Background
Treatment engagement and retention in residential drug rehabilitation settings for individuals with substance misuse and psychiatric comorbidities is variable [
1‐
4]. Studies have implicated mental health conditions such as social anxiety as having a negative impact on engagement and retention [
5]. As residential drug rehabilitation is often the last resort for people with severe substance misuse problems, effective management of negative prognostic factors would be a significant contribution to improved treatment outcomes. In Australia, as elsewhere, residential drug treatment is generally provided by Therapeutic Communities (TC). TCs are intense psychosocial based interventions which are grounded in the assumption that substance misuse is a dysfunction of lifestyle and character development involving the whole person, resulting in a range of inter and intra personal difficulties, which are most effectively addressed in long term treatment in a community comprised of therapists and peers (i.e., other uses) [
6,
7]. Treatment in TCs involves clients moving through a number of treatment stages over a six to 15 month period. Each stage is associated with specific learning objectives, which facilitate transition to subsequent stages.
Individuals attending residential drug treatment generally present with increased severity of symptoms e.g., [
8] and incidence of psychiatric co-morbidity e.g., [
9] relative to outpatients, and are more likely to have significant social problems [
10]. Unsurprisingly, individuals with co-occurring psychiatric disorders are at greater risk of poor treatment retention e.g., [
11,
12] than those without such comorbidities. Thus, given the complex nature of patients attending TCs, treatment dropout is high. For example, Polimeni, Moore and Gruenert [
13] examined dropout rates from an Australian TC over a 10 year period and found that approximately 45% of patients dropped out of treatment in the first five weeks. Australian Institute of Health and Workforce Studies revealed that four percent of patients stay in residential treatment less than one day, and 40% less than one month [
14]. This low rate of retention is mirrored in data collected in other Western countries [
15,
16].
Although the positive treatment efficacy of TCs has been demonstrated in a number of studies [
17,
18], the high drop-out rate of patients in the first few weeks of treatment is of particular concern, especially in light of a substantial body of evidence which has found that for residential treatments, client retention of at least three months is associated with improved outcomes [
16,
19‐
22]. It has been suggested that treatment factors affecting drop-out might differ between treatment modalities. In an analysis of predictors of client retention in TCs, Condelli and De Leon [
23] found that clients who spent a significant amount of time with large groups of people, rather than alone prior to drug use were more likely to stay longer in treatment. Similarly, in a retrospective review, Johns, Baker, Webster and Lewin [
15] found that clients who stayed in treatment less than 90 days were more likely to have reported spending their free time alone, or with one other person than clients who stayed in treatment longer than 90 days. It is unsurprising given the social nature of the TCs, that clients who have previously experienced difficulties interacting with small to large groups of people would find the TC environment challenging. Although the previous studies did not examine psychiatric comorbidities, it appears that social anxieties might be problematic for individuals remaining in residential treatment due to the highly social nature of the TCs. There is substantial evidence to suggest that social anxiety and substance use problems frequently co-exist, and that substance users with co-occurring social anxiety are significantly more impaired than substance users with no substantial social anxieties see [
24]. Thus, it is likely that social anxieties may impede actual entry into residential treatment for substance dependence. That is, unlike depressive symptoms which can sometimes increase the likelihood that an individual seeks substance misuse treatment [
25], it is possible that social anxiety may serve as a barrier to seeking treatment for substance misuse problems, particularly group or residential treatment. This is not surprising given the highly social nature of these two treatment modalities.
Social anxiety is pervasive and disabling and is characterised by intense, persistent fear and anxiety of situations involving social interaction, social evaluation and appraisal [
26]. It has been argued that social anxiety exists on a continuum, with social anxiety at the mild end of the continuum, manifesting as shyness, and at the extreme end as the clinical diagnostic categories of Social Anxiety Disorder (SAD) in the DSM-5 (APA, 2013) and Social Phobia in the ICD-10 (WHO, 1992) [World Health Organisation [WHO], 1992; [
27]. The DSM-5 defines SAD as a “marked fear or anxiety about one or more social situations in which the individual is exposed to possible scrutiny by others” (Criterion A; p. 202), with the individual fearing that they will act in a way or show anxiety symptoms that will lead to them being negatively evaluated (Criterion B). These social situations almost always provoke fear or anxiety (Criterion C), are avoided, or endured with intense fear or anxiety (Criterion D), with the fear or anxiety reaction being judged to be out of proportion with the actual threat posed by the social situation (Criterion E).
