Both cognitive behavioral therapy (CBT) and systemic therapy (ST) have a long tradition in the psychotherapeutic treatment of various disorders. However, CBT has a much stronger history of manualization and evaluation, especially for patients with anxiety disorders [
1]. It is also part of insurance-funded German psychotherapy, which ST is not. ST was approved as evidence based in Germany for a variety of disorders, but evidence did not include anxiety disorders [
2,
3]. Three randomized controlled trials (RCTs) for anxiety disorders are available [
4‐
6]. However, they either do not especially focus on social anxiety disorders but mood and anxiety disorders in general [
4], they do not use a multi-person ST but a combined therapy [
5], or they do not use a "gold standard" comparator but minimal supportive therapy (MST) [
6]. We therefore need more studies that investigate the efficacy of ST specifically for social anxiety disorders, and that use a multi-person ST compared to a well-investigated comparator, such as CBT.
Epidemiology
Social anxiety disorders are among the most prevalent anxiety disorders (prevalence of 7–16 %). In general, women are affected more frequently (3:2) [
7]. In clinical populations, we find women and men in equal shares [
8]. Both women and men are affected by the many significant psychosocial and occupational limitations that accompany social anxiety disorders [
9]. Those affected have lower income and education, higher unemployment, are more often unmarried and have fewer friends [
10]. The age of onset of social anxiety disorders averages 10–13 years [
7]; spontaneous remission in the first 2 years is below 20 % [
11,
12].
Clinical picture
Social anxiety disorders are interpersonal disorders. Symptoms of fear such as shaking or flushing arise when the affected person experiences that he or she may attract attention, and that symptoms such as shaking or flushing constrain his or her ability to build up social relationships [
13]. Socially anxious individuals display little emotionality, intimacy and secure bonding [
14]. The performance anxiety type describes, for example, fears being seized with stage fright, behaving awkwardly if aware of being scrutinized by others and/or getting flurried in larger groups when speaking publicly or eating in restaurants. The interpersonally anxiety type describes, for example, fear of giving an impression of being a bore, an odd fellow and/or difficult to get along with so that the individual is the only one out of keeping with company in everyday situations in medium-sized groups, for example, during lunch with work colleagues or at a garden party with neighbors [
15]. People with the specific subtype fear to speak or to eat in public in some specific social situations. Individuals with the generalized subtype fear such behavior in a greater number of situations. The generalized subtype shows stronger burden caused by the generalized socially anxious symptomatology and by more severe comorbidity [
7].
Psychotherapy
CBT explains the disposition of social anxiety disorders by cognitive schemas which stimulate misleading internal information processing. Socially anxious individuals show increased self-focused attention when interacting with others stimulated by the assumption that others can see their anxiety. Based on a linked reduction in observation of other people, they focus on negative reactions and tend to interpret ambiguous and neutral feedback in a negative way. This results in excessively negative inferences about how they appear to others. Incorrect cognitions induce avoidance of social situations and/or the extensive use of safety behavior (e.g. avoiding eye contact, speaking low). The intention is on the prevention of feared catastrophes. However, avoidance and safety behavior contributes to the maintenance of negative beliefs and the increase of feared symptoms. They make patients come across to others in ways that are likely to elicit less friendly responses [
8,
16,
17].
CBT aims to reverse the maintaining processes specified in the cognitive model of Clark and Wells [
17]. The goal is to establish a realistic self-perception. The CBT manual [
8] differentiates between five phases which are described in detail in the method section.
ST explains the development of social anxiety disorders by reciprocal interpersonal interactions. Symptoms of social anxiety serve the near-distance regulation when interacting with others. They represent the (non-verbal) communication of dyadic or multi-person “between-us” quality of relationship in private (e.g. family, couple, friends) or professional (e.g. work colleagues, superior-inferior) social systems. Social anxiety indicates that an actual development task has not yet been accomplished and that socially anxious behavior, feelings and thoughts appear to be the best solution currently available. ST thus interprets social anxiety as an individual’s sensitive reaction when fearing being scrutinized and socially rejected. However, maintaining social system structures reciprocally maintains social anxiety [
18,
19].
As we did not find a multi-person ST manual for treating specifically social anxiety disorders, JS and CH developed the first manualization of ST, integrating a broad range of systemic methods, for this type of disorder. We reviewed ST manuals for social anxiety disorders in child and adolescent psychotherapy [
19] and well-established ST manuals for different disorders [
20‐
22]. We used general ST concepts [
18], integrating constructivist, solution-oriented [
5,
23] and strategic [
24‐
27] methods, in addition to attending disorder-specific relational ST dynamics [
19,
28,
29]. According to the literature and our experiences of treating patients, the aim of ST is to contextualize symptoms of social anxiety by addressing an individual’s important private and/or professional social system. Social anxiety disorders indicate the disturbance of an entire social system and most if not all of its members. It does not rely on one single individual only. Consequently, the primary goal of ST is to identify and involve all important social system members. Thus, they will be invited to therapy sessions in addition to the individual patient. If they cannot attend physically, circular questions serve their inclusion into the room on a cognitive level. The analysis of transgenerational relationships, of past and present interpersonal interactions serve the development of a new understanding of the important roles, places and resources of all system members.
