Short-Term Psychodynamic Therapy of Obsessive-Compulsive Disorder: A Randomized Controlled Trial of Unmedicated Patients

Abstract

Introduction

Obsessive-compulsive disorder (OCD) is an often chronic and disabling mental disorder. Cognitive-behavioral therapy and serotonin-reuptake inhibitors and their combination are first-line treatments. However, about 50% of the patients do not achieve response or remission, although many patients received additional pharmacotherapy. These results highlight the need for additional evidence-based treatment options.

Methods

In a randomized controlled trial (RCT), a manual-guided method of short-term psychodynamic therapy (STPP) for OCD was tested against a waiting list condition. Thirty-seven unmedicated patients were randomly assigned to STPP, 30 to the waiting list. The Yale-Brown Obsessive Compulsive Scale (Y-BOCS) total score post-therapy assessed by trained and masked raters was the primary outcome. Secondary outcomes included response, remission, depression, interpersonal problems, and obsessive beliefs.

Results

In the intention-to-treat sample, STPP achieved a significantly lower Y-BOCS score (p < 0.0001) than the waiting list condition, associated with a large between-group effect size (d = 1.04). The mean difference of change in the Y-BOCS between baseline and post-therapy or waiting period assessment was −11.49 (95% CI −14.36, −8.63) for STPP and −1.10 (95% CI −4.29, 2.10) in the waiting list. Remission rates were 60% vs. 20% (p = 0.0006), response rates 64% vs. 16% (p = 0.0003) in favor of STPP.

Conclusion

In a first RCT, manual-guided STPP proved to be efficacious in OCD with a clinically meaningful improvement, thereby broadening the spectrum of evidence-based interventions. Of note, our study is one of the few trials using psychotherapy as a monotherapy in OCD, showing effects of psychotherapy alone. Further research is needed to confirm these first findings and to compare STPP with other established treatments.

Plain Language Summary

Obsessive-compulsive disorder (OCD) is an often chronic and disabling mental disorder. It is characterized by repetitive obsessive thoughts and ritualized compulsive actions. Psychotherapy and pharmacotherapy are recommended as first-line treatments. These treatments are efficacious for some patients, but the majority of patients does not benefit sufficiently from the available treatments. Patients who do not benefit from one treatment may do so from another. For this reason, it is important to offer patients with OCD a variety of efficacious treatments. Psychodynamic therapy has a long clinical tradition in the understanding and treatment of OCD. However, randomized controlled trials showing the efficacy of psychodynamic therapy in OCD are presently missing. For this reason, we developed a manual-guided method of short-term psychodynamic therapy (STPP) for OCD. The treatment procedures are described in a manual. This treatment was tested in a randomized controlled trial comparing STPP with a waiting list control condition. The treatment encompassed up to 24 weekly sessions. The patients in the waiting list waited for a corresponding time, that is, for 6 months. After the waiting period, they were offered STPP as well. Thirty-seven patients were randomly allocated to the treatment group and 30 patients to the waiting list. Established diagnostic instruments were applied at baseline (before therapy) and after therapy. After the treatment, the patients who had received STPP showed a statistically significant larger and clinically meaningful reduction in OCD symptoms than the patients in the waiting list condition. In the STPP group, 64% of the patients showed a response to the treatment defined as a reduction in symptom severity of 35% or more. In the waiting condition, this was only the case for 16%. Based on these results, our study provided empirical evidence that STPP is efficacious in OCD. With STPP, another evidence-based treatment is available for patients with OCD. The long-term outcomes 6 months after the end of treatment will be published as soon as they become available.

Introduction

Obsessive-compulsive disorder (OCD) is an often chronic and disabling mental disorder characterized by recurrent obsessions and uncontrolled compulsions such as repetitive behavioral or mental acts performed in response to an obsession [1]. Lifetime and 12-month prevalence rates of 2.7% and 1.3% were reported [2]. OCD is frequently associated with depressive and anxiety disorders and to a lesser degree with alcohol and drug dependence [3]. The disorder is characterized by severe social and occupational impairments, which are more marked than in depressive or anxiety disorders [3]. Evidence suggests that cognitive behavioral therapy (CBT) with exposure and response prevention (ERP) and selective serotonin-reuptake inhibitors (SSRIs) are efficacious in the treatment of OCD [4–7]. CBT, SSRIs, and their combination are recommended by both the UK NICE guidelines and the American Psychiatric Association guidelines as first-line treatments for OCD [4, 8]. In Germany, CBT is recommended as the first-line treatment for OCD [9]. SSRIs are recommended if patients reject CBT, or if they have made positive experiences with SSRIs or if the severity of OCD symptoms does not allow to carry out CBT [9].

