A Preliminary Community Study of Dialectical Behaviour Therapy (DBT) with Adolescent Females Demonstrating Persistent, Deliberate Self-Harm (DSH)



A study of dialectical behaviour therapy (DBT) for 16 older female adolescents with persistent and severe deliberate self-harm was undertaken. Independent assessments were carried out pre- and post-treatment, and at eight months follow-up. On average subjects completed more than 78% of sessions. There was a marked reduction in self reported depression (F = 12.8, df = 2, p < .001); hopelessness (F = 15.9, df = 2, p < .001); episodes of deliberate self-harm (F = 23.9, df = 2, p < .001); alongside an increase in general functioning (F = 22.9, df = 2, p < .001).

Key Practitioner Message:

  •  DBT appears to be a promising treatment for adolescents with severe and persistent deliberate self-harm.
  •  DBT can be delivered as a community treatment.
  •  Further evaluation is needed with a broader subject group, including both males and females.
  •  It is necessary to carry out further research, particularly randomised trials, before DBT can be widely recommended.


Deliberate self harm (DSH) and youth suicide represent major health concerns. In the UK, the rate of adolescent DSH in the community is high with 11.2% of females and 3.2% of males reporting an episode within the previous year (Hawton, Rodham, & Weatherall 2002). Suicide accounts for over 20% of deaths among UK male adolescents and young adults between the ages of 15 and 24, at a rate of 11/ 100,000 (http://www.samaritans.org/know/information/suicide_stats.shtm (date accessed: 19.06.07)). Although the rate for female youth suicide is lower at 3/100,000, young females who engage in repeated DSH are at particular risk of suicide (Zahl & Hawton, 2004). Despite this, the evidence-base for effective treatments for the adolescent exhibiting DSH is limited. Controlled trials of group therapy (Wood et al., 2001), family therapy (Harrington et al., 2000) and emergency room interventions (Rotheram-Borus et al., 2000) have shown some potential but with caveats; for instance, domiciliary family therapy (Harrington et al., 2000) did not benefit depressed adolescents presenting with DSH.

Available data on the demands DSH puts upon psychological services are difficult to quantify; its assessment requires considerable manpower to provide both routine and crisis responses (Kelvin, 2005). Kelvin has estimated that CAMHS serving a normal catchment area population of 350,000 are likely to see 300 cases of DSH per year, of which 43 will be repeat cases. Kelvin calculated that, if available, 22, would go into intensive treatment, requiring 660 hours of therapy with a multi disciplinary team plus 165 hours of family therapy, in addition to 525 hours for assessments.

For adults, the National Institute for Clinical Excellence (NICE) guidelines (NICE, 2004), a Cochrane review (Hawton et al., 1999) and the American Psychiatric Association (2001) recommend Dialectical Behaviour Therapy (DBT) for DSH. The NICE guidelines for repeated DSH in adolescents (NICE, 2004) recommend developmental group psychotherapy, based upon the evidence from just one randomised controlled trial (Wood et al., 2001). Compared to treatment as usual, DBT has been shown to reduce DSH and re-hospitalisation in groups of adult female patients with the diagnosis of borderline personality disorder (Linehan et al., 2006, 1994, 1991; Koons et al., 2001; Verheul et al., 2003; van den Bosch et al., 2005), although a Cochrane review was equivocal (Binks et al., 2006). As yet, however, the evidence for the use of DBT with adolescents is sparse. A study of 53 adolescents admitted to hospital following DSH and allocated to either DBT or treatment as usual found no differences with respect to depression scores, suicidal ideation, rates of re-hospitalisation and deliberate self-harm one year after admission. However, the number of self-harm incidents on the ward was reduced in the group receiving DBT (Katz et al., 2004). The latter is an important finding, as it is recognised that adolescents with persistent DSH and borderline traits can regress and develop a malignant dependency in a hospital setting.

The aim of this preliminary study was to evaluate the effectiveness of DBT in a community sample of adolescents referred for treatment as a result of persistent DSH and suicidal ideation.



DBT is a manualised treatment, originally designed for the treatment of adult borderline personality disorder. At its core, cognitive and behavioural techniques are employed within a Zen philosophical framework, with an emphasis upon mindfulness (Linehan, 1993a, 1993b). An important element is the attention that therapists pay to the dialectic between validation and acceptance of the patients as they are, and at the same time helping them to change (Swales, Heard, & Williams, 2000). In addition to self-harming behaviours, there is a focus upon helping the patients with any deficits they may have in the areas of emotion regulation, impulse control and problem-solving skills. As part of the DBT programme, therapy interfering behaviours are specifically addressed (Heard, 2002).

