The Impact of an Inpatient Treatment on the Psychodynamic and Symptomatology in Couples Concordant for Substance Use: An Exploratory Study

Background: Much literature deals with patients who use drugs and have partners who are drug-free. However, concordant couples, in which both partners are consuming drugs, are sparsely examined in the literature. This might be due to the fact, that couples are rarely treated together in healthcare services. Despite that fact we propose that it is feasible and clinically meaningful to treat concordant couples in the same ward. Consequently, this study pursues the goal to expand the body of knowledge in the treatment of concordant couples investigating the research question: is it feasible, clinically and prognostically meaningful to treat these patients in the same ward? Method: This exploratory study included ve concordant couples (ten patients), which were simultaneously treated between August 2013 and November 2014 in a specialised substance use ward at the Psychiatric Hospital Münsterlingen, Switzerland. All patients passed through a psychodynamic characterisation based on the OPD-II interview and the Structured Interview for Personality Organization. Symptom load was measured with the Brief Symptom Inventory at admission and termination of treatment. We calculated comparisons at the individual level using t-tests for paired samples. Results: We showed that it is feasible and clinically meaningful to treat couples concordant for substance use in the same ward. The psychodynamic characterization of the ve concordant couples revealed recurring patterns of collusion, involving divided roles between dependence and independence, caregiving and neediness, activity and passivity, control and submission, strength and deciency as well as superiority and inferiority. The patients didn´t change signicantly on the BSI between pre- to post-treatment, although men (d = 1.64) beneted to a greater extent than women (d = 0.10). Conclusion: Treating concordant couples together in the same inpatient setting is unusual, but feasible and clinically useful, because it makes it possible to take into account the couples dynamics in the treatment.

play a positive role by facilitating desistance from using substances [18]. In sum it can be said, that there are several arguments for including the dynamics of couples into therapy.
To the best of our knowledge, we are not aware of any study that investigated the psychodynamics and outcomes of concordant couples in inpatient treatment, that were treated simultaneously at the same ward. The rst aim of the current exploratory study was to show that it is feasible to treat concordant couples at the same ward. The second aim was to outline the relevance of understanding the psychodynamics and symptom change of these couples over the course of an inpatient treatment.

Ethics
The present study was performed according to Good Clinical Practice [19] on the basis of applicable legal framework conditions in Switzerland and in accordance with the Ethics Committee of Canton Thurgau. All procedures were conducted by trained staff in accordance with the Declaration of Helsinki [20].

Demographic and clinical variables
Demographic and clinical variables consisted of age, length of partnership, education, housing, substance use, comorbidity, presence of Hepatitis B and C and treatment length.
Length of partnership: Patients were inquired about the length of their partnership in years.
Housing: Patients were asked if they had a fixed abode or not.
Education: The patients were asked whether they have completed a vocational training, which requires a completed school education of nine years as admission.
Diagnoses and Comorbidity: Diagnoses were given on the basis of ICD-10-GM [21]. All patients were asked about the use of the following substances: Alcohol, Benzodiazepines, Cocaine, Heroin, Cannabis or Other. Diagnoses were given by the corresponding therapist.
Hepatitis B and C. The patients were tested by a standardized serum antibody screening test.
Treatment length: Treatment length was defined as the number of treatment days between pre-(t 1 ) and post-treatment (t 2 ) including holidays.

Process variables
Psychodynamic characterisation was accomplished on the basis of two psychodynamic interviews: The OPD-II interview [22], which is based on the model of the Operationalized Psychodynamic Diagnosis [23].
The Structured Interview for Personality Organization [24], which is based on the model of personality organization [25,26].
It included conscious and unconscious motives of the choice of the partner, collusive, regressive and progressive aspects of the partnership, dominance or equality within the couple, dominant affects, aggression as well as defence mechanisms. Psychodynamic characterisation was accomplished by the supervising registrar, which is a certified specialist in transference focussed psychotherapy [27]. During patients' treatment on the ward, he continuously monitored the patients and kept detailed records on their progress in therapy.