However, some limitations for the studies of CBT in OCD were recently noted [6]. Many patients treated with CBT and ERP also received additional pharmacotherapy [5, 6]. Furthermore, 78% of the studies of CBT with ERP showed a high risk of bias or concerns about bias, as assessed by the Cochrane risk-of-bias tool [6]. Reid et al. critically concluded that after 25 years of ERP trials in OCD currently fewer than 300 participants have received ERP in high-quality trials [6]. Thus, further high-quality trials are required. A recent meta-analysis reported response rates of 38% for CBT and of 43% for ERP [10], defining response by an at least 50% reduction of OCD symptoms.

SSRIs were found to achieve a medium effect size (SMD = 0.66) in comparison to placebo, and most studies were associated with a high risk of bias as well [5, 7]. The combined treatment of CBT and SSRIs was superior to pharmacotherapy alone (SMD = 0.73) [5] but not to CBT alone (SMD = 0.25) [11]. In these studies, risk of bias was often high, too [7]. Furthermore, many patients show residual symptoms after treatment [1]. Remission rates for pharmacotherapy are limited as well. Defining remission by a Yale-Brown Obsessive Compulsive Scale (Y-BOCS) score ≤12, an RCT found remission rates of 52%, 25%, and 58% for ERP, clomipramine, and the combined treatment, respectively [12]. Combining CBT with serotonin reuptake inhibitors achieved a remission rate of 33% (Y-BOCS ≤12) [13]. Another RCT reported remission rates (Y-BOCS ≤12) of 43% for CBT and of 13% for risperidone, with risperidone not being superior to placebo [14].

In sum, the remission and response rates of CBT, pharmacotherapy, and their combination in OCD are limited, ranging between 25% and 58%. In addition, a meta-analysis investigating treatment adherence found that 15.6% of the eligible patients refused CBT and another 15.9% dropped out of treatment [15]. Summing up, almost one-third of patients with OCD refused CBT or dropped out of the treatment. Thus, none of the existing treatments seems to fit all patients. For this reason, developing further efficacious treatments for OCD is an urgent clinical need. Patients who do not benefit from one therapeutic approach may benefit from another.

There is a long clinical tradition of describing and treating OCD from a psychodynamic perspective [16]. The psychodynamics of OCD have been frequently described starting with Freud’s early descriptions [17–20], which were followed by later contributions, including those of Fenichel, who emphasized the role of specific defense mechanisms in OCD, especially of isolation [21, 22]. Undoing, reaction formation, rationalization, displacement, and regression were described as further defense mechanisms characteristic of OCD [21, 22]. The superego of patients with OCD was characterized as rigid and hyperstrict, forbidding aggressive and sexual impulses. In contrast, the ego was described as weak, not being able to differentiate between thinking and acting (magical thinking), which Freud [18, p. 85] characterized in one of his famous patients (the “Rat Man”) as the “omnipotence of thoughts”, a concept later picked up by cognitive models as the so-called thought-action fusion (TAF) [23].

Kempke and Luyten summarized the classical model (e.g., Fenichel, 1945) as follows: “Stated differently, the ego would be trapped in a ‘double’ fight against intolerable impulses on the one hand and a hypermoral superego on the other” [16, p. 293]. From an object-relational perspective, representations of the self and others were described as ambivalent, with the self characterized by negative part properties (e.g., self-criticism) and with others viewed as critical and punishing. These representations are assumed to be associated with an impaired tolerance for ambivalence and a strong need for control and autonomy (ambivalent attachment style). Thinking is overemphasized and uncontrollable issues such as feelings and interpersonal relationships are avoided [16]. An empirical study corroborated the assumption that latent aggression is characteristic of OCD [24]. Another study confirmed that specific polarities are characteristic of patients with OCD (i.e., innocence-guilt, disgust-pleasure, good-bad, dead-alive, abstaining-becoming corrupt, and self-sacrificing-taking advantage) [25, 26]. These results are consistent with the psychodynamic assumption of conflicts between a hyperstrict superego and sexual or aggressive impulses. Interventions addressing the psychodynamics of OCD will be described below [27].

However, empirical evidence for psychodynamic treatments is presently scarce. Recent research on anxiety disorders, however, suggests that manual-guided STPP may represent a promising approach for OCD [28]. Building on STPP for anxiety disorders, a model of STPP specifically tailored to OCD was developed [27]. The randomized controlled trial presented here is the first RCT to test the efficacy of this method of psychodynamic therapy for OCD. For a newly developed treatment, a comparison with a waiting list condition is more appropriate than a comparison with treatment as usual or another specific treatment since considerably fewer patients are required, also reducing study costs. A non-inferiority trial, for example, comparing a new treatment to an established treatment would require large samples of patients with OCD, which are difficult to gain and is associated with high study costs due to the longer study duration. Furthermore, in routine clinical practice patients have to wait for several months for a treatment, too [29]. We expect the manual-guided form of short-term psychodynamic therapy (STPP) to be superior to a waiting list with regard to reducing OCD symptom severity assessed by the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) [30].

