Sexual abuse of children
Treatment - an overview
Translation: Roger G. Tanner

Contents
Preface
Summary
Introduction
Three emphases in treatment programmes
Four international treatment models
Child Sexual Abuse Treatment Program of Santa Clara County, USA
Sex Abuse Project, Ackerman Institute for Family Therapy, USA
Great Ormond Street Sexual Abuse Team, UK
Confidential Doctor Center, "Kind in Nood", Belgium
Summary of the treatment models
Common themes and components of treatment
BUP-Elefanten, Linköping
Treatment philosophy
Treatment programme
Summary
Evaluation of treatment
Experimental design
Quasi-experimental design
Before, during and after design
Summary concerning treatment
Key to instruments used in the text
Concluding remarks
Focus and development of treatment
Bibliography
Suggestions for further reading
Appendix

Preface
One of the tasks of the National Board of Health and Welfare is to support the scientific development of social services. It is doing so partly within the framework of a development project (the KUB Project) in which sexual abuse of children forms one of several thematic fields. Some 20 remits have been entrusted to experts in a variety of scientific fields. Their expert reports provide the National Board of Health and Welfare with documentation on which to base policy decisions concerning sexual abuse of children.

Carl Göran Svedin, Consultant/Associate Professor of Child and Youth Psychiatry and Project Leader at BUP-Elefanten, Linköping University Hospital, has been commissioned by the National Board of Health and Welfare to compile a survey of current knowledge and research concerning the treatment of children and young persons who have been sexually abused.

Prof. Svedin's account has been examined by Göran Bodegård, Consultant, of the Child and Youth Psychiatry Department, Karolinska Hospital, Stockholm, and by Kenneth Nilsson, Chartered Psychologist, of the Child and Youth Psychiatry Department, Uppsala University Hospital.

We hope that this report will contribute towards a growth of knowledge and understanding concerning both children's treatment needs and various models of treatment.

Under the rules applying to expert reports from the KUB Project, the authors alone are responsible for the content and conclusions presented.

Lars Pettersson
Deputy Director-General

Summary
International models influencing the development of treatment
This report opens with a brief presentation of four international models for the treatment of children and young persons who have been sexually abused. The models have been chosen on account of the very important part they have played in the development of treatment, both nationally and internationally. The programmes described are: the Child Sexual Abuse Treatment Program of Santa Clara County, USA, the Sex Abuse Project, Ackerman Institute for Family Therapy, USA, the Great Ormond Street Sexual Abuse Team, UK, and the Confidential Doctor Center "Kind in Nood", Belgium. Fuller descriptions of these programmes are appended to the main text of the report.

Although several of the centres described here are rooted in different schools of theory, they are many-sided, offering individual therapies for children and adults, couple therapies, family therapies, group therapies, self-help groups and combinations of different forms of therapy. On the strength of these many-sided activities, all of them can be termed more or less eclectic. One weakness is that most of them lack systematic scientific evaluations of the efficacy of their treatments.

Common characteristics of treatment
The section dealing with the international programmes is followed by a review of common characteristics of treatment. Literature, articles and books about the treatment of persons who have been involved in sexual abuse deal with various aspects of therapy content, regardless of whether the victim's therapy is individual or group-based. Some address different themes as and when they become relevant to the individual child, while others, above all in group therapy, employ themes which have been carefully prepared and timetabled in advance of each therapy session. This section describes various themes, regardless of how structured their introduction into the therapy may be.

BUP-Elefanten, Linköping
There follows a description of activities at BUP-Elefanten in Linköping, which is the first specialised unit in Sweden to work with children and young persons who have been physically maltreated and/or sexually abused, and with young persons who have committed sexual abuse. The theoretic content of this treatment model is eclectic, emphasising individual and family treatment, supplemented by group therapy. Its clearest approach is the importance of early intervention for the treatment process. Methods development, research and teaching are also an important part of day-to-day work.

BUP-Elefanten asserts the importance of treatment having a multi-generational perspective, both backwards and forwards. Gender and power perspectives are also important aids to achieving greater understanding of and a constructive approach to persons with "victim or perpetrator experiences". A developmental psychology perspective is applied to the individual and the family, especially when dealing with young persons and their families, so as to achieve an age-adequate response to the teenager and so as to meet the family in its life cycle. BUP-Elefanten takes the view that the people best placed to help the child are the child's/the family's facilitators, who may include professionals as well as the family and relatives.

Crisis theory and trauma theory are fundamental to BUP-Elefanten's eclectic model. Probably the most important factor of children's success in coping with abuse/maltreatment is the supportive capacity of parents. A crisis and trauma perspective on abuse means that it is not medicalised or psychiatrised. In keeping with this crisis and trauma perspective, BUP-Elefanten takes the view that help should be offered outside the hospital, on child-friendly premises which are readily accessible to families. The fact of the treatment programme being based on a knowledge of crisis and trauma means high priority for emergency measures and high accessibility, above all in cases arising out of an introductory consultation and any subsequent police interrogation. In practice this means the team being on standby for consultations and attendance at police interrogations. The purpose of attending an interrogation is to be able to offer child and family immediate contact. Child, parents and siblings may need both individual support and family counselling. Initially the child and parent/s are met separately, since, as indicated above, their crises can have different contents. Both child and parents need to be able to express themselves and their feelings without having to show consideration for anyone else.

Most often at BUP-Elefanten a combination of two or more treatment methods is called for. The methods available are:

• Individual therapy: counselling of young persons and adults, play therapy, symbol drama, picture therapy, EMDR.
• Family or parental therapy.
• Group therapy.
• Network therapy: both as a method in crisis treatment and as a recurrent element of ongoing work.

Treatment evaluation
The next section deals with treatment research concerning children and young persons who have been sexually abused. This is a field in which most of the work remains to be done. Many of the studies hitherto published suffer from few children being included, and from a high dropout rate in cases with protracted treatment periods. In addition, comparability is limited by the great variations between different studies in terms of gender and age structure, input symptomatology and treatment focus. Reference groups are often lacking, due to ethical considerations. As yet few writers have employed a more experimental design and compared different treatments.

Recent years have brought an encouraging growth in the number of experimental studies. Those with sexually abused children randomly distributed between treatment group and control group have presented a fairly unanimous picture, namely that of the treatment groups developing significantly better than the control groups. Many of these studies show what other studies of child psychotherapy have shown already, namely that psychotherapy has an effect and that treatment with cognitive behavioural therapy has a better effect on a number of different problems.

Focus and development of treatment
The majority today recommend cognitive behavioural therapy, above all for traumatised children. In Sweden and many other countries, this form of therapy has been mainly used for adult patients and to only a slight extent with children. The fact of cognitive behavioural therapy having begun to be applied to this target group, i.e. sexually abused children and young persons, is to be viewed in the light of modern knowledge concerning the functions of the brain in connection with trauma and extreme stress.

Without any doubt, the form assumed by the development of treatment during the 1980s and 1990s has been to develop treatment programmes and models based on multidisciplinary, specialised units, all of which have emerged in response to a growing need of therapeutic help for sexually abused children and young persons and to a growing need for development of science and methods in this field. This process also has the support of recommendations made by the Council of Europe to its member countries. In the light of international experience, development of more treatment centres in addition to BUP-Elefanten is to be recommended, and it should be possible for centres of this kind to be established at least in the three metropolitan regions of Stockholm, Gothenburg (Göteborg) and Malmö.

Training measures needed
Training measures are also needed, and above all training programmes are called for in cognitive behavioural theory for children generally, but especially with reference to trauma. Experience of BUP-Elefanten also argues strongly in favour of the benefits of combining treatment with research and teaching. More resources need to be devoted to evaluation research concerning both the overarching objectives of these centres and more specific treatment objectives and inputs.

Introduction
The treatment of sexually abused children and young persons is an important field but a difficult one to master. Like all other forms of psychotherapy, it presents a multiplicity of theories, treatment models and programmes.

The present report, which cannot in any way be termed an exhaustive account, is divided into four sections. The first section contains short descriptions of activities at four internationally known treatment centres. These treatment models have variously impacted, both nationally and internationally, on the development of methods and discussions of treatment, but scientific documentation of their efficacy is extensively lacking. Even so, given the decisive influence which these models have exerted on developments in the field, and the qualified, experienced clinical expertise manning them, they merit the fairly exhaustive description appended to this report.

Both the working approach of the Giaretto Institute, the Ackerman Institute and Great Ormond Street are models which are basically applicable in Sweden, since they all operate under legislation similar to Sweden's, i.e. under a duty of notification (mandatory report) towards the social welfare authorities (Swedish law: Section 71 of the Social Services Act). The Belgian-Dutch model, represented here by "Kind in Nood", Brussels, is alone in preserving anonymity, which is not allowed in Sweden. The Hans och Greta (Hansel and Gretchen) Project in Köping comes closest to the Brussels model by encouraging anonymity and adopting a neutral stance in relation to social service reports to the police (Westerström et al, 1999).

The second section describes what the international literature regards as important themes or topics in work with sexually abused children and young persons.

The third section presents the activities of BUP-Elefanten, starting with the treatment philosophy on which activities are based. This is followed by a review of the treatment programme. BUP-Elefanten applies a crisis and trauma perspective to abuse, which accordingly is neither medicalised nor psychiatrised.

Lastly a description is given of research findings, i.e. of the current state of knowledge concerning the outcome of the treatment of children who have been sexually abused.

Three emphases in treatment programmes
Cohn (1979) writes that programmes for sexually abused children and young persons have three identifiable approaches: a lay or voluntary approach, mainly offering peer counselling and supportive groups, a group approach offering group therapy and training, and a approach based on social work and mainly offering individual counselling/treatment. MacFarlane (1983) also identifies a fourth systematic approach based on the importance for effective treatment of co-operation between public authorities and mediation. Keller et al (1989) write that in practice the majority offering treatment have adopted an eclectic treatment model which is often based on the humanist model of family treatment developed, for example, by Giaretto (1982). This model combines voluntary efforts in the form of counselling and support with professional individual treatment and group therapy in a co-ordinated multidisciplinary activity.