Outcome variables
Symptom load: Symptom load was measured with the Brief Symptom Inventory [BSI; 28,29]. It includes 53 items from the patient's perspective, measuring physical and psychological impairments. Answers are given on a 5-point Likert scale ranging from "not at all" (0) to "extremely" (4). The inventory consists of a global value (Global Severity Index, GSI) and the subscales Somatization, Obsession-Compulsion, Interpersonal Sensitivity, Depression, Hostility, Phobic Anxiety, Paranoid Ideation and Psychoticism.
Construct validity was supported, the correlations between the BSI and the long version, the Symptom-Checklist-90-R [31], ranged between r = .92 and r = .99 [29,30]. Subsequent therapy: We defined seeking subsequent therapy after treatment termination as a positive treatment outcome because of three reasons: First, addicted patients need an intensive therapy, but have a negative treatment prognosis and are at high risk to drop-out from treatment [32]. Second, therapy motivation and insight into illness are critical for a positive prognosis of the disorder [33,34]. Third, relapses are very common in addicted patients, which can lead to persistent psychological strain [35], where a subsequent therapy may help.

Measurement time points of the study
The study design was based on a prospective, naturalistic analysis of five concordant couples, who were treated in an inpatient psychotherapy. Outcome data were collected within two days after admission (t 1 ) of therapy, post-treatment data within two days before treatment termination (t 2 ) and follow-up data 12 months after treatment termination (t 3 ) via telephone. Psychodynamic characterisation was done between the time range t 1 and t 2. Table 1 shows the measurement time points of the study. Treatment concept: Our treatment model consisted of a specialised substance use treatment including detoxification, withdrawal treatment and a psychodynamic psychotherapeutic concept. Psychodynamic theory and psychotherapy is not only eligible to understand the genesis and maintenance of substance abuse [37], but it also highlights the unconscious conflictual processes of the disorder, for example collusive relationships [38]. Our treatment model is influenced by Yalisove et al. [39], who recommended the use of modified psychoanalysis as a treatment for addiction. He suggested modifications to traditional psychoanalysis for persons with addiction including an initial phase of treatment that is supportive and didactic, followed by more psychoanalytically oriented treatment; a therapist who is active rather than passive; the forestalling of transference; and the recommendation for participation in a self-help group. Important for the concept of the ward is also Khantzian's self-medication hypothesis of substance use disorders [40].
Prochaska et al. [41] state that change processes traditionally associated with experiential, cognitive and psychoanalytic traditions are most helpful during contemplation and precontemplation stages whereas change processes associated with behavioural and existential processes are most suitable during action and maintenance stages. In the review on evidence-based practice for substance use disorders McGovern and Caroll [42] found psychodynamic supportive-expressive psychotherapy with cognitive elements as developed by Luborsky [43] to be an effective intervention for opiate use disorders, especially when delivered by skilled therapists.
The unique characteristic of our ward is that we applied a psychodynamic model to treat patients with substance use disorders not only at the individual level, but also on the couples level, which allows to study the psychodynamics and outcomes of couples simultaneously.
Treatment plan: All patients were assigned to the same therapy scheme: Regular individual consultations with physicians and nursing personnel, sports therapy, ergotherapy, qualified pharmaceutical treatment, acupuncture by the protocol of the National Acupuncture Detoxification Association (NADA), and various group therapies (e.g., social skills, dialectic-behavioral therapy, addiction education). Moreover, couples therapy was provided by a specialised therapist upon request. The weekly schedule at the ward is shown in Table   2. The typical duration of treatment at the ward was four to twelve weeks.   Table 4, 5, 6, 7 and 8.

