Ambiguous Loss in Families of Children WithAutism Spectrum Disorders*

Marion O’Brien**

Abstract: Learning that a child has a lifelong developmental disorder is stressful and challenging to any family, yetit is clear that some families adapt and adjust more readily than others. In this article, it is proposed that a diagnosisof an autism spectrum disorder (ASD) is especially likely to be experienced as ambiguous loss. Interviews withmothers of children with ASDs are used to identify whether mothers express feelings of ambiguous loss when talk-ing about their child. Then, a specific hypothesis derived from ambiguous loss theory—that higher levels of identityambiguity in mothers are linked to higher levels of depressive symptoms and perceived stress independent of theseverity of the child’s diagnosis—is tested and found to be supported. Recognition of ambiguous loss in families ofchildren with ASDs would help professionals provide more effective support and assistance to families.

Key Words: ambiguous loss, autism, developmental disability, family adjustment.

An extensive literature exists on the outcomes forparents when children have developmental disabil-ities. It has been commonly reported that parents,especially mothers, of children with disabilities expe-rience more depressive symptoms (Noh, Dumas,Wolf, & Fisman, 1989; Sloper & Turner, 1993) andhigher levels of stress (Goldberg, Morris, Simmons,Fowler, & Levison, 1990; Kazak & Marvin, 1984;Pelchat et al., 1999) as compared with parents ofchildren who are typically developing. However, incontrast to this bleak picture, some researchers havenoted that many families of children with disabilitiesadapt successfully to their child’s needs and manageto function well (e.g., Hastings & Taunt, 2002;Hauser-Cram, Warfield, Shonkoff, & Krauss, 2001).

Efforts to understand why some families accom-modate more readily to a child’s disability thanothers have focused largely on two factors: childcharacteristics, especially type and degree of severityof the child’s disability; and social support availableto the family. For the most part, the severity ofa child’s disability has not been consistently linked

to family functioning (Baker, Blacher, Crnic, &Edelbrock, 2002; Wallander & Varni, 1998): althoughsocial support has been found to be helpful in reducingpsychological distress of parents (Dunst, Trivette, &Cross, 1986; Minnes, 1998), it does not provide anexplanation for the initial variation in distress.

In this article, it is proposed that ambiguous losstheory (Boss, 1999, 2004) can be used to helpunderstand differences in parental responses tolearning that their child has a lifelong and pervasivedisability. The specific situation examined here isthe diagnosis of an autism spectrum disorder(ASD) in a child. In brief, it is suggested that manyparents of children with ASDs experience a sense ofambiguous loss. The child they thought they had isnot the child they must learn to live with. Thus,parental expectations must change. Further, par-ents’ ability to tolerate and manage the ambiguityof their child’s disability is proposed as a key factorin their experience of ambiguous loss and therelated confusion over family members’ identities,both of which contribute to emotional distress.

*This work was supported in part by a grant from the University of Kansas Research Development Fund. I particularly thank the families who gave so generously of

their time. Thanks also to Dr. Julie Daggett, Dr. Donna Dailey, Dr. Matthew Reese, Dr. Michal Nissenbaum, and Kera Watts whose efforts were instrumental in

accomplishing the research; to Lauren Keel Shinn for coding assistance; and to Cheryl Sarratt for help with manuscript preparation.

**Marion O’Brien is a Professor and Director in the Family Research Center, Department of Human Development & Family Studies at University of North Carolina

at Greensboro, P.O. Box 26170, Greensboro, NC 27402 (m_obrien@uncg.edu).

Family Relations, 56 (April 2007), 135–146. Blackwell Publishing.Copyright 2007 by the National Council on Family Relations.

Ambiguous Loss in Autism

The theory of ambiguous loss (Boss, 1999, 2006)arose out of family stress theory, which posits thatstress results whenever there is change within a fam-ily. According to Boss, the most severe stressors arethose changes that are not clear-cut but are ambigu-ous. In these cases, resolution of the situation is notpossible and the outcome not predictable. In exam-ining family response to chronic illness and toAlzheimer’s disease, Boss and colleagues (Boss &Couden, 2002; Caron, Boss, & Mortimer, 1999)have identified several areas of ambiguity that con-tribute to family distress: (a) a lack of clarity indiagnosis, (b) difficulty in predicting outcomes, (c)day-to-day changes in functioning of the ill personthat affect family relationships, (d) the fact that theill individual may give an outward appearance ofhealth thus raising expectations for his or her behav-ior and functioning within the family, and (e)fear that important emotional relationships will beaffected by the illness. All of these areas of ambiguityapply directly to the situation of parents whose chil-dren are diagnosed with ASDs.

Lack of clarity in diagnosis. ASDs are consideredto have a neurological basis, but there are no knownbiological markers. Therefore, diagnosis must rely onclusters of behavioral symptoms, primarily in theareas of communication, social interaction, and repet-itive and restricted interests (American PsychiatricAssociation, 2000). As a result, the diagnosis of anASD is never straightforward, and the process of iden-tifying the child’s disorder is often prolonged. Inaddition, the etiology of ASDs remains unclear.Genetic links that are evident to researchers whoexamine the big picture (Shastry, 2005) are notalways discussed fully with parents. Many parentscontinue to believe earlier or more skilled medicalintervention at the time of birth or an avoidance ofchildhood immunizations might have altered theirchild’s course.

It does appear that autism results from an interac-tion between a genetic propensity and some environ-mental trigger occurring either pre- or postnatally(Rutter, 2000). The identification of candidate genesfor the disorder will be only the first step in what maybe a long process of understanding what causes autism.In the meantime, many parents remain confused anduncertain about their child’s ASD diagnosis.

Difficulty in predicting outcomes. ASDs take asmany forms as there are children with the disorder

(Coplan, 2003). No two children show the samepattern of symptoms or strengths, and the specificevents that challenge one child appear to be handledeasily by another. Similarly, the long-term outcomefor individuals with ASDs is highly variable(Howlin, Goode, Hutton, & Rutter, 2004). Someadults who appeared to be rather severely impairedas children are able to hold jobs and live with mini-mal support, whereas others require intense supportand intervention throughout their lives. Parentstruly do not know what their child with an ASD willbe like as an adult.