Methods

Trial Design

This is a randomized controlled observer-blind two-center study (Clinic of Psychotherapy and Psychosomatics, University of Göttingen, Sigmund-Freud-Institute Frankfurt) with a two-arm design (STPP vs. waiting list). The trial was pre-registered at DRKS00011256 on 14 February 2017 (https://drks.de/search/de/trial/DRKS00011256). The study protocol can be found in online supplementary Material 1 (for all online suppl. material, see https://doi.org/10.1159/000550421). The CONSORT checklist can be found in the online supplementary material.

Procedures, Randomization, and Masking

Patients were recruited in the centers of Frankfurt and Göttingen between February 2018 and October 2022. Using computer-generated random numbers, patients fulfilling the inclusion criteria were randomly assigned (1:1) to the two study conditions by use of sequentially numbered opaque sealed envelopes. Random assignment to the study groups was carried out by a person not included in the assessment of outcome, assuring allocation concealment. Outcome assessments were carried out by masked raters who were neither involved in treatment nor in allocation of patients.

Inclusion and Exclusion Criteria

Inclusion Criteria. Patients with a primary diagnosis of OCD according to SCID I and ADIS-IV-clinical severity rating (CSR) [31, 32]. By ADIS-IV, the severity of the disorder is rated on a 0–8 scale, with higher scores indicating more severity [31]. For inclusion in the study, OCD had to be the most severe that is the highest rated disorder. Age: 18–65. Comorbid disorders were permitted if they were not the primary disorder; that is, the ADIS-IV-CSR rating of comorbid disorders must have been lower than that of OCD.

Exclusion Criteria. Psychosis, acute substance dependence, organic mental disorder, somatic disorders not compatible with psychotherapy, cluster A and B personality disorders (SCID II [33]), patients at acute risk for self-mutilating behavior or acute risk of suicide, concomitant psychotherapy or pharmacotherapy.

Treatment

A manual-guided method of STPP specifically tailored to OCD was used in this study [27]. The treatment encompasses up to 24 weekly treatment sessions, corresponding to a treatment duration of 6 months. STPP for OCD consists of 12 modules, including both supportive and expressive (interpreting) interventions [27]. It is based on Luborsky’s supportive-expressive therapy [34]. Supportive interventions include, for example, informing the patient about the disorder and treatment or establishing a secure alliance. Expressive interventions include identifying and working through the core conflictual relationship theme (CCRT) underlying the symptoms of OCD. A CCRT includes (unconscious) wishes, reactions of others, and response of self. Working through responses of the self includes a focus on OCD symptoms, defense mechanisms (e.g., reaction formation or undoing), and avoidance. Encouraging the patient to face the feared situation was strongly recommended by Freud for the treatment of both agoraphobia and OCD ([35], p. 166). Consistent with Freud, Fenichel described the procedures as follows ([22], p. 244): “As far as the clinical technique is concerned, it may be said that – in agreement with Freud’s observations – patients suffering from various forms of phobias and compulsions must gradually and systematically be encouraged and helped to expose themselves to the situation they fear, to refrain from their defense procedures, and to face the anxiety without flinching until it subsides.” The experiences made by the patient during self-exposure serve as material for working through the three CCRT components, which is a difference to ERP in CBT [27]. Whereas CBT patients are desensitized against the (conscious) fear associated with the OCD symptoms (e.g., that a fire could start if he or she does not repeatedly control whether the electric stove is switched off), the psychodynamic procedure aims at also desensitizing patients toward the anxiety provoking wish or affect (e.g., his or her own destructively experienced aggression or sexuality, a “fire” in the example given above) included in the CCRT. Further treatment modules address ambivalence, resistance, and potential nonresponse as well as preparing treatment termination and relapse prevention, including booster sessions [27].

For participating in the study, therapists were required to have an advanced or completed training in psychodynamic therapy. Thus, all therapists held a degree in clinical psychology or were physicians and had completed a training in psychodynamic therapy or were in advanced psychodynamic training. Therapists were specifically trained by experts in STPP of OCD following the treatment manual [27]. During the whole trial, they received group supervision by experts weekly or biweekly in the beginning, monthly after that. Before inclusion in the trial, the therapists were required to treat one patient competently. These training cases were not included in the analysis. All treatment sessions were audiotaped. Adherence and competence of therapists were assessed by trained raters using specific checklists. Thus, treatment integrity was established by training and supervision of therapists as well as by empirical investigation of sessions.

