Pictorial representation of attachment: measuring the parent-fetus relationship in expectant mothers and fathers

Over the past decades, attachment research has largely focused on the quality of the attachment relationship that infants develop in the course of the first year of life with their parents, in particular the mother [1]. Also studies and knowledge about the representations of attachment relationships that adolescents or adults have of their own parents are increasing [2]. Far less is known about the development of feelings of attachment in parents towards their children. Klaus and Kennell [3] introduced the term ‘maternal bonding’ and defined the bond between mothers and their newborns as a biologically based emotional investment in the infant. Maternal emotional investment in the infant, however, does not start from birth onwards. It is recognized that this bond already starts to develop before birth [4–6] and that building a bond or ‘relationship’ with the unborn child is a key developmental task for a successful psychological adjustment of pregnant women [7].

Prenatal attachment can be described as the parent’s emotions, perceptions, and behaviors that are related to the fetus. These unique emotions and behaviors may be represented by an affiliation and interaction with the unborn baby and the desire of the parent to know and to be with the unborn baby [8]. Analysis of the literature suggests that prenatal attachment during pregnancy is not the same kind of attachment relationship as Bowlby [9] and Ainsworth et al. [10] have defined it, but a multi-faceted construct guided by the caregiving system [11]. Although attachment behaviors are essentially about eliciting care from others, the substance of maternal–fetal relationships deals with the development of feelings of parental love and protection. These relationships can be seen as a strong emotional tie. Therefore it seems reasonable to refer to them as attachment, as a generation of researchers has done before [4, 11–13]. The extent to which women feel attached and emotionally connected to the unborn child has been shown to be related to various pre- and postnatal parent and child outcomes, such as feelings of competence in infant feeding and care [14] parental mood [15] and the quality of mother’s interaction with the infant after birth [16].

Various self-report questionnaires have been developed to assess the quality of the mother-fetus relationship or antenatal attachment. The Maternal Fetal Attachment Scale (MFAS: [12], the Prenatal Attachment Inventory (PAI: [13]), and the Maternal Antenatal Attachment Scale (MAAS: [4]) are among the most frequently used scales which all tap overt (i.e., conscious) thoughts, feelings, attitudes and behaviors of the parents towards the fetus. All three questionnaires proved to be psychometrically sound instruments in Caucasian middle-class samples [17]. For research purposes -aimed at clarifying and identifying underlying mechanisms- these questionnaires were proven to be valid and reliable. For routine use in clinical settings, however, these measures are less feasible, because they require a certain reading level and more time to complete. In addition, it is unclear to what extent the answers are influenced by social desirability. Earlier studies revealed a tendency of pregnant women to respond in a socially desirable manner on questionnaires about mother-fetus relationships [18, 19]. In clinical settings there are several examples of simple, often non-verbal instruments, which can easily and quickly be applied to obtain information for the clinician, such as pain and wellbeing measures, and a pictorial measure of suffering [20]. We here propose an alternative, non-verbal tool that is fast and easy for expectant parents to complete, and which can be easily administered at antenatal routine visits by obstetricians and midwives.

The Pictorial Representation of Attachment Measure (PRAM: [21]) is a tool designed to measure parental representations of the (antenatal) parent-infant relationship in (expectant) fathers and mothers. It is inspired by the Pictorial Representation of Illness and Self Measure (PRISM) [20] a measure originally developed to assess the burden of suffering due to illness into a patient’s life, and the Inclusion of Other in the Self Scale (IOS: [22]) a single-item non-verbal measure for the structure of interpersonal closeness. The PRISM requires individuals or patients to put a disk representing their illness in relation to a disk representing the self. The distance between the self and the other (‘illness’) disk is measured and represents the impact of the illness on the self. Various studies have demonstrated the reliability, validity and clinical and research usefulness of the original PRISM or adapted versions of this tool in patients with cancer [23], dermatological problems [24], chronic pain [25], psoriasis, and rheumatic arthritis [20]. A recent study also showed the validity of the PRISM as a method for the assessment of suffering after trauma in PTSD patients [26]. In addition, one previous study has used this instrument to assess suffering in a non-physical-illness condition [27, 28]. Parents having experienced the loss of a premature child were asked to indicate the place of their lost child in their lives. The results of that study -using an adapted paper-and-pencil version of the PRISM- showed that a shorter distance between the ‘self’ circle and a cross representing the child, correlated significantly with greater grief due to loss of the child.

The PRAM also shows a similarity with pictorial measures designed to assess interpersonal relationships [23, 29] and attachment networks in adults [30], which have shown convergent validity. We here introduce the PRAM and expect it to be a useful tool to determine the place of the (unborn) child in the lives in (expectant) parents. A preliminary study with 76–80 participants showed promising results [21].

