Depression is a significant public health problem. It affects twice as many females as males (
1,
2 ). Risk factors involved in its development and clinical course include a history of childhood physical and sexual abuse (
3,
4 ). Patients with an abuse history often present with a severe and complex psychopathology, including posttraumatic stress disorder (PTSD), chronic symptoms of anxiety and depression, impulsiveness, and interpersonal difficulties (
5,
6,
7,
8 ).
The complex clinical picture observed among adults who have a history of childhood trauma has been explained as a process of neurobiological and psychological vulnerability (
9 ). It has been argued that traumatic events should be taken into account in the treatment of these patients (
10 ). However, early trauma and its consequences are not always addressed when patients consult health services (
11 ). Therefore, it is important to develop effective treatments and evaluate their impact (
12 ).
In the United States and Canada, several treatments have been developed to deal with the consequences of early trauma (
12,
13,
14,
15,
16 ). The clearest empirical evidence of effectiveness in treating the symptoms associated with PTSD has been found for the cognitive-behavioral approach (
17,
18 ). This approach has also been shown to be effective in treating women with alcohol dependence and PTSD (
19 ) and in treating depressive symptoms among adults with a primary diagnosis of PTSD (
20 ).
A U.S. multicenter study of patients with depression showed that those with early trauma responded better to cognitive-behavioral-analytic therapy than to pharmacotherapy (
21 ). However, the study did not consider the impact of this therapy on other consequences among adults of childhood trauma (
22 ).
The efficacy of a cognitive-behavioral orientation in addressing psychosocial difficulties among persons with complex symptoms resulting from childhood trauma has not yet been demonstrated (
23 ). In addition, the effectiveness of a psychodynamic orientation in the treatment of these individuals has not been empirically validated (
24 ).
In Chile depression is the second-highest contributor to disability-adjusted life years among women (
25 ). Findings of Chilean national epidemiological studies have confirmed those of international studies with respect to the relationship between early trauma and affective disorders (
26,
27,
28 ). However, standardized programs in Chile for the treatment of depression do not include routine screening for traumatic childhood experiences and analysis of their consequences (
29 ).
In 2002 at Curicó Hospital, located 180 km south of Santiago, we found a prevalence rate of childhood sexual abuse of 42% among patients with chronic depression (
30 ). A majority of cases featured important psychosocial factors that had not been considered in the patients' treatment (
31 ). In view of such evidence, we developed an intervention for this patient group. In addition to standard pharmacological treatment, the model incorporated a three-month structured intervention by a multidisciplinary team. Its objectives were to actively explore early traumatic experiences among women with depression and to focus treatment on interpersonal difficulties with use of a psychodynamic orientation (
32 ).
The objective of the study was to compare the efficacy of this intervention and a standard intervention recommended in clinical guidelines issued by the Chilean Department of Health for women with severe depression and childhood trauma.
Methods
Sample and procedures
The study was approved by the hospital's Ethics Committee and was authorized by the Committee of Ethics of the Health Service of Maule. Between April and August of 2006, all women over 20 years of age who were given a diagnosis of severe depression in a primary care clinic or other health service were sent by the clinic or service to the Curicó Hospital for an outpatient evaluation (N=154). They were randomly assigned to one of three senior psychiatrists. Each psychiatrist had experience in diagnosing and treating affective disorders and worked in the mental health unit of the hospital. One of the psychiatrists was a member of the team providing the intervention.
During an initial psychiatric interview the psychiatrist assessed the severity of depression according to
ICD-10 (
33 ). Interpersonal difficulties were also assessed. The need for pharmacological treatment was determined in accordance with the practice guidelines established by the Chilean Ministry of Health (
29 ). The guidelines offer algorithms about prescription of antidepressants (fluoxetine, sertraline, paroxetine, and venlafaxine), mood stabilizers (lithium, valproate, and lamotrigine), and second-generation antipsychotics (risperidone, olanzapine, and quetiapine).
Intervention
Patients with severe depression according to
ICD-10 were referred to a clinical psychologist, who inquired in a climate of safety and confidence about whether they needed to speak about their history and link past traumatic experiences to present interpersonal difficulties. A total of 136 patients agreed to be evaluated by the psychologist, who administered the Hamilton Depression Scale (Ham-D) (
34 ) and the screening scale for early trauma developed by Marshall and colleagues (
35 ) and validated in Chile by one of the authors (
36 ). This scale asks whether an individual has memories of having one or more of the following traumatic experiences before the age of 15: traumatic separation from a parent or caregiver, alcohol or drug abuse of a family member, physical injury associated with punishment, and forced sexual contact with a relative or a nonrelative.