Minimal Attention Waiting List Group

As a comparator, we used a minimal-attention waiting list group. Patients assigned to this group waited 24 weeks for treatment, corresponding to the duration of STPP. To ensure ethical standards, minimal attention was provided through regular phone calls every 4 weeks to monitor patients’ status. In order to maintain masking, telephone calls were carried out by a person neither involved in outcome assessment nor in the treatment of study patients.

Primary and Secondary Outcomes

Primary Outcome. the primary outcome was the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) [30], a clinician-administered interview by which we assessed OCD symptom severity at the end of treatment or, respectively, the waiting period. The Y-BOCS encompasses 19 items, of which 5 refer to obsessive thinking and another 5 to obsessive actions. The items are rated on a scale from 0 to 4. The total score of these 10 items is used, ranging from 0 to 40.

Secondary Outcomes. secondary outcomes included response defined as a ≥35% reduction of the Y-BOCS [30, 36], remission (Y-BOCS ≤12 [36]), depression (Beck Depression Inventory II, BDI [37]), anxiety (Beck Anxiety Inventory, BAI [38]), interpersonal problems (Inventory of Interpersonal Problems, IIP [39]), obsessive beliefs (Obsessive Beliefs Questionnaire, OBQ [40]), self-reported OCD symptoms (Obsessive-Compulsive Inventory (OCI-R) [41]). The BDI II and the BAI are well-known 21-item self-report questionnaires for assessing the level of depression and anxiety [37, 38]. The IIP is a reliable and validated self-report instrument for assessing interpersonal problems [39]. The OBQ is a 24-item self-report instrument addressing dysfunctional beliefs [40]. Self-reported symptoms of OCD are assessed by the 18-item OCI-R [41].

Treatment Integrity

All sessions were audio-recorded, and three sessions per case – randomly sampled from the initial (sessions 1–8), middle (sessions 9–16), and late phases (sessions 17–24) of each treatment – were assessed for treatment integrity. Sessions were evaluated with the PDT-OCD Adherence and Competence Scale [42], which was specifically developed for this study and covers the twelve modules of the treatment manual [27]. Adherence and competence were rated with a 14-item, four-point Likert scale, ranging from 0 (adherence not observable; competence inadequate) to 3 (pronounced adherence; very good competence). An additional 15th item captured the global level of adherence and competence exhibited during each session. In addition, the Comparative Psychotherapy Process Scale (CPPS), which assesses generic psychodynamic-interpersonal and cognitive-behavioral interventions, was applied [43]. The ratings were carried out by two doctoral students with a master’s degree in clinical psychology. Of the rated sessions, 20% were double rated for determining interrater reliability (intraclass correlation coefficient [3,1]).

Assessments

Assessments with regard to the primary outcome were carried out by trained and masked raters using the Y-BOCS. Assessments were carried out for primary and secondary outcomes at baseline, after end of therapy or waiting period, and 6 months after end of therapy (follow-up).

Sample Size and Power Calculation

A meta-analysis of CBT reported large between group effect sizes of 1.67 for the comparison of CBT with waiting list condition [44, p. 36]. In order to detect a difference of 1.67 by 2-sided tests with the power of 0.80, 7 patients per condition are required [45]. As a sample size of 2 × 7 patients is very small and would not ensure that patient differences are randomly distributed between the study conditions, we planned our study more conservatively. We expected a between-group effect size of 0.70 for whose detection 2 × 33 patients are required in order to achieve a power of 0.80 by a two-sided test at alpha = 0.05 [45]. Data of dropouts were compensated for by multiple imputation of missing data.

Statistical Analysis

To test for differences in the primary outcome, the post-therapy Y-BOCS scores were compared between study groups by linear mixed-effect models with study group, baseline Y-BOCS, and study center as covariates. For other continuous measures, similar models were used with the baseline scores of the measure as covariates. Remission and response variables were analyzed by logistic regression models with study group, baseline scores of the respective measure, and study center as covariates. For continuous measures, effect sizes were calculated from the test statistics of the models (e.g., t values). Analyses were carried out in the ITT sample of all randomized patients. Missing values were handled by multiple imputation with 100 repeats by fully conditional specification. Results were combined by Rubin’s rules. For the completer sample, complementary analyses were performed. For the secondary outcomes, no adjustments for multiple testing were made. Statistical analysis was carried out by an independent statistician of IZKS Mainz (C.R.) by means of SAS (SAS Institute Inc., Cary, NC, USA, version 9.4).