Until now, research on prenatal attachment has mainly focused on expectant mothers, rather than on fathers. Nevertheless, not only mothers but also fathers may feel more or less attached or connected to their unborn child [31], since for them also, pregnancy is a time of psychological preparation. Condon [32] described father-fetus bonding as a subjective feeling of love for the unborn child, which is at the heart of a man’s experience of early parenting.

Although one might assume that the way women and men interact with the unborn child may differ, evidence demonstrating gender differences in feelings of antenatal attachment is not consistent. Studies comparing maternal and paternal antenatal bonding show contradictory results with some studies showing mothers to have more and stronger feelings of bonding towards the fetus compared to their partners [33–35] whereas other studies failed to demonstrate such differences [36] or even yielded opposite findings [37]. Moreover, only few studies have evaluated both the feelings of antenatal attachment in women and their partner concurrently [34, 35].

Studies on prenatal attachment as experienced and perceived by primiparous and multiparous pregnant women and their partners have also yielded inconsistent results. Cranley [12] failed to find a relationship between prenatal attachment and parity, whereas Ferketich and Mercer [38] and Van Bussel, Spitz, and Demyttenaere [18] showed that multiparous women had lower attachment scores compared to primiparous women. Sorensen and Schuelke [39] further demonstrated that prenatal fantasies about the unborn child were more prevalent in primiparae than in multiparae. However, there is a lack of studies comparing parents expecting a first child compared to those already having a child with respect to the felt relationship with the fetus [40]. Finally, there is some evidence that demographic variables such as age of the parent show a negative relationship with feelings of attachment, although these findings are not consistent [12, 35, 41–43]. A recent study by McMahon et al. [44] showed that women having their first baby at an older age, appear to have some psychological advantages over younger women, because they are more resilient and better adjusted. This may result in less preoccupation with the fetus (i.e., as reflected in lower maternal-fetal attachment scores). Although the reasons for inconsistent results are not always clear, the fact that maternal antenatal attachment was measured during different time periods of pregnancy, with different instruments -each stressing slightly other aspects of the prenatal attachment relationship- in pregnant women whose ages varied across the different samples, may have played a role.

This study is the first to examine whether a newly developed non-verbal (pictorial) measure of parental representations of the antenatal relationship with the fetus relates to a widely used valid and reliable verbal self-report measure of parental representations of the antenatal relationship. If the PRAM is measuring feelings of bonding and connectedness, it should be associated significantly and meaningfully with self-report measures of antenatal attachment such as the Maternal Antenatal Attachment Scale (MAAS: [4]). The MAAS consists of items all focusing upon feelings, attitudes and behaviors towards the fetus, in contrast to scales such as the MFAS and PAI, which contain a number of items related to the pregnancy state and motherhood role. Moreover, Condon [32] also developed a paternal antenatal attachment scale (PAAS) similar to the MAAS. The PRAM specifically addresses the question “where do you place the baby in your life at this moment” and is therefore also specifically directed at the fetus per se. The present study focuses on (1) whether verbal and non-verbal (i.e., pictorial) representations of parent-fetus attachment are significantly correlated with each other and (2) whether gender of the parent, parity, and age are systematically connected with feelings of bonding or attachment as assessed by both the PRAM and verbal (MAAS and PAAS) measures.

The present study is part of a larger prospective longitudinal project called ‘Expectant Parents’ on prenatal risk factors and postnatal infant development [45]. More details about recruitment of the participants and the data collection used in this study have been reported earlier [45] and are briefly summarized here. At the first routine antenatal visit (between 9–15 weeks gestational age) 835 women were invited to participate by their midwives, who informed them about the project and handed out a leaflet with written information about the aim and design of the study. Parents were asked to participate in a longitudinal study, in which parents were followed from pregnancy (15 weeks gestational age) until 24 months postpartum. The aim of the larger prospective study was to gain more insight into feelings, emotions, behaviors and expectations about the pregnancy and the fetus, and about experiences with and development of the child after birth. The purpose is to investigate the quality of parent-infant relationships from parents’ perspectives, both in the prenatal and postpartum period. This was also communicated to parents [45]. For the purpose of the present study we informed parents that we aimed to gain more insight into feelings and emotions with the pregnancy and with the developing fetus. Women with a poor understanding of Dutch or English language and those expecting multiple births were excluded from participation.