Ambiguity can be amplified when parents reador hear about children whose autism has been‘‘cured’’ through intensive intervention or a new die-tary or medical therapeutic approach. The challengefor parents whose children have ASDs is to acceptthe reality of the permanent nature of their child’scondition while maintaining optimism about poten-tial gains that can be made through effective inter-vention (Boss, 2006; O’Brien & Daggett, 2006).

Day-to-day variability in functioning. Childrenwith ASDs have good days and bad days. What iseasy for them to do at one time may create chaosand conflict later that day or the next time the situa-tion arises. This variability in behavior makes itdifficult for parents to plan activities or respondeffectively to their children’s behavior. Anotheraspect of ASDs that contributes to ambiguity is thefact that many children, even those who are rela-tively severely impaired in some domains, show con-siderable abilities in other domains (Gillberg, 2002;Ozonoff & Rogers, 2003). The inconsistency inchildren’s performance levels can keep family mem-bers feeling off-balance, never quite certain whetherthe child is truly unable—or just unwilling—tolearn certain basic skills.

Appearance of health. Children with ASDs donot look obviously different from healthy childrenin their early years. Many parents whose children arediagnosed with ASDs believed for a number of yearsthat their children were developing typically or atleast that their children were basically healthy if abit odd or especially difficult (O’Brien & Daggett,2006). Thus, at the time they receive a definitivediagnosis, most parents must alter their image oftheir child and of their own relationship with thechild. Furthermore, the fact that children with ASDslook just like typical children often encouragesextended family members and friends to discount

Family Relations � Volume 56, Number 2 � April 2007136

the child’s diagnosis and convey messages to parentsthat they are overreacting or being alarmist in pursu-ing a diagnosis (O’Brien & Daggett). Such messagesundermine parents’ self-confidence and increasetheir sense of ambiguity.

Loss of relationships. The nature of autism itself,especially the difficulties of individuals with ASDsto perceive others’ feelings and share meaningfulexperiences with others, creates a situation of consid-erable ambiguity within the family. The child withan ASD is clearly still present in the household; infact, the child with an ASD is likely to require anincreased amount of parental attention and familyresources over those devoted to a child who is devel-oping typically. Yet, psychologically, the child is notentirely present, at least not according to traditionalviews of the ways a child participates as part of a fam-ily. Accordingly, the nature of the parent-child rela-tionship must change—in this sense, ASDsrepresent a situation of ambiguous loss to parents.

Identity Ambiguity in Families of ChildrenWith ASDs

A key aspect of ambiguous loss that has been identi-fied in a number of situations is boundary ambiguity(Boss, 1988, 1999, 2004; Caron et al., 1999; Kaplan &Boss, 1999). Although most commonly described asconfusion over who is inside the family and who isnot, boundary ambiguity can also be experienced asambiguity regarding separation between familymembers’ identities (Boss & Greenberg, 1984). It isthis latter component of boundary ambiguity, herelabeled identity ambiguity, which is applied in thisstudy to families of children with ASDs. Whena child is diagnosed with autism, the identities ofthe child and parent can become unclear, and theparent-child relationship changes. Some parentsrespond to this identity ambiguity by taking onresponsibility for all aspects of their child’s life, blur-ring the boundaries between themselves and theirchild. If parents of children with ASDs are unable toseparate their own life situation from their child’sdisability, they can be described as experiencingidentity ambiguity. Similar responses have beenidentified in other populations. For example, confu-sion over roles and responsibilities has been foundto mediate the relation between the difficulties ofcaring for an aging parent with dementia anddepressive symptoms in the caregiver (Boss, Caron,Horbal, & Mortimer, 1990) and to relate to

depressive symptoms in spouses of patients withAlzheimer’s disease who were institutionalized(Kaplan & Boss).

To date, the concept of identity ambiguity hasnot been applied to families of children with ASDs,but it is relevant to the responses of parents toa child’s diagnosis. Theoretically, parents who are sopreoccupied with their child’s autism that they areunable to view their own lives as independent fromthe difficulties their child experiences are failing tomaintain clear boundaries between themselves andtheir child. Identity ambiguity has been linked toimmobilization, feelings of being overwhelmed andunable to change the situation, and difficulty adapt-ing to change (Boss, 1999). Thus, parents of chil-dren with ASDs who experience identity ambiguitywould be expected to have difficulty making deci-sions about their child’s intervention services andalso would be ineffective in analyzing their ownoptions regarding such activities as employment,socializing with friends, and community involve-ment. As a result, when identity ambiguity resultsfrom the ambiguous loss represented by their child’sautism, parents are predicted to experience higher orincreased levels of personal distress following theirchild’s diagnosis.

Although no research has specifically examinedidentity ambiguity in families of children with ASDsor other disabilities, some investigators have exam-ined constructs that are somewhat similar. Forexample, Ireys and Silver (1996) found that motherswho perceived their child’s chronic illness to havea greater impact on the family were more likely toexperience mental health problems than those whoviewed the illness as having less impact, regardless ofthe actual medical severity of the child’s illness. Sim-ilarly, Trute and Hiebert-Murphy (2002) developeda measure of positive and negative appraisals of dis-ability and found that mothers who reported rela-tively more negative than positive appraisals showedhigher levels of parenting stress 7 years later.

The Present Study

The research reported here included two compo-nents. In the first, semistructured interviews with 63mothers of children with ASDs were examined toexplore whether issues related to the theory of ambig-uous loss come up spontaneously when parents talkabout their experiences. The goal of this qualitativeportion of the study was to determine whether the

Ambiguous Loss and ASD � O’Brien 137

application of the theory of ambiguous loss to the sit-uation of families of children with ASDs appears tohave face validity. Such evidence is needed prior toaddressing specific research questions arising fromthe theory of ambiguous loss to these families.