Two trained psychology students who were blind to the results of the trauma screening scale administered the section of the Composite International Diagnostic Interview that is used to determine an
ICD-10 diagnosis of posttraumatic stress disorder (PTSD) (
37 ). Patients who met
ICD-10 diagnostic criteria for severe depression were invited to participate in the study. Criteria for exclusion included organic symptoms, hypoacusia (hearing loss), current substance abuse, psychosis, and receipt of psychiatric treatment in the year before study intake. Patients with an Ham-D score of 21 points and three or more positive answers on the Marshall scale were admitted to the study after providing written informed consent (N=87). They were randomly assigned either to the intervention (N=44) or to standard treatment (N=43).
Each patient assigned to the intervention was asked to meet with a multidisciplinary team, which used a psychodynamic orientation (
24,
38 ) to determine whether the patient had experienced as an adult an established pattern of mistreatment and aggression. Such a pattern was seen as a repetition of childhood traumatic experiences. After the initial meeting with the team, the patient met for a weekly session with one of the team members for a psychological intervention that focused on developing a cognitive understanding of personal characteristics and behaviors that allowed the repetition of traumatic experiences in the present. Behavioral changes that would alter the relationship between the victim and aggressor were addressed. Sexual trauma was validated as a real experience, and recollected memories were not challenged. Instead, the focus was on changing feelings of guilt and shame that could trigger new situations of abuse.
The psychiatrist conducted a monthly checkup that incorporated psychoeducational elements and monitored symptom change, adherence to pharmacological treatment, and the presence of self-destructive behaviors. The psychiatrist told the patient that these behaviors were a repetition of past mistreatment in the present. The multidisciplinary team met weekly with the therapist to address possible transference or countertransference. The social worker made home visits or telephone calls as needed. At any stage patients who did not want to talk about their traumatic experiences could choose to continue standard treatment with the same team.
Standard treatment
Standard treatment followed the recommendations in clinical guidelines issued by the Chilean Ministry of Health. Psychotherapy was supportive and was provided by a psychologist according to the patient's needs. The therapy was not routinely focused on traumatic experiences and their consequences. Team meetings were held only in more severe cases. Home visits by a social worker were performed when the team requested them.
Monitoring
An evaluation team not associated with the study assessed all participants at the beginning of the study, at three months (either after the intervention ended or after three months of standard treatment), and at six months. A clinical psychologist on the external team administered the Ham-D at each evaluation. Psychology students, blind to treatment assignment and Ham-D scores, administered Lambert's Outcome Questionnaire (OQ-45.2) (
39 ) and the Post-traumatic Stress Treatment Outcome scale (PTO 8) (
40 ). The OQ-45.2 can be completed by the patient or administered by another person. It consists of 45 items divided into five categories. The items assess the patient's state in three areas: symptoms, interpersonal relationships, and social role functioning. The established cutoff score for the Chilean population is 73 (
41 ), with higher scores representing a dysfunctional population. The PTO 8 is an eight-item scale used to assess the results of PTSD treatment. The items focus on the three symptom criteria for a
DSM-IV diagnosis of PTSD. A cutoff score of 12 or higher indicates the presence of PTSD (
42 ).
Adherence to treatment and attrition
There was some early attrition from the intervention group. Of the 44 patients in the group, four dropped out shortly after the baseline interview with the psychiatrist and could not be located. One patient who did not want to talk about her traumatic experiences dropped out after the baseline interview with the psychologist. Thus a total of 39 patients remained in the intervention group at the end of the first month. In the subsequent months two patients dropped out of the intervention group because they moved out of the city, and four dropped out when they felt better. No statistically significant differences were found in sociodemographic characteristics and baseline test scores between those who dropped out of the intervention group and those who remained. The 44 patients assigned to the experimental group were included in the intent-to-treat analysis.