Three-hundred fifty two pregnant women and their partners (n=268, 78%) agreed to participate and gave their written consent. The mean age of the women was 31.5 years (SD=4.4), ranging from 17 to 44 years. The mean age of the partner was 34.0 years (SD=4.6), ranging from 22 to 49 years. The mean gestational age was 26 weeks (SD = 1; Range 23–30 weeks). Most participants were white Caucasian women with the Dutch nationality (84%). The other participants have another Western (7%) or non-western (9%) ethnic background (e.g., Polish, Moroccan, Turkish, Surinam or Asian). Both women and men had completed an average of 16 years of formal education (SD=4.1). Fifty-three percent of the women (n= 187) and 53% of the men (n=143) were primiparous. The Medical Ethical Committee of the Elisabeth Hospital Tilburg approved the study protocol.

The present study examined whether a simple non-verbal (i.e., pictorial) measure of attachment (Pictorial Representation of Attachment Measure: PRAM) is a valid and feasible instrument to measure parental representations of the antenatal relationship with the fetus in a non-clinical sample. Moreover, we examined gender, parity, and age effects.

PRAM-SBD was found to be significantly related to a self-report questionnaire of antenatal attachment in both expectant mothers and fathers. Results showed significant correlations for mothers and fathers between PRAM-SBD and the reported feelings of attachment as assessed with antenatal attachment questionnaires [4]. Parents who placed the sticker of the Baby closer to the Self also reported more feelings of attachment and more preoccupation with the fetus on a self-report questionnaire. The PRAM-SBD was related to all three dimensions of antenatal attachment (i.e., Global Attachment, Quality of Attachment and Intensity of Preoccupation) to the same extent. This suggests that the PRAM-SBD may be a reflection of a general feeling of bonding and connectedness with the unborn child at that moment.

Interestingly, significant positive relations were found between women and their partners for all study variables. This may suggest that at different time periods of pregnancy when mothers feel less or more attached with the fetus their partners also show less or more attachment behaviors towards the fetus. However, when comparing the mean scores of the pregnant women and their partners, the maternal attachment scores on the questionnaire were significantly higher and the PRAM-SBD was smaller than the scores of their partners on the questionnaires and the reported paternal PRAM-SBD. This may be a result of women experiencing the physical signs of pregnancy more directly. Quickening (i.e., the moment that refers to the initial motion of the fetus in the uterus as it is perceived or felt by the pregnant woman) also enhances feelings of attachment towards the fetus and may be more eminent to the mother than to her partner. Condon [32] demonstrated that fathers and mothers are nearly equivalent in their thoughts and feelings about the unborn child, but that fathers are less likely to express these feelings in behavioral ways. The use of pictorial measures, like the PRAM, might have diminished this bias, but we also found a significant difference between mothers and fathers on the PRAM-SBD, also suggesting that at the end of the second trimester, fathers feel less attached, connected or bonded with the fetus than mothers. These differences between men and women may disappear at the end of the pregnancy in the third trimester [36], since a growing affection of both parents towards the unborn child during the course of pregnancy has been found in previous studies [49]. Empirical studies support this notion and have shown that prenatal attachment increases during pregnancy [18], although individual scores on prenatal attachment seem to be relatively stable from the first to the third trimester [19]. Especially in the second trimester, fetal growth is more observable and the child begins to feel more real to most women. Ultrasound technology has contributed significantly to recognizing the reality of the pregnancy and the developing fetus [50]. Especially for expectant fathers the effect of ultrasound was stronger for the recognition of the pregnancy and bonding than feeling the baby’s movements and the growing abdomen [51, 52]. Although differences between parents may disappear at the end of the pregnancy, Ustunsoz et al. [35] found similar significant gender differences in the third trimester suggesting that men remain less attached and oriented towards their fetus compared to women. Administering the PRAM at different time points during pregnancy may yield more insight into the development of bonding over time.

Parity was found significantly related to bonding as measured by the PRAM, MAAS and PAAS. Primi- and multiparae showed significant different scores on the MAAS and PAAS (sub)scales and also showed different PRAM-SBD scores. However, this latter applies only to women. The PRAM and questions -as included in the MAAS and PAAS- about how often parents think about the fetus and whether they feel affection, confidence, or annoyance by the fetus might be highly colored by postnatal experiences they have with other children and may be affected by having had these experiences before in a former pregnancy. Further studies, however, are needed to substantiate these assumptions.

We further found younger women showing stronger feelings of attachment towards the fetus than older women. This is in accordance with studies of Ustunsoz et al. [35], where younger parents showed more affect and were more preoccupied with the unborn child than older parents. Also a study by McMahon et al. [44], recently showed that women having their first baby when older, appear to have some psychological advantages over younger women, because they are more resilient and better adjusted. This may result in less preoccupation with their pregnancy and the fetus (i.e., as reflected in lower maternal-fetal attachment scores). However, according to meta-analytic studies [47], demographic predictors such as age -with low effect sizes in relation to antenatal attachment measures- are considered as factors that are less useful for strict theory building but merely need to be included in studies as potential confounding factors.