In the second, quantitative part of the study,a hypothesis derived specifically from ambiguousloss theory was tested in these same 63 mothers:whether high levels of identity ambiguity are directlyrelated to high levels of perceived stress and todepressive symptoms in mothers of children withASDs. Analyses also controlled for the level of sever-ity of the child’s ASD symptoms because stress anddepressive symptoms might be expected to be higherin mothers of children with more severe disabilitiesand/or more difficult behavior. It was predicted thatidentity ambiguity would account for significantvariance in stress and depressive symptoms ofmothers of children with ASDs and that symptomseverity would not moderate this relation.

Method

Participants

The 63 participating mothers (including 2 adoptivegrandmothers and 6 other adoptive mothers) rangedfrom 23 to 54 years (M ¼ 35.8). The majority ofmothers (81%) were White, and about half themothers (54%) were college graduates. As is com-mon for mothers of children with special needs(Booth & Kelly, 1999), most of the mothers (58%)were not employed outside the home. All but 11 ofthe mothers were married or partnered (83%). Thechildren’s fathers (or grandfathers in two cases)ranged from 25 to 57 years; half of them had collegedegrees and all but two were employed. Familyincomes ranged from less than $1,000/month tomore than $10,000/month, whereas the median fam-ily income was between $2,500 and $3,000/month.

The children of participating mothers rangedfrom 2 to 13 years (M ¼ 70.8 months), although halfwere 5 years or younger. The vast majority of thechildren (87.3%) were boys. The children werelargely in good physical health; relatively few hadbeen born prematurely or had serious perinatal prob-lems, and their mothers reported the current healthstatus of 90% of the children to be good or excellent.

Just about half the children (n ¼ 32) were diag-nosed with autistic disorder (AD), the ASD subtype

that includes significant deficits in all three of thecore areas: communication, social interaction, andrestricted/repetitive interests (American PsychiatricAssociation, 2000). The other half (n ¼ 31) werediagnosed with other ASD subtypes that are definedprimarily by the presence of fewer symptoms or lesssevere symptoms than required for a diagnosis ofAD. Within this group, 16 children were diagnosedwith pervasive-developmental disorder not otherwisespecified (PDD-NOS), 7 with Asperger syndrome,2 with high-functioning autism, 3 with atypicalautism, and 3 with mild autism. For the quantitativeanalyses, the families were divided into these twogroups, which differed by the number and severityof children’s ASD symptoms.

Procedure

Families were contacted through several diagnosticclinics at a major Midwestern medical center. Lettersdescribing the study were sent to families by clinicstaff after parents had scheduled or completed a devel-opmental evaluation for their child. Parents interestedin participating returned contact information to theresearch team, who then sent them survey packets.Two waves of data were collected; in the first wave,letters were sent to 50 families and in the secondwave, letters were sent to 268 families. Postcards indi-cating interest in participating were returned from149 families (47%). Informed consent forms andquestionnaire packets were then sent to all of thesefamilies; a total of 122 complete packets werereturned (82%). After receipt of their survey packets,mothers were interviewed by telephone about theirexperiences with their child prior to and during theperiod surrounding their evaluation and diagnosis.Of the 122 mothers who participated in the largerstudy, 63 had children who were diagnosed withASDs; these are the mothers included in this report.All procedures were approved by the InstitutionalReview Boards of both the medical center and theaffiliated university.

Interviews

Mothers were interviewed over the telephone byfemale graduate students in child clinical psychol-ogy, developmental and child psychology, and coun-seling psychology, all of whom had experienceworking with children with developmental disabil-ities and their families. The interview protocol wassemistructured, intended to capture the nature of

Family Relations � Volume 56, Number 2 � April 2007138

the mothers’ experiences with their child beforea diagnosis was received, during the evaluation anddiagnostic process, and in the time (approximately6 – 12 months) since receiving a firm diagnosis atthe medical center. All interviews were recorded andtranscribed verbatim.

The interview questions were not designed specifi-cally to identify elements of ambiguous loss theorybut to provide insight into family issues that wouldhelp professionals gain a more complete understand-ing of the concerns and needs of parents of childrenwith ASDs (see O’Brien & Daggett, 2006). In read-ing and analyzing the transcripts, individual differen-ces in mothers’ responses to autism became apparent.Efforts to place these differences in theoretical con-text led the author to consider the theory of ambigu-ous loss. The transcripts were then reexamined todetermine whether mothers referred to their experi-ences and feelings in terms that suggested they werestruggling with issues of ambiguity. Thus, the ana-lytic approach used was deductive qualitative analysis(Gilgun, 2005) in which the interviews were studiedwith a specific question in mind. This question con-cerned whether aspects of ambiguous loss theorycould be identified in any of the mothers’ conversa-tions about their children and their experiences deal-ing with their children’s ASDs. In other words, thetranscripts were analyzed for spontaneous expressionsof mothers’ feelings of ambiguous loss in theirexperiences surrounding their child’s diagnosis.Such expressions would indicate that the theory ofambiguous loss appeared to have meaning to thesemothers.

Two readers knowledgeable about the theory ofambiguous loss independently identified interviewsegments as expressing ambiguity, defined (Boss,1999, 2006) to include expressions of conflictingthoughts and feelings (ambivalence); hopelessness orhelplessness (loss of mastery) alternating with hopeor being in control; doubt about one’s identity, role,or relationship with the child; self-blame or incom-petence; anxiety about the future; and uncertainty orconfusion about the child’s diagnosis. Only thoseexcerpts that both readers identified as indicators ofambiguity are used as examples in the present report.

The goal of this deductive qualitative analysiswas to provide evidence regarding the meaningful-ness of the concept of ambiguous loss in mothers’reports of their own and their family’s response toa child’s diagnosis of an ASD. In short, this analysiswas an important first step prior to testing specific

hypotheses derived from ambiguous loss theory inmothers of children with ASDs.

Questionnaire Measures

Demographics. Demographic characteristics ofthe families, including family composition, income,education and employment status, and child charac-teristics, were obtained via maternal report.

