Mothers’ narratives about having a child with cancer: A view through the attachment lens

Childhood cancer

Based on figures from 2006 to 2008, an average of 1,550 children were diagnosed with cancer each year in the United Kingdom, a figure which includes non-malignant brain tumours (Cancer Research UK, 2011). Although childhood cancer is not common, receiving a diagnosis is serious for families (Björk, Wiebe, & Hallström, 2009). Fortunately, a cancer diagnosis no longer represents a death sentence for a child. Mortality rates for childhood cancer have significantly decreased, falling by over half since the 1960s in Great Britain (Cancer Research UK, 2009). Despite these improvements, cancer remains the leading cause of death from disease in children in the age group of 1–14 years in the United Kingdom and accounts for approximately a fifth of all deaths in this age group (Cancer Research UK, 2012).

The onset of cancer might be sudden and unexpected as, in contrast to adults, childhood cancers tend to have short-latency periods and often grow rapidly (Dang-Tan & Franco, 2007). The treatment of cancer usually involves a series of intensive and invasive treatments over a period of up to several years (Schweitzer, Griffiths, & Yates, 2012). Childhood cancer, therefore, remains an emotionally traumatising experience for parents (Da Silva, Jacob, & Nascimento, 2010). Children are embedded within their family system, and parental psychological and family functioning are viewed as critical factors influencing the interaction between the cancer course and the child’s developmental process and subsequently the child’s physical and psychological functioning (Pai et al., 2007). Consideration of attachment theory is relevant due to its contributions to our understandings of relationships, how information about dangers and threats are recognised and processed, and the strategies used for protecting oneself and one’s child from such dangers (Crittenden & Landini, 2011). Despite its relevance, this is an area that has not yet been researched.

Attachment theory

Attachment theory is not just relevant to childhood but across the lifespan (Crowell, Fraley, & Shaver, 1999). John Bowlby argued that a child forming an attachment to his or her parent, or other caregiver, is a fundamental and biologically based inevitability affording protection from predators (Bowlby, 1969). When threatened, an infant will engage in various forms of behaviour to obtain and maintain proximity to his or her caregiver. How the caregiver responds to this is key in shaping and developing the child’s attachment system and future attachment-related behaviours. It is suggested that children construct internal working models from these interactions, which can be understood as being similar to a computer simulation program that uses stored information about experiences to interpret and make predictions about the child’s own and caregiver’s attachment behaviour (Dallos, 2006). Into adulthood, these models become increasingly complex and use the apparatus of perception, memory and affect (Steele & Steele, 2008).

It is argued that while the attachment system is not directly observable, the working models of attachment can be revealed by activating the attachment system and then viewing through the lens of a particular assessment (George & West, 1999). A semi-structured interview known as the Adult Attachment Interview (AAI) was designed for this purpose (George, Kaplan, & Main, 1985). The AAI not only accesses salient content but facilitates analysis of how the individual speaks about it, as this gives insight into the processes of the working model. In other words, the AAI assesses how able the speaker is to access and evaluate attachment-related memories while remaining coherent, collaborative and plausible (Hesse, 1999).

The AAI has been developed further in recent years as the Dynamic-Maturational Model of discourse analysis (DMM-AAI; Crittenden, 2007). The DMM-AAI uses a modified version of the AAI protocol to explore adult concerns relating to danger, reproduction and threat to offspring. Analysis within the DMM-AAI continues to focus on how individuals speak about their experiences but frames them as functional mental processes rather than language-specific discourse markers. The DMM-AAI suggests a number of classifications organised into three major patterns of attachment: Type A, Type C and Type B (Crittenden & Landini, 2011). Type A patterns are represented by a dismissal of the perspectives, intentions and feelings of the self and a preoccupation with those of others. Cognitive information about the consequences of previous behaviour is used to understand others’ perspectives. Type C patterns are shown by a preoccupation with the perspective and justification of the self and a dismissal of others as valued people and sources of valid information. Subjective feelings or affect are used as a source of information. Type B patterns are understood as a balanced integration of affect and cognition. The continued reliance on analysis of how a person speaks about their experiences in the DMM-AAI means that a person’s attachment experiences may be expected to influence their construction of narratives when their attachment system is activated.