According to National Comorbidity Survey data [
28,
29], SAD is the most commonly reported anxiety disorder. In a nationally representative Australian population study, Teesson, Hall, Lynskey and Degenhardt [
30] found that 6.5% of Australian adults met ICD-10 criteria for alcohol use disorder. Of these individuals, 3.7% met criteria for SAD, and among individuals who met criteria for SAD, 16.7% met criteria for an alcohol use disorder. Among adults seeking drug or alcohol treatment, rates of SAD are considerably higher, with comorbidity rates of up to 56% reported [
31,
32]. While the causal relationship remains unclear, concurrent social anxiety seems to lead to social isolation and depression in those recovering from substance dependency, which in turn increases the likelihood of drug or alcohol relapse and the reliance on drug use in social situations i.e., self-medication and tension-reduction; [
33]. Thus, individuals with comorbid SAD are at greater risk of poor treatment retention [
11,
12] and relapse to substance misuse than those with substance misuse problems alone. Furthermore, although depressive symptoms can be ameliorated to some extent by treatment for substance misuse alone see [
34], social anxiety disorder remains at clinical levels after drug treatment despite reductions in substance use [
35] and has been shown to be risk factor for relapse [
36].
Although most research and treatment models tend to categorize social anxiety as a categorical phenomenon, it has been argued e.g., see [
37] that social anxiety is better represented dimensionally. Preliminary evidence supporting a dimensional conceptualization of social anxiety came from studies which attempted to form subgroups based on symptom severity. For example, although Stein, Torgrud and Walker [
38] were able to classify individuals into three groups based on the number of social situations they feared (e.g., 1 to 3 social fears, 4 to 6 social fears and 7 to 12 social fears) they found that disability was related to the number of situations feared in a continuous manner, which indicated that viewing the number of social fears dimensionally was potentially more appropriate than viewing them categorically. Similarly, Vriends, Becker, Meyer, Michael and Margraf [
39] found that when socially anxious individuals were classified into groups according to symptom severity, that there was no clear boundary between the subtypes.
More conclusive evidence for the dimensionality of social anxiety has come from taxometric analyses. Using data from the National Comorbidity Survey Replication (NCS-R), Ruscio [
40] conducted a number of taxometric analyses (MAMBAC, MAXEIG and L-Mode) and found strong support for a dimensional, rather than taxonic structure. These results were replicated by Crome, Baillie, Slade and Ruscio [
41] whose analyses of both the NCS-R dataset and the Australian National Survey of Mental Health (ANSMH) dataset also strongly supported the dimensional nature of social anxiety. Importantly, support for the dimensional nature of social anxiety has also been found in clinical populations, with Kollman, Brown, Liverant and Hofmann [
42] finding support for the dimensional nature of social anxiety in a sample of 2,035 outpatients presenting for treatment at a US based anxiety disorder treatment centre.
From a clinical perspective, the usefulness of considering social anxiety dimensionally was demonstrated by Ruscio [
40] who found that dimensional SAD was a better predictor than categorical SAD of a number of outcomes, including subsequent suicide ideation, mood disorders, and treatment seeking behaviours. From the perspective of alcohol use, Crum and Pratt [
43] found that individuals with subclinical social anxiety symptoms were more likely to have drinking problems than those with a clinical diagnosis of SAD, indicating consideration of social anxiety as a dimensional, rather than categorical phenomenon might be especially important in research exploring social anxiety and alcohol and substance use disorders.
Taken together, the results of these studies is that differences between shyness, social anxiety, and avoidant personality disorder are quantitative, rather than qualitative (although for an alternative position, see [
44]). These results have implications for clinical research on social anxiety and suggest that an exclusive focus on individuals who meet DSM criteria for SAD is not clinically warranted. Moreover, as Kollman, Brown, Liverant and Hofmann [
42] note, it also suggests that social anxiety ought to be assessed using continuously, rather than forced choice measurements.
Currently, social anxiety in individuals seeking treatment for substance misuse social anxiety is poorly detected [
45] and inadequately treated within alcohol and drug treatment settings [
46]. Indeed, very few studies have attempted to address social anxiety in individuals presenting for substance use disorders [
33]. This highlights the possible need for targeted treatment for individuals with social anxiety and comorbid substance misuse. Given that social anxiety may play a role in an individual deciding not to commence or continue residential treatment, it is argued e.g., [
47] that addressing social anxiety symptoms prior to and/or at the early stages of treatment could improve treatment retention in individuals with comorbid social anxiety and substance misuse problems. Considering the complex nature of the relationship between SAD and substance misuse, it is unclear how to proceed when treating SAD in individuals with substance misuse problems. Specifically, it is unclear if SAD should be treated first or concurrently with treatment for substance misuse problems. Currently, only two RTCs have explored the efficacy of treating SAD in substance use populations [
5,
48].