The USA has also developed what are called "Children's Advocacy Centres" (CACs), which can mainly be described as child-focused units in which police, prosecutors, social services, doctors and therapists work together as a team dealing with cases of child abuse. Theirs is primarily an investigative approach, but the majority of teams offer crisis processing and many of them continue treatments individually, for the family or in groups. The overriding objective of CACs is to ensure that children are not further traumatised by the system which basically exists for their protection. The element of voluntary work in the form of "Child Advocates" is pivotal. These Child Advocates are the children's representatives. They keep them occupied in waiting rooms, help them to cope with judicial systems, collect money and, not least, work with public opinion and with lobbying of politicians and decision-makers. In 1998 the USA had more than 400 of these centres, 190 of which were affiliated to a national network called the National Children's Alliance.

In a survey in 1986, Keller et al (1989) identified upwards of 2,250 different programmes in the USA. Of these, 553 replied to a questionnaire. Most of these programmes were either part of a private organisation (43 per cent) or (30 per cent) of a public activity such as medical care or social welfare. Over 88 per cent of the units concentrated their attention on the victim and the victim's family, and most of them had a family-oriented treatment programme. More than half of them (56 per cent) offered individual, family, pair (child-parent, parent-parent) and group treatment. Only a fraction of the units offered only individual treatment (3 per cent) or group treatment (3 per cent). There was a wide variety of techniques, such as insight therapy, play therapy, behavioural modification, picture therapy and more educational techniques such as psycho-educative treatment, heterosocial skill training, anti-recidivist techniques and cognitive treatment. Only 27 per cent of the programmes had post-treatment follow-up routines.

No conspectus of treatment programmes in Europe has been found, but in 1998 the Council of Europe appointed a working group tasked among other things with reviewing treatment programmes in the 40 member countries.

Four international treatment models
Child Sexual Abuse Treatment Program of Santa Clara County, USA
The overarching philosophy of this programme is to offer something different from traditional psychotherapy, which is based on a medical model of curing disease. The aim is to promote, through a humanistic approach, the individual's self-awareness and capacity for self-management and for becoming a responsible member of society.

The first object of the training is to provide immediate counselling to the child and its family. More often than not, families where sexual abuse has occurred are in a state of dissolution, as a consequence both of the original familial dysfunction and of the disclosure. Usually the child, mother and father need to be treated separately before family therapy can achieve any results. The therapists, counsellors, are often in close contact with judicial authorities, since the treatment may be prescribed in the offender's sentence. The counsellor reports regularly on how well the offender is maintaining contact.

The programme can be termed highly composite, based as it is on humanistic psychology, early intervention, individual help, family work and above all constellations of different groups with a strong element of voluntary work and self-help orientation. The groups and voluntary measures are referred to as a powerful support in the process of family resocialisation.

No scientific presentation of treatment outcomes has been traceable.

Sex Abuse Project, Ackerman Institute for Family Therapy, USA
The Ackerman Institute's model is based on two schools of theory in family therapy: social constructionism and feminism (Sheinberg, 1992).

The Ackerman Institute itself describes its therapeutic approach as a multimodal programme, meaning that various modalities, such as individual, group and family therapy, are integrated in the treatment (Sheinberg, True, Fraenkel, 1994). The programme is client-driven, which means that all the time it is framed according to the individual child's unique experience. In the programme-driven models which are often used in group treatment, for example, different themes are worked through in a structured manner, it being known that they are usually of great importance to children who have been subjected to abuse.

The treatment is based on the assumption of people having "multiple self-images", depending on the context and relations in which they see themselves (Gergen, 1991). Accordingly, in the therapy the child and other members of the family are offered different experiences by alternating between different therapy contexts or modalities. This working approach has been developed into a recursive flow of information whereby information from one therapeutic context - individual therapy, for example - is introduced into another context, e.g. family therapy, and then back again in a continuous flow. By making different questions, subjects, difficulties, conflicts etc. migrate between the different modalities, greater understanding is achieved of the unique experiences of each member of the family.

No scientific evaluation of the therapeutic outcome has been presented.

Great Ormond Street Sexual Abuse Team, UK
Treatment by this team has been influenced from many quarters, not least by parallel developments in family therapy. Bentovim (1987, 1988) describes, on the basis of a family-systemic model, how a behaviour of sexual abuse can start and continue. The family can be meaningfully described and understood through seven different levels, viz:

1. Elements of interaction.
2. Cycles of interaction.
3. Active meaning systems.
4. The dimensions of family life in sexual abuse.
5. The formulation of the family.
6. The potential for treatment, actual and ideal.
7. The generalities of family life.

The Great Ormond Street programme has made an important difference to European developments in this field. The programme has attempted to integrate two main approaches - family therapy and group therapy - within the framework of clear co-operation with social services and criminal welfare. Great importance is attached to an initial assessment of each individual case.

The programme has been evaluated through a follow-up study (Bentovim, van Elburg, Boston, 1988), in a descriptive study and in a treatment study (Monck et al, 1996). The latter is presented in the fourth section, Treatment evaluation, of the present report.

Confidential Doctor Center, "Kind in Nood", Belgium
The Confidential Doctor Center (CDC) questions the ability of the traditional legal and medico-psychosocial models to help children at risk and their families. Both models are liable to be repressive and controlling instead of understanding, supportive and helpful. Harmful treatment of children, it is felt, has not hitherto been observed or understood as a socio-psychological, cultural, political or gender stereotype problem. Anyone, it is claimed, can become a child abuser, depending on special relational and social circumstances.Child abuse and failure of care differ only in quantity, not in quality, from adult attitudes to children.

At CDC the central problem when working with children at risk (i.e. physical maltreatment and sexual abuse) is felt to be the dual position of the person treating them, regardless of whether that person is a social welfare officer or a therapist, i.e. that the act of balancing between sympathy and control and dependence on the legal system creates confusion of a kind which can easily result in lack of credibility and trust.

The model is based on offering a differentiated array according to the needs of the individual child or the individual family. This presupposes efficient social services in the community at large, e.g. access to free or inexpensive medical care and day nursery care.

Practical work is organised with reference to three functions: direct assistance to children and parents, guidance for others encountering child abuse, and prevention.

The treatment programme is based on offering help instead of punishment, on respecting confidentiality instead of exercising control, on showing solidarity instead of writing reports, on mobilising the family's resources instead of keeping them passive, and on co-operating with other professionals instead of competing with them. The treatment programme covers both physical maltreatment and sexual abuse. Contact with CDC is entirely voluntary and is based on an assurance that psychotherapy can be offered with full confidentiality, thus allowing the client to express his or her innermost thoughts and feelings.

The uniqueness of this programme can be said to consist in its being dominated by an approach to the families whereby confidentiality is put at the centre of things. The advantages claimed are that the risk of secondary traumatisation is minimised, that parents develop in their parental role and that more people, offenders included, request help for themselves and their families. In this respect the Center has succeeded quite well. Of the 3,858 children coming there between 1986 and 1994, 37 per cent did so as a result of parents themselves getting in touch with the Center. The corresponding figure before 1986 was a mere 3 per cent.

No concise scientific account of treatment outcomes has been traceable.

Summary of the treatment models
As has already been made clear, the four models presented above were chosen because they have played a very important part in the development, both national and international, of treatment for sexual abuse of children and young persons. Their weakness is that most of them lack systematic scientific evaluations of the efficacy of the treatments.

Although several of the centres described above are rooted in different theories, they are many-sided, offering individual therapies for children and adults, couple therapies, family therapies, group therapies, self-help groups and combinations of different forms of therapy. On the strength of these many-sided activities, all of them can be termed more or less eclectic. Therapy research is a difficult and complicated field which begs ethical questions and involves practical difficulties. Moreover, research into models like these is made difficult by their involving inputs of so many different kinds. Nonetheless, efforts to study scientifically the treatment measures based on these models - both in their entirety and with reference to different components - should be intensified. Therapy research is extremely time-consuming and expensive, but this should not be an obstacle if support is forthcoming from funding agencies. Very often, unfortunately, this is not the case, with the result that activities have been based more on "proven experience" than on science. But is fair to say that the models described here have a high degree of what is termed "face validity", meaning that, judging by the general state of science, they appear to be both serious and cogent. Further to this point, the reader is referred to the section on therapy evaluation.

Common themes and components of treatment
Literature, articles and books about the treatment of persons who have been involved in sexual abuse deal with various aspects of therapy content, regardless of whether the victim's therapy is individual or group-based. Some address different themes as and when they become relevant to the individual child (Boatman, Borkan, Schetky, 1981, Sgroi, 1982, Sheinberg, 1996, Marnaffe, 1997), while others, above all in group therapy, employ themes which have been carefully prepared and timetabled in advance of each therapy session (Hildebrand, 1988, Mandell et al, 1989, Nelki, Walters, 1989, Karp, Butler, 1996). The following section describes various themes, regardless of how structured their introduction into the therapy may be. The description is not intended to cover every possible theme which may come up. Instead it will be combined to the majority of themes dealt with in literature on the subject. The main focus of the description will be on psychological consequences of an abuse and on the theme of treatment, but also on symptomatic behaviours clearly linked with abuse, such as sexualised behaviour, post-traumatic stress syndrome (PTSD) and dissociation.

Guaranteeing security and protection for the child
The child's security and protection, even if they are the responsibility of the social welfare authorities, are often the introductory and most important theme of the treatment and something which has to be resolved initially (Jones, 1997, Kempe, 1997). Both initially and subsequently, protection of the child goes before all other treatment (Bentovim, 1988). The question of protection is of pivotal importance, regardless of whether it is resolved in a more traditional manner, through the social services in societies with mandatory reporting, or whether work is subject to confidentiality (Marnaffe, 1997).

