Reactive Attachment Disorder: A Disorder of Attachment or of Temperament?

It seems that there should be difficulty in assessing Reactive Attachment Disorder (RAD) because attachment theory itself may be flawed. There is no way of telling whether the symptoms of RAD are the result of the attachment style or of the child's genetic behavioral disposition. Those trying to diagnose RAD have the assumption that the parent-child relationship actually does influence attachment style.

The difficulties experienced in trying to conceptualize RAD may be because there is no way of linking the symptoms with their cause. RAD may not be an attachment disorder because the symptoms described may result from the child's genetic composition. The sensitive or insensitive treatment by the parents during infancy may not even play a role in the child's development of RAD. It may be merely be a behavior resulting primarily from genetic dispositions.

If it is the case that the symptoms described in the DSM are not caused by attachment then RAD would not be an attachment disorder but just a genetic result of personality. There should be consideration of how genes and the treatment of the primary care giver interact. The methods of genetic transmission should be looked at in more detail if we are even to consider RAD an attachment disorder.

There are three models of genetic transmission that can occur (Reiss, 1997). If two of these models are true then this would mean that attachment theory cannot work and RAD would be the result of genetic behaviors and not an attachment disorder. The three models of genetic transmission are briefly described below.

First is the passive model. In the passive model of transmission the genes that influence the parent's sensitivity are passed down to the child. Those same genes are responsible for the child's attachment style. The genes are the reason for both occurrences and attachment style is not caused by the parent's sensitivity. The tendency for the parent to be sensitive to the child is because of their genes and not the environment. The attachment style of the child is because of genes and not because of the parent's treatment of the child. The sensitivity of the parent and the attachment style of the child are completely independent.

Second is the child-effects model. In this model the child's attachment style is caused by genes. It is then the child's attachment style that causes the parent to be sensitive or insensitive. In this model the parent's sensitivity is caused by the child but the child's attachment style is still due to their genetic composition. The child's attachment style is the result of purely genetics the parents sensitivity is only a response to the child's attachment style. Therefore the child causes the parents sensitivity but this does not have any influence on the child's attachment style.

Third is the parent-effects model. In this model the child's behavior is caused by their genes. The child's behavior causes the parent to act either sensitive or insensitive. This in turn will influence what attachment style the child will have. The child's behavior is cause by their genetics but their attachment style is not. The child's behavior evokes the sensitive attitude from the parent. The sensitivity of the parent does determine what the child's attachment style is. This is the only model that allows a parents behavior to mold a child's attachment style and is the only possibility for attachment theory.

Because there are three different ways genetics and sensitive responding by the parents can result in attachment style, it is uncertain which model is accurate. The cause of RAD is theoretically the result of a disruption of the attachment process. There is much debate as to whether the attachment process even has an effect on personality.

The symptoms observed may be the result of genes. This may lead us to believe that RAD is not an attachment disorder at all. Its symptoms are possibly caused by other sources than childhood attachment and it may be false to label RAD as an "attachment disorder." There is poor evidence supporting that attachment is the true cause of RAD.

The difficulties in conceptualization of RAD are because its causes are uncertain. It may not be feasible to diagnose RAD because we are unsure of its true cause. It may be false to classify a disorder with little understanding of the concept and evidence that is inconclusive. The DSM may be presumptuous its definition of RAD.

It is obvious from Wood's analysis that reactive attachment disorder manifests itself during the infancy. But what separates these infants when they are older from the rest of us? How do their relationships and development differ from ours?

A child who has a secure attachment style will explore more of the surrounding world. The child has the confidence that someone will be there to protect him or her if a threat arises during exploration. As Wood stated, children suffering from RAD do not have a secure attachment style and do not freely explore their world, as they are afraid no one will be there for them.

Larose, Bernier, and Tarabulsy (2004) set out to answer those questions. Sixty-two college students were interviewed and the relation between their attachment style, learning dispositions, and academic performance during the transition to college were evaluated. Toddlers with insecure attachments are less enthusiastic, less effective, and show less endurance during a challenging task than secure toddlers (Larose et al., 2004). In previous studies, infants with secure attachments had better academic performance during adolescence, children who were securely attached at age 7 got higher school grades than insecure children from ages 7 to 17, and an insecure attachment style was related to a drop in school marks during the transition from high school to college.

