Transitional Objects of Grief

1. Introduction

Attachment behavior enables a person to attain or maintain proximity to an identified individual who promotes their sense of safety and security. [1] Adversity and stress activate proximity-seeking for that attachment figure. [2] Secure attachment also supports separation from an infant’s attachment figure as a child becomes developmentally able, facilitating the initiation of relationships with the external world. As this separation occurs, transitional objects appear in an infant’s life, and are thought to contribute to the security necessary for this separation. [3] (Figure 1) A blanket or a stuffed animal, for example, is thought to internally represent features of the trust and safety the attachment figure offers, providing qualities that support the development of autonomy and independence. An element of this illusion is that the attachment object remains under the infant’s control and is ever available to provide comfort in displacement of the actual attachment figure. Although attachment theory has principally focused on an infant’s relationship with its mother, attachment behavior continues throughout the lifecycle. [1] Grief is theorized to be a manifestation of attachment, [4–7] and ensues when a person is separated from an attachment figure due to death.

Figure 1. Features of D.W. Winnicott’s Transitional Objects.

Defining features of transitional objects. After Winnicott, D.W., Transitional objects and transitional phenomena.¹¹

Attachment behavior is also reciprocal. In describing the profound attachment characteristics of the infant-mother relationship, Winnicott observed the critical role an infant played for the mother’s sense of self, and a mother’s self-identification with the attentive and nurturing environment she provided for her dependent and responsive infant. [8] Following the death of a young child at a time when attachment behavior is strong, the continuation of this attachment-oriented sense of self might be expected to be especially significant for parents. Indeed, grief research shows that the grief responses of parents following the deaths of their young children are uniformly severe. [9]

While conducting home-based interviews with parents who had experienced the death of their child, [10] some version of a shrine to the deceased was almost always found, consisting of photographs and occasionally keepsakes. Upon questioning, some parents, especially mothers, reported drawers or boxes of the child’s objects that had been instinctively collected, often including special items of clothing, a blanket, or a favored toy. They reported visiting these objects frequently, to hold and smell them, typically in private. They described comfort and also distress when they accessed them. Many described an effort necessary to go to them due to the intensity of the experience, and heartache as they noticed the smell of their child disappearing from them. These may be considered objects of displaced attachment similar to those noted by Winnicott, referred to here as transitional objects of grief. We hypothesize that such objects appear and serve a role for parents in their adaptation to loss after a child’s death.

Psychosocial care around bereavement shows evidence of such transitional objects and suggests a potential therapeutic role. These efforts implicitly recognize the significance of continuing bonds and meaning-making in post-loss adaptation. Research on Prolonged Grief Disorder (PGD), which under various guises designates a persistent and debilitating form of grief, [11–15] has demonstrated the role of continuing bonds and their contribution to meaning-making, [16] as well as its influence on coping. [17] Continuing bonds, both concrete or symbolic efforts to maintain an ongoing relationship with the deceased, are crucial to post-loss adaptation [18–20] and are especially prominent in the coping of bereaved parents. [21] Various aspects of meaning-making are described that affect grief outcomes in parents, including role confusion, the disorienting loss of familiar structures in the life of a young family, and the responsibility to protect the worth of the deceased child’s life. [22, 23] Clinical models have been developed that address these factors as therapeutic targets [24] and pediatric hospitals increasingly provide personal objects of remembrance to parents when their children die, motivated by the potential contributions of meaning-making and continuing bonds to post-loss adaptation. [25, 26] However, like stuffed animals given to a newborn, the nursery may be filled with objects that identify it as the infant’s space, but only certain, specific objects become selected, nurtured and embedded with more personal significance. These are transitional objects of grief. They take on specificity within the context of continuing bonds and meaning-making.

To advance understanding of the phenomenon of transitional objects of grief, we conducted a prospective study to document the extent of their existence, some of their features, and dynamics of related behavior in mothers after their infants died from the Sudden Infant Death Syndrome (SIDS). Rates of PGD have been shown to be extremely high in bereaved mothers following the sudden and unexpected deaths of their infants from SIDS. [27] We further examined this behavior as it related to PGD.