Building a therapeutic alliance with the child
Building a therapeutic alliance with children who have been abused or traumatised is considered a very important initial goal of the therapy (Hauggard, Repucci, 1998, Karp, Butler, 1996). This, however, is considered difficult on account of several factors. In the first place, the child has not spontaneously requested therapy but has been brought there by a parent or the social welfare authority. Thus the child may feel that he or she is being forced and, sometimes, is being punished for what has happened. Therapy, as Karp and Butler (1996) observe, can also be interpreted by the child as ushering in yet another "secret" relationship, at least one such relationship previously having led to abuse. Then there is the child's reluctance to share or talk about unpleasant feeling like shame, guilt and self-contempt. Lack of self-confidence and deficiencies in the child's bonding capacity, moreover, often make this contact-forming phase of the therapy a good deal more protracted than in connection with other states for which children and parents seek therapy. In their model for "abuse-focused therapy", Karp and Butler (1996) suggest introductory exercises which create contact, building bridges between child and therapist. Similar exercises occur in the majority of group therapy models (Mandell et al, 1989, Monck et al, 1996). Deblinger and Heflin (1996) claim that the best way is to try unreservedly to approach and enter into the child's world. Lanktree (1994) maintains that everything depends on the child feeling secure and receiving support and encouragement when proceeding to explore and process the abuse.

Developing the child's capacity for trust
As a consequence of abuse and of growing up in a situation of neglect, the child's lack of confidence in its surroundings and in adults forms an inevitable theme in its relation to the therapist or other "helpers", e.g. social welfare officers. The child tests the therapist's good intentions, dependability, honest and interest (Kempe, 1997). Distrust can sometime be so powerful as to impede the commencement of therapy (Bannister, 1992). Continuity, regularity, becomes important and the end of every therapy session can be experienced by the child as an ominous separation.

Abuse breaks down children's capacity for trust and harms their basic instincts and their perception of other people, making them emotionally vulnerable and augmenting the risk of further abuse. The feeling of betrayal is judged to be greatest if the abuser has been close to the child, and especially if he or she is the child's parent (Finkelhor, Browne, 1986, Kendal Tacket et al, 1993).

Trust nearly always becomes an important theme in the therapeutic relation, i.e. in transference and counter-transference. But capacity for trust also needs to be given wider scope in therapy, e.g. as regards reliance on other adults and on peers. This can be done through counselling and exercises, both individually (Karp, Butler, 1996) and in a group (Hildebrand, 1988, Mandell et al, 1989).

The child's fear - an introductory theme of treatment
Fear of what is going to happen after a disclosure is the child's first dominant feeling (Giaretto, 1982) and thus an important introductory theme. Questions which the child asks itself are whether it will be believed, how other people will react, what will happen to the abuser or whether the child will be taken away from home. The child needs to be kept informed of the procedure and what is going to happen, step by step.

Fear of being punished, rejected and abandoned is not uncommon (Terr, 1983, 1990) and has to be dealt with over a long period of time, usually until the legal process has been concluded and the non-abusing parent or parents can confirm the child's needs and create security.

The child's experiences during the actual abuse and during the period in which it has occurred can have made it anxious and easily frightened and induced separation anxiety, and may give rise to situational and panic anxiety. All these things can be a part of PTSD, but they can also be viewed as more isolated symptoms. Situational anxiety can occur in situations reminiscent of the abuse or when meeting people reminiscent of the abuser and the abuse. Sometimes in PTSD, flashbacks and obtrusive memories are incentives for situational anxiety and panic attacks.

Sensory perceptions such as smell, taste and sound are believed to be more significant among younger children than in older ones, to whom words, contexts etc. become more important.

For specific anxiety and fear, Cohen and Mannarino (1993) recommend a technique of suspension of thought, positive conceptualisation and relaxation exercises. This technique can also be taught to parents. Through play, role play etc., the child can successively begin to devise ways of coping with its anxiety and be assured in everyday life of stable, dependable adults who can offer security.

Abandonment and sense of loss
Following the disclosure of abuse, the child may have a strong sense of abandonment and loss (Jones, 1997). It may, for example, miss its parental home if placed elsewhere, or it may sometimes miss the attention and affection received from the abuser. Realisation by the child that its parents have been unable to protect it or cater to its needs can lead to grief and, sometimes, depression.

Shame and guilt
Shame, according to Tomkins, is one of the nine biologically congenital affects, and is widely regarded as one of the most neglected affects (Tomkins, 1991, Nathansson, 1992). Shame is important for the young child's process of socialisation into an adult, social being. Seen in a socio-cultural/anthropological perspective, shame in the majority of cultures plays a central part in the control of sexuality, e.g. through the taboo on incest.

Shame is also one of the most important affects for the development of self-esteem. Its function is to regulate exaggerated risk-taking, to make us stop to think and to prevent our positive affects (pleasure/ecstasy, for example) leading us into dangerous paths. Its function of interrupting positive states therefore makes shame the companion of negative self-esteem.

Since sexual abuse with manipulation of the genitals has the effect of stimulating them, it automatically generates purely physiological reactions which the child cannot control. This stimulus can give rise to feelings of sexual pleasure which in turn generate feelings of shame. Closely allied to the feeling of shame is a feeling of confusion, and this feeling of confusion is easily reinforced by the child having difficulty in distinguishing between feelings of displeasure, pleasure and shame.

Guilt - feelings of guilt - comes later in the child's development and is more an effect of the child's socialisation in relation to its surroundings communicating what is right and wrong. Guilt presupposes a cognitive understanding of right and wrong. The child's experience of participation in being together with the abuser, who has often lulled the child into believing that the sexual activity is normal, frequently becomes bound up with feelings of guilt- not least during and after the process of disclosure.

Henry Giaretto (1982) maintains that when fear begins to subside, feelings of guilt take over. But feelings of guilt are often held back, due to sexuality being a taboo area. The child, in this instance a girl, may believe that she was a participant, that she seduced the abuser, and this is often reinforced if sexual feelings are aroused, i.e. if she feels that she has displaced her mother and believes herself to have destroyed the mother's relationship with the father or step-father; she ought to have rebuffed him more effectively, she feels responsible for the break-up of the family.

If the abuser accuses the child of responsibility for what has happened, claiming for example that the child was seductive, this often compounds the child's feeling of guilt (Briere, 1992). Sgroi (1982) sees three causes of feelings of guilt: feelings of responsibility for the sexual behaviour, for the disclosure and for the ensuing consequences. Giaretto (1982) also argues that feelings of guilt are indirectly manifested in a variety of ways. A child with fairly good self-esteem can react angrily to official interference, angrily against the person who abused her and angrily against the mother who failed to protect her. This profound anger can also assume manifestations of a more self-destructive character, with self-injurious behaviour, outbursts at home and in school, and criminal behaviour. Another and - Giaretto maintains - more serious way of dealing with feelings of guilt is to enclose them in an inhibited, reticent and silent behaviour. It is only through careful response, requiring a great deal of patience, that the therapist can create an accepting secure environment in which the child/young person can begin to feel trust and start talking. The therapist must gradually convince the child that the fault is not his or hers and that the anger the child feels is justified. In a controlled and supportive environment, either individually or in a group, the child can express its anger. Karp and Butler (1996) take the view that, before a child can divest itself of feelings of guilt and shame, he or she must first be able to identify them and put a name to them. Admission of shame is the first step. Shame is born of secrets. By starting to talk about the secrets, shame and guilt are defused. Karp and Butler (1996) regard this as perhaps the most difficult and important part of the therapy an of the healing process, "the core of the damaged self".

Bentovim (1992) believes describing the sexual feelings aroused by the abuse to be the most difficult aspect for the victim, both in counselling and in treatment. These physiological sensations are often interpreted by the child as participation in, cause of and blame for what has happened.

Hildebrand (1988) underlines the importance in group therapy of making clear who is responsible for the abuse. Support and reflections both from the group leaders and from the other children help the child to understand that the guilt lies with the adult abuser.

It is important to note that a therapist may be liable to focus excessively on relieving the child of guilt and feelings of guilt, at the expense of the more difficult areas which experiences of pleasure and shame can imply for the child.

Verbal and non-verbal expression of feelings
Many abused children have difficulty in expressing their feelings in words and actions. Sometimes underlying feelings and needs are expressed in undifferentiated anger which the child does not properly understand or is unable to put into words. Sometimes words fail the child completely. This may be because the child has grown up in an environment which has not stimulated this side of things, but most often difficulties in expressing feelings are connected with their having been a part of the child's psychological strategy for survival after a trauma. Often, moreover, the child has met with a response of such a kind that its feelings, instead of being confirmed, have been invasively explained away or denied by adults (Karp, Butler, 1996). This in turn has created uncertainty, sometimes chaos, in the child's emotional life. If so, it is important to teach the child that feelings are neither "right" nor "wrong" (Deblinger, Helfin, 1996). Being abused by someone you are fond of or have relied on can easily create emotion confusion/chaos, due to the conflicting feelings, e.g. of anger/hatred and longing/love which exist.

Hildebrand (1988) writes that the basic intention of group therapy is the pooling of feelings and information in the group. Mandell et al (1989) regard the child group as an excellent forum for making children more aware of their feelings, putting names to them and seeing the connection between a feeling and its expression in the body or through actions. Comparing notes or encouraging one another augments the child's possibilities of differentiated self-expression. Both Mankell et al (1989), Karp and Butler (1989) and Deblinger and Heflin (1996) argue that counselling and exercises in this field should precede the work of penetrating experiences of abuse more deeply.