The findings of the Larose et al. (2004) study indicated that secure attachment style protects the student against the negative impact the college transition has on learning style. The results varied based on the type of insecurity. For example, securely attached students experienced more fear of failure at the middle of the first semester in college than at the end of high school, felt less comfortable seeking help from teachers, and gave less priority to their studies. Additionally, insecurely attached students had a decrease in exam preparation and quality of attention. The qualities that pertain to securely attached students relate to emotional and social components, whereas those that pertain to insecurely attached students relate to practical, school-related behaviors.

To summarize the findings, the security of attachment in late adolescence favors security of exploration by providing the student with emotional, cognitive, and behavioral resources that have been shown to favor college success. Students with a more secure attachment style have more characteristics that favor college success than students with an insecure attachment style. To further support this, Larose et al. (2004) found that students with an insecure attachment style had lower grades throughout the first three college semesters.

The Larose et al. (2004) study illustrates how not only the future relationships of an infant with RAD are affected, but also how academic performance is affected in college. The study did not focus on children suffering from RAD specifically, but rather children who were insecurely attached for any reason. Because infants who suffer from RAD are insecurely attached, it can be inferred that RAD in infants predicts poor academic performance in future years.

In "Reactive Attachment Disorder: A Disorder of Attachment or of Temperament?" Wood raised many valid arguments for this new disorder, which is not yet completely understood. The disorder itself is new and is fairly undefined. It is generally difficult to diagnose infants and gage whether or not their behaviors are abnormal. Unlike older people, infants cannot report to the psychologist (or whomever) if they believe that they are experiencing symptoms of RAD. Moreover, maladaptive caretaking can be very difficult to detect or control.

Both the DSM-IV and ICD-10's criteria for the disorder are vague and do not draw definite--or even moderately clear--distinctions for when infants' behavior crosses from normal to abnormal. Not only are the criteria for RAD vague, but also they are subjective. The infant clearly cannot state whether he or she experiences symptoms such as a "lack of comfort seeking for distress," "emotion regulation difficulties," or " a willingness to go off with relative strangers," to name a few. The paper also addresses the issue of the temperament of the child and its relation to RAD. What may seem like RAD may have nothing to do with possible abuse, but may be just part of the child's natural temperament. Some children who naturally possess more difficult temperaments may show signs of RAD. The vague criteria can lead to false diagnoses, such as in cases of hypersensitive caretakers or children who originally possess more problematic temperaments. Moreover, given that the diagnosis for RAD includes the presence of a maladaptive caregiver, if a child develops symptoms that are characteristic of RAD but has not suffered any form of abuse, what diagnosis would the child receive?

I was also confused as whether RAD is an infant disorder or a childhood disorder that can be applied to children starting as early as infancy. The behavorial symptoms (some of which apparently include cruelty to animals and other people, destruction of property, gorging on food, abnormal speech patterns, inappropriate sexual behavior) are very hard for infants to perform, so I assume that RAD is not merely a disorder of infancy but also of childhood. Does RAD only develop in infancy, or is infancy merely the typical time of onset? Can RAD develop after infancy, in older children, or even adolescence? It would be interesting to see whether gender plays a role in the development of RAD.

The paper stated that RAD is based on the inability of children to form normal attachments but does not specify whether the children have difficulty forming attachment exclusively with their primary caretaker or if the difficulty extends to the other family members and peers as well. The paper also did not specify if the child is unable to form normal attachments only in relation to the maladaptive caregiver, all other caregivers and family members, or with whomever else they come in contact on a regular basis.

Another loophole of this paper was the lack of discussion pertaining to treatments or therapies. Treatment is a very important aspect, and it was odd that such an important part of psychology was almost completely excluded from the paper. There was also no mention of possible hereditary or genetic influences, which are also very important and would give great insight to the disorder.

The limitations of this paper, though, may not have been Wood's fault. Because RAD is a new disorder, perhaps the literature on RAD just does not cover aspects such as treatment, heredity, prevalence, gender, or age of onset. I think that RAD, as a disorder, needs to undergo much more research to be understood as a valid disorder. Though the nature of this disorder makes it difficult to study, perhaps future classification modifications and research will give more concrete insight into this new disorder.

"Reactive Attachment Disorder: A Disorder of Attachment or of Temperament?" provided an interesting view of RAD summarized by its analogy to the diathesis-stress model. The focal point of the relatively small amount of research on RAD is, in most cases, on infants and young children. This paper mentioned a question raised by Zeanah (1996), specifically "When do risk factors (insecure attachments) become clinical disorders (attachment disorders)?" Though the paper did not fully addressed this point, the question warrants further investigation. Specifically, when combined with the question of what happens to individuals diagnosed with RAD after childhood, it leads to a theory that a link must exist between RAD and personality disorders.