The importance of narratives

Narrative theory argues that stories are fundamental to the human experience and we live our lives through the creation and exchange of stories (Murray, 2008). Through narratives, we can bring a sense of order to a seemingly disordered world, and define ourselves as having temporal continuity and as being distinct from others (Murray, 2008). When individuals experience threats or other adversities, they often feel compelled to re-examine and re-fashion their narratives about themselves to maintain their sense of identity (Bury, 2001). Narratives have thus often been considered in relation to illness and how this is incorporated into a person’s biographical framework (e.g. Frank, 1995; Hyden, 1997).

A number of studies have used qualitative methods to explore the experiences of parents of a child with cancer. These have found, for example, that when their child was first diagnosed, parents talked about feeling vulnerable, not in control, aware of their child’s mortality and that everything ‘normal’ had disappeared (Björk, Wiebe, & Hallström, 2005; Schweitzer et al., 2012). During treatment, the focus was on changed relationships, the practical demands of treatment and trying to maintain hope, gain control and create a new normality (Björk et al., 2005; Björk et al., 2009; Clarke-Steffen, 1997; Schweitzer et al., 2012; Young, Dixon-Woods, Findlay, & Heney, 2002). Parents were often compelled to re-evaluate their priorities and the way they lived their lives (Schweitzer et al., 2012). After treatment had been completed, parents felt their lives were forever changed in how they saw the world and how safe they felt, and thus the construction of a new normality would be required (Woodgate, 2006; Woodgate & Degner, 2002).

Returning specifically to the topic of narratives, it is argued that narratives not only mirror ways of seeing the world but are active constructions that shape our world and how we live our lives (Murray, 2008). In ontological terms, they can be defined as existing and giving an insight into these individual and group processes. While attachment theory and narrative theory have historically distinct origins, it can be argued that both approaches draw on forms of discursive analysis to explore how individuals represent threatening and dangerous episodes in their lives (Dallos, 2006). Attachment theory typically pays greater attention to unconscious and defended emotional processes in contrast to the emphasis placed on performative aspects of self-presentation in narrative approaches.

Attachment theory and narratives about cancer

Although some researchers would disagree that attachment patterns influence discourse outside of an AAI interview (Hughes, Hardy, & Kendrick, 2000), it seems reasonable to consider that there may be an influence on narratives that are generated about events considered to be an attachment ‘threat’. Attachment in adults can be extended to relationships, information processing and strategies for protecting both the self and one’s offspring (Crittenden & Landini, 2011). It has been suggested that when parents receive a diagnosis of cancer for their child, doubts about survival are topmost in their minds and many parents might exhibit signs of anticipatory grief (Al-Gamal & Long, 2010; McCubbin, Balling, Possin, Frierdich, & Bryne, 2002). It therefore seems reasonable that childhood cancer can be considered a ‘threat’ to the attachment relationship between child and parent.

The primary aim of this study is to explore how a mother’s pattern of attachment influences her experience of having a child with cancer and how this is processed. This will be investigated through the exploration of mothers’ narratives about their experience and related to their patterns of attachment assessed using the DMM-AAI. The secondary aim is to investigate whether narrative methodologies can be used to provide information about attachment-related strategies used by mothers when faced with a life-threatening illness in their child.

Participants and recruitment

Six mothers in the age group of 36–53 years agreed to take part in this study. At the time of the study, the children were in the age group of 8–21 years, though all had been under 18 when diagnosed with cancer. Three participants were girls and three were boys. One child had received a non-cancerous diagnosis of a serious blood disorder; however, since this illness was also potentially life-threatening and was treated with medical protocols common to cancer, this mother’s experience was considered equally relevant for the purposes of this study.

Participants were recruited through the Paediatric Oncology Team at a large district general hospital. Parents were considered eligible for the study if their child had completed active treatment within the previous 2 years and survived. The Clinical Psychologist for the team exercised clinical judgement regarding parents’ capacity to meet the task demands of the study and to deem them at low risk of becoming excessively negatively affected through participation. All parents who met these criteria were contacted by the Clinical Psychologist. Those who expressed an interest in the study were sent an information sheet by S.B. Meetings were then arranged for those wishing to take part. Only mothers agreed to participate.

Design and procedure

The study comprised two interviews. The first of these explored the mother’s narrative about her experience of having a child with cancer. The second used the DMM adapted AAI protocol (Crittenden, 2007). The interviews were completed in one or two meetings depending on the preference of the participant. The Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) was given prior to commencing the first interview both to provide a descriptive measure and to facilitate discussion with participants about their current emotional state and how taking part in the study might impact on them. During the interviews, if participants displayed undue distress, the researcher did not continue further. Basic demographic information was also gathered.