2. Methods

Bereaved mothers were recruited during the period 2–36 months post-loss using two sampling methods. Subjects in the Prenatal Alcohol in SIDS and Stillbirth (PASS) Research Network, a prospective study following 12,000 pregnancies with heightened risk for SIDS in the areas of Cape Town, South Africa and the northern Great Plains, South Dakota USA, were recruited after their infants died from SIDS between May 2013 and July 2016. Surveys were administered by trained project professional staff. A second sample was recruited from bereaved mothers in parent support organizations associated with the International Society for the Study and Prevention of Perinatal and Infant Death (ISPID), publicized with website or newsletter announcements. Surveys were completed online between September 2014 and July 2016. Written, voluntary informed consent was obtained from all subjects/patients. This research was approved by institutional review boards at Dana-Farber Cancer Institute (protocol 13–207); the Eunice Kennedy Shriver National Institute of Child Health and Human Development in Bethesda, Maryland; Stellenbosch University in Stellenbosch, South Africa; Sanford Health in Sioux Falls, South Dakota; the Oglala Sioux Tribe Research Review Board; and the Great Plains Institutional Review Board.

Basic demographic data and characteristics of the loss were collected. Poverty rates in South Africa were determined using the South African upper-bound poverty line, [28] an estimated income level below which the purchase of food and daily essentials puts a household in jeopardy of not meeting the daily minimum energy intake of 2100 kcal (R753 [South African rands] or $57 [US dollars] per month) determined by family size. For others, mean reported incomes were converted to US dollars using 2015 exchange rates and poverty was defined as household income less than $24,000. If multiple surveys were administered to a participant, the responses corresponding to the longest post-loss interval before 36 months were used.

Prolonged Grief Disorder was assessed using the Parental Bereavement Questionnaire (PBQ), an 18-item survey based upon the PG-13, [29] modified for use with parents of young children. The PBQ measures diagnostic criteria for PGD (Prigerson 2009 criteria [12]) with items addressing separation distress (yearning and emotional pain), and cognitive, emotional, and behavioral symptoms, including confusion about one’s role in life or diminished sense of self, difficulty accepting the loss, avoidance of reminders about the reality of the loss, inability to trust others since the loss, difficulty moving on with life (e.g. making new friends and pursuing interests), emotional numbness since the loss, and feeling stunned, dazed, or shocked by the loss. The adaptations in the PBQ that incorporated parent-specific language into the PG-13 were pilot tested on a sample of Massachusetts parents, [10] and input from focus groups was used to increase reliability for local language preferences specific to Cape Town. Subjects were asked to respond to items describing defined symptoms of PGD [12] using a 5-point Likert scale for degree or frequency of the symptom. PG-13 items have a sensitivity of 1.00 and specificity of 0.99 for PGD. [12] Internal consistency of the PBQ is α = 0.92 and criterion validity (Kuder Richardson) was 0.72 for PGD criteria. [30]

We included questions in the survey about maternal behavior related to transitional objects of grief. We introduced the topic by stating: “Mothers sometimes keep things to remind them of their lost child. Examples of this are pictures, toys, blankets, or footprints. Have you kept any of these kinds of things?” Other items asked whether the objects were hidden or public, the frequency of their visits, behavior related to visiting them, and whether they provided comfort and/or distress. Two separate items asked about comfort and distress and were not mutually exclusive. In order to gain details about other loss-oriented behavior, additional items surveyed communication with others about the deceased child, including the ease with which a mother spoke of her child and her beliefs that she needed to hide her feelings from others. Avoidance (a 2-item measure) and meaninglessness (a 1-item measure) were assessed in the PBQ as part of PGD criteria. Participants were not asked to quantify or refer to explicit objects in their responses. Our hypothesis did not include an expectation that the number of objects identified was a critical distinction; rather, our focus was on the use of any transitional object related to emotional distress in bereaved mothers. Subsequent questions were posed relative to whatever items the participant identified without regard to their quantity.

2.1. Analysis

Responses on a 5-point scale were dichotomized, with the two highest values being coded positive (e.g., “Quite a bit” or “Overwhelmingly” defined our comfort and distress measures). We used the Pearson’s chi-square statistic for bivariate comparisons and logistic regression for multivariate comparisons. Effects of comfort and distress on loss-oriented behavior by PGD status were assessed using logistic regression with terms for PGD, comfort, distress, PGD by comfort interaction, and PGD by distress interaction. Alpha was set at 0.05 and all tests were two-tailed. All model assumptions were met for Pearson’s chi-squared test (independence, expected cell counts) and logistic regression (binary outcome, no evidence of significant collinearity among independent variables in the logistic regression model).