Safety and Ethical Considerations

The study protocol was reviewed and approved by the Ethics Committees of the Universities of Giessen and Göttingen (reference numbers 136/16 and 13/7/17, respectively). Patients were informed of procedures, gave written informed consent, and could withdraw at any point with no disadvantage. Since patients in the waiting list group were offered the same treatment following the waiting period, all participants ultimately received an active intervention. Risk for patients was relatively low; however, suicidal tendencies were checked regularly. If the need arose, inpatient services were offered. For patients who could not be included in the trial, the centers provided further care.

Results

A total of 159 patients were screened, and 67 patients were included in the study (Fig. 1). Thirty-seven patients were randomly allocated to the STPP condition, 30 to the waiting list. Seven patients dropped out of both STPP and the waiting list condition. Reasons for dropping out of treatment were dissatisfaction with the therapist (N = 2), starting pharmacotherapy (N = 1), and occurrence of exclusion criteria (N = 4). Reasons for dropping out of the waiting list were leaving the city (N = 3), withdrawing consent (N = 1), or others (N = 3). In the ITT analysis, 37 patients were included in the STPP condition, 30 in the waiting list. Patient characteristics are displayed in Table 1. The STPP condition included 22 women and 15 men, the waiting list condition 19 women and 11 men. This difference was not statistically significant (chi square = 0.105, p = 0.7463, df = 1). Mean age of the patients was 36.24 (SD = 13.63) in the STPP condition and 37.97 (SD = 13.58) in the waiting list condition (t = −0.52, df = 62.25, p = 0.6079). Data on severity of OCD symptoms based on the Y-BOCS following the benchmarks suggested by Cervin et al. [46] as well as data on prior psychiatric or psychotherapeutic treatments are presented in Table 1.

Fig. 1.

Flowchart. A total of 159 patients were screened, and 67 patients were included in the study. Thirty-seven patients were randomly allocated to the STPP condition, 30 to the waiting list. Seven patients dropped out of both STPP and the waiting list condition. Reasons for dropping out of treatment were dissatisfaction with the therapist (N = 2), starting pharmacotherapy (N = 1), and occurrence of exclusion criteria (N = 4). Reasons for dropping out of the waiting list were leaving the city (N = 3), withdrawing consent (N = 1), or others (N = 3). In the ITT analysis, 37 patients were included in the STPP condition, 30 in the waiting list.

Table 1.

Patient characteristics

​	STPP	Waiting list
N (ITT)	37	30
Mean age (SD), years	36.24 (13.63)	37.97 (13.58)
Women, %	50	63
Y-BOCS 0–13, n (%)	1 (3)	3 (10)
Y-BOCS 14–21, n (%)	13 (35)	13 (43)
Y-BOCS 22–29, n (%)	14 (38)	13 (43)
Y-BOCS 30–40, n (%)	9 (24)	1 (3)
Major depressive disorder, n (%)	7 (19)	4 (13)
Persistent depressive disorder, n (%)	6 (16)	7 (23)
Anxiety disorder, n (%)	9 (24)	12 (40)
Obsessive-compulsive personality disorder, n (%)	6 (16)	6 (20)
Avoidant personality disorder, n (%)	1 (3)	6 (20)
Eating disorder, n (%)	1 (3)	4 (13)
Prior psychotherapy, in- or outpatient, n (%)	21 (57)	19 (63)
Prior psychiatric treatment, in- or outpatient, n (%)	12 (32)	13 (43)
Prior psychotherapeutic and/or psychiatric treatment (in- and/or outpatient), n (%)	23 (62)	21 (70)

Fifteen therapists participated in the trial in Frankfurt and 8 in Göttingen. Mean age of therapists was 48 years (SD = 14.78). Of the 23 therapists, 20 were female.

Treatment Integrity

Interrater reliability (intraclass correlation coefficient [3,1]) was high to very high for adherence and moderate to good for competence (see also online suppl. Table 1). Therapists’ adherence to the treatment manual and competency in applying the treatment modules were examined by the PTD-OCD Adherence and Competence Scale [42]. Since it was not possible to implement all treatment modules within a single session, adherence was defined as a total score of ≥10, corresponding to high adherence on more than three items or moderate adherence on several items. Overall, therapists’ adherence was found to be sufficiently high, with 75% of sessions rated as showing moderate or high adherence (M = 0.96, SD = 0.35, range: 0.36–1.86; see online suppl. Table 1). Global adherence ratings were also favorable, with 23.2% of sessions rated as high, 45.3% as moderate, 30.5% as low, and 1.1% as showing no adherence (M = 1.89, SD = 0.75, range: 0–3). Competence in applying the treatment modules was similarly high, with global competence rated as very good in 21.0%, good in 62.1%, moderate in 15.8%, and not applicable (due to missing adherence) in 1.0% of the sessions (M = 1.96, SD = 0.34, range: 1.14–2.57). On the CPPS, therapists scored high on the use of psychodynamic techniques (M = 2.83, SD = 0.73, range: 1.30–4.60) and low on cognitive-behavioral interventions (M = 0.69, SD = 0.44, range: 0–2.10), which is consistent with the findings from the PDT-OCD Adherence and Competence Scale. The difference in the CPPS (2.83 vs. 0.69) corresponds to a large effect size (d = 2.35), confirming that our treatment used primarily psychodynamic interventions.