All participants were interviewed alone at a private venue of their choice. Each participant gave informed consent to take part in the research. The narrative interview began with completion of a family tree to provide contextual details. Participants were then asked an initial open question designed to elicit a narrative of their experience – ‘Please tell me as much as you can about your experience of [child’s name] having cancer, his/her treatment and how things are now. Begin wherever feels right to you’. Once participants reached a natural conclusion, they were prompted for more detail in any areas they had not fully discussed. These prompts included the various time periods of the illness, how the child and other family members had experienced events, emotional experiences of themselves and others and the impact on relationships.

Analysis

The DMM-AAI interview protocol has an associated coding method. This was completed by R.D. who is a trained and experienced administrator. The DMM-AAI is recognised as a well-established method of assessing patterns of attachment in adults. For a discussion of its validity, see Crittenden and Landini (2011).

The narrative analysis began with multiple readings of the transcripts. Each transcript was coded for potential themes. These initial codes were grouped together with similar ones within the same transcript to identify narrative themes for each person. The narratives were then each considered in terms of their overall structures and performative aspects (Riessman, 2008). Once these analyses had been completed for each individual transcript, the transcripts were grouped together according to classification of attachment patterns on the DMM-AAI to consider within-group similarities and between-group differences with each of the analyses. As a final layer of analysis, the narratives were explored to see whether there was evidence in the transcripts of attachment-related strategies.

It is acknowledged that the narrative analysis process is not passive and that researchers will bring certain assumptions and beliefs to the analysis (Murray, 2008). For this reason, S.B. used ‘bracketing interviews’ at key stages in the research process to explore and bring to attention some of these assumptions and beliefs (Rolls & Relf, 2006). To increase reliability, a complete narrative transcript was reviewed by an independent researcher in terms of the narrative analyses detailed above. A narrative transcript was also reviewed by R.D. in terms of identifying attachment strategies. These reviews broadly concurred with the original analysis.

Ethical considerations

It is important that research asking participants to discuss traumatic events in their lives is conducted with careful consideration and sensitivity. Ethical approval for this study was granted by the National Research Ethics Service (NRES) Committee South West–Central Bristol. Following each interview, S.B. debriefed participants to ascertain whether they required any follow-up support.

HADS

Only one participant, Nicola, scored highly on this measure. Her answers were discussed to ensure she felt emotionally able to participate in the research.

DMM-AAI classifications

Table 1 shows the distribution of major attachment pattern classifications on the DMM-AAI for the sample. Tracey moved between Types A and C strategies within her DMM-AAI transcript. In contrast her narrative transcript suggested that she was predominantly using Type A strategies in relation to the threat of her child’s illness. She was therefore grouped with the mothers classified as Type A.

OPEN IN VIEWER

Table 1.

DMM-AAI classifications.

	Classification
Gemma	Type B – Earneda,b
Beverly	Type B
Emma	Type A
Nicola	Type A b
Naomi	Type A
Tracey	Mixed, Type A and Type C

DMM-AAI: Dynamic-Maturational Model of discourse analysis–Adult Attachment Interview.

a

‘Earned’ – patterns of attachment reorganised to Type B due to personal effort in adulthood to understand developmental process of the self and parents.

b

Classified as additionally having unresolved loss or trauma from childhood.

Thematic narrative analysis – Type B

There were two superordinate themes particular to this group: ‘Protecting my child’ and ‘Questioning and evaluating the experience’.

Performative narrative analysis – Type B

In performative terms, Gemma and Beverly positioned themselves in their narratives as capable and assertive women and active agents of change within their lives.They made frequent use of direct speech in their narratives. In narrative analysis terms, direct speech can be used to build credibility of what is being said and pull the listener into the narrated moment (Riessman, 2008). With these transcripts, it also appeared to bring depth and richness to the characters within the story. Both used direct speech to report conversations with their daughters.

Because she used to be quite argumentative, you know, ‘I’ve not got a temperature, I’ve not got’. And she was sat there shaking with rigours. I said, ‘yes you have, you’ve got to go into the hospital’. And she didn’t want to go in the hospital and I felt like the big bad wolf. You know, ‘yes I’m taking you, you’ve gotta go’. ‘Oh but I wanted to do so and so’. ‘No you’re not, you’re going back up the hospital’.

I shouted at her for some reason or other and she said ‘oh thank goodness for that’, and I said ‘what do you mean?’, she said ‘well you’ve been treating me funny since I’ve been diagnosed’ and I said ‘no I haven’t’. And I suppose I had, I’d been tip toeing around her you know. And I said ‘no you make it like I’m sort of shouting at you all the time’ and she said ‘no but it just feels normal if you tell me off every now and again’. So I did tip toe round her for a bit.