Identity ambiguity. Subscales from two question-naires were combined to create an index of identityambiguity: the Impact on Parent subscale of the Ill-ness Perception Questionnaire (Weinman, Petrie,Moss-Morris, & Horne, 1996) and the six-item Par-ent Control subscale of the Health-Specific Locus ofControl measure (Rau & Ware, 1981). The Impacton Parent subscale includes six items assessing theextent to which mothers report themselves to bepreoccupied with their child’s ASD and its impacton themselves. The Parent Control subscale includessix items measuring mothers’ beliefs in their ownresponsibility for and ability to control the outcomefor their child. Both scales were modified so thatitems referred to the child’s disability rather thanthe respondent’s own illness. Each item used a5-point scale ranging from 1 ¼ strongly disagree to 5¼ strongly agree; higher scores indicated greater iden-tity ambiguity. One item from the Parent Controlsubscale (‘‘What I did when my child was a baby haslittle or no relation to his or her problem’’) wasfound not to be correlated with the other items andwas therefore dropped. The alpha for the remaining11 items was an acceptable 0.69. The 11 itemssummed to create the identity ambiguity score listedin Table 1.

Depressive symptoms. Mothers reported on theirsymptoms of depression within the last week usingthe Center for Epidemiological Studies-Depressionscale (Radloff, 1977). This widely used scale lists 20symptoms of depression, with the frequency of expe-riencing each rated on a 4-point scale ranging from0 ¼ less than once to 3 ¼ 5 or more days per week.Cronbach’s a was 0.89 for this sample.

Child-related parenting stress. The parent-childdysfunctional interaction and difficult child sub-scales of the short version of the Parenting StressIndex (Abidin, 1990) were used to assess mothers’perceptions of stress linked directly to their child.This 24-item measure consists of statements describ-ing mothers’ feelings about their child (e.g., ‘‘Mychild doesn’t seem to smile as much as other

Ambiguous Loss and ASD � O’Brien 139

children,’’ and ‘‘My child turned out to be more ofa problem than I expected’’). Each item was scoredon a 5-point scale ranging from strongly disagree tostrongly agree. Higher scores indicated a maternalperception of how difficult and unrewarding it is toparent this particular child. Cronbach’s a for thismeasure in the present sample was 0.91.

Quantitative Analyses

Descriptive data, demographic correlates, and meancomparisons by diagnostic group were analyzed ini-tially. Then, the study hypothesis was tested usinghierarchical multiple regression techniques in whichdiagnostic group, identity ambiguity, and the inter-action between diagnostic group and identity ambi-guity were entered sequentially to predict mothers’reports of depressive symptoms and child-relatedstress.

Results

Indications of Ambiguity in Mothers’ Descriptionsof Their Experiences

The deductive qualitative analysis used to identifystatements suggesting that mothers were experiencingfeelings associated with ambiguous loss revealed fre-quent expressions of conflicting or contradictoryemotions and ideas about the child and about ASDs,confusion about the child’s and their own future,and alternations between hope and hopelessness, feel-ing in control and feeling helpless. More than half

the mothers interviewed made statements that werecoded by both readers as representing ambiguity.Only a few illustrative examples of the mothers’descriptions of their feelings of ambiguous loss arereported herein.

Ambivalence, or the expression of contradictoryfeelings in situations of ambiguous loss, wascommon among respondents. One mother of a7-year-old child recently diagnosed with Aspergersyndrome and sensory integration disorder said,‘‘I’m a little more hopeful now that there’s some-thing that we can do about it . . . [but] some days Iget pretty depressed because I know he’s alwaysgoing to have it.’’ Another mother of a preschool-ageson with AD accompanied by language delays andhigh levels of aggressive behavior expressed her ambi-guity this way: ‘‘I still feel sorrow for my son, and Istill am hopeful.’’ Quite a few mothers expresseda wide range of emotions, similar to those describedby this mother of a 5-year-old boy with AD:

[My feelings] kind of go back and forthdepending on the situation. I mean, sometimesI’m very frustrated when he’s not doing whathe’s supposed to do. Sometimes I feel sadbecause he doesn’t have the social skills thatother kids his age have. And sometimes [I feel]just thrilled because he’s made wonderful prog-ress in the last year.

This mother was typical of many who clearly recog-nized a loss but were uncertain exactly what that lossmeant for them and for their child.

Table 1. Items Making up the Measure of Identity Ambiguity

Consequences for parent subscale of the modified Illness Perception Questionnaire (Weinman et al., 1996)

My child’s condition strongly affects the way I see myself as a person

My child’s problem strongly affects the way others see me

My child’s problem has serious economic and financial consequences for me and my family

My child’s condition is difficult for me to live with

My child’s problem does not have much effect on my life (reverse scored)

My child’s problem will have major consequences on my life

Parent Control subscale of the modified Health-Specific Locus of Control scale (Rau & Ware, 1981)

There is a lot I can do to control my child’s problem

My child’s problem is my own fault

I have a lot of confidence in my ability to help my child overcome his or her problem

There is little I could have done to prevent my child from having his or her condition (reverse scored)

What I do can determine whether my child’s problem gets better or worse

Family Relations � Volume 56, Number 2 � April 2007140

Another aspect of ambiguous loss that was identi-fied in these mothers’ narratives had to do with theuncomfortable feeling of being on a roller coaster,not knowing on any given day whether things wouldgo well or be difficult. ‘‘Some days are good, andsome days are bad’’ was the summary of one motherof a 3-year-old child with significant delays anda diagnosis of PDD-NOS. The ambiguity that sur-rounds the diagnosis of ASDs was reflected bythe mother of a relatively severely affected 3-year-old child who was adopted shortly after birth;she said, ‘‘Autism is just such a Pandora’s box. Thereare just a million different variables every singleday.’’

When children displayed widely varying skills indifferent domains, mothers found themselves con-stantly off-balance, not knowing what level of func-tioning to expect from their child. One mother ofa second grader diagnosed with atypical autismdescribed her son’s inconsistency this way: ‘‘Youknow, autism’s weird. He’s doing math, reallyadvanced math, he’s doing algebra, he’s reading twoyears ahead of his grade, and then he’ll say, �Whichway’s left?’ I don’t understand.’’ The difficulty andambivalence involved in adjusting expectations tothe child’s abilities were at the heart of the concernsof a mother of an 8-year-old child with Aspergersyndrome:

I’ve got real mixed feelings. Part of me may beexpecting too much of him, but then there’spart of me that says he’s very, very bright andhe should be expected to do what is expectedof his age. I don’t want him thinking he canget by in life with less, but if this is all that wecan expect from him, then maybe I am expect-ing too much.