Though little research has been conducted to determine the existence of RAD in adults, it is possible to draw connections between its symptoms and those of personality disorders that have already been recorded in adults. Although RAD may not persist through adolescence and into adulthood, perhaps due to the decrease in a need for attachment to parental figures, the lasting effects may manifest in different forms. Taking the argument one step further and considering the connection drawn between RAD and temperament results in the possibility that temperament is the underlying cause of both RAD and later personality disorders. In other words, as an individual with a temperament that acts as a diathesis to RAD ages beyond childhood, the temperament remains consistent, although it is analyzed and diagnosed differently.

To further illustrate this concept, it may be helpful to consider the similarities between the primary subtypes of RAD and Cluster A and B personality disorders. Children diagnosed with the inhibited form of RAD tend to display apathy toward comfort offered by caregivers and are generally withdrawn. Often these symptoms result from abuse. The tendency to withdraw may be a result of a fear of repeated abuse or of a mistrust of caregivers. These symptoms seem to mirror the characteristics of Cluster A personality disorders, primarily paranoid and schizoid personality disorders. Although paranoid personality disorder accounts for mistrust of others, the most obvious link is to schizoid personality disorder, which is defined by an impaired ability to form proper social relationships due to a lack of desire to form attachments to others.

Similar connections are evident when comparing the disinhibited subtype of RAD to Cluster B personality disorders. As described by Wood, disinhibited individuals tend to display the same high level of comfort and social interaction with all adults, including strangers. This behavior is similar to the relationship behaviors demonstrated by individuals with histrionic personality disorder. Primarily the connection can be drawn between the disinhibited individual's ability to generate a form of attachment with any adult, and the histrionic individual's tendency to consider relationships to be deeper than they actually are.

The concept relating RAD and temperament could be further illustrated by analyzing the ways in which that existing temperament may persist and result in other personality disorders later in life. This theory is clarified further when connections are made between the two major subtypes of RAD and Cluster A and B personality disorders. More specifically, inhibited subtypes tend to display characteristics similar to schizoid personality disorder, and disinhibited subtypes share a connection to histrionic personality disorder.

In the paper "Reactive Attachment Disorder: A Disorder of Attachment or of Temperament?" Wood maked some very strong conclusions, not all of which are supported by the paper. First, Wood advocated using Zeanah's suggestions to modify the DSM diagnostic criteria for Reactive Attachment Disorder. One of Zeanah's main suggestions was that the requirement that the disordered attachment be exhibited across a wide variety of social contexts be dropped. Zeanah suggested that attachment may be expressed differentially in relationships. Wood stated that some children may have no attachments at all, whereas others may form disturbed attachments to caregivers. Having no attachment and disturbed attachment are both signs of an attachment disorder. Part of the severity of Reactive Attachment Disorder comes from the fact that it affects most of the relationships a child forms or is not able to form as a consequence of the disorder. If the requirement that disordered attachments be exhibited across a wide variety of social contexts were dropped, then a child who failed to form a secure attachment with an abusive caregiver but had many secure relationships with friends would be a candidate for RAD. This however may not be an example of RAD, as it would be expected that a child with an abusive or neglectful parent would form a disordered attachment to the others too. The extreme pervasiveness of the disorder, causing disrupted relationships outside the family, is at the heart of Reactive Attachment Disorder. It may not be a wise decision to drop the disordered attachments across a wide variety of social contexts requirement from the DSM, as this would lessen the severity of the disorder.

Second, Wood made somewhat contradictory conclusions with regard to temperament. The first conclusion was that temperament acts as a diathesis for RAD, whereas the second conclusion was that temperament is a deciding factor between which subtype of RAD is developed. If temperament is a diathesis to developing RAD, it would most likely be a difficult temperament according to Thomas and Chess's model of temperament. If a difficult temperament would put a child at greater risk for RAD, then as Zeanah also suggested a temperament characterized by positive affect could serve as a protective factor for RAD. Wood suggested that children with a difficult or more negative temperament when exposed to pathogenic caregiving would develop RAD. This suggests that children without this poor temperament would not develop RAD in the same circumstances. This conclusion is fine on its own, but the next conclusion was that temperament is the determining factor as to which subtype of RAD is developed. If only children with difficult temperaments are susceptible to RAD, then how can temperament determine the type of RAD developed? This would suggest that the subtype associated with a negative temperament would be much more common than the other because these children are more susceptible to RAD. This however is not the case. Wood gave no indication as to what kind of temperament would develop which subtype. A more thorough exploration as to which dimensions of temperament play the most role in developing RAD would be needed to support this conclusion. Additionally, it was stated earlier in the paper that most children who grow up in institutions have disinhibited RAD, whereas children with abusive parents tend to have inhibited RAD. This suggests that the type of RAD developed is most likely caused by the type of stress, either neglect or abuse, not the type of temperament of the child.