These extracts suggest that both mothers were able to hold their child’s experience in mind throughout and consider how they might have thought and felt about events.

Thematic narrative analysis – Type A

Transcripts of two mothers, Tracey and Emma, were chosen to provide illustrative examples representative of the group. There were two superordinate narrative themes for this group: ‘Being strong and positive’ and ‘Playing down feelings’.

Performative narrative analysis – Type A

In performative terms, Emma, in particular, presented herself passively in the story. Action was taken by others such as the medical professionals or her husband. Tracey presented an increasingly active character although she described action as being carried out jointly by herself and her husband: we said ‘hang on a minute, stop’.

In the interview, Emma was concerned about whether she was answering appropriately and queried this several times: Is that a good answer? I don’t know if I’m answering the way you want me to or if I’m trailing off on to things I shouldn’t talk about?

In contrast to Beverly and Gemma, the characters in the narratives of Tracey and Emma, including themselves and their children, had less depth and richness. There was also less detail about how the experience might have been for others, such as their child.

Unresolved loss or trauma from childhood

Gemma and Nicola showed evidence in their DMM-AAI of having some loss or trauma from their childhood that they had not fully resolved. Their narratives showed particular difficulties with story construction that were not present for those without the unresolved loss or trauma.

For Gemma, this effect was less pronounced. The beginning of her narrative was well structured and ordered in time and place; however, she lost this structure at an emotionally salient place when discussing whether she should have immunised her daughter. She stopped talking and was prompted by the interviewer to continue with her story. Following this prompt, she continued to talk but although the rest of the narrative was coherent, it moved between memorable events that were no longer temporally ordered:

What else sticks out?

She managed to recover towards the end of her narrative and finished with a paragraph reflecting on her experience and bringing it back to the present:

So all in all, I look back now and I think ‘she was really fortunate’. She wasn’t fortunate to get cancer but actually, when I look at families that I’m dealing with now, she went through really well, she was well most of the time, she didn’t end up in the hospital.

The difficulties in constructing a narrative were considerably more pronounced in Nicola’s narrative. In response to the initial open question, she gave a few sentences about her daughter being diagnosed and having treatment, and then prematurely brought the account to an inconclusive end:

I can’t think what else to say really, it’s not very often I talk about it.

She was given another open prompt encouraging her to construct a narrative. She gave more brief details about her daughter’s treatment, and then brought it to a premature end again:

What else can I say really? I don’t know what else to say really. What else would you like to know?

The interviewer attempted to scaffold the narrative construction by giving a structure of time periods within the story. Nicola continued to struggle with constructing a story that was ordered in time and place. The remaining interview comprised Nicola answering specific questions asked.

Attachment lens

An attachment lens was taken to the narratives to examine whether attachment-related processes could be seen in action.

The narratives of Gemma and Beverly suggested they were using strategies consistent with a Type B pattern of attachment. Their narratives were relatively coherent and well structured. Beverly’s narrative was particularly well structured, ordered in time and place and easy to follow. Gemma’s narrative lost temporal ordering part way through as discussed above.

Gemma and Beverly discussed their experiences in emotionally open terms, describing and labelling the emotions of themselves and others. Difficult experiences were acknowledged and talked about in a realistic and non-minimised way and were balanced against positive experiences. Reflective functioning was evident in both narratives, demonstrating their capacity to understand other’s behaviours in terms of their thoughts and feelings. They both showed this in relation to their daughters. For example, Gemma discussed a difficult experience her daughter had at a party:

She had a really hard time at the party and she ended up under the bed crying at this party . . . And I think that was just the after effects of the blimming steroids and just feeling different from everybody, not feeling like she fitted in.

The narratives of Tracey and Emma, in contrast, showed strategies consistent with a Type A pattern of attachment. They tended to be emotionally closed in their narratives, showing little acknowledgement or labelling of their own or others’ emotions. This was evident despite Emma presenting during interview as tearful. They frequently minimised the severity of distressing subjects, used distancing language and quickly moved the narrative on:

At that time, we didn’t even know if it was survivable with Jonathan. I mean you have the worst fears ever. But we did have a lovely guy, he was our Consultant up there. He was brilliant.

As soon as they said it’s a form of cancer, that was the worst sort of, you know, the worst news ever. So that was how we first found out about everything and we were like, ‘oh right ok how do we deal with this’.