Both group-oriented (Hildebrand, 1988, Mandell, 1989) and individually structured programmes (Karp, Butler, 1996) use various exercises to develop the child's capacity for identifying, controlling and expressing feelings.

Expressing and controlling anger
Deblinger and Heflin (1996) regard anger as one of the most important feelings for a child to express and gain control of, since anger occurs commonly and can create problems in various ways. They also maintain that the venting of aggressiveness by punching a cushion or punchbag is unsupported by research (Taris, 1989), because if anything it sustains unhealthy physiological reactions associated with anger. Instead they recommend helping the child to find other solutions to provocative situations or, failing this, getting the child to explore and express its anger through various therapeutic activities. Through discussion, writing or drawing - that is, by communicating its anger in various ways - the child can gain an insight into its anger and the roots of it. This can also mean writing (but not posting) a letter to the abuser, expressing how angry one feels.

Deblinger and Heflin (1996) recommend a cognitive approach. This can include, for example, encouraging children to replace aggression-provoking thoughts with aggression-reducing ones such as "I won't let him (the abuser) upset me again."

Describing the abuse
Children who have been abused have often kept the matter to themselves, as a secret, for varying lengths of time, without talking to anyone. Others have kept completely quiet about the sexual abuse until it has been revealed, e.g. through the confiscation of child pornography (Svedin, Back, 1996). Efforts to reveal what is going one have perhaps not met with any response, and the child has successively learned that these are things one does not talk about. It is not uncommon, after a disclosure, for people around the child (parents, relatives, foster-parents) to encourage the child not to think any more about it, not to talk about it and instead to consider it over and done with.

Most writers see a central component of treatment in the child, at its own speed and individually or in a group, starting to talk about the abuse (sexual, physical and mental) which it has been subjected to (Mandell et al, 1989). Deblinger and Heflin (1996), in their cognitive treatment model, recommend a gradually hierarchic approach to the questions of abuse, beginning with more general information and then gradually moving on to more detailed descriptions and experiences.

Being allowed to talk about what has happened, what the abuser did, what the child did, what the abuser said, how the child reacted afterwards, what things were like, what relationship the child had with the abuser etc., and linking different types of emotion and reaction to this, are deemed important for reducing the risk of symptomatic behaviour (Mandell et al, 1989).

Being able to study the children's explanatory model, as described for example by Mandell et al (1989) in the exercises "I think it happened to me because…" and "I believe … did it so as to …" helps the therapist to understand how the child thinks, how the child apportions the blame, how the abuse acquires a contextual meaning. For every situation, it is important that the child should be allowed to ponder who, in the child's opinion, is responsible. If the child has a distorted picture of responsibility, Cohen and Mannarino (1993) recommend what is termed cognitively reconstructive interventions, e.g. in the form of re-wordings. This is done by letting the child first see or imagine everyday situations, followed by quasi-abusive situations and, finally, the real situation.

In various ways - out of ignorance, due to emotional ties or by dint of threats - the child has been induced to keep silence. Secretiveness has become both a means of survival and a prison, while at the same time contributing towards the child's feelings of shame and guilt. The child is hiding something, which means a tie to the abuser (sharing a secret) and a distance from others, e.g. the non-abusive parent or parents. Getting the child to describe how the secrets were initiated and how they could be maintained therefore becomes a theme of the therapy. In therapy the child can learn what are good secrets and what are bad ones, and whom it can tell them to (Karp, Butler, 1996).

Independence, self-assertion and self-esteem
Poor self-esteem is so common among children who have been sexually abused that Sgroi (1982) is of the opinion that a therapist can take it for granted. Giaretto (1982) believes that further abuse is prevented partly by strengthening the child's independence, self-assertiveness and self-esteem. Bentovim (1988), Sgroi (1982) and others write that group therapy is particularly suitable for self-confidence training. Mandell et al (1989) call this phase of the treatment "taking care of myself".

Abused children have in various ways found their limits infringed. They have been deprived of their sense of control in their private sphere; a feeling of powerlessness to protect their own personal sphere can result. Often, as a result of the abuse, the child has not learned its right of maintaining an emotional distance and a personal sphere (Karp, Butler, 1996).

Most programmes include learning to avoid risk situations, learning strategies and learning to define limits and protect oneself in future, the aim being to strengthen the child's feeling for other people's and its own limits and personal space and to train the child to say No in various situations, as well as teaching the child the difference between suitable and unsuitable ways of touching. A whole variety of exercises for this purpose are described (Hildebrand, 1988, Mandell et al, 1989, Karp, Butler, 1996). A certain amount of material has been produced in Swedish (e.g. Centervall, Fabricius 1991, Granvik, 1994).

Cohen and Mannarino (1993) find thematic painting books an excellent aid to teaching children to identify and distinguish between safe and unsafe situations, both generally speaking and as regards sexual abuse.

Dangerous situations can also be portrayed or re-created by means of doll play. Through doll play/role play, alternative paths can be created and new solutions found to potentially dangerous situations.

Sexuality and body image
Children who have been sexually abused are liable to acquire a distorted image of their own body and of what is normal sexuality. They may imagine that their body has been destroyed for ever. To a girl, for example, this can mean thoughts of no one wanting her in future because, for example, she is not a virgin - the "damaged goods syndrome" (Sgroi, 1982). Others again may feel impure and dirty, and may sometimes have a greater need of washing - compulsive symptoms.

As regards the child's fear of having been harmed or its experience of being ruined, it is important to arrange for a medical examination to be carried out by a doctor who is accustomed to examining children who have been sexually abused and who understands their emotional reactions (Sgroi, 1982).

Giaretto (1982) attaches importance to the prevention of subsequent difficulties affecting the child's emotional and sexual life. Hildebrand (1988) regards group therapy as a very suitable forum for teaching children the difference between suitable and unsuitable sexual contacts between children and adults.

Bentovim (1992) sees a difference between boys' and girls' ways of coping with a sexual trauma. Boys try more often to cope with flashbacks and memories of abuse by acting them out, e.g. by identifying with the abuser. This can lead to sexualisation and an added risk of subjecting others to abuse. Girls on the other hand react more often as victims and have a greater tendency to internalise their traumatic experiences. They can feel that the abuse was their fault and develop poor self-esteem, self-contempt and a self-destructive attitude with can lead to self-injurious behaviour, anorectic patterns and a sexualised attitude including promiscuity.

Gil (1993) argues that if children's deviant behaviour has the character of abuse, e.g. towards other children, then this problem needs to be raised at a relatively early stage of therapy. If it is of a more harmless kind, such as masturbation, one can wait until the child is ready to raise the question and talk about it. Gil (1993) underlines that the child often has difficulty in talking about sexualised behaviour, above all if it has been of an abusive nature. The child is ashamed, often has feelings of guilt and paints itself into a corner by denying the behaviour if the therapist is over-hasty.

Kernerg and Chazan (1991) and Gil and Jonson (1993) recommend a combination of individual supportive and expressive play therapy, parental training and group play therapy for sexualised behaviour in children. If the child does not involve other children in its sexualised behaviour, Gil (1993) favours a combination of individual therapy and family therapy, but Gil also maintains that, if the child as an abusive behaviour towards other children, then group therapy is to be preferred.

Both with and without sexualised behaviour after sexual abuse, most therapists recommend including sex education as a stage of the treatment. Supportive use is made of various kinds of material, such as pictures, anatomical models and anatomical dummies, books and video films.

Post-traumatic stress syndrome
Sometimes a trauma is followed by post-traumatic syndrome (PTSD). Trauma is defined by van der Kolk (1996) as an extremely strenuous event/situation which cannot be evaded or handled with the individual's available resources. Terr (1981) maintains that a mental trauma is a personality injury which appears when a sudden, unforeseeable anxiety overwhelms the individual's capacity for dealing with the situation of defending himself. Sexual abuse is one type of trauma. Terr (1994) divides traumas into Type I and Type II, instancing Type I with an isolated event, such as abuse unaccompanied by threats or violence, while Type II is more serious and can be instanced with repeated abuse. The symptoms of PTSD can be divided into three categories. The first is re-experiencing of the traumatic occurrence, which can take the form of obtrusively painful flashbacks, thoughts, nightmares about the occurrence or unpleasant caused by internal or external stimuli reminiscent of the occurrence. The second category comprises avoidance of things which can remind one of the occurrence and loss of interest in taking part in ordinary activities or indifference to others, coupled with incapacity for feelings and for having hopes of the future. The third category is concerned with heightened stress symptoms, such as sleep difficulties, irritability/aggressiveness, concentration difficulties, exaggerated vigilance and timorousness.

Gil (1993) points out that it is important for therapists to understand the underlying mechanism of PTSD, since unpleasant flashbacks and symptoms like anxiety can suddenly engulf the patient. Gil (1993) recommends, as part of the treatment, teaching the child that these symptoms are the body's way of remembering and informing it about previous experiences.

Dissociation
Dissociation is a defence mechanism which protects the individual from threats or injury. Putnam (1993) describes dissociation as a psycho-physiological process which alters thoughts, feelings and actions in such a way that, for a period, certain information is not integrated with other information which it would normally and logically be integrated with otherwise. In DSM-IV (American Psychiatric Association, 1994), the most important trait of dissociation (dissociative disorders) is described as a failure of functions which are otherwise integrated as regards consciousness, memory, identity or perception of one's surroundings. Dissociation covers a wide spectrum from normality, e.g. daydreaming because we are bored, to intermediate forms under severe stress, or trauma, when the perception of self, memory capacity and consciousness are impaired. In rare and severe states, above all in adults, what is now termed dissociative identity disturbance (formerly "multiple personality) can be seen in DSM-IV. Terr (1994) maintains that a repetition of Type II trauma is needed in order for a picture of clinical dissociation to occur.