3. Results

3.1. Demographics.

We assessed 294 SIDS-bereaved mothers, 248 affiliated with SIDS parent-support organizations in the United States, United Kingdom, Australia, New Zealand, and the Netherlands, and 46 living in areas at high risk for Sudden Infant Death Syndrome, economically impoverished areas near Cape Town, South Africa and the northern Great Plains of the US. (Table 1) Average time from loss for the total sample (n=294) was 17.9 +/− 11.0 months; for mothers with PGD using 6 month post-loss diagnostic criteria (n=120) was 19.5 +/− 9.6 months; and for all mothers reporting symptomatic criteria for PGD without regard to time post-loss (n=145) was 16.7 +/− 10.7 months. As previously reported, rates of PGD were not statistically different between groups. [27] In addition, PGD was not significantly associated with any of the characteristics in Table 1.

Table 1.

Demographic variables

	Total	PGD+	PASS	ISPID
Number	294	120	46	248
Mean age at loss (yrs)	29.8	30.1	26.5	30.6
White race (%)	79.4	79.5	7.1	92.1
Education beyond Highschool (%)	66.0	62.9	10	77.1
Poverty (%)	23.3	26.5	68.4	16.1
Previous loss of child or pregnancy (%)	31.6	30.5	37.5	30.6
Months Since Loss (n)
2-<6	48	---	1	47
6-<12	55	40	2	53
12-<18	49	23	13	36
18-<24	37	19	2	35
24-<30	41	23	6	35
30–36	64	49	22	42

Sample characteristics including age at loss, race, education, poverty, previous loss and post-loss interval. The two subgroups were significantly different with regard to age, education, race, rates of poverty and previous child or pregnancy loss.

3.2. Transitional Objects.

Essentially all the mothers (98.6%) reported the existence of transitional objects and 28.2% kept them privately. Nearly two-thirds, 63.2%, visited the objects more frequently than once a week and this response was used as the frequency variable for subsequent analysis. (29.5% visited the objects several times each day.) The behavior reported with the objects was not significantly associated with time since loss, although there was a trend towards less private visiting and decreased frequency of visiting over time. In a subset, 91% of the mothers reported smelling and touching the objects when they visited them. The frequency of visits, the private versus public location, and the sensory and tactile experience were not statistically different relative to a diagnosis of PGD. (Table 2) Data on the number of objects was not explicitly collected, although no participant identified a need to address the use of multiple transitional objects. Examples of private objects included baby clothes, stuffed animals, pacifiers, locks of hair, quilts made from baby clothes, or the infant’s mobile. Typical more public objects included photos, hand- and footprints, cremation ashes, and memorial tattoos.

Table 2.

The presence of transitional objects and related behavior

	Total	PGD−	PGD+	χ2(1)	P value	2–12 months Post-loss	13–24 months Post-loss	25–35 months Post-loss
Presence	288 (98.6)	147 (98.7)	141 (98.6)	0.00	0.97	107 (99.1)	90 (97.8)	91 (98.9)
Private	81 (28.2)	38 (25.8)	43 (30.7)	0.84	0.36	38 (35.8)	22 (24.4)	21 (23.1)
Public	206 (71.8)	109 (74.2)	97 (69.3)			68 (64.2)	68 (75.6)	70 (76.9)
Visit more frequently than once a week	180 (63.2)	85 (58.6)	95 (67.8)	2.61	0.11	68 (64.8)	60 (66.7)	52 (57.8)
Look, touch, smell#	42 (91)	27 (93)	15 (88)	0.32	0.57	3 (75)	15 (94)	24 (92)

Results shown in Number (percent).

^#PASS cohort only

3.3. Comfort, distress, loss-oriented behavior and associations with PGD.

Overall, 35.1% reported that the objects provided high levels of comfort and 26.4% reported that the objects provided high levels of distress. A minority, 5.1%, reported high levels of both comfort and distress. The presence of PGD was associated with distress and certain loss-oriented behavior. (Table 3) PGD was significantly associated with reporting distress when accessing the objects (p<0.001). Mothers with PGD reported more distress when accessing the objects either publicly or privately (p=0.005 and p<0.001, respectively). The association between PGD and distress was much stronger in mothers who visited the objects in private (interaction effect, p=0.04). In separate testing, the association between PGD and experiencing higher levels of distress did not vary by time since loss.