Primary Outcome

Linear mixed-effect models with study group and study center as fixed effects and baseline Y-BOCS as a random effect showed that STPP achieved a significantly lower Y-BOCS score at the end of treatment than the waiting list condition at the end of the waiting period (t = −4.24, p < 0.0001). The between-group effect size was large (d = 1.04). The mean difference of change in the Y-BOCS between baseline and post-therapy assessment was −11.49 (95% CI, −14.36, −8.63) for STPP and −1.10 (95% CI, −4.29, 2.10) in the waiting list group, corresponding to pre-post effect sizes of d = 2.12 and 0.11. In STPP, the Y-BOCS values were reduced by almost half; in the waiting list, the Y-BOCS score remained almost unchanged (Table 2). The reduction of OCD symptom severity in the Y-BOCS by −11.94 points achieved by STPP is almost double the size of the minimal clinically important difference (MCID) of 6 points determined for the Y-BOCs [47], with the lower limit of the 95% CI (−8.63) clearly exceeding the MCID, thus indicating a clinically important difference.

Table 2.

Primary and secondary outcomes for STPP and waiting list in the ITT sample at the end of treatment/waiting period

​	STPP	MA waiting list
	mean (SD)	mean (SD)
Y-BOCS
Baseline	23.97 (5.99)	20.30 (5.20)
Post	12.01 (10.83)***	19.69 (8.58)
BDI-II
Baseline	18.09 (10.75)	16.69 (9.97)
Post	13.37 (11.09)	15.23 (12.47)
BAI
Baseline	15.58 (9.78)	12.85 (8.16)
Post	10.60 (9.49)	12.47 (10.64)
OBQ
Baseline	4.58 (1.03)	4.47 (0.97)
Post	3.92 (1.22)	4.25 (1.26)
OCI-R
Baseline	30.10 (13.52)	23.06 (12.31)
Post	21.79 (17.00)*	22.55 (14.58)
IIP-total
Baseline	1.51 (0.59)	1.55 (0.45)
Post	1.44 (0.63)	1.50 (0.55)

*p < 0.05.

***p < 0.0001.

Secondary Outcomes

STPP achieved a remission rate (Y-BOCS ≤12) of 60% at the end of treatment compared to 20% in the waiting list condition (OR = 0.065, 95% CI 0.02, 0.28, p = 0.001). Response rates defined by an at least 35% reduction of OCD symptoms were 64% and 16% (OR = 0.06, 95% CI 0.06, 0.24, p = 0.0003). If response is defined by an at least 50% reduction [10], the rates are 53% vs. 16% (OR = 0.11, 95% CI 0.03, 0.42, p = 0.003). At the end of treatment, STPP achieved a significantly lower severity of self-reported OCD symptoms (OCI-R) than the waiting list (Table 2, p = 0.049). There were no significant differences post-therapy between STPP and the waiting list in the BDI-II, BAI, OBQ, and IIP (p ≥ 0.8). Analysis of the completer sample corroborated the results for the primary and the secondary outcomes, with one exception: STPP was significantly superior to the waiting list post-therapy in the BAI (9.74 vs. 12.19, N1 = 27, N2 = 26, p = 0.04, F = 4.39).

Discussion

In this study, STPP for OCD proved to be superior to the waiting list control condition with a large and clinically meaningful effect size. Rates for remission (60%) were comparable to those reported for ERP (52%) alone and for ERP combined with clomipramine (58%) [12, 48]. Response rates of STPP defined by a 35% or 50% reduction of OCD symptoms were 64% and 53%. The between-group effect size of 1.04 achieved in our study is slightly smaller than that reported by Reid et al. [6] (1.27) for CBT with ERP but of the same magnitude. However, in contrast to many studies of CBT with ERP, our treatment effects represent effects of psychotherapy alone. It can be regarded as a strength that in contrast to many other studies [5, 6] the patients included in our study did not receive concurrent pharmacotherapy. Thus, the outcomes can be attributed solely to the effects of psychotherapy. The fact that patients using medication who were not willing to stop the medical treatment could not be included and the restrictions of the COVID-19 pandemic contributed to the relatively long time of recruitment.