There are four stages to DBT as summarised in a review by Blernnerhasset and O'Raghallaigh (2005). In stage 1 the primary and crucial focus is on stabilising the patient and achieving behavioural control. The target behaviours are: (a) decreasing life-threatening suicidal behaviours; (b) decreasing therapy interfering behaviours; (c) decreasing quality-of-life interfering behaviours; (d) increasing behavioural skills. Stage 1 is expected to take at least a year. In stage 2 the focus is on treating problems related to any past trauma. In stage 3 the emphasis is on the development of self-esteem and the effective management of problems of daily living. Stage 4 aims to optimise the capacity of the individual to develop and gain new experiences.

In this study, DBT was delivered in the community. This report is on stage 1 of the treatment only. The treatment package consisted of a once-weekly skills training group, a once- weekly hour-long individual session and telephone support. During the skills group, which lasts for one and a half hours, the therapist helps the participants learn essential skills such as emotional regulation and problem solving. There are four modules: core mindfulness (from aspects of Zen Buddhism), distress tolerance focusing on acceptance, interpersonal effectiveness focusing on change and finally, emotion regulation. These modules are then cemented within the weekly individual psychotherapy sessions, which use techniques from cognitive-behaviour therapy. Each individual also has access to telephone consultation and support. If the subjects phoned they would be supported to use the skills they had learnt in groups. If a psychiatric emergency arose that could not be dealt with by telephone support, the subject would be referred to the child and adolescent mental health team. The stage 1 was divided into two six-month blocks to allow review of progress. While the participant is in DBT, psychotropic medication is allowed, but any other psychotherapy is discouraged. Casework continued and was required to help link with agencies, parents and other parties.


The study was approved by the Oxford Mental Healthcare Trust Audit Committee and written consent was obtained from all the participants. Subjects were referred from the outpatients clinics in the child and adolescent mental health services (CAMHS) covering Oxfordshire, with a population of 650,000. To be included in the study subjects had to have a history of more than six months of severe and persistent DSH, defined after Hawton et al. (2002) as: an act with a non-fatal outcome in which an individual deliberately did one or more of the following: self cutting; jumping from a height which they intended to cause harm; ingesting a substance in excess of the prescribed or generally recognised therapeutic dose; ingesting a recreational or illicit drug that was an act that the person recognised as self harm; ingesting a non-ingestible substance or object. Exclusion criteria included a diagnosis of schizophrenia, bipolar disorder, autism, autistic spectrum disorder and those with moderate and severe mental impairment. Participants had been referred to the DBT clinic over a two-month period by community psychiatric nurses and psychiatrists. All had extensive histories of DSH over 18 months to two years which had not responded to other psychiatric treatments including medication, individual psychotherapy, family therapy and where necessary, focused abuse work. Twenty-five adolescents were referred, however, only sixteen agreed to participate in the study. All were female between the ages of 15 and 18 years. There was no difference in terms of age, family situation, levels of DSH between those who agreed to participate in the study and those who refused.


Independent assessments were carried at the start and end of treatment, and at follow-up (mean length of follow-up 268 (SD 108) days). The Structured Clinical Interview for DSM-IV 11 (SCID-II) was used to assess and quantify borderline personality traits, rather than make a diagnosis of borderline personality disorder (BPD). The diagnosis of BPD is not formally made until adulthood, and the SCID-II is not validated for this purpose in this age range. However, all the subjects would have qualified for a diagnosis of BPD if they had been aged 18 years or older. All participants underwent a SCID II interview, however, due to resource limitations only the first nine were re-interviewed at the end of treatment. Other assessment instruments included the Beck Depression Inventory (BDI) (Beck, 1979), the Beck Hopelessness Scale (BHS) (Beck et al., 1974) and the DSM-IV Global Assessment of Functioning (GAF) (APA, 1994). The number of episodes of DSH per week was determined by clinical interview. For those who did not complete therapy (n = 2) the last observations were carried forward and analysed.


Statistical analyses were performed using SPSS for Windows (12.0.1) (2003). A General Linear Model (GLM) repeat measures analysis was undertaken of BDI, BHS, GAF, DSH over three time periods: pre- treatment, post treatment and follow-up. The significance level (α) was .05 (2 tailed).


There were 16 participants, all female, average age 16.4 years (SD 1.2). Their domestic situation was often problematic with only five living in intact families, half in one parent families or with step-parents and three living independently in hostels with little or no day-to-day support. There were two with documented histories of abuse following child protection investigations by the Local Authority. Eight (8/16 50%) were abusing drugs including crack, heroin and cannabis and two were on methadone. All had been seen in child and adolescent mental health services (CAMHS) and nine had been prescribed an antidepressant, with two also receiving low dose antispychotics - risperidone and chlorpromazine.