Table 3.

Comfort, distress and loss-oriented behavior in those with and without PGD

	Total	PGD−	PGD+	χ2(1)	P value
Provides-comfort	101 (35.1)	57 (38.8)	44 (31.2)	1.81	0.18
Private – comfort	18 (22)	10 (26)	8 (19)	0.69	0.40
Public – comfort	83 (40)	47 (43)	36 (37)	0.77	0.38
Provides distress	76 (26.4)	20 (13.6)	56 (39.7)	25.26	<0.001
Private – distress	32 (40)	5 (13)	27 (63)	20.79	<0.001*
Public – distress	44 (21)	15 (14)	29 (30)	7.96	0.005*
Unable to talk about the infant	19 (6)	8 (5)	11 (8)	0.60	0.44
Unable to talk about infant without feeling overwhelmed	117 (40)	39 (27)	78 (54)	23.11	<0.001
Unable to talk about infant without crying	61 (21)	16 (11)	45 (31)	18.69	<0.001
Believes she needs to hide feelings from others	112 (38)	32 (21)	80 (55)	35.38	<0.001

Results in Number (percent).

^*significant PGD by location interaction effect (p=0.04)

PGD was not associated with a difference in whether a mother was able to talk about their deceased infant, but mothers with PGD were less able to talk about their deceased infant without crying (p<0.001) or feeling overwhelmed (p<0.001). Mothers with PGD were more likely to report that they needed to hide their true feelings from others (p<0.001).

3.4. Loss-oriented behavior and effects of comfort and distress by PGD status.

Mothers with PGD reporting distress from the objects visited them less frequently (p<0.05). Mothers who found comfort in the objects visited them more frequently (p<0.01), but this did not reach significance in mothers with PGD despite a large effect size. (Table 4) Among mothers with PGD, experiencing distress from the objects was associated with being less able to talk about their infant without feeling overwhelmed (p<0.05). Conversely, mothers with PGD experiencing comfort were more able to talk without feeling overwhelmed. Fewer mothers with PGD who found comfort in the objects found life unfulfilling or meaningless. (p<0.02) Avoidance was not associated with distress or comfort in mothers regardless of PGD status. Pearson goodness-of-fit tests indicated that all models in Table 4 were a good fit to our data.

Table 4.

Loss-oriented behavior and effects of comfort and distress by PGD status

	PGD− Comfort	Distress	PGD+ Comfort	Distress
Visit more frequently than once a week	3.04 [1.46,6.33]**	1.37 [0.50,3.71]	2.32 [0.95,5.65]	0.44 [0.21,0.92]*
Role confusion	0.72 [0.35,1.47]	1.38 [0.46,4.11]		0.87 [0,33.75]
Anger	0.75 [0.35,1.58]	2.38 [0.90,6.29]	1.46 [0.48,4.50]	2.75 [0.84,9.01]
Diminished trust	1.44 [0.60,3.42]	2.96 [1.02,8.54]*	1.18 [0.56,2.50]	1.01 [0.50,2.05]
Avoidance	1.49 [0.68,3.24]	1.18 [0.39,3.58]	1.09 [0.50,2.36]	1.53 [0.73,3.22]
Life is unfulfilling, meaningless	0.85 [0.32,2.17]	1.37 [0.41,4.62]	0.41 [0.19,0.87]*	1.34 [0.63,2.85]
Unable to talk about infant	0.93 [0.21,4.13]	0.89 [0.10,7.81]	0.17 [0,1.10]	1.83 [0.40,9.49]
Unable to talk about infant without feeling overwhelmed	0.97 [0.45,2.08]	1.19 [0.42,3.39]	0.28 [0.13,0.61]***	2.35 [1.12,4.91]*
Unable to talk about infant without crying	1.06 [0.35,3.19]	2.43 [0.68,8.65]	0.72 [0.31,1.63]	1.38 [0.66,2.91]
Believes she needs to hide feelings from others	0.80 [0.35,1.83]	0.59 [0.16,2.20]	0.61 [0.29,1.26]	1.49 [0.73,3.00]

Odds ratios and 95% confidence intervals for the effects of comfort and distress by PGD status, examining the outcomes of loss-oriented behavior. Constituent PGD symptoms of difficulty avoidance and meaninglessness (meaning-making) are included because they represent loss-oriented behavior. Role confusion, anger and diminished trust are included because they are the symptoms of PGD most frequently reported in this sample.