After the baseline interview with the psychologist, only one patient in the intervention group continued medication treatment solely for depression. After the baseline interview with the multidisciplinary team, two additional patients in the intervention group continued medication treatment solely for depression. All other patients in the group were taking at least one additional type of medication and attending weekly therapy. At the end of the first month, all 39 patients in the intervention group were adherent to pharmacotherapy. At the three-month assessment, 30 patients in this group were adherent, and at the six-month assessment 26 of these patients were adherent.
There was also some attrition in the control group. Of the 43 patients in the control group, three discontinued treatment after the first interview. During the first month four patients dropped out of treatment for unknown reasons, six patients continued only pharmacological treatment, seven attended only one therapy session and continued pharmacological treatment, and the rest received at the same time pharmacological treatment and supportive psychotherapy
Analyses
The primary outcomes of interest were a reduction in depressive symptoms as measured by the Ham-D and improved total score on the OQ-45.2. Data were analyzed with SPSS, version 12.0. The intent-to-treat analysis used the averages of each group for multiple analyses of variance (MANOVAs). Between-group differences were assessed with Student's t test (significance level of .05). Seventy-nine participants (91%) were evaluated by the external team at three months (39 in the intervention group and 40 in the control group). At six months 71 (81%) were evaluated (36 in the intervention group and 35 in the control group).
A second analysis considered remission as a total OQ-45.2 score below 73 and improvement as a decrease in total score of 17 points from the baseline score. A Ham-D score below 8 was considered to indicate remission, and improvement was a decrease of at least 50% from the baseline score. For the PTO 8, patients with a baseline score over 12 points were considered to have improved when their score decreased by at least 40% since baseline; patients were considered to be in remission when PTO 8 scores at the two follow-up assessments were below 7 points (
39 ).
Results
Table 1 presents baseline data for the two groups. No statistically significant differences were found in sociodemographic characteristics, presence of childhood trauma, and mean test scores. The prevalence of PTSD did not differ between groups.
During the six months of follow-up, seven patients in the intervention group (16%) were hospitalized for a mean±SD of 6.0±5.0 days and six patients in the control group (14%) were hospitalized for a mean of 21.3±8.2 days. The mean number of visits with a psychiatrist during the six-month period was 7.0±3.7 for the intervention group and 2.5±2.0 for the control group. The mean number of visits with a psychologist over six months was 5.7±3.8 for the intervention group and 4.5±4.9 for the control group. Over the study period no serious suicide attempts were made by participants in either group.
Table 2 shows the mean scores at baseline and the two follow-up assessments. At three months significant differences were observed in favor of the intervention group in Ham-D scores (t=2.69, df=85, p<.001) and in the OQ-45.2 total score (t=1.98, df=85, p<.05). At six months MANOVA results indicated that total Ham-D scores improved significantly in the intervention group (Pillai's trace=.49, F=38.85, df=2 and 42, p<.01) as well as in the control group (Pillai's trace=.51, F=21.39, df=2 and 41, p<.01). The Ham-D scores indicated significantly greater improvement for those in the intervention group (t=2.41, df=85, p<.01).
Total scores on the OQ-45.2 also indicated improvement at six months for the intervention group (Pillai's trace=.42, F=18.60, df=2 and 42, p<.001) and the control group (Pillai's trace=.21, F=5.70, df=2 and 41, p<.01), with a significant difference in favor of the intervention group (t=2.16, df=85, p<.05).
The OQ-45.2 scores for interpersonal relationships and social role functioning showed significant improvement at six months only for the intervention group (for interpersonal relationships, Pillai's trace=.37, F=12.71, df=2 and 42, p<.01; for social role, Pillai's trace=.32, F=9.99, df=2 and 42, p<.01). The between-group differences in scores at six months favored the intervention group only in the area of social role functioning (t=1.97, df=85, p<.05).
Scores on the PTO 8 indicated significant improvements in PTSD symptoms at six months for the intervention group (Pillai's trace=.36, F=11.82, df=2 and 42, p<.01) and the control group (Pillai's trace=.33, F=10.38, df=2 and 41, p<.01), with no significant differences between the groups (p=.058).
The number of patients who met criteria for remission and improvement on the basis of OQ-45.2, Ham-D, and PTO 8 scores is shown in
Table 3 . At six months a significantly greater proportion of patients in the intervention group met criteria for remission as measured by the OQ-45.2 (39%, compared with 14% for the control group,
χ 2 =5.48, df=1, p<.05) and the Ham-D (22%, compared with 5% for the control group
χ 2 =3.99, df=1, p<.05).