Statistics
All calculations were computed using IBM SPSS Version 25. We ran t-tests for paired samples for women and men separately. The Shapiro-Wilk test of normality showed that all scores were normally distributed in the sample of women (with p-values ranging from .161 to .995). Also, for the men all scores were normally distributed (with p-values ranging from .571 to 1.000) except for the subscales Depression, Hostility and Paranoid Ideation. For the interpretation of the results of the BSI we calculated the Reliable Change Index [RCI,44].
According to Franke [29] we classified the results of the therapy as follows: (a) clinically significant improvement: improvement of at least 0.53 points in the GSI between admission and termination of therapy and crossing the cut-off-value between clinical and non-clinical samples (cut-off-value: 0.62), (b) reliable improvement: improvement of at least 0.53 points in the GSI between admission and termination of therapy, (c) no change: change less than 0.53 points in the GSI between admission and termination of therapy and (4) reliable deterioration: decrease of at least 0.53 points in the GSI between admission and termination of therapy. The Cohen's effect sizes for paired samples d z were calculated using the formula by Rosenthal [45] with T for the t-value und df for the degrees of freedom: The effect sizes were interpreted as follows: d = 0.20 small, d = 0.50 medium and d = 0.8 large [46]. Case study of Couple 1: At admission of therapy we detected a collusion between the female and male patient. The female patient largely took care of the male patient. She also took the lead and organized their social life. During clinical visits, she often spoke for him.
Furthermore, she often cut him short during conversations or commented on his ideas as ridiculous. The dominant affects of the male patient were insecurity and subliminal anger.
Due to his insecure self-perception he handed over the responsibility to his partner, simultaneously he felt worthless, insufficient and dependent of his partner, which made him angry. The female patient enjoyed to be needed from her partner. She felt self-confident, when she could support him. However, she felt burdened because she had to organize everything. He seemed socially dependent on her, which seemed to increase her self-esteem.
The psychodynamic characterization revealed a characteristic pattern: the male patient seemed to have the position of inferiority, neediness, submission, deficiency and passivity and the female patient the position of caregiving, activity, control, strength and superiority. Table 4 shows the demographic and clinical variables as well as the outcome changes for Couple 1. Both patients were in early adulthood and three years together. Both completed a vocational training, but had no stable housing. They used multiple-drugs, had no comorbid disorder and were Hepatitis B and C negative. According to the GSI the female patient didń t change reliable or clinically significant, but she made substantial improvements in Somatization, Paranoid Ideation and Psychoticism [29]. For the male patient we couldn´t calculate the RCI due to data loss. Follow-up data showed that both partners received subsequent therapy after treatment termination at the ward. One year after treatment termination Couple 1 was still together and raised a family with two children.   positive and Hepatitis C negative. The GSI of both partners was in the subclinical range at admission and termination of therapy. For the female patient the GSI didn´t change reliable [29], but tended to deteriorate in Obsession-Compulsion, Interpersonal Sensitivity, Phobic Anxiety and Paranoid Ideation. The GSI of the male patient also didn´t change reliable [29], no substantial changes were detected in all scales of the BSI. Follow-up data showed, that both partners received subsequent therapy after treatment termination at the ward. The couple got engaged after leaving the inpatient therapy.  Case study of Couple 3: In the therapy process it became obvious that the female patient took care of both partners. In a crisis, the roles switched, and he protected her from suicide. During this crisis, it came to a massive confrontation with verbal and physical aggression. The dominant affects were anger and fear of loss in both partners. He was highly dependent on her, whereas he was the main motivating factor behind abstinence and treatment motivation. The psychodynamic characterization revealed a role switch between care and neediness as well as between control and submission. Table 6 shows the demographic and clinical variables as well as the outcome changes for Couple 3. The couple was middle-aged, three years together and had no stable housing.
Both used multiple drugs, had a comorbid personality disorder and were Hepatitis C negative. In contrast to the female patient the male patient completed no vocational training and was Hepatitis positive. The GSI of the female patient improved clinically significant with substantial effects for all subdomains of BSI except for Paranoid Ideation [29]. The GSI of the male patient didn´t change reliable or clinically significant [29], but he improved in Obsession-Compulsion and Interpersonal Sensitivity. Follow-up data showed, that both partners received subsequent therapy after treatment termination at the ward. In addition the data showed that the couple separated after being discharged from the ward.
Both partners died, the male was found dead and the female died by suicide.   as well as between activity and passivity, whereas the female patient primarily filled a role of caregiving and the male patient the role of neediness and consumption.  [29]. She tended to deteriorate in GSI, Obsession-Compulsion, Interpersonal Sensitivity, Hostility, Paranoid Ideation and Psychoticism. The GSI of the male patient also didn´t change reliable or clinically significant [29], but he tended to improve in Interpersonal Sensitivity, Phobic Anxiety and Psychoticism. Follow-up data showed, that both partners received subsequent therapy, but separated after treatment termination at the ward.    On average the patients didn´t change significantly on the BSI between pre-to posttreatment. Only the patient 3W improved reliable and clinically significant. On the level of the couples, Couple 3 seemed to benefit most with regard to symptom reduction. Table 9 presents the symptom change between pre-and post-treatment for women and men separately. None of the effects gained significance. For the men, all effect sizes were at least small in size. Large effects were found for the GSI, Interpersonal Sensitivity, Depression, Paranoid Ideation and Psychoticism. For women the effect sizes were also at least small in size, with the exception of the GSI and Obsession-Compulsion. The largest effect was found for Somatization. However, it should be noted, that for women Interpersonal Sensitivity, Hostility and Paranoid Ideation increased during inpatient therapy. Note. BSI= Brief Symptom Inventory, t 1 = admission of therapy, t 2 = termination of therapy, n = 4 for women. n