The mean Y-BOCS baseline score in the STPP group (23.97) of our study was comparable to those of most studies of CBT + ERP [6], showing that our patient sample was representative of OCD patients included in other studies. We did not apply a minimum entry criterion of OCD symptom severity, which may be regarded as a limitation, possibly leading to the inclusion of patients with the diagnosis of OCD showing less severe symptoms. However, a correlation between initial symptom severity and treatment outcomes in OCD has not been shown [44]. Thus, it is unlikely that the lack of a minimum entry criteria of OCD symptom severity affected our results.

STPP for OCD aims at working through the components of the CCRT associated with the symptoms of OCD. Working through the responses of the self includes encouraging the patient to expose to the feared situation which was strongly recommended by Freud and Fenichel for the treatment of both agoraphobia and OCD [35, p. 166, 22, p. 244]. However, as described above, there are important differences between self-exposure in STPP and ERP in CBT. Self-exposure is used on a psychodynamic background to work through the CCRT related to the symptoms of OCD [22, 27, 35]. As shown by the data on treatment fidelity, the STPP treatment used primarily psychodynamic interventions. Interventions of CBT were significantly less frequently used; the difference corresponds to a large effect size.

A comparison with a waiting list condition is a less stringent test than comparisons with an active treatment, TAU or placebo, representing a limitation of our study since it does not control for nonspecific therapeutic factors. It is true, however, that many studies of CBT in OCD used a waiting list control, too. In the meta-analysis by Olatunji et al. [44], this was true for 56% of the studies. Although no active comparator was included in the studies evaluated by Olatunji et al., hardly a CBT researcher would argue that the effect size of g = 1.39 reported by Olatunji et al. was only due to nonspecific factors such as attention or expectancy. This applies to our study of STPP, too, and can be supported by empirical evidence, listed in the following: (1) The suggestion that a waiting list condition is a nocebo has been critically discussed and refuted [49, p. 3]. (2) For OCD, a meta-analysis showed that patients in waiting list control conditions did not deteriorate but showed small improvements in symptoms (g = −0.14) [50], which is comparable to the effect size achieved by the waiting list group in our study (d = −0.11). These data suggest that waiting list comparisons in OCD do not result in “inflated” effect sizes as claimed by some researchers [51]. (3) Specifically for OCD with its often chronic course and persistent nature, it is rather unlikely that the symptoms improve substantially just by expectancy or attention. (4) For studies of CBT in OCD, a meta-analysis evaluated the effects achieved by the comparators [52]. Active but not disorder-specific comparators (e.g., anxiety management, relaxation) achieved pre-post effect sizes of g = 0.26 and response rates of 15% [52]. The pre-post effect size of 2.03 and the response rate of 64% achieved in our study by STPP clearly exceed these effects of active (non-specific) treatments. (5) The pre-post reduction of OCD symptoms (−11.94) is almost double the size of the MCID of 6 points [47], thus clearly indicating a clinically important difference. (6) Furthermore, the 50%-reduction response rate of 53% achieved by STPP is comparable to that reported for ERP (43%) [10] and the remission rate of 60% is comparable to that of ERP (52%) and to ERP plus clomipramine (58%) [12]. For all these reasons, it is highly unlikely that the effects of our study can be primarily explained by factors such as attention or expectancy. The arguments listed above may be considered as quasi-experimental controls [53]. However, we agree that an experimental control of unspecific therapeutic factors by including an active comparator is a more stringent test of efficacy. The lack of an active comparator may be seen as a limitation of our study. However, as stated explicitly by Mohr et al. waiting list conditions are “potentially useful for first evaluations of novel interventions.” This applies directly to STPP for OCD [54, Table 1, p. 278].

The question whether a treatment can be designated as “efficacious” needs to be discussed in a broader context. The American Psychological Association has made several proposals to address this question, starting with the seminal article by Chambless and Hollon who required at least two RCTs showing that a treatment is superior to no treatment, placebo or alternative treatments or as efficacious as an already established treatment ([55], p. 18). At present, with our study only one RCT for STPP in OCD provides evidence for efficacy. Thus, according to Chambless and Hollon, STPP would be designated as ”probably efficacious” since only one RCT fulfilling their criteria exists and STPP has not yet proved to be more efficacious than placebo other another active treatment, required for the designation as “efficacious and specific” ([55], p. 18]). The criteria for empirically supported treatments were updated by Tolin et al. [56] who put the focus on meta-analyses providing evidence for efficacy. As soon as several RCTs of STPP in OCD are available, meta-analyses may be carried out. These meta-analyses may include comparison with waiting list conditions, unspecific treatments (e.g., TAU), and comparisons with active treatments, addressing different questions of efficacy.