^*p<0.05,

^**p<0.01,

^***p=0.001.

A significant interaction effect was found only for PGD and comfort when predicting “unable to talk about infant without feeling overwhelmed” (p=0.02).

4. Discussion

Virtually all bereaved mothers reported the presence of concrete objects whose access and proximity was involved in coping with their infant’s death, items we refer to as transitional objects of grief. Features in Winnicott’s description of transitional objects were evident: the objects were chosen, unchanging, used with great frequency in a tactile and sensory manner, and evocative of a range of emotions. Similarly, the bereaved mothers interacted with these spontaneously saved objects not only as physical objects but recognizing deeply personal and emotional aspects related to their child. The objects often produced comfort and, when they did, they seemed to confer protective benefits. Fewer mothers with PGD reported that life was unfulfilling and meaningless when they found comfort in the objects.

The existence of these objects is consistent with essential constructs in the attachment literature. In this sense, transitional objects of grief serve a similar adaptive function in bereaved mothers as Winnicott’s transitional objects provide the individuating infant: a way to hold onto an absent and essential attachment object to assist them in their new and challenging state. This research also illustrates PGD as a disorder of attachment. Taken together, a picture emerges of bereaved mothers with PGD who are impaired in their ability to access the common, generally positive, restorative aspects that transitional objects and their displaced, specific infant characteristics might provide. In cases of mothers with PGD, private, frequent and distressing visits to the objects seemed to impede, rather than facilitate, the natural oscillatory process of adaptation to loss described in the context Dual Process Model of coping with bereavement. [31] Rather than benefitting from the restorative aspects, their distress with transitional objects seemed to accompany impeded access to positive aspects of the relationship. Moreover, distress associated with the transitional objects was not transitory in those with PGD, also consistent with the recalcitrance of the condition.

Anecdotally, asking mothers about these objects seemed to be affirming to them, regardless of whether they met criteria for PGD. In contrast to the avoidant behavior generally described in PGD, [32] they recounted details of the objects and how they visited them in the manner of sharing a comforting secret, and would later describe the conversation as positive and liberating. Although surprised at the question, they seemed proud of their private objects, and presented them as precious, important and meaningful. Their comments about the objects typically reflected an ongoing, special relationship with their deceased infant. Our conceptualization has centered on the emergence of the behavior towards these objects as attachment behavior, but the meaning-based aspects of this symbolic behavior also seem to illustrate the role of micro-ritual in sustaining continuing bonds with the deceased infant.[33] However, the degree of stigma and social disenfranchisement around these deaths suggests that the behavior reveals less of the sociocultural elements of ritual than a psychological response, particularly well reflected in the association between private visitation with the transitional object and emotional distress.

Others have concluded that avoidance of reminders of loss is an important component of the functional impairment of prolonged grief, [2] and avoidance is a durable component of the independent construct of PGD. [12] Our data about transitional objects and avoidance presents a more complicated picture. [34] With regard to PGD and avoidance-related behavior, subjects with PGD were not different in whether they kept objects, the frequency of their visits, the location of the objects, the associated sensory behavior, or their rates of avoidance. However, the emotional climate of their access to the objects had an effect. Mothers with PGD who experienced distress visited the objects less frequently while, in general, those with PGD who found comfort visited them more frequently. This indicates a potential therapeutic target to overcome in PGD, [2, 35] using techniques such as Acceptance and Commitment Therapy which targets experiential avoidance to address feelings of stress and distress. [36]

Moreover, the interaction between keeping the objects private and distress may qualify our understanding of avoidance, since many mothers welcomed the private experience of intense emotions that occurred. Another observation supporting the role of emotion in the experience was the association between finding comfort in the objects and finding life more fulfilling and meaningful in mothers with PGD. Additionally, PGD was associated with the inability to talk about their child without feeling overwhelmed and not simply being able to talk about their child, similarly reflecting blocked access to restorative functions of memory-making and remembrance. PGD is not only a set of symptoms but also patterns of thought that leave the bereaved stalled and unable to engage in positive adaptational behavior. These observations additionally reflect the impaired post-loss adjustment predicted by social constraint theory when intolerance of this behavior in the social environment is perceived. [37, 38] Indeed, the private experiences of the mothers coupled with their limited opportunities to share their thoughts and feelings without fear of judgment would be an illustration of how constraints may impede post-loss adaptation.