Differences between men and women
Concerning the GSI, it is very interesting, that males bene tted more than females. While gender differences are well examined concerning substance-related epidemiology, demographic, the progression of dependence, comorbidity, treatment entry, retention, and completion [48][49][50], there aren´t many studies, which investigated the differential effects on symptom load between men and women. Kosten et al. [51] and Mc Hugh et al. [52] reported comparable treatment effects for men and women. These results are contradictory to our ndings. The gender differences found in our study warrant further research.
Concerning the different scales, large effects were found for decreased Interpersonal Sensitivity, Paranoid Ideation and Psychoticism in men and for Somatization in women. Improvements in Interpersonal Sensitivity, Paranoid Ideation and Psychoticism in men are in line with the nding, that couples therapy is effective in reducing dyadic adjustment and intimacy [53] as well as in reducing partner violence, which is usually a problem of the male partner [54].

Limitations of the study
The inferences which can be drawn on the current study have several limitations. As this study used a naturalistic design without randomization or control groups, no conclusions can be drawn about the causality of the reported effects. Furthermore, the differential impact of the various treatment elements and the speci c setting of the inpatient treatment remain unclear. It also remains unclear how sustainable these effects are, since no follow-up examination of symptom load was performed. We did not collect addiction-related outcomes, as we had an open setting, where no drug searches were routinely performed, so we don´t know how the patients developed in relation to their addictions. The study was further constrained by its small sample size, which may lead to limitations regarding reliability and representativeness of the results. In addition, no satisfactory study completion rate was achieved in this study. Low response rates are problematic as they may distort the results, reduce external validity and undermine the scienti c credibility of conclusions from clinical trials [55,56]. We didn´t systematically collect the causes of missing data, but we assume, that non-completion of study measures was primarily caused by a lack of motivation and compliance. It should also be noted critically that psychodynamic characterization was not a standardized procedure. This may have distorted interrater reliability. Another major point is that only self-assessment tools were used to measure symptom load. The advantages of these measures can be found in their economy, practicality and cost e ciency. At the same time, they may be fraught with substantial sources of errors [57].

Conclusion and outlook
The investigation of concordant couples is a relatively new research eld, since most literature deals with patients who use drugs and have partners who are drug-free. Treating concordant couples at the same time and in the same setting is challenging for the team and fellow patients. Our results provide preliminary evidence that it is feasible and clinically useful to treat concordant couples in the same ward.
This study demonstrated that inpatient therapy might help to address the speci c problems in the relationship of concordant couples, which may lead to a better prognosis. We showed that inpatient therapy offers the possibility to examine and work through the complex psychodynamics of concordant couples. Our study provides preliminary evidence that concordant couples show characteristic patterns of collusion, which are related to substance abuse. Our patients bene ted not statistically signi cant from the treatment, but showed symptom reductions at treatment termination. These effects were larger for the men than for the women.
Future studies should focus on the mechanisms of change of the treatment of concordant couples to determine which treatment is most effective for which individuals with speci c de cits. Future studies may want to contrast the outcomes of concordant couples treated simultaneously and separate. Also, it would be interesting to examine possible gender differences when concordant couples are treated together. In future studies, other measures should be included to assess for instance relationship dynamics, social support and substance use. We recommend to use the Alcohol, Smoking and Substance Involvement Screening Test [ASSIST, 58].