The lack of clarity regarding outcomes for chil-dren with ASDs was also reflected in the mothers’reports of their concerns. Some found themselvescontrasting the likely future for their child againstwhat their expectations had been for their childprior to receiving the diagnosis of an ASD. Onemother whose preschool-age son had only recentlybeen diagnosed with severe autism said that

We were driving the other day and it dawnedon me that, you know, he may never drive.You realize things, like what his life may be

like. And yet you have hope, and you want tomake his life as good as it can be.

Other mothers focused on the realization that thingswould never be the same for anyone in the family.An adoptive mother of a 4-year-old son shared herfeelings of dismay and fear when she said, ‘‘I wentthrough the stage of �I can’t believe this is happen-ing. I don’t want to spend my whole life doingthis.’’’ Another mother, whose kindergarten-age sonwas in the midst of receiving a diagnosis of an ASD,was articulate in describing the ambiguous nature ofher family’s situation:

The frustrating part of it is that I know thatthere isn’t anything that I can do about it, andI feel bad, but then I feel better because I knowthere is nothing that can be done about it.

Finally, the ambiguity led many mothers to feelthey were somehow to blame for the presence of anASD in the family. A typical feeling was expressedby this mother whose son had been diagnosed withAD: ‘‘When I got the diagnosis of autism, I took itpersonally, thinking, did I do something wrongwhen I was pregnant?’’ Another mother clearly feltresponsible for making things better for her 11-year-old son but saw no clear path to follow: ‘‘You feellike it’s the parents’ job and you’re supposed toknow what to do. You’re supposed to be able to helpthis and you can’t.’’ One mother seemed to sum upwhat many were feeling when she said, ‘‘I wonder. . . what things would be like if he were different orsomething.’’

Relation of Identity Ambiguity to Stress andDepressive Symptoms

The qualitative portion of the study indicated thatmothers of children diagnosed with ASDs have feel-ings that fit constructs arising from the theory ofambiguous loss. The next step involved testing a spe-cific question derived from ambiguous loss theory.Therefore, quantitative analyses were conducted todetermine whether identity ambiguity, or the blur-ring of roles and responsibilities between motherand child, was related to mothers’ perceptions ofstress related to parenting their child with an ASDand to their reports of depressive symptoms. Forthese analyses, families were divided into two groupson the basis of the number and severity of ASD

Ambiguous Loss and ASD � O’Brien 141

symptoms displayed by the child. Children diag-nosed with AD had significant impairments in allthree core areas of autism; the children in the‘‘other’’ group had fewer symptoms and their symp-toms were less severe.

Descriptive data. Demographic factors were notconsistently related to the constructs of interest. Theonly significant correlations between demographicfactors and study variables were for depressive symp-toms; mothers of older children (r ¼ 20.25, p ¼.048), older mothers (r ¼ 20.27, p ¼ .034), andmothers with higher levels of education (r ¼ 20.41,p ¼ .001) reported fewer symptoms of depression.Child gender, family income, and partner statuswere not related to any of the study variables.

Descriptive data on measures of identity ambigu-ity, depressive symptoms, and stress by diagnosticgroup are shown in Table 2. Comparisons of themeans using t tests indicated no statistically signifi-cant differences, but a marginal difference in depres-sive symptoms, with the mothers of more severelyaffected children (the ‘‘AD’’ group) showing some-what higher depressive symptoms. The mean scorefor this group was higher than the score of 16defined as the cutoff for further clinical evaluation(Radloff, 1977). However, the lack of significantdifferences between the groups indicated that symp-tom severity was not a major factor in mothers’experience of identity ambiguity.

Tests of the study hypothesis. Results of the testsof the study hypotheses are shown in Table 3. Theanalysis for depressive symptoms included child ageand maternal education as control variables becausepreliminary analyses indicated these factors to becorrelated with reports of depressive symptoms(maternal age was correlated with both child age andeducation level and so was not included). In bothmodels, diagnostic group was entered first, thenidentity ambiguity, and finally, the interactionbetween diagnostic group and identity ambiguity.

The hypotheses tested were that identity ambiguitywould be linked to depressive symptoms and stress,and that child symptom severity, defined by diag-nostic group, would not moderate the effect. Thatis, the prediction was that identity ambiguity wouldbe related to negative outcomes equally in bothgroups.

The analysis for depressive symptoms indicatedthat identity ambiguity contributed significantly todepressive symptoms, and this effect was the samefor both diagnostic groups. Identity ambiguityaccounted for 14% of unique variance in the model.Results of the analysis for child-related stress weresimilar; identity ambiguity again was a significantpredictor of stress across both groups of mothers,accounting for 23% of the variance. The findingthat the severity of the child’s ASD was not relatedto stress or depressive symptoms and did not moder-ate the relation between identity ambiguity andthese outcomes, strongly suggests that maternal dis-tress is linked directly to identity ambiguity arisingfrom the situation of ambiguous loss.

Discussion

In this study, mothers’ narrative reports and theirsurvey responses were used to examine the utility ofthe application of the theory of ambiguous loss tofamilies of children diagnosed with ASDs. Bothqualitative and quantitative results indicated that thetheory of ambiguous loss has direct relevance to theexperiences of these families. The inherent ambigu-ity of autism, related to the lack of knowledge sur-rounding its etiology, symptoms, treatment, andoutcome, suggests that parents whose children arediagnosed with ASDs experience conflicting realitiesand emotions. Coming to terms with the discrepan-cies between their initial expectations for their childand family and the altered outlook that is required

Table 2. Mean Scores for Measures of Identity Ambiguity, Depressive Symptoms, and Child-Related Stress byDiagnostic Group

Autistic Disorder (n ¼ 32) Other ASD Diagnosis (n ¼ 31)

M SD M SD t (63)

Identity ambiguity 35.15 6.49 32.52 6.12 1.66, p ¼ .103

Depressive symptoms 16.89 11.06 11.73 10.98 1.86, p ¼ .068

Child-related stress 64.13 15.42 62.21 15.07 0.50, p ¼ .620

Note. ASD ¼ autism spectrum disorder.