Discussion
This study found that an outpatient intervention developed for depressed women with a history of childhood abuse that used brief psychodynamic psychotherapy was more effective than standard treatment in improving depressive symptoms, interpersonal relationships, and social role functioning. We think that the improved response to treatment is a product not only of the structured nature of the intervention but also of its psychotherapeutic orientation. Both the intervention and standard treatment view the symptoms and behaviors of these patients as resulting from a history of abuse and that patients can relive traumatic interpersonal experiences in the present (
43,
44 ) and experience retraumatization, even in a therapist-patient relationship (
45 ).
The study sought to demonstrate the importance of a multidisciplinary approach to treatment, in terms of providing patients with structured treatment and emotional healing and proving therapists with support and help in managing transference and countertransference problems (
46 ). These aspects of treatment are not included in standard guidelines in Chile for the treatment severe depression (
29 ).
We also found that an approach that actively screened for and focused on childhood traumas and explored their connections with present psychosocial difficulties resulted in patients' improved understanding of current conflicts. This strategy was effective in improving interpersonal relationships and social role functioning in the intervention group. The results also suggest that therapeutic work on interpersonal problems effectively augments the use of antidepressants to treat depressive symptoms but not PTSD symptoms.
Symptoms of PTSD improved in both groups over six months, with no significant between-group difference. The improvements may have resulted from pharmacological treatment alone, because the medications used by patients in this study have been shown to be effective not only for depression but also for PTSD (
47 ). The steady downward trend over six months in PTO 8 scores for the intervention group, compared with the apparent leveling off of scores for the control group, may mean that PTSD symptoms would have improved significantly for the intervention group if a longer period of observation had been used. In light of these results, we believe that the impact of this intervention on PTSD symptoms requires further investigation and that the intervention should be compared with the cognitive-behavioral treatments.
Even though the treatment provided in the intervention was more effective than standard treatment, 61% of patients in the intervention group showed no or minimal clinical improvement. This might be explained by the brevity of the intervention and the relatively short observation period. These findings are in accordance with those of studies showing that depression treatments have little effectiveness when the primary outcome considered is remission of depressive symptoms (
48 ). A lack of remission is associated with the chronicity of depressive pathology and comorbid PTSD (
49 ). In our sample 53% of the patients had a diagnosis of PTSD according to the
ICD-10 (
33 ). Poor response to brief treatments among individuals with childhood trauma was also observed by van der Kolk and colleagues (
50 ) in an evaluation of the effectiveness of eye movement desensitization and reprocessing for adults with PTSD and childhood trauma.
The principal strength of this study is its external validity. The intervention was developed with existing professional resources under normal working conditions in a public health service in a developing country. The exclusion criteria for sample selection were minimal. Patients in the sample were those who are encountered in routine treatment settings and have severe chronic symptoms and a risk of suicide. Many of them had been routinely excluded from studies conducted by the academic center. The intervention emphasized outpatient management. In contrast to the study by Stalker and colleagues (
16 ), our study found no exacerbation of symptoms after the intervention ended. Another methodological strength of this study was its evaluation of various patient outcomes, including psychosocial outcomes, its use of instruments validated for the Chilean population, and its use of intent-to-treat analysis.
Among the limitations of the study was the lack of structured treatment for the control group, which made it difficult to identify components of the intervention that accounted for its relative effectiveness. In addition, although the Ham-D was administered by an evaluation team not associated with the study, the psychologist who administered it was not blind to the type of intervention to which the patient was assigned. Both of these limitations should be addressed in future investigations to improve the internal validity of the results.
Conclusions
A brief intervention developed for severely depressed women with a history of childhood trauma and delivered in an outpatient setting with the usual resources available at a public hospital was more effective than standard treatment. The intervention actively screens for patients' experiences of childhood trauma and focuses on these experiences and patients' current interpersonal difficulties, with the understanding that current difficulties may be repetitions of past traumatic situations. It is important to continue to develop, implement, and validate interventions for this patient subgroup that can be delivered in public health settings.
Acknowledgments and disclosures
This study was supported by grant Sa05i20037 from the Fondo Nacional de Investigación y Desarrollo en Salud of Chile. Additional financial support was provided by Curicó Hospital. The authors thank Ignacio L. Iturria, M.A., and Carolina R. Nuñez, M.A.
The authors report no competing interests.