Gil (1993) maintains that it is important for therapists to recognise and be able to deal with dissociation in the therapeutic situation, because otherwise parts of the therapy are liable to have no effect on the child. Karp and Butler (1996) point out that it can be helpful to find a common language with the child and in this way make the child notice that it sometimes dissociates. They also consider it important to point out to the child that this was a survival skill which worked well before but no longer does so, the aim being to help the child to develop other skills, so that it will stop using dissociation as a strategy.

Stages of treatment
In most programmes the treatment can be seen, explicitly or implicitly, to follow a set order of stages or phases, each of which contains various themes. One such example is Stuart House (abuse-focused therapy, Lanktree, 1994), where the introductory clinical evaluation is integrated with or followed by the first rapport-building phase. Next come techniques/exercises to develop communicative skills and the capacity for expressing feelings. The child will next explore, in greater depth, subjects and feelings which have to do with the child's experience of itself and its identity, including self -perception and self-esteem. A more stable ego strength and an experience of inward capacity are needed before any processing of the more trauma-related occurrences and feelings can take place. Many children and young persons need to revert to earlier stages/themes of the treatment as they approach more frightening material or more abuse-related feelings (Lanktree, 1994).

Other programmes, based for example of cognitive behavioural therapy, describe themes in modules (Deblinger et al, 1990). In this programme the content of module one (two sessions) is described as helping the child to modulate and express feelings and teaching the child different ways of coping with its feelings (anger, anxiety). Module number two (six sessions) contains the real abuse-related material, to which the child is gradually exposed. The concluding module (two sessions) consists of education and preventive training.

As regards the order in which the components of different forms of therapy are to come, the majority today - Lanktree (1994), for example - are of the opinion that before group therapy, if any, children, just like adults, must initially be allowed a long enough individual contact for talking about the abuse.

Work with parents
The above description is based on themes commonly occurring in therapy with children and young persons. Nearly all writers underline the importance of close co-operation with parents/care providers in order for therapy to be as effective as possible. There are many reasons - getting the child to come to the sessions and complete the therapy, for example - for obtaining the parents' knowledge and insights concerning the child and for involving parents in the therapeutic process (Deblinger, Heflin, 1996).

Work with parents takes place above all in family therapy (e.g. Bentovim, 1988, Marnaffe, 1997, McCarthy, Byrne, 1988), parallel to (e.g. Guiaretto, 1982, Furniss, 1991) or alternating with individual or group counselling (e.g. Fraenkel, Sheinberg, True, 1996, Deblinger, Helfin, 1996).

Furniss (1991), however, maintains that, even when working individually with the child, one often has to have a family therapy focus, referred to as "family therapy by proxy".

Criteria for concluding the treatment
Few therapists describe the grounds on which, and the juncture at which, one is able to decide that the treatment is sufficient and can be concluded. At the Ackerman Institute (Fraenkel, Sheinberg, True 1996) a child's treatment is considered complete when the child has someone in the family to talk to when it is uneasy, if the child can talk about a wide spectrum of feelings in therapy and outside of it, if the child is able to cope in school and achieves as well as before the abuse, and if the child is asymptotic (e.g. is not troubled by sleep difficulties, nightmares or bedwetting).

At the Giaretto Institute (Giaretto, 1982), four questions decide whether the treatment (for incest only) can be concluded:

• Is there a risk of abuse recurring? That is to say, have the parents' relationship and home conditions generally improved sufficiently to prevent a recurrence and to assure the child of a secure domestic environment?
• Has the abuser accepted responsibility for his/her behaviour and become aware of his/her impulses and of how to control the impulses which led to abuse?
• Does the child not only feel secure in the home with the father living there but also wish to re-establish the child-father relationship?
• Have feelings and conflicts between all members of the family been dealt with openly and adequately, so that the family has attained a satisfactory level of harmony?

Suggestions for further reading
The above review of literature on themes commonly occurring in treatment naturally reflects the difficulties which children can have after being subjected to sexual abuse. The review gives only sporadic, fragmentary descriptions of therapeutic approaches and processes. For the more interested reader there is today a host of literature in this field describing various approaches. Suggestions for further reading are given in conjunction with the bibliography.

BUP-Elefanten, Linköping
BUP-Elefanten was launched jointly in 1995 by the Östergötland County Council, the National Board of Health and Welfare, the Linköping Social Insurance Office and the Municipality of Linköping, in response to an upsurge of reported cases of child assault, but above all of sexual abuse, occurring at the beginning of the 1990s. These cases made up a large proportion of the psychiatric case load among children and young persons. Experience showed that they were usually very complicated and that there was a great need for knowledge concentration/specialisation and methods development so that better help could be given to children and families. In addition, experience and earlier studies had shown that abuse problems were often left unprocessed (Svedin, Gustafsson, 1994) and that families felt badly treated in the contacts to which the suspected abuse gave rise (Svedin, Gustafsson, Sundqvist, 1991).

BUP-Elefanten is a specialised unit for psychotherapeutic assistance when sexual abuse and/or assault is suspected or established. Psychotherapeutic help is also given to young abusers and their families and, to a limited extent, to women and men who suffered abuse during their formative years. In addition to treatment, BUP-Elefanten also engages in training, consultation, tutoring and research and methods development in the fields of aggressiveness/sexuality and child abuse. The unit is operated by the Child and Youth Psychiatry Department, Linköping University Hospital.

Treatment philosophy
BUP-Elefanten has been guided and influenced by the international treatment programmes already described, through the medium of lectures and supervision (Bentovim, Sheinberg), field trips (Brussels) and literature studies. Other international visiting lecturers have comprised Finkelhor, Seablom, van der Kolk, Black and Johnson. The treatment model developed can be termed eclectic, but a number of important approaches deserve to be highlighted, in particular the overarching view taken of society, the family and the individual, the view taken of abuse and treatment, and the view taken of sexuality and sexual abuse.

The view of society, family and individual
Society today embraces a diversity of values, rules and laws which are changing all the time. Added to this, all societies have many different "sub-cultures" with views, partly their own, as to what constitutes a good life. The technical inventions of the past century, including everything from the filament bulb to the Internet, have also created and are re-creating the experience which most people have of themselves, their relation to others and their place in the overall scheme of things. The culture of which family and kin form part and which the child grows up in is a vehicle and creator of context and meaning.

BUP-Elefanten argues the importance of a multi-generational perspective, both backwards and forwards in time. Gender and power perspectives are also important aids to greater understanding and a constructive approach in our encounters with people having "victim or culprit experiences". A developmental psychology perspective is applied to the individual and the family, especially when dealing with young persons and their families, so as to achieve an age-adequate response to the teenager and so as to meet the family in its life cycle.

Each individual is unique, every family, kinship circle and family network is distinctive. The problems, abuse included, for which the family seeks help need to be given a unique response, adapted to the child and to the history and preconditions of the family. BUP-Elefanten takes the view that the people best placed to help the child are the child's/the family's facilitators, who may include professionals as well as the family and relatives.

The view of abuse - crisis and trauma theory
Crisis theory and trauma theory are fundamental to BUP-Elefanten's eclectic model. Trauma is defined, e.g. by van der Kolk (1996), as the individual being engulfed by stimuli which he cannot cope with. The individual lacks previous experience of or defence against this stimulus, which creates a situation of maximum insecurity. The individual can feel helpless, vulnerable and bereft of protection and control. The way in which the individual is affected by/experiences the trauma hinges on the extend to which he has previously lived in conditions of security or has been severely neglected. The bonding process, therefore, is crucial (Bowlby, 1969, Ainsworth, 1969).

Bonding is defined as an emotional tie between two persons uniting them relatively permanently in time and space. Primarily it refers to the child's ties with its parents and other important persons around it. Being firmly linked to the custodian means a good "vaccination" for stresses and strains occurring later on. Inadequate bonding and failure of care often go together. Both greatly affect the way in which the individual experiences the trauma which he or she is subjected to. A child with poor bonding and/or failure of care is also more liable to be subjected to abuse. After the abuse has been disclosed, these children do not always meet with the understanding response from people close to them which could alleviate the harmful effects of the trauma.

View of sexuality and sexual abuse
BUP-Elefanten shares the WHO view of sexuality as an integral part of the personality of all human beings - men, women and children. Children and young persons are sexual beings who from birth have "sexual" encounters with other people. They need positive preparation in the field of sexuality, just as in all other important sectors of life. The most effective sexual learning takes place in the family. Just as with all other important human relations, healthy sexual relations are based on the fundamentals of honesty, gender equality and responsibility. It is through honesty, gender equality and responsibility that we are best assured of avoiding sexual coercion and /or sexual violence. Human sexuality and sexual expressions, including sexual violence, are caused by a complex interplay of biological, psychological and social factors.

Sexual abuse is not a disease or a diagnosis but an experience of a more or less traumatic nature which can lead to both transient and long-lasting injuries/impairment. The way in which children are affected depends among other things on whether abuse occurred on one or more occasions, whether the child had a close relationship with the abuser or the abuser was an unknown person, how long the abuse has been going on, whether the abuse has been documented, whether or not there were elements of violence involved, whether the child's body has been interfered with or the child was just a witness to sexual activity, and how strong the child's own resources are.

Probably the most important factor deciding how children succeed in coping with abuse/maltreatment is the supportive capacity of parents. A crisis and trauma perspective on abuse means that it is not medicalised or psychiatrised. In keeping with this crisis and trauma perspective, BUP-Elefanten takes the view that help should be offered outside the hospital, on child-friendly premises which are readily accessible to families.

Treatment programme
Crisis response
The fact of the treatment programme being based on a knowledge of crisis and trauma (Cullberg, 1975, Dyregrov, 1997, van der Kolk et al, 1996) means high priority for emergency measures and high accessibility, above all in cases arising out of an introductory consultation and any subsequent police interrogation. In practice this means the team being on standby for consultations and attendance at police interrogations. The purpose of attending an interrogation is to be able to offer child and family immediate contact.