Family Relations � Volume 56, Number 2 � April 2007142

for a family to continue to function well after a childreceives an ASD diagnosis is a process that few fami-lies can be expected to negotiate easily.

According to ambiguous loss theory, when peo-ple experience ambiguous loss, they have difficultymaking decisions, are unable to alter family rolesand routines even when these old ways of doingthings do not fit their new situation, question theirbeliefs and competence, and become exhausted(Boss, 1999, 2004, 2006). It is the ambiguity of thesituation, not the psychological characteristics of theindividual that contributes to perceptions of stressand inadequate functioning (Boss, 1999, 2006). Inthe narratives of mothers of children with ASDs,expressions of stress arising from ambiguity werecommon. Many mothers, when asked about theirfeelings at different points after receiving theirchild’s diagnosis, reported conflicting and contradic-tory emotions: fear and hope, helplessness and deter-mination, frustration and joy. When looking tothe future, mothers tended to cling as well to theirimages and expectations from the past. Althoughnot all the mothers interviewed expressed feelings

that could be described as arising from ambiguousloss, more than half did so. Thus, it seems evidentthat the theory of ambiguous loss has direct rele-vance to many families of children with ASDs.

The second component of this study testedhypotheses derived from ambiguous loss theory:Identity ambiguity would be related to mothers’depressive symptoms and perceptions of stress con-nected to her child with an ASD, and this relationwould not vary with the severity of the child’s disor-der. For this analysis, identity ambiguity was definedas the extent of overlap mothers perceived betweentheir own identities and those of their children withASDs. As predicted by ambiguous loss theory,mothers with high identity ambiguity also reportedmore symptoms of depression and perceptions ofhigher stress related to their child with an ASD thanthose who were less personally identified with theirchild’s condition.

The construct of identity ambiguity as definedhere is distinct from perceptions of stress or ofpersonal well-being. Identity ambiguity is viewed asconfusion between one’s own identity and that of

Table 3. Hierarchical Linear Regression Analyses Examining Identity Ambiguity and Psychological Distress for Mothers ofChildren Diagnosed With Autistic Disorder and Another ASD Subtype

Depressive Symptoms

B SE B b DR2

Controls

Child age 20.09 0.04 20.23*

Maternal education 23.00 0.86 20.40** 0.222**

Diagnostic groupa 23.10 2.62 20.14 0.018

Identity ambiguity 0.68 0.19 0.39*** 0.143***

Diagnostic group �Identity ambiguity

0.04 0.39 0.03 0.000

Adjusted R2 ¼ 0.33, F(5, 57) ¼ 7.07, p , .001.

Perception of Child-Related Stress

B SE B b DR2

Diagnostic groupa 21.92 3.84 20.06 0.004

Identity ambiguity 1.16 0.27 0.49*** 0.230***

Diagnostic group �Identity ambiguity

20.56 0.60 20.37 0.013

Adjusted R2 ¼ 0.21, F(3, 59) ¼ 6.45, p , .001.

Note. ASD ¼ autism spectrum disorder. Betas reported at point of entry.aDiagnostic group: 1 ¼ autistic disorder , 2 ¼ other ASD diagnosis.*p , .05. **p , .01. ***p , .001.

Ambiguous Loss and ASD � O’Brien 143

another family member. Although the scales usedhere to index identity ambiguity were not originallydeveloped to measure this construct, the items aredescriptive of a situation in which a mother is notclear where her identity ends and her child’s begins.Some of the items used, especially those from thelocus of control measure, may involve issues of mas-tery (Boss, 2006) rather than identity ambiguity.Further development of measures to examine con-structs that are described by the theory of ambiguousloss, including boundary ambiguity, identity ambi-guity, and mastery in families of children withASDs, would be helpful.

The reality of ASDs is that they are lifelong andpervasive; nevertheless, with effective interventionand appropriate supports, many individuals withASDs are able to function at home, in school, andin the community (Howlin et al., 2004). It is veryclear that ASDs do not result from poor or inade-quate parenting but instead are genetic disorders,possibly triggered by pre- or postnatal environmen-tal events that are out of the control of parents(Rutter, 2000). Mothers are therefore not the causeof their children’s ASDs nor can they alone affectthe outcomes for their children. Thus, a mother isexpressing identity ambiguity when she feels herown life is so entwined with that of her child thatshe is somehow responsible for his or her ASD orshould be able to determine the outcome for herchild.

Given the limitations of this study, including itscross-sectional nature, the fact that all of the mea-sures were collected from mothers, and the lack ofa measure developed specifically to index identityambiguity, the findings reported here can only beginto suggest the potential value of a continuing pro-gram of research on ambiguous loss in family adjust-ment to the diagnosis of ASDs. It is clear thatlongitudinal work, with independent and multiplemeasures of both identity ambiguity and psychologi-cal well-being, is needed to examine change inmothers’ perceptions of their role as related tochanges in stress and depressive symptoms and alsoto parents’ decision making regarding interventionservices for their child. Research must also beextended beyond mothers to include fathers, sib-lings, and extended family members, all of whomcan be expected to experience ambiguous loss tosome degree and to influence the responses of oneanother. In addition, the measure used in this study,although it addresses important aspects of the

construct, was not originally intended to assess iden-tity ambiguity. The development and testing ofmeasures of identity ambiguity and other aspects ofboundary ambiguity that address issues specific toparents of children with ASD would be an impor-tant step toward clarifying the experience of ambigu-ous loss in this population.