They also showed little evidence of reflective functioning even when directly asked about others’ experiences. It has been suggested that people with Type A patterns may struggle to recall many details of their experience, but what is recalled, will be presented coherently (Crittenden & Landini, 2011). Both narratives were generally coherent and temporally ordered; however, Emma was frequently vague and reported not remembering various details. It should be noted that the attachment strategies discussed appeared to be used more at particularly emotionally salient junctures of the narratives.

Main findings

As stated above, it was beyond the scope of this article to include all the narrative themes; however, the full body of themes was similar to those found by previous researchers (e.g. Schweitzer et al., 2012).

The narrative themes of those classified as Type B, balanced integration of affect and cognition, were ‘Protecting my child’ and ‘Questioning and evaluating the experience’. With an attachment lens ‘Protecting my child’ could be understood in terms of the ability of the mother to not only hold her child in mind but also to be able to understand the thoughts and feelings that her child might be experiencing. Through this she is able to think about her child’s needs and act accordingly. This is supported by the strong presence and identity that the child has in both narratives. The narrative theme of ‘questioning and evaluating the experience’ could be thought about with relation to how able the mother is to hold in mind and acknowledge both the positive and negative aspects of their experience. Being able to do this in a realistic and non-minimising way could be considered an essential part of evaluating the experience. This is consistent with their Type B classification.

The narratives of those classified as Type A, or showing predominantly Type A strategies in their narrative, showed different characteristics. The narrative themes particular to this group were ‘being strong and positive’ and ‘playing down feelings’. A key, and potentially adaptive, strategy used by those with Type A patterns of attachment is to minimise difficult experiences particularly in terms of their emotional impact in order to be strong and get through difficult experiences. A finding from the wider narrative analysis was the less well-developed character of their child within the story. This might suggest that they found it more difficult to hold their child in mind, at least in relation to the recounting of their lived experiences. This could be due to them finding it more difficult to reflect on the thoughts and feelings of their child. It might also be painful emotionally to reflect on their child’s experience therefore excluding some of this from their story could be a self-protective strategy.

Apart from these findings, there appeared to be characteristics specific to the narratives from mothers who had loss or trauma from their own childhood that they had not yet fully resolved. This seemed to have a particular impact on the construction of coherent and temporally ordered narratives. One interpretation could consider whether it is more difficult to process a new trauma effectively, and thus construct a narrative about it, when still holding a previous unresolved trauma. The effect might be mediated by the individual’s attachment patterns more generally as the mother who was otherwise classified as Type B attachment seemed more able to construct and hold her narrative together than the mother who was classified as Type A. It was interesting that Nicola was the only mother who scored highly on the HADS measure of current mental state and she produced the most incoherent narrative. It could be useful to explore this further with a larger sample.

A final key finding of this article relates to the visibility of attachment strategies being used in the narrative telling of the mothers’ experiences. As discussed previously, it has been suggested that these strategies are only evident in transcripts of interviews completed using an AAI protocol (Hughes et al., 2000). This study offers contradictory findings suggesting that these strategies are evident in narratives about events that can be considered an attachment threat, such as the potential loss of a child. It should be noted, however, that these strategies might be subtle at times and, as would be expected, more clearly used at emotionally salient parts of the story. It must also be considered that an individual might be using emotionally minimising Type A strategies while presenting as upset and tearful, as in the case of Emma. This might be more common early in the experience when emotions are understandably raw. Over time, it might be expected that these strategies would have been used to process and narrate the experience in a manner that minimised the emotional response. One mother, Tracey, who moved between Type A and Type C strategies in her AAI, however, showed predominantly Type A strategies in her narrative interviews. It is important to note that within this classification, respondents can shift between the different strategies depending on the person or situation they are speaking about (Crittenden & Landini, 2011).

Clinical implications

While this study focussed on mothers who had a child with cancer, the findings are relevant for other populations. It is hypothesised that the perceived threat to life is the relevant factor as it is this which represents the threat to the attachment relationship. With this in mind, it is not the reality of whether the illness is a threat to life but how the parent perceives it that is important.