For the sixteen participants, the average percentage of sessions attended was 78% (SD 18.6%) with a range of 36–100%. There was only one extra consultation with the psychiatric services over a period of one year and no participant required psychiatric hospitalisation during the study period or during the 6 months post-treatment. Three, however, went on to need adult psychiatric care. Medication was not altered during the course of the DBT, and afterwards medication levels remained similar (6/14).

A GLM repeat measures analysis was undertaken of the BDI, BHS, GAF and number of episodes of DSH over three time periods: pre- treatment, post treatment and follow-up (table 1). There was a marked reduction in self-report depression scores (BDI) (F = 12.79, df = 2, p < .001), hopelessness (BHS) (F = 15.97, df = 2, p < .001), episodes of deliberate self-harm (F = 23.95, df = 2, p < .001), and an increase in general functioning (GAF) (F = 22.95, df = 2, p < .001).

Table 1.   Pre- and post- dialectical behaviour therapy treatment scores
MeasureNumberPre Mean (SD)Post Mean (SD)Follow-up Mean (SD) F df = 2P value
  1. Note: Measures include: the number of incidents of deliberate self-harm per week (DSH); Beck Depression Inventory (BDI); Beck Hopelessness Scale (BHS); General Assessment of Functioning (GAF); and the Structured Clinical Interview for DSM-IV II (SCID-II)

DSH163.0 (2.0)1.0 (1.5)0.53 (0.89)24.0<.001
BDI1642.0 (9.6)29.5 (19.3)20.4 (17.3)12.8<.001
BHS1615.3 (4.6)8.6 (4.6)6.4 (8.7)15.9<.001
GAF1653.8 (14.8)74.5 (9.5)80.3 (10.6)22.9<.001
SCID-II77.5 (3.2)1.5 (2.0) 53.7 df = 1<.001


The study examined the effectiveness of DBT in a community sample of female adolescents who displayed high rates of DSH or persistent suicidal ideation. Overall, there was a marked improvement in the scores on all measures post DBT treatment, in line with studies of adults (Koons et al., 2001; Linehan et al., 1991, 1994, 2006) and the inpatient study of adolescents (Katz et al., 2004). There are, however, several important limitations to this study. Firstly, this was an evaluation of a treatment outcome and not a randomised controlled trial. The sample was small, and evaluation referred only to stage 1 of DBT. Without controls it is not possible to compare the natural remission rate for the DSH and associated symptoms. However, the level of symptoms and DSH displayed by this group of subjects were stable at a high rate over a period of 12–18 months before the study, and had shown little or no improvement with other forms of treatment. Secondly, the referred subjects were all female so it is not possible to extrapolate the findings to males. Nevertheless the all female sample may reflect differing patterns of referral to CAMHS with a psychiatric presentation, as an ongoing audit of DBT for adolescents in Local Authority Care and youths involved in the Criminal Justice System has revealed around 30% males with histories of severe DSH (Oxfordshire DBT Service).

Overall, the treatment was well accepted in a group that is notoriously difficult to engage in psychotherapy. The average percentage of sessions completed was 78% of stage 1, although two dropped out before the end of treatment.

Part of the DBT programme consists of helping the patient learn skills and develop competencies to overcome the emotional dysregulation, which is seen as a fundamental and characteristic deficiency in those who repeatedly engage in DSH. For the first seven participants scores on the SCID-II reduced significantly, indicating that they may have learnt to manage and control their self-harming behaviours and emotions. The episodes of DSH reduced significantly in all participants dropping from an average of 3.5 to <0.5 incidents per week (F = 20.5, df = 2, p > .001). They did, however, cite DBT skills such as chain analysis as being particularly useful in coping with emotions that would usually lead them to DSH. Notably, as in the study by Koons et al. (2001), the measures of depression and hopelessness (BDI and BHS) were significantly reduced post treatment. This is important as both are related to repetition of DSH (Hawton, Kingsbury, James, et al., 1999; Hawton, Kingsbury, Steinhardt, et al., 1999), severity of DSH, and the likelihood of suicide (Evans, Hawton, & Rodham, 2004), while depressive affect is a predictor of treatment non-response in suicidal youth (Huey et al., 2005). Besides these improvements, there was an improvement in the participants’ overall functioning as reflected in the GAF scores. Notably only half were in normal education or employment pre-treatment, whereas post treatment 13 out of 15 (one unknown) were in education or employment, a considerable improvement. The improvement seen at the end of treatment on all scales of psychopathology, global functioning and deliberate self-harm was maintained at follow-up. Indeed, there is a linear pattern of reduction suggesting that any benefit derived during the treatment phase is not only maintained but continues. Further studies could usefully determine whether this continued improvement is due to longer-term changes in the patient's dysfunctional thought processes, rather than immediately reinforced behaviour.