Limitations

The theory of ambiguous loss has generally beenapplied to families in which an adult member ismissing from the family, either physically or psycho-logically (Boss, 1977, 1980; Boss & Couden, 2002;Boss et al., 1990; Caron et al., 1999), and less oftento circumstances in which a child is the focus. Inone child-focused interview study, DeYoung andBuzzi (2003) examined differences in emotional andcoping responses between parents of children whohad been murdered and parents of children who hadbeen abducted and were still missing 6 – 16 yearslater. Parents of murdered children reported them-selves to feel ‘‘better off’’ than those of missing chil-dren because they could resolve their loss and sharetheir grief with other parents in similar situations.Parents of missing children reported the need toremain hopeful while at the same time moving for-ward with their own lives. In another study focusingon parents, Golish and Powell (2003) used the con-struct of ambiguous loss to describe the emotionalresponses and communication patterns of parentswhose infants were born prematurely. Prematurebirth is typically accompanied by a period of timewhen parents are unsure about the child’s eventualhealth, and even survival; thus, the time surroundingearly birth is clearly one of ambiguity for parentsand other family members.

The diagnosis of an ASD in a child is similar insome ways to each of these situations but has uniqueelements as well. When a child has an ASD, thechild is not physically absent but psychologically dif-ferent from the child the parents expected, and theparents are called upon to adapt to those differenceson a day-to-day basis. By comparison with prematu-rity, an autism diagnosis is not a time-constrainedsituation but one in which parents must live withthe contradiction inherent in recognizing the irre-versible nature of their child’s condition while main-taining a long-term commitment to the child’supbringing. Thus, families of children with ASDs

Family Relations � Volume 56, Number 2 � April 2007144

can be expected to experience ambiguity and lossover many years and to need support in their ef-forts to live with the emotions that often emerge ateach new phase of their child’s development.

Implications for Practice

Labeling the experience of ambiguous loss for parentsof children with ASDs is important, because, in mostcases, parents are not aware that their distress, confu-sion, and immobility are located in the inherentlyambiguous situation and not in themselves. As hasbeen often noted by Boss (1999, 2004, 2006; Boss &Couden, 2002), identifying the source of the diffi-culty in the loss that people have experienced can bea major step toward helping individuals reorganizetheir thinking and thus begin to employ their usualcoping mechanisms rather than remaining ‘‘stuck’’ ina pattern that is no longer appropriate to theirchanged family circumstances. The theory of ambigu-ous loss predicts that parents’ ability to make deci-sions with regard to gaining access to services fortheir child could be impaired as long as they areunable to come to terms with the ambiguity of ASDs.In such cases, helping parents increase their tolerancefor living with ambiguous loss so that children canreceive appropriate services may have long-termimplications for children’s outcomes.

An understanding of ambiguous loss could helpprofessionals who work with families of childrenwith ASDs to be more supportive and empathetic.Even today, despite considerable empirical workshowing that ‘‘grief work’’ is not effective in helpingpeople cope with traumatic loss (see review inBonanno, 2004), many professionals who workwith families of children with disabilities expect par-ents to progress through grief stages (Shuchter &Zisook, 1993). Another common belief amongpractitioners is that psychologically healthy parentswill eventually achieve resolution following a child’sdiagnosis (Pianta, Marvin, Britner, & Borowitz,1996). Both of these models assume linear move-ment from psychological distress to a form of clo-sure accompanied by a return to stability andpositive functioning. Many professionals alsoassume that all families will follow a similar path,and, when a family does not meet this expectation,the family is often labeled dysfunctional or ‘‘indenial’’ (Stroebe, 2001).

As noted by Boss (2004, 2006), however, there isno closure or resolution in cases of ambiguous loss.

Although parents may accept and cope with thepresence of an ASD in their lives, they will neverentirely let go of their image of their child beforethey learned the diagnosis. To make the best choicesamong educational and intervention services fortheir child, parents must be aware that autism isa severe and lifelong condition while at the sametime holding out hope for improvement in theirchild’s symptoms. These beliefs are paradoxical butcan energize parents to seek ways to move towardnew goals for their child rather than becomingimmobilized and isolated.

The diagnosis of an ASD in a child presents fam-ilies with difficult to resolve loss issues. Using thetheory of ambiguous loss as a framework for thestudy of families of children with disabilities holdsgreat promise for advancing our understanding ofthe challenges these families face and their resiliencein learning to live with ambiguity.

References

Abidin, R. R. (1990). Parenting stress index (3rd ed.). Charlottesville, VA:Pediatric Psychology Press.

American Psychiatric Association. (2000). Diagnostic and statistical manualof mental disorders (4th ed., text revision). Arlington, VA: Author.

Baker, B., Blacher, J., Crnic, K., & Edelbrock, C. (2002). Behavior prob-lems and parenting stress in families of three-year-old children with andwithout developmental delays. American Journal on Mental Retardation,107, 433–444.

Bonanno, G. A. (2004). Loss, trauma, and human resilience: Have weunderestimated the human capacity to thrive after extremely aversiveevents? American Psychologist, 59, 20–28.

Booth, C., & Kelly, J. (1999). Child care and employment in relation toinfants’ disabilities and risk factors. American Journal on Mental Retar-dation, 104, 117–130.

Boss, P. (1977). A clarification of the concept of psychological father pres-ence in families experiencing ambiguity of boundary. Journal of Mar-riage and the Family, 39, 141–151.

Boss, P. (1980). The relationship of psychological father presence, wife’spersonal qualities, and wife/family dysfunction in families of missingfathers. Journal of Marriage and the Family, 42, 541–549.

Boss, P. (1988). Family stress management. Thousand Oaks, CA: Sage.Boss, P. (1999). Ambiguous loss. Cambridge, MA: Harvard University

Press.Boss, P. (2004). Ambiguous loss research, theory, and practice: Reflections

after 9/11. Journal of Marriage and Family, 66, 551–566.Boss, P. (2006). Loss, trauma, and resilience: Therapeutic work with ambigu-

ous loss. New York: Norton.Boss, P., Caron, W., Horbal, J., & Mortimer, J. (1990). Predictors of

depression in caregivers of dementia patients: Boundary ambiguity andmastery. Family Process, 29, 245–254.