Child and family often find themselves in two crises at once. The first crisis, which above all concerns the child, relates to the child's experiences of the abuse. For the non-abusing parent or parents, the crisis concerns feelings of having failed to protect the child. Where intrafamiliar abuse is concerned, the crisis also includes the non-abusing parent's ambivalence as to who she (it is usually the mother) is to believe and feelings of betrayal by the partner. The other crisis (the disclosure crisis) concerns, even if the suspicion of abuse cannot be confirmed, the consequences of the actual disclosure. What happens now? Can the family stay together? Will the child be placed in care? Will the abuser go to prison? Who will know? What shall we say to friends and relatives?

Child, parents and siblings may need both individual support and family counselling. Initially the child and parent/s are met separately, since, as indicated above, their crises can have different contents. Both child and parents need to be able to express themselves and their feelings without having to show consideration for anyone else.

In crisis interviews one meets the child unconditionally on his or her own ground. The child describes what has happened and expresses his or her thoughts and feelings. The child is offered various modes of self-expression - in words, play and pictures. By the same token, the non-abusing parent or parents are given the opportunity of expressing themselves. After these one or two introductory interviews, the important thing is to set about strengthening the relationship between child and non-abusing parent, especially if this tie has previously been weak or has intermittently failed. Seeing and supporting the child and assuring it of continued protection becomes an important theme for the parent in these interviews. The initial contact also includes instruction about crisis, crisis reactions and how people usually react in crisis or after a trauma.

The early contact occurs parallel to the social and police inquiry. If necessary, BUP-Elefanten helps the child and family with these contacts and co-operates with these authorities on the child's and/or family's behalf. Networking meetings, attended by the social welfare officer responsible, are common during this phase.

For some children and families, the crisis response can be all the help they need. Others need treatment of varying duration. The early contact has two advantages: it captures children and families when they are most motivated and desirous of help, and it creates a possibility of assessing any need for further treatment.

Assessment
The need for treatment is always assessed, irrespective of whether BUP-Elefanten meets child and family when they are in crisis or later on.

General questions

• What symptoms does the child have and how badly is it suffering from them?
• What is the child's level of development and what are the child's resources?
• What are the prospects of the child benefiting from therapy?
• What sort of therapy is suitable in the individual case?
• What difficulties and resources are present in the family?
• Are the parents capable of perceiving the child's needs?
• Are the members of the family motivated for taking part in the treatment process?
• In what way is the network - including day nursery, school, contact family etc. - to take part in the treatment process?

Questions relating specifically to the abuse

• At what age and during what period did the abuse occur?
• Which person or persons committed the abuse and who if anyone was acquainted with them?
• The nature of the abuse, e.g. how much coercion or violence, pleasure, degradation has the child experienced?
• How has the child explained the matter in its own mind, where does the child put responsibility and blame?
• Sexuality in the family.

The assessment leads to a recommendation concerning the form and content of treatment. The main focus of the treatment is on the actual trauma and its implications for the child's mental health and development, the child's functioning and adjustment and the child's relations. If the abuse and the situation connected with it have affected the child over a protracted period during its formative years, the treatment usually needs to be broader and more generalised. It is important that different forms of treatment should all have the same objective. A case officer in the team has the task of linking together all the measures taken at BUP-Elefanten with those possibly taken by the social services or in school. Network meetings are an important method of co-ordination and of giving everyone a feeling of participation and responsibility.

Trauma-focused therapy
For some children, focused therapy of brief duration is all the help they need. Describing the abuse and variously expressing and processing one's thoughts and feelings about it and what it has entailed are an important part of the treatment. Where younger children are concerned, BUP-Elefanten on the whole complies with the BUSA model developed by the boys' clinic at the Swedish save the Children Federation in Stockholm (Nyman, Svensson, 1995). BUSA is short (in Swedish) for four central areas of treatment: Describing the abuse, Expressing feelings, Speaking out and Accepting.

Describing the abuse is considered "to make it really real" and thus accessible for processing together with the therapist. The interview is enough for some children, but often a child needs help with different types of play material such as ordinary dolls, anatomical dolls, puppets, soft toys of different kinds or drawing materials.

During the processing it is important that the child should be enabled in various ways to express the experiences and feelings connected with the abuse. These can, for example, take the form of feelings of sadness, fear, abandonment and anger. Once again, the child can express this in the interview, but things are often made easier for the child if it can use various kinds of play material. These can symbolise different themes which the child often has to wrestle with, such as good and evil, fear and security. For some children who have difficulty in expressing their feelings, ready-made sentences or pictures expressing different feelings can be a useful aid.

In between, but above all at a later stage of the treatment, it is important for the child to find a way of speaking out and protecting itself from abuse in future. Role play and study materials, for individual or group use, can come in very handy here. Self-assertion training and the corporeal and spatial significance of integrity and limits are worked with, as well as training to say Yes and No. Other strategies reviewed are not having bad secrets and knowing whom one can talk to.

Finally, it is important for the children to be able to move on, establish peer contacts, go in for leisure activities - in short, come to terms with what has happened, accept it and not allow the abuse to govern the rest of their childhood and adolescence.

Treatment of older children and teenagers follows much the same theme, but play material is less relevant here and group therapy tends more often to be an option.

Some children and young persons develop PTSD, post-traumatic stress syndrome. The symptoms can be nightmares, recurrent, painful flashbacks, avoidance of things associated with the trauma, anxiety and panic reactions, concentration difficulties and sleep difficulties. Signs of dissociation can also occur. BUP-Elefanten has begun working with symbol drama and EMDR (Eye Movement Desensitization Reprocessing, Shapiro, 1995) to make it possible for teenagers and adults to start processing the emotions and memories which they feel haunted and tormented by.

Overarching aims of the treatment
The overarching aims which BUP-Elefanten has defined for its therapy work are: for the individual to protect himself and set limits, to be able to identify and handle his thoughts, feelings and behaviour, to rely on others, to achieve greater self-esteem and a better self-image, to develop social skills, not to sexualise relationships, not to assume the role of victim in relationships, and to find his or her own sources of enjoyment and happiness.

Most often at BUP-Elefanten a combination of two or more treatment methods is called for. The methods available are:

Individual therapy

• Counselling of young persons and adults.
• Play therapy.
• Symbol drama.
• Picture therapy,
• EMDR.

Family or individual therapy

Group therapy
Groups have comprised teenage girls, adult women, young, sexually assertive boys (aged 10-14), young sexual abusers (aged 15-19), foster ("family home") parents.

Network therapy
Network therapy is used both as a method in crisis treatment and as a recurrent element of ongoing work. To facilitate and encourage communication between different therapeutic contexts, use is also made of the techniques described by Sheinberg in the multimodal model (1994).

Owing to the difficulties involved in continuously assembling groups of individuals with similar problems and with the right age structure, group therapy is not usually offered until after a brief introductory contact for trauma processing or prolonged therapy.

Summary
BUP-Elefanten is the youngest of the units described in this report and the first specialised unit in Sweden to work with children and young persons who have been physically maltreated and/or sexually abused, and with young persons who have committed sexual abuse. The theoretic content of this treatment model is eclectic, emphasising individual and family treatment, supplemented by group therapy. Its most distinctive characteristic is the importance of early intervention for the treatment process. Methods development, research and teaching are also an important part of day-to-day work.

Evaluation of the therapy work is in progress but has yet to be published.

Evaluation of treatment
The section dealing with the evaluation of treatment is based on the studies reported in Finkelhor and Berliner's conspectus article published in 1995 (Finkelhor, Berliner, 1995) and in Reeker and associates' meta-analytical study of group treatment (Reeker, Ensing, Elliot, 1997), together with supplementary data searches of the well-known Medline and Psychlit databases for 1993-1998. Only studies with more than 15 participants are included. A study with fewer participants than this can in principle be both interesting and informative, but there were three reasons for choosing this limit. Firstly, studies with few participants are limited by low statistical power. That is to say, with few participants it can be hard to demonstrate differences between groups, e.g. a treatment group and a control group. The second reason is that in group therapy a lower limit of 15 will automatically mean the evaluation being based on at least two therapy groups and not just one, which in certain aspects can be regarded as a single case-study. The third reason was a desire to keep the volume of this report within bounds. Only the studies published in scientific articles or books have been included. These are presented in three groups according to design. Studies with an experimental design have been put before studies with a quasi-experimental design, which in turn come before studies with a simple before, after and during design. The reason for this division is that the scientific value of a study is greatest with a randomised control group procedure and least with a simple before and after measurement of a treatment period.

In the following review of different studies, the rating form used has been replaced with a number (x). The forms used are explained on page 32.

Experimental design
In these studies, the individuals examined have been either randomly placed in a treatment or control group or else randomly distributed between two groups receiving different treatment.

Group treatment of girls - control group
Verleur, Hughes and Dobkin de Rios (1986) compared group treatment of 16 girls aged between 13 and 17 with a control group of 14 girls of the same age, all of them institutionalised. The group treatment continued every week for 6 months. Two questionnaires were used before and after the treatment, one measuring self-esteem (5) and one referring to awareness of sexual anatomy/physiology (1).

Both groups had significantly better self-esteem at follow-up, but the therapy group had better self-esteem than the control group. Only the therapy group significantly improved its sexual awareness/knowledge.

Group therapy with girls - waiting list group
Burke (1988) made a randomised sample for a 6-week group therapy with 12 girls aged 8-13 or alternatively a waiting list group of 13 girls.

The children receiving group therapy reduced their symptoms of depression (3), anxiety (24, 7), abuse-related malaise and internalised symptoms (2) compared with the waiting list group.