Given the positive associations with gaining comfort from the objects, it may be helpful to consider transitional objects of grief in the context of exposure therapy, where processing memories of the deceased has been shown to reduce grief-related symptoms and improve social and psychological functioning. [39–41] Incorporating transitional objects might include targeting secluded distress for intervention, utilizing a graded exposure hierarchy from least distressing to most distressing, or varying and facilitating the time spent in contact with the objects. Such an approach may provide an opportunity for the bereaved mother to safely metabolize the loss and decouple the threat from the comfort. Therapeutic opportunities may occur by inquiring about these objects and their related behavior and emotions. Such discussion may offer validation of their deep personal connection to the deceased infant and the continuation of maternal behavior, while gaining useful material for the characterization and treatment of PGD. It is important to recognize, however, that continuing bonds behavior is not uniformly beneficial. In some cases, frequent interaction with loss-related objects may be perseverative, divert from reckoning with the finality of death and may actually increase PGD over time. [18, 20, 42, 43] In fact, some exposure treatments already contain techniques to reduce such behavior. While our intent is to be descriptive, the interaction between this behavior and grief-related psychopathology is multidimensional and warrants further study.

Limitations of the manuscript include the determination of PGD by the use of a survey. Despite this being common practice in the literature, perhaps a more accurate label would be to describe mothers with “probable PGD” vs. “non-probable PGD”. Also, the nosology and symptom structure of PGD have been modified in the time since the study was conducted and, although a validated metric was used in this research, when we reanalyzed data using proposed ICD-11 criteria [13] without including survivor guilt, since we did not collect explicit data on survivor guilt, 32/169 probable PGD cases (18.9%) were gained. When we re-analyzed using proposed DSM-5-TR criteria [44] without including loneliness, since we did not collect explicit data on loneliness, 20/165 (12.1%) probable PGD cases were gained. In both cases, the 2009 criteria were more stringent. The reader is cautioned to interpret findings accordingly, since we chose not to present re-analysis using measures that are not implemented and for which we lacked relevant data. The lack of complete longitudinal data limited our ability to examine this behavior over time. A further limitation is an increased risk of type I error due to the number of statistical tests. We believed that the cost of a type I error was small compared to the cost of a type II error in this novel research area and did not correct for multiple comparisons. These findings should be replicated in other research. Future research may also explore explicitly whether the number of transitional objects influences the observed outcomes.

The implications of this research may also be limited by the specificity of the bereavement relationship, since the attachment behavior between a mother and her young, dependent infant is singularly strong and prototypical. Although SIDS is not rare (it is the leading cause of death for infants between one month and one year of age in high-income economies [45] ) and there is little reason to believe that these psychological dimensions do not apply in other settings, some caution is warranted with regard to generalizability. We have no data on paternal behavior or mothers experiencing different deaths of their child. A further limitation is the lack of data on the attachment style in the infant-mother relationship, which might provide a more discriminant analysis. [32, 46] Moreover, we have questioned elsewhere if the severity of grief after the death of a young dependent infant documented in the literature warrants its own special considerations as a category of loss. [9]

5. Conclusions

Transitional objects of grief are present, prevalent and an active dimension of processing a profound loss. Our results suggest a potential therapeutic role for transitional objects in a bereaved mother’s process of coping. Engagement around transitional objects of grief and efforts to address distress associated with the objects may be a novel addition to grief-directed therapy in PGD.

Highlights.

• Transitional objects aid in establishing separation from an object of attachment

• Nearly all parents save objects from their deceased children

• Mothers with Prolonged Grief Disorder were more distressed when visiting them

• Their comfort correlated to talking about the child and finding meaning in life

• These objects may provide a therapeutic role when addressing PGD