Boss, P., & Couden, B. (2002). Ambiguous loss from chronic physicalillness: Clinical interventions with individuals, couples, and families.Psychotherapy in Practice, 58, 1351–1360.

Boss, P., & Greenberg, J. (1984). Family boundary ambiguity: A new vari-able in family stress theory. Family Process, 23, 535–546.

Caron, W., Boss, P., & Mortimer, J. (1999). Family boundary ambiguitypredicts Alzheimer’s outcomes. Psychiatry, 62, 347–356.

Ambiguous Loss and ASD � O’Brien 145

Coplan, J. (2003). Atypicality, intelligence, and age: A conceptual model ofautistic spectrum disorder. Developmental Medicine and Child Neurol-ogy, 45, 712–716.

DeYoung, R., & Buzzi, B. (2003). Ultimate coping strategies: The differen-ces among parents of murdered or abducted, long-term missing chil-dren. Omega, 47, 343–360.

Dunst, C., Trivette, C., & Cross, A. (1986). Mediating influences of socialsupport: Personal, family, and child outcomes. American Journal ofMental Deficiency, 90, 207–218.

Gilgun, J. F. (2005). Qualitative research and family psychology. Journal ofFamily Psychology, 19, 40–50.

Gillberg, C. (2002). A guide to Asperger syndrome. Cambridge, UK:Cambridge University Press.

Goldberg, S., Morris, P., Simmons, R., Fowler, R., & Levison, H.(1990). Chronic illness in infancy and parenting stress: A comparisonof three groups of parents. Journal of Pediatric Psychology, 15,347–358.

Golish, T., & Powell, K. (2003). Ambiguous loss: Managing the dialecticsof grief associated with premature birth. Journal of Social and PersonalRelationships, 20, 309–334.

Hastings, R., & Taunt, H. (2002). Positive perceptions in families of chil-dren with developmental disabilities. American Journal on Mental Retar-dation, 107, 116–127.

Hauser-Cram, P., Warfield, M., Shonkoff, J., & Krauss, M. (2001). Chil-dren with disabilities: A longitudinal study of child development andparent well-being. Monographs of the Society for Research in Child Devel-opment, 66, 1–131.

Howlin, P., Goode, S., Hutton, J., & Rutter, M. (2004). Adult outcomefor children with autism. Journal of Child Psychology and Psychiatry, 45,212–229.

Ireys, H., & Silver, E. (1996). Perception of the impact of a child’s chronicillness: Does it predict maternal mental health? Developmental andBehavioral Pediatrics, 17, 77–83.

Kaplan, L., & Boss, P. (1999). Depressive symptoms among spousal care-givers of institutionalized mates with Alzheimer’s: Boundary ambiguityand mastery as predictors. Family Process, 38, 85–103.

Kazak, A., & Marvin, R. (1984). Differences, difficulties and adaptation:Stress and social networks in families with a handicapped child. FamilyRelations, 33, 67–77.

Minnes, P. (1998). Mental retardation: The impact upon the family. InJ. Burack, R. Hodapp, & E. Zigler (Eds.), Handbook of mental retarda-tion and development (pp. 693–712). New York: Cambridge UniversityPress.

Noh, S., Dumas, J., Wolf, L., & Fisman, S. (1989). Delineating sources ofstress in parents of exceptional children. Family Relations, 38, 456–461.

O’Brien, M., & Daggett, J. (2006). Beyond the autism diagnosis: A professio-nal’s guide to helping families. Baltimore: Paul H. Brookes Publishing.

Ozonoff, S., & Rogers, S. (2003). From Kanner to the millennium: Scientificadvances that have shaped clinical practice. In S. Ozonoff, S. Rogers, &R. Hendren (Eds.), Autism spectrum disorders: A research review for practi-tioners (pp. 3–33). Washington, DC: American Psychiatric Publishing.

Pelchat, D., Ricard, N., Bouchard, J., Perreault, M., Saucier, J.,Berthiaume, M., & Bisson, J. (1999). Adaptation of parents in relationto their 6-month-old infant’s type of disability. Child: Care, Health andDevelopment, 25, 377–397.

Pianta, R., Marvin, R., Britner, P., & Borowitz, K. (1996). Mothers’ reso-lution of their children’s diagnosis: Organized patterns of caregivingrepresentations. Infant Mental Health Journal, 17, 239–256.

Radloff, L. (1977). The CES-D scale: A self-report depression scale forresearch in the general population. Applied Psychological Measurement,1, 385–401.

Rau, R., & Ware, J. F., Jr. (1981). Refinements in the measurement ofhealth-specific locus of control.Medical Care, 19, 1147–1158.

Rutter, M. (2000). Genetic studies of autism: From the 1970s into the mil-lennium. Journal of Abnormal Child Psychology, 28, 3–14.

Shastry, B. (2005). Recent advances in the genetics of autism spectrum dis-orders: A minireview. British Journal of Developmental Disabilities, 51,129–142.

Shuchter S., & Zisook, S. (1993). The course of normal grief. In M.Stroebe, W. Stroebe, & R. Hansson (Eds.), Handbook of bereavement:Theory, research, and intervention (pp. 23–43). New York: CambridgeUniversity Press.

Sloper, P., & Turner, S. (1993). Risk and resistance factors in the adapta-tion of parents of children with severe physical disability. Journal ofChild Psychology and Psychiatry, 34, 167–188.

Stroebe, M. (2001). Bereavement research and theory: Retrospective andprospective. American Behavioral Scientist, 44, 854–865.

Trute, B., & Hiebert-Murphy, D. (2002). Family adjustment to childhooddevelopmental disability: A measure of parent appraisal of familyimpacts. Journal of Pediatric Psychology, 27, 271–280.

Wallander, J., & Varni, J. (1998). Effects of pediatric chronic physical dis-orders on child and family adjustment. Journal of Child Psychology andPsychiatry, 39, 29–46.

Weinman, J., Petrie, K., Moss-Morris, R., & Horne, R. (1996). The illnessperception questionnaire: A new method for assessing the cognitive rep-resentation of illness. Psychology & Health, 11, 431–455.

Family Relations � Volume 56, Number 2 � April 2007146