It may be regarded as a limitation of our study that an imbalance in the number of patients allocated to STPP vs. waiting list (37 vs. 30) occurred. However, using a simple randomization (1:1) an uneven split may occur by chance. The observed allocation reflects chance variation rather than protocol deviation. Furthermore, we included the baseline outcome scores as covariates in the analyses. As another limitation, adverse events were not systematically collected. Nevertheless, all treatments were closely supervised throughout the trial, and neither therapists nor participants reported any serious adverse events during the intervention. Furthermore, many studies of CBT in OCD did not report adverse events either [57–60]. Another limitation of the present study is the absence of patient or public involvement in the design, planning, or conduct of the randomized controlled trial. Although this approach followed conventional research practice at the time, we recognize that involving individuals with lived experience can provide valuable perspectives, enhance the relevance and acceptability of interventions, and ultimately improve the quality of research. Future studies would benefit from incorporating co-production principles to ensure that patient voices are represented throughout the research process. However, we included findings on lived experience [61], by interviewing seven study therapists and two supervisors on their experiences in taking part in the study and applying the therapy manual. The therapists had treated 30 patients with OCD as part of the present RCT, including those patients first randomized to the wait list condition. Among others, the research questions we addressed included what it was like working with the therapy manual, i.e., what was helpful, what was challenging, and which mechanisms of change were responsible for the observed improvements in symptoms according to the study therapists. It became apparent that therapists initially held rather modest expectations regarding what could be achieved within 24 sessions. They were, however, surprised to observe that some patients experienced a noticeable reduction in symptoms despite the limited treatment duration – particularly when the focal approach enabled insight into the patients’ inner obsessive-compulsive worlds and helped establish meaningful connections between warded-off affects, anticipated reactions from internal and external objects, and the manifestation of obsessive-compulsive symptoms. The therapists also reported that, after studying the manual, they initially believed that the therapy was conceptually quite similar to CBT. After all, it included elements such as exposure and a somewhat more structured approach than is typical for psychodynamic therapies. Over time, however, this impression faded for all of them as it became clear that the decisive factor was the focus on the psychodynamics of warded-off (defended against) affects. These affects became increasingly evident and accessible to therapeutic work as the therapy progressed – also because the patients’ tolerance for them grew.

Summing up, our study provided evidence for another method of psychotherapy to treat OCD, thereby broadening the spectrum of evidence-based interventions. It is one of the few studies of psychotherapy as a monotherapy in OCD, showing pure psychotherapy effects. However, further research is needed to replicate these findings and to compare STPP with other empirically supported treatments. Additionally, future studies are needed to identify which patients are most likely to benefit from which evidence-based therapeutic approach.

Conclusion

In a first RCT, manual-guided STPP proved to be efficacious in OCD with a clinically meaningful improvement, thereby broadening the spectrum of evidence-based interventions for OCD. Of note, the study is one of the few trials of unmedicated patients in OCD, showing effects of psychotherapy alone. Further research is needed to replicate these first findings and to compare STPP with other established treatments.

Acknowledgments

We thank Prof. Dr. Marianne Leuzinger-Bohleber for her support in initiating the study, Dr. Constanze Rickmeyer for her help in preparing the study, and Prof. Dr Eric Leibing for handling the administrative matters in the study center of Göttingen.

Statement of Ethics

This study protocol was reviewed and approved by the Ethics Committees of the Universities of Giessen and Göttingen (reference numbers 136/16 and 13/7/17, respectively). Patients were informed of procedures, gave written informed consent, and could withdraw at any point with no disadvantage.

Conflict of Interest Statement

The authors declare no conflicts of interest.

Funding Sources

The study was supported by grants of the DGPT and the Heidehof-Stiftung. Neither DGPT nor Heidehof-Stiftung had any influence on the design, the results, or the presentation of the study.

Author Contributions

Falk Leichsenring, Christiane Steinert, and Heinz Weiss conceived, designed, and supervised the study. Günter Reich and Annabelle Starck served as main local investigators at the study centers in Göttingen and Frankfurt, respectively. Joachim Biskup and Reinhard Kreische supervised and managed the treatments in Göttingen and contributed to the development of the treatment manual. Britta Nimis coordinated patient recruitment, treatment administration, and data collection in Göttingen. Lila Feix contributed to the design and administration of the study. Fatima Nöske and Nikolas Heim substantially contributed to the assessment of treatment integrity, and Christian Ruckes was responsible for the statistical analyses. All authors critically read and commented on the manuscript.

Funding Statement

The study was supported by grants of the DGPT and the Heidehof-Stiftung. Neither DGPT nor Heidehof-Stiftung had any influence on the design, the results, or the presentation of the study.

Data Availability Statement

The data supporting the findings of this study are not yet publicly available since further analyses will have to be made but will be made available from the corresponding author upon request.