Coercive Restraint Therapies: A Dangerous Alternative Mental Health Intervention

Jean Mercer, PhD

In This Article


Investigation of the sources described above revealed sharp contrasts between evidence-based treatment and CRT practices. There is a systematic theoretical background for CRT and CRTP, but it is severely at odds with either accepted theory or research evidence about the nature of child development. The research evidence offered by CRT advocates in support of their practices is so flawed in design as to be useless.

The use of physical restraint and other coercive practices by CRT advocates stands in the sharpest possible contrast to conventional mental health practices. However, other contrasts also exist and have been noted by CRT proponents (Attachment Disorder Site). Generally, CRT views emphasize the authority of the adult and reject any active decision-making role to be played by the child. For example, parents are to establish behavioral goals and the child is not to participate in this process. Children are to be told the words to say that are thought to express their emotions; adults do not wait or follow the child's lead in this matter. All information is to be shared with the family; the child does not talk privately with a therapist. Finally, wraparound services are rejected on a number of grounds, including the idea that children may be given rewards that the parents do not approve of.

CRT advocates claim that their belief system is derived from the theory of attachment developed by Bowlby and Ainsworth,[12] but examination of CRT materials shows little relevance except for the use of the term "attachment." In fact, CRT beliefs appear to derive from a combination of fringe systems, including the work of Wilhelm Reich,[13] Arthur Janov,[14] Milton Erickson,[15] and the various body therapy proponents (for example, Soul Song).

Many CRT and CRTP advocates assume that each cell of the body can carry out mental functions, such as memory and the experience of emotion (for example, Official Site of Dr. Bruce Lipton). This belief implies that physical treatment, such as restraint or compression, can alter thinking and attitudes. In addition, body cells may contain memories that interfere with processes, such as emotional attachment, and physical treatment can erase those memories so that the individual is free to develop loving relationships. Another implication is that a sperm or ovum, as a cell, is able to store memories and emotional responses.

Many CRT and CRTP advocates assume that personality functions and attitudes date back to the time of conception or before (Emerson Training Seminars). According to this view, a fetus, or even an embryo, stores memories of events, including the mother's emotional response to the pregnancy. If her feelings are positive, the unborn child begins to develop an emotional attachment to the mother; if she is distressed by the pregnancy or considers abortion, the unborn child responds with rage and grief over this rejection and cannot form a normal attachment.

CRT and CRTP advocates assume that all adopted children, even those adopted on the day of birth, experience a profound sense of loss, grief, rage, and desire for the vanished birth mother. This emotional pattern interferes with attachment to an adoptive mother.

CRT and CRTP advocates assume that anger and grief must be removed through a process of catharsis. The child must experience and express these negative feelings in an intense manner. He or she can be helped to do this by a therapist or parent who initiates restraint and physical and emotional discomfort in order to stimulate expression of feeling.

Unlike conventional child development researchers, CRT and CRTP advocates believe that normal attachment follows an attachment cycle[1] consisting of experiences of frustration and rage, alternating with relief provided by the parents. On the basis of this assumption, they posit that emotional attachment in the adopted child can be achieved through the alternation of distress and gratification of infantile needs, such as sucking and the consumption of sweets. Some CRT proponents warn that conventional therapy, with its emphasis on following the child's communicative lead, will in fact worsen an adopted child's emotional status.

CRT and CRTP advocates believe that cheerful and grateful obedience to parents is the behavioral correlate of emotional attachment, and that this is true for children of all ages. A parent's sense that the child is aloof and unaffectionate is the best indication of disordered attachment.

A comparison of these CRT points to conventional theory and evidence-based views of early development shows little or no overlap beyond the idea that emotional attachment occurs in infancy and has some impact on behavior. Cells outside the nervous system are not conventionally believed to be capable of memory or experience, nor are memories considered to go back to preconception or even to the embryonic or early fetal stage. Although a mother's emotional state and stressful experiences during pregnancy do appear to have some effects on development, these effects have never been specifically related to her attitude toward the pregnancy, nor is that attitude easily isolated from postnatal events. Emotional attachment is generally considered to be a process beginning after the fifth or sixth month after birth and resulting from pleasurable, predictable social interactions with a small number of interested caregivers. Attachment behaviors vary with age and developmental status and at some stages include negative actions, such as tantrums or arguing. Attachment disorders are not easy to define or to diagnose, but, like most early emotional problems, they are best treated through techniques that facilitate the child's enjoyment of social play and mutual social interaction, as well as by treatment of factors, such as maternal depression.

The difficulties of clinical outcome research are obvious, but professionals working with outcome issues have set out criteria for effective work of this type.[16] One useful approach has involved the concept of levels of evidence, which can be used to define the conclusions that can legitimately be drawn from different research designs.