Finally, Wood made no mention of treatment for Reactive Attachment Disorder. Learning about what treatments have been effective can aid in developing proper theories for the disorder. If no treatments have been effective in treating the disorder, then maybe Reactive Attachment Disorder is more like a personality disorder in children. The two subtypes described sound eerily similar to common adult personality disorders. The inhibited subtype is characterized by being emotionally withdrawn and rarely responding to or seeking comfort. This sounds very similar to schizoid personality disorder. The disinhibited subtype is characterized by being overly sociable and eliciting comfort and affection non-selectively. This sounds very similar to histrionic personality disorder. If no treatments are shown to be effective for treating RAD, then maybe RAD could better be characterized as a personality disorder of children rather than an Axis I DSM disorder.

Balduf brought up the argument that RAD may not be caused by attachment issues at all but by genetic influences. He presented three models of genetic transmission that could explain the causality of the disorder as being entirely genetic. Although I think that genetics may a very important factor in the development of RAD, I disagree that attachment is not a factor. Also, I think that the best way to improve the conceptualization of RAD is for psychologists to make more definite decisions as to which symptoms must be observed for the disorder to be diagnosed. Although determining causal factors may provide enlightenment into the development of the disorder and possible treatments, for purposes of reliable diagnosis the disorder must first be operationally defined. I hope subsequent editions of the DSM will provide improved definitions of the disorder and what it entails.

Howard's look at a study of adolescents with insecure attachment provided some good insight into possible long-term effects that may befall children who have RAD. I think the findings of the study show how important it is to address RAD, as it has significant implications for the well-being of these children into adulthood.

Huang brought up a variety of important concerns regarding my presentation of RAD and whether it is a valid disorder. Current definitions of RAD as published in the DSM-IV and ICD-10 are somewhat unclear and thus do not provide clinicians with optimal reliability in diagnosing it as a disorder. Certainly, children so young are unable to easily speak for themselves with regard to difficulties with emotion regulation or other symptoms associated with the disorder; diagnoses are most likely based on parent reports and clinical observations made by psychologists. As for whether RAD is a disorder of infancy or childhood, I think that it is both. What is unique about this disorder is that it can be applied to individuals so young. I think the onset of the disorder is likely to occur during infancy and continues to manifest itself through a variety of symptoms during childhood. I agree with Huang's statement that treatment of the disorder is very important, as are genetic influences on the pathogenesis of RAD. The reason these issues were not addressed in my paper is that I simply did not find any articles addressing them. The majority of research on RAD to date deals with characterizing the disorder operationally and with examining groups that seem to have a higher prevalence of the disorder. The only information I found that addresses the possibility of genetic influence is the article written by Zeanah and Fox (2004) that merely suggested the possibility of temperament as a causal factor, although no empirical research has yet been published to address this.

Marino suggested that RAD may be a precursor to personality disorders. His comparison of the two subtypes of RAD to Schizoid Personality Disorder and Histrionic Personality Disorder were very interesting and certainly warrant further investigation. Perhaps this is something that could be explored in a long-term study that follows children diagnosed with RAD into adulthood. Because of the high comorbidity with many psychological disorders, I think it is very likely that RAD may be a risk factor for the development of such personality disorders.

McConky expressed concern about my conclusions as to the influence of temperament in the pathogenesis of RAD. Perhaps my paper was not clear enough in my explanation for how I think RAD may be affected by temperament. Certainly, temperament is not a black and white issue--there are gradients and individual variations in temperament from one individual to the next. A certain range of temperaments may put an infant at greater risk for developing RAD, and within that range there may be some variants that lead to one subtype of RAD or the other. Again, as I mentioned in response to Huang's commentary, I did not address the issue of treatment for RAD because I did not find any articles providing information on treatment.