This research has implications for the availability of mothers to meet the emotional needs of their child during a time of great distress. Researchers have consistently found that a mother’s capacity to regulate and organise her thoughts and feelings about own experiences of receiving care is linked to her ability to regulate, organise and respond sensitively to her child’s needs for proximity, safety and comfort in life-threatening situations (Main, 2000; Van Ijzendoorn, 1995). Those mothers using Type A strategies might find that they effectively reduce their own distress enabling them to keep going and attend to the practical demands of their child’s illness and treatment. A consequence of this might be that they find it more difficult at times to attend to the emotional needs of their child. It might also be considered whether Type A strategies, while effective in helping the person getting through the trauma, might inhibit the ability to fully process their experience. This can be linked to the lesser degree of evaluation in the narrative transcripts of these mothers. A lack of processing might be problematic if they experience a subsequent trauma such as a relapse of their child’s illness.

Understanding how different attachment patterns might present themselves could be useful for all clinicians working with families. It is hypothesised that difficulties might be less easily recognised in those with Type A patterns due to use of minimising strategies. This research suggests that while it would not be expected or appropriate for clinicians to attempt to classify parental attachment patterns on the basis of talk about their child’s illness, they may be able to have a sense of the types of strategies being used and the potential implications.

This research has implications for psychological therapy that might be undertaken with families struggling with a child’s illness. As described previously, narratives are a way that we describe and make sense of our world, and are drawn upon at times of adversity. Dallos (2006) has some useful suggestions for integrating attachment theory with clinical work using narratives. For people using Type A strategies, he suggests the aim would be to help them access information about their feelings, to develop narratives incorporating these and integrate the narratives with their wider life story. In clinical work, this might utilise therapeutic techniques of role-play, demonstrating emotional processes, empathic questioning and internalised other interviewing. For people using Type C strategies, the aim would be to create emotional distance and strengthen semantic cognitive processes. This might include creating timelines of the narrative asking the person to locate key events in terms of time and place to build temporal order. It might also include questions about how other people see the events, and their beliefs and cognitions about it. It is important to recognise that attachment strategies are developed within close relational systems to optimise survival and self-protection in response to dangerous situations. Attachment informed interventions are, therefore, more likely to succeed in the dynamic context of working with the family system as a unit so that mutually adaptive care-seeking and care-giving interactions can be encouraged. Further research would be helpful in this area.

Limitations and suggestions for future research

As an initial exploration of the topic, this study was based on a small sample of six mothers. Apart from the small size of the sample, it did not appear representative of the general population in terms of attachment classifications. Research using the original AAI suggested that the typical distribution of classifications in non-clinical samples of North American mothers is 58% secure, 23% dismissing and 19% preoccupied (Bakermans-Kranenburg & Van Ijzendoorn, 2009). While the original AAI and the new DMM-AAI classifications do not exactly correspond, it can be hypothesised that Type C patterns might represent a similarly small proportion of the population. The lack of clear Type C patterns in this study might therefore be due to a small number in the general population; however, it is clear that Type A patterns were over-represented in the sample. One explanation for this might be that these mothers feel less distressed about their experiences due to their use of minimising strategies and therefore feel more able to participate in such research. It might also be possible that some mothers with expressive Type C patterns were appropriately excluded through the process of clinical gatekeeping, since they may have been perceived as emotionally too vulnerable to participate in the study by the liaison psychologist. It will be important that this is considered in future research. The research also only interviewed mothers. While both parents in families were invited to take part, no fathers agreed to participate. In addition, there were a range of ages of children included in this study. While there may be a difference in the nature of the care-giving provided at these different ages, it is proposed that the impact on the attachment relationship of a life-threatening illness would be the same across the age span. This is supported by bereavement literature concerning parental responses to the loss of their child (e.g. Harper, O’Connor, & O’Carroll, 2013). The mothers were interviewed after their child completed treatment with a successful outcome. It is possible that different stories would have been shared had they been interviewed during the treatment or if the treatment had not been successful. It would be useful for future studies to explore this further.

This study has highlighted the relevance of attachment theory to considering parental experience of having a child with a life-threatening illness. In considering future directions, this study could be replicated with a larger sample, including fathers and those with a Type C pattern of attachment. The research could also be expanded to consider assessments of attachment for both parents in a family and their child, as well as exploring wider family functioning in relation to this.

Attachment-based research is often time-consuming, complex and requires a researcher trained in analysing measures of attachment. This research has demonstrated the utility of narrative methods in attachment research. While these methods cannot be used to classify habitual attachment patterns of individuals, they offer an attachment lens through which attachment-based strategies may be identified in relation to life-threatening events. This is of immediate clinical relevance and further research into attachment informed narrative analysis could lead to developing methods of clinical assessment that would provide a simpler, less time-consuming and pragmatic alternative to formal measures of attachment.