Interestingly, the only other study of DBT with adolescents in inpatient settings (Katz et al., 2004) found a similar reduction in DSH and depression scores, but not greater than treatment as usual. However, the authors pointed out that within the framework of DBT, the practitioners were more confident and this, in turn, was reflected in a reduced number of incidents of DSH in the hospital setting. The nurse practitioners in this study felt similarly empowered and confident to deal with a demanding and worrying group of patients without recourse to any patient being hospitalised.

Criticisms of DBT (Scheel, 2000) have included the small number of controlled trials, small sample sizes, and a lack of persistent effects at one-year post treatment (Linehan et al., 1994). These concerns, have been answered, in part, by a recent large scale study of adult patients (n = 101), over a two-year period, which demonstrated considerable benefits of DBT versus standard treatment with lower rates of suicide attempts, re-hospitalisation and drop-out from therapy (Linehan et al., 2006). There have been few studies of DBT with adolescents and clearly, this is only a small, preliminary study. To be able to make more confident conclusions about the effectiveness of DBT in the treatment of DSH, the next stage would be a randomised control trial. DBT is an extensive treatment involving telephone support and at least two therapy sessions per week. The question, therefore, of a comparison treatment to disentangle the effects of therapy versus support, is important. Further, if DBT is a successful treatment, which posits the need for the patient to learn skills to overcome emotional dysregulation, then it will be important to determine whether adolescents have, indeed, learnt skills, rather than solely relying on outcome measures of depression and rates of DSH.

Adolescent DSH is often associated with family dysfunction (Kerfoot et al., 1996) and, therefore, it will be important to develop DBT in conjunction with family treatments, as proposed by Miller et al. (2002). In this study all of the participants had received standard outpatient treatments, which included family therapy. In the context of repeated DSH, however, family relationships were often very strained, with high levels of expressed emotion (EE), often militating against family meetings. Although family therapy was not undertaken during the course of this DBT study, a psycho-education group for parents and carers has been initiated for subsequent DBT groups.

In adult studies the lack of an effect of DBT on features of the borderline pathology, other than suicidality, for example interpersonal instability, chronic features of emptiness and boredom, has led some to suggest that DBT is a treatment for life threatening impulsive disorders, rather than borderline personality disorder per se (Verheul et al., 2003). While the diagnosis of BPD was not used in this study, it is noteworthy that there appears to be a reduction of these characteristic behaviours. It may be that DBT does, indeed, improve the adolescent's pathology sufficiently to prevent the development of borderline personality disorder, however, other possible explanations need to be examined. Firstly, it is possible that this group of young people were not sufficiently disordered to be considered as candidates for developing a borderline personality disorder in adulthood. This is unlikely because these older adolescents were the most disturbed in the CAMH service serving a population of 650,000, and each one had a long history of borderline type pathology. Second, there is a natural recovery in these types of behaviours and evidence that borderline states do improve with time (Zanarini et al., 2006). It is also likely that there is some plasticity of personality development in the adolescent period. Clearly, a small trial with limited follow-up time is not going to answer the question whether DBT undertaken in adolescence can prevent the development of borderline personality disorder. A larger controlled trial, with treatment as usual as the control condition, and sufficient follow-up, could shed light on this very interesting question.

A recent psychotherapeutic advance has been the development of a mentalisation-based psychotherapy (Bateman & Fonagy, 2004). This has been shown, within a partial hospitalisation programme, to be effective for adults with borderline personality disorder (Bateman & Fonagy, 2004), however, it has not been adapted for adolescents with severe and persistent DSH. This may be possible, and although it shares many elements with DBT, it is probably less intensive and may represent an alternative treatment choice. Another possible treatment option, reported in adult patients with borderline personality disorder is the use of olanzapine, an atypical antipsychotic medication, in conjunction with DBT (Soler et al., 2005). There may be concern, however, about the use of antipsychotic medication in such a young, non-psychotic population. A recent trial comprising of 10 sessions of CBT (n = 120) (Brown et al., 2005) for adults with repeated DSH was successful, with 41% (n = 23) in the control group compared to 24% (n = 13) in the intervention group (p = .045). The question arises whether CBT alone is sufficient or whether there is a need for the structured DBT package, which includes many aspects of CBT.

The NICE guidelines, a Cochrane review and the APA all recommend DBT for DSH in adults. Should these recommendations apply to adolescents? At the moment there is insufficient evidence to answer this question. The lack of a robust evidence-base for treatments for DSH in adolescence is worrying, given the frequency of DSH and its potential serious consequences, and it highlights the need for randomised control treatment studies. This study gives an indication that a larger trial may be warranted.


We would like to thank the patients and the DBT team for their help and co-operation.