A limitation of this study is that we did not take factors such as mental state of the parents and social support into account, which have been found to affect attachment feelings. Alhusen [53], for example, found that having a supportive spouse positively affected feelings of prenatal attachment. Moreover, in the present study, only pregnant women without medical risk participated. When the pregnancy is high risk and mothers are referred to obstetricians, it may be more difficult for expectant women to cope with their pregnancy. Feelings of attachment towards the fetus may then be adversely affected, when being overwhelmed by one’s own problems [54]. Extension of the evaluation of this measure to clinical and high-risk samples is recommended and may further add to our knowledge of optimal bonding and the usefulness of PRAM-SBD for clinical purposes.

Furthermore, we assumed that a smaller PRAM-SBD is more positive and reflects more attachment. This is also confirmed by our findings. However, we do not know yet whether there is an optimal distance between Baby and Self. In other words, one could argue that the relationship between PRAM and attachment/bonding is not linear, but rather curvilinear. Although we generally expect larger distances to be related to less feelings of attachment or bonding, we need to be cautious by interpreting the results in such as way. One might argue that placing the sticker (i.e., the fetus) too close to the ‘Self’ with a minimal PRAM-SBD is not the same as feeling optimally attached. It rather might reflect strong preoccupation with the child. Because of a lack of norm-scores, we have to be careful with interpreting the PRAM-SBD. In a recent study with very preterm (born <31 weeks gestational age) and moderate preterm infants (born between 32–37 weeks gestational age) [55]), longitudinal analyses revealed that mothers’ PRAM-SBD scores decreased after moderately preterm delivery, whereas decreases in PRAM-SBD scores were observed in both parents after very preterm delivery. As lower PRAM-SBD scores represent stronger feelings of parent-infant connectedness, these findings suggest a higher degree of bonding after premature childbirth. These results were in line with outcome measures operationalized with other instruments, as parents of preterm infants reported less bonding problems on a questionnaire compared to parents of full-terms. These findings may support the hypothesis that in affluent countries with adequate resources, bonding in parents of preterm infants on average may be higher than in parents of full-term infants. Future studies with various clinical populations, in developing countries or in parents with very limited resources, as well as later follow-up measurements, are still needed to clarify the development and process of parent-infant bonding. Moreover, future studies are needed to gain more insight into optimal distances between Self and Baby.

Since it is not possible to conduct item analyses or inter-item consistency measures on a single-item scale [56], no reliability analyses have been performed. Nevertheless, this study showed evidence of convergent validity of the PRAM with other measures of prenatal attachment, even when the response mode varied considerably. Moreover, it could be argued that the significant positive correlation that we found between the PRAM-SBD and the MAAS or PAAS might be indicating the desire of some participants to respond in a socially desirable way. The range of scores (i.e., distance between self and baby), however, indicates that parents felt free to put the sticker close to themselves or at larger distance. Finally, we did not counterbalance the assessment of the M/PAAS and the PRAM and therefore were not able to control for order effects. Future studies using the PRAM with a focus on validity and reliability aspects of this instrument (i.e., discriminant and predictive validity, short-term test-retest reliability) are strongly recommended and are also in progress by our research group.

The clinical usefulness of an instrument further depends on its capacity to identify clinical cases, so that it can be used for screening purposes, to detect parents, for whom the bonding process may not adequately develop. In addition, sensitivity to change is another important asset of a clinical instrument. Can it be used to evaluate the effects of an intervention, aimed to stimulate the bonding process? A recent study by Wittmann et al. [26] showed a similar instrument to be sensitive to change in PTSD patients. We recommend further research in clinical groups.

The present findings provide support that the PRAM is as a valid, quick, and easy-to-administer instrument of parent-infant bonding. However, further research focusing on its capacity as a screening instrument (to identify parents with serious bonding problems) and its sensitivity to change (necessary for the use in evaluation of intervention studies) is needed, in order to prove its clinical value.

For now, the results showed that the PRAM can be used as an effective instrument of parent-infant bonding in research.

HvB: Department of TRANZO, Tilburg University, Tilburg, the Netherlands, Centre for Infant Mental Health, Dimence, Deventer, the Netherlands and Herlaarhof, Centre for Child and Adolescent Psychiatry, The Netherlands. AM, CV, Department of Developmental Psychology, Tilburg University, Tilburg, the Netherlands. AV: Department of Medical and Clinical Psychology, Tilburg University, Tilburg, the Netherlands.