CRT advocates in the 1970s showed little concern for research evidence,[17] but in more recent years have become aware of the commercial value of claiming an evidence basis. Internet sites offering CRT frequently include claims that a favored treatment "works" and that conventional treatments not only fail to "work," but cause exacerbation of problems. A small number of empirical studies of CRT have been published or posted on the Internet; these are critiqued below. Surprisingly, there are no CRT studies at the lowest level of evidence, the case study level, although there are scattered anecdotes about cases. Of no surprise, there are also no randomized, controlled trials, and, considering the deaths and other problems associated with CRT, it seems unlikely that an institutional review board will ever permit such research. Available research reports are at the second level of evidence, with quasi-experimental designs, and can thus not be used to support conclusions about causality. It should be noted that there are a number of confounded variables in all of these studies; children who receive CRT usually are separated from their parents for a period of time, and they experience CRTP carried out either by foster parents or by the adoptive parents.

The use of a paper-and-pencil instrument, the RADQ, is frequent in research reported by CRT proponents.[4] An understanding of the development and nature of this instrument is a necessary beginning for a survey of CRT research.

The RADQ is a questionnaire that is to be answered by a parent or another adult who has spent a great deal of time with the child. Diagnosis of an attachment disorder (reactive attachment disorder, or the CRT-posited attachment disorder, depending on the investigator) is based on the adult's responses to statements about the child. These statements uniformly refer to undesirable behaviors or attitudes; there is no check for response bias, so an adult who agrees with every statement creates the highest possible attachment disorder score. The items on the RADQ were not derived from empirical work. A number of them actually come from a questionnaire that has been in existence for decades, at one time being used as a measure of child sexual abuse, but originally coming from a survey meant to detect masturbation.[18,19]

A major problem of the RADQ is that it has not been validated against any established objective measure of emotional disturbance. Validation was against a Rorschach test administered and scored by the creator of the RADQ, who also administered and scored the RADQ.[4] A degree of spurious respectability has been given to the RADQ in the last few years as a result of psychometric studies concentrating on the internal reliability of the test, but this does not, of course, speak to validity issues.

The RADQ and other ad hoc questionnaire measures used in studies of CRT outcomes are thus inadequate evaluative devices. Similarly, there is no evidence to support claims that a child's movement patterns can be interpreted to yield an attachment disorder score.[20] There is 1 empirical study of CRT published in a peer-reviewed journal.[9] This report, based on a doctoral dissertation at a distance-learning institution with problematic accreditation, has a controlled clinical trial design with serious flaws in the comparison group. The investigation studied children whose families had contacted the Attachment Center at Evergreen and expressed their wish to bring the children for treatment because of behaviors categorized as disorders of attachment. All the parents were asked to respond to a questionnaire about the children soon after their initial contact. One group brought the children for a 2-week intensive treatment, during which time the children had little contact with the parents and stayed in therapeutic foster homes for CRTP, while the parents themselves often vacationed. The comparison group in this study was comprised of families who had made the initial contact with the Attachment Center, but for reasons of their own had not brought the child for treatment. Both groups were asked to respond to a second identical questionnaire about a year after the initial contact had been made. The investigators concluded that the treatment group improved more than the comparison group in the course of that year.

This study has been used by CRT advocates as evidence supporting the efficacy of their practices. However, one would expect some degree of improvement in the course of a year, both because of maturation and regression to the mean. The difference in amounts of improvement could result from the many variables confounded with the treatment variable: the reason for the comparison group's failure to attend treatment (marital disagreement over the decision, financial concerns, physical or mental health needs of other family members, or employment problems); the effect of separation from the parents on the children in the treatment group; the effect of separation from the children on the parents in the treatment group; the parents' vacations and travel experiences; and cognitive dissonance factors encouraging the parents to believe that there must have been a positive outcome resulting from this expensive and disturbing experience, or a negative effect if they were unable to come for treatment. Design problems thus make it impossible to accept this study as evidence supporting CRT.

Two simple before-and-after studies claiming to support CRT have been posted on the Internet ( and Attachment Treatment & Training Institute). The first, by Becker-Weidman, administered the RADQ and a behavior checklist to parents of 34 children before and after CRT. Becker-Weidman concluded that CRT had caused changes in the children, basing this statement on significant differences between test scores. However, the treatment variable in this study was confounded with simultaneous maturational change. In addition, natural variations in behavior and attitudes may be involved, because parents are most likely to bring children for mental health treatment when their behavior is at its worst, so that spontaneous improvement occurs during the time of treatment but not because of treatment.

The second, similarly designed study by Levy and Orlans is difficult to follow because of the lack of detail in the Internet posting, but its conclusion that CRT is effective appears to be subject to the same criticisms as the Becker-Weidman work.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.