The first study [
5] involved individuals seeking treatment for alcohol use problems who met criteria for both alcohol misuse and SAD. Individuals were randomised into one of two treatment conditions, receiving either a 12 week individual cognitive behavioural therapy (CBT) intervention targeting the alcohol problem only or a 12 week individual CBT treatment targeting both alcohol misuse and social anxiety problems (which they referred to as the dual group). In the alcohol only intervention group, treatment sessions lasted 60 minutes. In the dual group, both alcohol and social anxiety was covered in each session (beginning with alcohol and ending with social anxiety). Unexpectedly, there were no differences between the alcohol only and dual intervention groups on social anxiety symptom severity. Furthermore, the dual intervention group reported poorer alcohol-related outcomes compared to the alcohol only intervention group. The authors noted that a limitation of their methodology was the somewhat segregated approach to treating both disorders. Specifically, the concurrent sessions were divided into two sections during the one treatment session: treatment of the alcohol problem, followed by CBT for social anxiety. The effectiveness of this temporal within-session sequence was questioned by the authors and it was noted that the literature is unclear on the best approach.
A second study [
48] compared the efficacy of an intensive psychosocial relapse-prevention program delivered on its own or in combination with an anxiety treatment program comprising CBT and optional pharmacotherapy (i.e., selective serotonin re-uptake inhibitors; SSRI). Although the addition of the anxiety treatment program did result in significant reductions to anxiety symptoms and avoidance, the program was not associated with concomitant reductions in alcohol relapse rates. Despite this, there was a non-significant trend for the combined treatment to be associated with a greater likelihood of abstinence of at least 30 days relative to the alcohol only intervention group, with an effect size of 0.13. It has recently been noted, however that the small sample size might have meant that the study was underpowered [
49].
Taken together, there is some evidence to suggest that treatment of social anxiety in treatment seeking substance users results in reductions in anxiety symptoms, however, the effects on substance misuse are less clear. Currently, there are no studies which examine whether addressing social anxiety prior to treatment improves rate of treatment entry and subsequent treatment retention. This is important, because as noted earlier, length of tenure in TC treatment is a significant predictor of improved treatment outcomes. Thus, the focus of the current study is on examining whether addressing social anxiety symptoms prior to entry for residential treatment improves the likelihood that individuals will enter and stay in treatment. A secondary goal, in line with previous clinical studies, is to explore if treatment of social anxiety results in concomitant reductions in other indices of psychological distress (e.g., depression and general levels of anxiety).
The current study utilises a randomised control design which adheres to CONSORT guidelines [
50]. The intervention group receives four sessions of treatment for social anxiety symptoms, plus boosters, and the control group remain on the waiting list for entry into residential treatment (as a treatment as usual group). The proposed treatment program for SAD is based on Rapee’s [
51] program “Overcoming Shyness and Social Phobia: A Step by Step Guide”. This treatment approach involves a range of standard empirically validated methods for addressing social anxiety disorder symptoms, including: attention training, cognitive restructuring, exposure to feared situations and realistic feedback of social performance. Rapee’s [
51] program focuses on setting ‘homework’ for clients, primarily on restructuring beliefs about social situations into more realistic terms. It is designed to be undertaken as either ‘pure self-help’ or ‘therapist augmented self-help’ where clinicians help ‘problem-solve’ the program’s concepts with a specific focus on each client’s personal context [
52] Two RCTs support the effectiveness of the program in reducing social anxiety symptoms and other related psychological problems [
52,
53].
The current study sought to examine whether treatment of social anxiety prior to entry to treatment would increase the likelihood that individuals would enter, and subsequently remain in treatment for at least three months. Additionally, the current study focusses on whether brief treatment of social anxiety would result in significant reductions in social anxiety symptom severity over time, and significant reduction in other indices of psychological distress.
Authors’ contributions
PS. Design of the study, development of the intervention, manuscript preparation, final approval of manuscript. MK. Design of the study, development of the intervention. JW. Manuscript preparation, data collection. NK. Design of the study, statistical design. AH. Development of the intervention. SG. Design of the study, development of the intervention. All authors read and approved the final manuscript.