate levels, could prove beneficial to clinicians.

Validation of a simpler tool like the FOBS or a single question is required as there is a need for a tool that is easy and quick for women to fill in, as not all clinicians have time to administer the longer W-DEQ. Newly-developed, untested scales cannot easily be compared with other studies, and should not be used in clinical practice without further testing.

We recommend that future studies on FOC should use either the W-DEQ tool with a cut-off point of ≥85, or a more thoroughly tested version of either the FOBS scale with a higher cut-off point, or a single question such as ‘Are you afraid about the birth?’ with a three- or un-dichotimised four-point Likert response scale. In this way, valid comparisons can be made between countries and other studies.

Further research is also needed into reasons why FOC might differ in different countries and whether care for women with FOC needs to be made culturally specific. Measurement of FOC needs to include aspects such as fear of abandonment by staff during birth, fear of medical interventions, loss of autonomy and control, as well as fear of mistreatment and obstetrical violence. This more focused research agenda to guide future studies will result in more meaningful results that can be used to improve care provided for all women with fear of childbirth.