Play therapy group - individual play therapy - control group
Perez (1988) compared three groups of sexually abused children aged 4-9 years with each other: children in group play therapy (21 children), individual play therapy (18 children) and a control group (16 children). Two questionnaires (PSCI) and a Locus of Control Scale (LCS) were administered before and after the treatment.

The results showed that the children who had received play therapy improved their scores on the measuring instruments. There was no difference between group and individual play therapy.

Group therapy with girls - matched control group
McGain and McKinzey's (1995) study comprised 30 girls aged 9-12 who had been sexually abused. The girls were matched to 15 pairs and randomised to a treatment group and a control group. The treatment group received group therapy once a week for six months, while the control group was on the waiting list during the same time. Two behavioural instruments (23 and 6) were used as dependent variables.

A positive and significant improvement occurred on practically all scales/sub-scales of both instruments after completion of treatment.

Cognitive behavioural therapy and ordinary abuse-specific treatment - ordinary abuse-specific treatment
Berliner and Saunders' (1996) study comprised 80 children aged 4-13 who had been randomly distributed between a treatment group (48 children) and a control group (32 children). The control group received ordinary abuse-specific treatment, while the special treatment group also received cognitive behavioural therapy specially targeting anxiety and angst. Both groups had 10 therapy sessions. Children and parents were assessed before and after the therapy and, respectively, one and two years after the conclusion of their treatment. The survey employed a number of parental and self-response questionnaires (2, 3, 4, 7, 24, 25).

The results showed that the children improved in nearly all respects but that results did not differ between the two treatment options.

Cognitive behavioural therapy - municipal treatment alternative
Deblinger, Lippman and Steer (1996) made a study of 90 children who had been sexually abused and displayed a complete or partial post-traumatic stress syndrome. The children, aged 7-13 years, were randomly divided into three treatment groups and a control group. The treatment was based on cognitive behavioural therapy and the treatments offered were individual treatment only for the child, individual treatment only for the non-offending mother and a combined treatment for mother and child. The control group only received the service which the municipality normally offers. The instruments used before and after the treatment were designed to measure general behaviour (2), the occurrence of depressive symptoms (3), anxiety/malaise (30), PTSD (10) and parenting capacity (18).

The survey showed that the children who had received therapy (individual only, or combined with therapy for the mother) had fewer PTSD symptoms after treatment than the other groups (individual treatment for the mother and controls). The mothers who had received individual therapy reported better parenting capacity, plus fewer externalised symptoms and depression in their children.

Family therapy - family therapy and group treatment
Monck et al (1996) studied 47 children, aged 4-16. and their families, who had randomly received either family therapy (25 children) or family therapy combined with group therapy (22 children). Interviews and questionnaires were used before the treatment started and 12 months later.

At whole-group level, there were noticeable improvements in the mothers' reporting the children's symptoms and the children's own reporting of depressive symptoms (3). No difference could be seen as regards the children's self-esteem or other reported symptoms, nor did teachers' ratings of the children indicate any difference. The mothers' self-esteem and health (8) showed a significant improvement, while the rating of the mother-child relation did not show any change between observation points. But the therapists rated a number of clinical improvements concerning predefined family therapy targets. A comparison between the therapy groups revealed no differences, except in the therapists' ratings of family therapy targets, which showed a better outcome for family therapy combined with group therapy than for family therapy alone.

Cognitive behavioural therapy - non-intervening supportive therapy
Cohen and Mannarino (1996a) studied the treatment outcome for sexual abused pre-school children, aged 3-7. Altogether 67 children were included in the study. The children were blindly randomised to two forms of treatment: cognitive behavioural therapy for sexually abused pre-school children (CBT-SAP) and non-intervening supportive therapy (NST). These forms of therapy were given individually in 12 sessions and lasted for 12 weeks. The week before the treatment started, the children were given a picture/symptom test (19), while the parents rated the occurrence of the children's symptoms/behavioural disturbances (2), sexual behaviour (4) and a continuous recording of the child's behaviour (32). This was repeated at the end of the treatment.

Intra-group comparisons showed that while the NST group receiving non-intervening supportive therapy (28 children) did not improve significantly in terms of symptomatology between observations, most measurements improved positively and significantly for the CBT-SAP group which received cognitive behavioural therapy (39 children). In a statistical comparison between the therapy groups, CBT-SAP showed significantly better results with regard to CBCL (total and internalising) and WBR. There was no difference in the children's self-reporting (19).

In a follow-up of the above study 6 and 12 months later, 43 children took part: 28 who had received CBT-SAP and 15 who had received NST (Cohen, Mannarino, 1997). The results showed significant differences in favour of the CBT-SAP group regarding most of the follow-up measurements CBCL (total. internalisation, 32). There was no difference between the groups as regards sexual behaviour (4), but 12 children from the NSYT group (none from the CBT-SAP group) had had to discontinue the treatment on account of sexually assertive behaviour, so that abuse-focused cognitive behavioural therapy was construed as being clinically preferable to a non-intervening supportive therapy.

The analysis of family data and family questionnaires showed the emotional status of the non-offending parent (15) at the end of treatment to be the strongest predictor of the treatment outcome, while parental support (21 and 14) was the most important predictor in follow-up after 6 and 12 months (Cohen, Mannarino, 1998a).

Cognitive behavioural therapy - non-intervening supportive therapy
Cohen and Mannarino (1998b) compared the treatment outcomes of two forms of therapy: abuse-specific cognitive behavioural therapy (SAS-CBT) and non-intervening supportive therapy (NST). The study included 49 children aged 7-14 who had recently been sexually abused. The children were randomly allotted the two forms of therapy, which were given individually for 12 sessions spread out over 12 weeks. Owing to uneven dropout in the original material, 30 received SAS-CBT and 19 received NST. One week before the treatment started and at the end of the final session, the occurrence was measured of behavioural disturbances (2), depressive symptoms (3), anxiety (30) and sexual behaviour (4) in the children.

The results showed that the children who had received abuse-specific cognitive behavioural therapy reduced their depressive symptoms significantly more (3) and improved their social competence significantly more (2) than those receiving non-intervening supportive therapy. A clinical comparison at ± 1SD on the various measuring instruments showed the percentage improved > + 1SD to be significantly greater on most measuring instruments for the children who had received SAS-CBT, while a deterioration <-1SD was more common among the children who had received NST. The study also showed that SAS-CBT was possibly more successful in reducing sexual behaviour in the children. The dropout rate (discontinuation of treatment) was also lower in the group receiving SAS-CBT than in the group receiving NST.

Quasi-experimental design
This group contains studies which used comparison groups but without any random distribution between survey group and comparison group.

Psychodynamic therapy - supportive therapy
Downing, Jenkins and Fisher (1988) compared treatment outcomes between two groups of children receiving either a psychodynamically oriented treatment (12 children) or a supportive therapy mainly addressed to their parents (10 children). The children in the groups were aged between 6 and 12. Parents and teacher reports were used, but no standardised instruments.

According to the parents' ratings, the children receiving supportive therapy improved more with regard to sleep difficulties, sex games with other children, enuresis and generally assertive behaviour, while neither group showed any improvement with regard to sexual self-stimulation.

Crisis-oriented short term treatment - crisis-oriented treatment and continued treatment
Gomez-Schwartz and associates (1990) studied 156 children aged up to 18 at the commencement of therapy (Family Crisis Program) and after 18 months. At follow-up one group had received introductory treatment while another had received introductory and continued FCP treatment (15 per cent) and a third was receiving introductory treatment, followed by continued treatment outside FCP (20 per cent). The research report does not show the sizes of the different groups. The introductory one is best described as crisis-oriented short-term treatment. At the commencement of the therapy and at follow-up measurements were taken of the children's general behaviour (11 and 2) and of their self-esteem (17).

The majority of children showed a reduction of general behavioural problems and an improved self-image after the therapy. The children who had received continued FCP therapy showed the biggest improvement. It is hard to evaluate the outcome, because the report does not give any comparative figures between the different groups.

Many-faceted group therapy - control group
Rust and Troupe (1991) examined 25 girls, aged 9-18, who had undergone a 24-week many-faceted group therapy programme, comparing them with a control group of girls with approximately the same background. An achievement test (26) and a test measuring self-esteem (16) were used before and after the treatment.

The results showed that the treatment group, i.e. the sexually abused children, had significantly lower academic achievement and inferior self-esteem compared with the control group before the treatment. The treatment group improved their results significantly between the observations, as regards both academic achievement and self-esteem.

Individual therapy - control group
Sullivan et al (1992) examined 72 deaf children between the ages of 12 and 16 who were institutionalised and had all been sexually abused. One group of 35 children received therapy while another group of 37 children, whose parents declined therapy, was used as a control group. The house wardens made blind ratings of the children's general behavioural status (2) before and one year after the treatment had been inaugurated. The treatment consisted of 2 hours' individual therapy for 36 weeks.

At follow-up the boys showed a significant improvement on the aggregate CBCL scale, the scales for externalisation and internalisation and another 7 sub-scales, while the girls showed a significant improvement on the aggregate CBCL scale, the externalisation scale and another 3 sub-scales.

Therapy - control group
Oates et al (1994) present results from a study of 84 sexually abused children aged between 5 and 15. The children were examined at the beginning of the therapy and 18 months later, and were compared with a control group of 76 children. The 18-months follow-up was completed for 64 children. Only 65 per cent of the abused children underwent training lasting on average for 9 months. Occurrence of depressive symptoms (3), general behavioural disturbances (2) and self-esteem (20, 13, 16) were used as measuring variables.

Statistical processing showed that there was no connection between treatment and outcome.

Before, during and after design
This group contains studies which used before and after measurements but lack comparison groups. Some, moreover, have observations during the treatment. The quasi-experimental or experimental studies, of course, are also studies with before -and-after design.