The systemic model of Schweitzer and Hunger [
30] combines three different therapy settings with different participants in a multisystemic therapy concept. The first two therapies, and the majority of all therapy sessions alike, are individual but with a strong focus on relationship issues. They are combined with sessions with partners, parents or closest friends, and with one 3-hour group therapy session bringing together project patients and therapists. The ST manual differentiates between four phases which are described in detail in the method section.
Efficacy
The efficacy of CBT for social anxiety disorders is well established [
31]. A meta-analysis of 29 RCT studies showed a general effect size of 0.70 [
32]. The largest German multicenter RCT investigating social anxiety disorders is the Social Phobia Psychotherapy Network (SOPHO-NET;
n = 495) [
18]. The SOPHO-NET was initiated by the lack of evidence for the efficacy of a new psychotherapy method for social anxiety disorders. In the SOPHO-NET, this was psychodynamic therapy (PD) which was compared to CBT. Based on the Liebowitz Social Anxiety Scale (LSAS), the authors investigated remission (LSAS score ≤30) and response (LSAS score reduction of at least 31 %) demonstrably comparable to an improvement of ≤2 in the Clinical Global Impression (CGI) [
33]. The authors expected superiority of CBT with a small effect (Cohen’s
h = 0.30; i.e. CBT response rate of 70 %, PD response rate of 55 %) [
34]. With smaller effect than expected, results demonstrated superiority for remission (36 % CBT, 26 % PD;
h = 0.22) but not for response (60 % CBT, 52 % PD) [
35]. Again with small effect sizes, secondary interpersonal outcome measures such as the Inventory of Interpersonal Problems (IIP) [
36] also showed significant differences in favor of CBT. At 6, 12 and 24 months after the end of treatment, significant between-group differences were no longer found in any outcome. Contrariwise, some other studies found that CBT patients tend to continue with problems in relationship formation after the end of therapy [
37,
38].
Considering ST, we found only three RCTs for anxiety disorders [
4‐
6], and only two of them investigated social anxiety disorders. One RCT (
n = 120, Poland) compared 10 weeks of brief strategic therapy (BST) with MST. The outcome relied on the Interpersonal Sensitivity, Anxiety and Global Severity Index of the Symptom Checklist-90-R (SCL-90-R) [
39,
40]. This trial demonstrated BST superiority for patients with social anxiety disorders only (response: 50 % BST, 20 % MST) but no significance of any treatment for patients with comorbid personality disorders (response: 7 % BST and MST) [
6]. The second RCT (
n = 83, Germany) compared a combined resource-oriented cognitive-behavioral therapy (ROCBT) with CBT. The main outcomes were the Social Interaction Anxiety Scale (SIAS) and Social Phobia Scale (SPS). SIAS and SPS are original German scales and were developed by the leading German CBT researcher for social anxiety disorders [
41‐
43]. This trial demonstrated ROCBT superiority (Cohen’s
d = 1.29, for both SIAS and SPS) compared to CBT (
d = 0.86, for SIAS;
d = 0.97, for SPS). Therapy effects were stable over the 2- and 10-year follow-up in both treatment conditions [
5]. The third RCT (
n = 326, Finland) compared long- and short-term PD with solution-focused ST. This study included patients with both mood and anxiety disorders. Results indicated a statistically significant reduction of symptoms on all outcomes in all three treatment groups during the 3-year study period, including the SCL-90-R anxiety subscale and the Hamilton Anxiety Rating Scale (HAM-A). Both instruments measure anxiety but not social anxiety [
44]. The reduction was faster in short-term psychotherapies, including the solution-focused ST during the first year of follow-up. However, the reduction continued during the 3-year follow-up only for the long-term PD [
4].
Consequently, none of the above described ST RCTs used the LSAS, though it is most often used for the assessment of social anxiety disorders [
45]. The SOPHO-NET used the LSAS as the main outcome for that reason [
35]. Willutzki and colleagues decided to use the original German SIAS and SPS, which highly correlate with the LSAS [
5]. We therefore miss information regarding the efficacy of multi-person ST on established social anxiety assessments, such as the LSAS, SIAS and SPS. Additionally, we also miss information regarding the efficacy of multi-person ST when compared to a well-established comparator, the CBT.
Pilot study
RCTs are regarded as the “gold standard” of present day evidence-based research in psychotherapy. They are often complex, time-consuming and expensive. Before a large RCT is undertaken, a pilot study should be conducted that mimics all the major essentials of the planned larger study [
46]. As this study is the first trial that includes both a multi-person ST, integrating a broad range of systemic methods, and CBT, especially focusing on social anxiety disorders, we will have to conduct this pilot RCT before planning in detail the future main trial. Due to the mismatch of this study with previous investigations of ST for social anxiety disorders, reliable information that is needed for the sample size calculation for the main RCT is missing. Thus, this pilot study will help us to pay careful attention to sample size calculation, it will aim to save costs and use patients efficiently in the future RCT [
47].