Cognitive behavioural therapy for sexually abused children with PTSD
Deblinger, McLeer, Henry (1990) examined the treatment outcomes of 19 girls aged 3-19 who had undergone a programme of cognitive behavioural therapy developed for sexually abused children with post-traumatic distress syndrome. The children were investigated by interviewing their parents about the occurrence of post-traumatic stress symptoms in the children, the parents reporting the children's general behaviour (2) and children aged over 6 filling in a depression form (3) as well as a form concerning the occurrence of anxiety (3) at the time of assessment, at the commencement of treatment, 2-3 weeks later, and after completing 12 therapy sessions.

During the baseline period of 2-3 weeks, all measures were unchanged. After the treatment period, all measurements had undergone a statistically confirmed improvement.

Multimodal treatment
Friedrich et al (1992) examined 33 boys (out of originally 42) aged between 4 and 16 who had completed a multimodal treatment consisting of a combination of group therapy, individual therapy, parental training and family therapy over an 8-month period. The parents filled out questionnaires concerning the children's general behaviour (2) and the occurrence of sexual behaviour (4) in the children. The children themselves completed a depression questionnaire (3) and two questionnaires measuring self-esteem (16 and 12) before and after treatment.

Statistically reliable changes were seen in the parents' rating of the children's general behaviour and sexual behaviour. No difference was seen regarding the children's rating of depression and self-esteem. The therapy outcome was connected with the degree of depression in the mother, the mother's social support of the child, the degree of family conflict, the gravity of the abuse, previous treatment and time in therapy.

Abuse-focused therapy
Lanktree and Briere (1994) investigated the treatment outcome for 105 boys and girls who had been sexually abused and had undergone an abuse-focused therapy. The children were between the ages of 8 and 15, and 71 of them completed at least 3 months' treatment. The children's health development and behavioural development were rated every three months with self-response questionnaires for trauma symptoms (31) and depression symptoms (3).

Scores on the five sub-scales for depression, anxiety, post-traumatic stress, dissociation and anger declined significantly after 3 months, but those for sexual problems did not. Scores on all sub-scales except dissociation continued to improve in subsequent follow-ups, while the score on the sub-scale for sexual problems did not respond until later in the treatment.

Group therapy
Richardson (1994) studied 69 girls aged between 6 and 16 who had undergone a group therapy programme lasting for between 6 and 12 weeks. In a pre-post design the children's general behaviour was assessed by the child, parents and teachers (2, 33).

The programme was most successful with the older girls, and the self-reported symptoms of depression and criminal behaviour, together with the teacher-reported symptoms anxiety, withdrawal, depression, lack of initiative and aggressiveness.

Group treatment, cognitive behavioural therapy
Sinclair et al (1995) evaluated a 20-week group therapy programme based on cognitive behavioural therapy. The population consisted of 6 therapy groups totalling 43 girls between the ages of 12 and 18, all of them institutionalised. Observations were made within a week of the first group session and a few weeks after the conclusion of group therapy. The youngsters' general behaviours were measured (2, 33), as well as sub-scales of the same measuring post-traumatic stress syndrome (YSR-PTSD and CBCL-PTSD). In addition, the youngsters themselves filled out questionnaires concerning depressive symptoms (22) and self-esteem (29).

Significant improvements could be seen concerning general symptoms (YSR; internalisation, externalisation, total and PTSD and CBCL; internalisation and PTSD) and self-esteem (29). On the other hand there was no difference between the observations where depressive symptoms were concerned (22).

Group treatment, cognitive behavioural therapy for children and mothers
Stauffer and Deblinger (1996) evaluated the effect of an 11-week group treatment programme for children aged between 2 and 6 and their non-offending mothers. 19 children were included in the treatment, which was based on cognitive behavioural therapy. The parents' mental health and stress were measured with (5, 9, 28), the parents' attitudes to their children were assessed (18), and so were the children's general mental health and behaviour through (2) and sexual behaviour (4). Measurements were taken at the first contact, at the commencement of treatment, on conclusion of treatment and three months later.

The results showed a significant improvement in parental health and stress and in parental attitudes to the children, as well as reduction of the children's sexual behaviour. This outcome could be referred to the treatment, no difference being visible during the baseline measurement before the treatment began.

Multicentre study of 6 community-based treatment programmes
In a multicentre study in England, Monck, E (1997) examined 239 sexually abused children between the ages of 4 and 16. Measurements were taken at the commencement of the therapy, at its conclusion and 12 months later. The treatment offered varied between 6 and 35 times (13 sessions on average) and can be described as a combination of community-based individual therapy, family therapy and group therapy. The parents completed questionnaire form CBCL (2) concerning their child, and children over 11 completed CDI, a Self-Esteem Inventory, YRS and FSSC-R.

The questionnaire scores for depression (3) and self-esteem (16) declined between the two observation points, whereas no change occurred regarding fear/anxiety (7). Parents and children both reported significantly reduced general symptoms and behavioural disturbances (2, 33) at the end of the therapy. The parents reported significant reductions of both the internalisation and externalisation factors (2), while the children themselves only reported a significant change in the internalisation factor (33).

At the first observation 163 mothers and 144 children replied, but at the end of the treatment only 52 (32 per cent) mothers and 79 (55 per cent) children. At the twelve-month follow-up, the number was so small that no results were presented.

Abuse-focused treatment
Gothard and Heinrich (1999) are continuously evaluating the outcome of treatment at The Center for Child Protection in San Diego, which offers abuse-focused treatment. Through multiple measurements, at the commencement and every six months until the end of treatment, the children's and parents' development is being measured on a host of scales which for the children include Youth Self report (33), Trauma Symptom Checklist (31), Child Depression Inventory (3), and for the parents Child Sexual Behavior Inventory (4) and Child Behavior Checklist (2).

After 6 months in therapy the children had improved their scores on a number of different scales, above all with regard to internalised symptoms, anxiety, depression and PTSD.

Summary concerning treatment
As can easily be seen, therapy research concerning sexually abused children and young persons is an area in which most of the work still remains to be done. Many of the studies which have been published suffer from the inclusion of only a few children, and dropout rates are high in cases where treatment continues over a long period. Comparability, moreover, is limited by great variations between the studies regarding gender and age structures, initial symptomatology and focus of therapy. Control groups are often lacking, for ethical reasons, and those using a more experimental design and comparing different treatments are, as yet, few in number.

It is encouraging to note that more experimental studies have been reported in recent years. The experimental studies which have randomly distributed sexually abused children between treatment and control groups have shown a fairly unanimous picture, namely that the treatment groups have developed significantly better than the control groups (Burke, 1988, Perez, 1988, Verleur et al, 1986, McGain, McKinzey, 1995).

The experimental studies randomly allocating children between two different forms of treatment have presented a somewhat different picture (Berliner, Saunders, 1993, Deblinger et al, 1996, Cohen and Mannario, 1996a, 1997, 1998b, Monck et al, 1994, Perez, 1998). Several of these studies have failed to demonstrate any difference between different treatment models, such as anxiety versus non-anxiety-reductive programme elements (Berliner, Saunders, 1993), group treatment versus group treatment/family therapy (Monck et al, 1996), and group play therapy versus individual play therapy (Perez, 1988).

Cohen and Mannario's studies (1996a, 1997, 1998b), however, show statistically reliable differences in favour of abuse-focused cognitive behavioural therapy compared with non-intervening supportive therapy. Comparative behavioural studies have the inherent weakness of more being needed to demonstrate a difference between two treatments, each of which may be effective, than to demonstrate a difference between a treatment group and a control group receiving no treatment.

The quasi-experimental studies support the picture of children being improved by therapy and not just as a spontaneous effect over time. The weakness of several of these studies, however, is that the control groups have not been fully comparable with the treatment groups. Several studies comparing sexually abused children who have received therapy with those who not received any, except what the community provides, have failed to show any differences. As Finkelhor and Berliner (1995) point out, this may be because those children received good parental support and were not felt to be in need of therapy.

Most of the studies can be described as pre-post, and all such studies show that, on one or more criteria, children have improved significantly by the time of follow-up after treatment. But those studies cannot answer the question of whether this is an effect of therapy or a spontaneous improvement over time, though some studies have tried to better this by having an observation period between the first assessment and the commencement of therapy (Deblinger et al, 1990, Stauffer, Deblinger, 1993).Both these studies showed that no spontaneous improvement took place during the observation period, which corroborates the efficacy of the treatment. One of the studies, moreover, had an observation period corresponding in length to the ensuing treatment period (Stauffer and Deblinger, 1993).

One problem reported, for example, by Monck (1997) and Gothard and Heinrich (1999) where clinically ongoing data gathering is concerned, is the difficulty of getting therapists to administer and complete the collection of data in a research project, not least in a multi-centre study. There are probably several reasons for this. Inexperience of research and research methodology and of meaningfulness, i.e. of the potential benefit to therapeutic work, are probably two impeding factors. Pressure of time at work and space for research administration not being given or not being adequately motivated are other explanations.

Fear of research and the amount of particulars involved by data gathering disrupting or harming the therapeutic relationship is another factor. In Monck's study (1997) of, initially, 239 children, only 66 reports, corresponding to 28 per cent, were completed by the therapists at the conclusion of treatment. In Monck's (1997) analysis of difficulties, the heads of teams in the treatment units included were able to identify four factors. The first reason was that the therapists believed that the evaluation of the children's development in therapy would lead to an evaluation of the therapists. Secondly, some therapists felt that the evaluation process, e.g. the children having to fill out self-response questionnaire forms, would amount to a continuation of abuse. The third identifiable reason was that some therapists did not consider the filling out of data gathering questionnaires to be part of their duty as therapists. Another reason, finally, was that research was something foisted on them from outside. Similar difficulties have been reported by others (Wiffen, 1994). Both Monck (1997) and Gothard and Heinrich (1999) argue the importance of researchers speaking the same language as the clinics, supporting and motivating them not only to participate but also to c