(p. 149) Assessment of Social Anxiety Disorder
Assessment and therapy are inexorably intertwined in our treatment approach. This appendix focuses on what we perceive as being an adequate pretreatment and posttreatment assessment. Readers seeking a more thorough review of assessment and psychometric issues in social anxiety disorder are referred to the work of Fernandez, Piccirillo, and Rodebaugh (2014); Herbert, Brandsma, and Fischer (2014); and McNeil and Quentin (2014).
Assessing for Clinically Significant Social Anxiety
In most clinical settings, the client’s first session consists of an unstructured clinical interview. Because individuals with social anxiety disorder do not typically seek treatment unless they have a comorbid diagnosis (Schneier et al., 1992) and because healthcare professionals often fail to detect clinically significant social anxiety, we recommend that all intake interviews routinely include questions screening for problems with social anxiety. Research has revealed that clients obtaining a total score of 6 or more on three statements derived from the Social Phobia Inventory (SPIN; Connor et al., 2000) are likely to meet diagnostic criteria for social anxiety disorder (Connor, Kobak, Churchill, Katzelnick, & Davidson, 2001; Katzelnick et al., 2001). These items, referred to as the “Mini-SPIN,” are
■ “Fear of embarrassment causes me to avoid doing things or speaking to people”;
■ “I avoid activities in which I am the center of attention”; and
■ “Being embarrassed or looking stupid are among my worst fears.”
(p. 150) Each item is rated on a 5-point scale as follows: 0 = not at all; 1 = a little bit; 2 = somewhat; 3 = very much; 4 = extremely. At the Adult Anxiety Clinic of Temple University, we have used the Mini-SPIN as a brief telephone screening instrument for some time, and it has proven to have substantial utility. Furthermore, among callers who later came to the clinic for evaluation, the Mini-SPIN was significantly correlated with other measures of social anxiety and functional disability, but not with measures of generalized anxiety/worry, depression, or anxiety sensitivity (Weeks, Spokas, & Heimberg, 2007).
We present the Mini-SPIN as one method that may be used to initially screen for the presence of clinically significant social anxiety. Naturally, once you suspect that social anxiety is a problem for any given client, a more thorough assessment is necessary to arrive at a diagnosis of social anxiety disorder according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria. The clinician-administered and self-report instruments described in this appendix can be helpful in clarifying whether the social anxiety is within the normal range or suggestive of clinically significant difficulties. Semi-structured clinical interviews can also be helpful in terms of detecting social anxiety disorder as well as assisting with differential diagnosis and assessing comorbid conditions that may affect the course of treatment. Commonly used semi-structured clinical interviews include the Anxiety and Related Disorders Interview Schedule for DSM-5: Lifetime Version (ADIS-5-L; T. A. Brown & Barlow, 2014) and the Structured Clinical Interview for DSM-5 Disorders, Clinician Version (SCID-5-CV; First, Williams, Karg, & Spitzer, 2015). Although in many settings unstructured interviews are favored over semi-structured interviews due to financial, time, and training constraints, we recommend having copies of one or more of these interviews on hand for use with clients with particularly complex presentations.
At times it may be difficult to determine whether a client’s symptoms are best characterized as features of social anxiety disorder or some other disorder. One of the more challenging differential diagnostic decisions can be between social anxiety disorder and panic disorder. Individuals with (p. 151) social anxiety disorder may occasionally or routinely experience panic attacks in social situations (Potter, Drabick, & Heimberg, 2014; Potter, Wong, et al., 2014; Versella, Potter, & Heimberg, 2016). Individuals with panic disorder may also experience panic attacks in social situations and feel embarrassed about how others may perceive them if they abruptly leave due to panicky feelings or if their anxiety symptoms are noticeable. Despite these overlapping features, these disorders can be distinguished. Making the correct diagnosis is important because different approaches to treatment may be indicated.
One question to consider is whether the individual is primarily afraid of the symptoms themselves or that the symptoms will lead to negative evaluation. Individuals with panic disorder would be expected to be concerned about the anxiety symptoms even if they were alone and no one witnessed their panic attack. Individuals with social anxiety disorder, in contrast, would be expected to be most concerned that others will see their anxiety symptoms and evaluate them negatively on that basis.
Another question to consider is whether panic attacks occur in nonsocial situations or, conversely, whether panic attacks occur exclusively in certain social situations. A diagnosis of panic disorder is more likely to be warranted if the individual is afraid of having a panic attack in at least some situations in which social evaluation is unlikely (e.g., when alone, during sleep, going over a bridge). Additionally, social anxiety disorder should be considered if the panic attacks occur exclusively during or when anticipating a feared social situation (e.g., giving a speech). Importantly, social anxiety disorder and panic disorder can co-occur, and some individuals meet criteria for both disorders. In these cases, an agreement should be reached between the client and the clinician regarding which disorder should be the initial focus of treatment.
We are occasionally asked how to differentiate social anxiety disorder from paranoid personality disorder or “paranoia” in general. It is true that some socially anxious clients report social-evaluative concerns that seem quite improbable to people without the disorder. For example, it may seem “paranoid” for socially anxious clients to fear that they are being laughed at by strangers in restaurants, being criticized by people walking down the street, or being otherwise scrutinized by strangers in what are innocuous situations for most people. This hypervigilance to negative evaluation may seem paranoid to most people who do not (p. 152) view others as sources of painful rejection or who do not view themselves as fundamentally unacceptable to others. Such reports from socially anxious clients, especially ones with more generalized fears, are not uncommon and can be conceptualized as part of the presentation of social anxiety disorder rather than as an alternative or additional diagnosis. Furthermore, in most cases, there is at least some recognition that these types of fears are either excessive or unreasonable. Clients meeting criteria for paranoid personality disorder would be expected, in contrast, to have less concern about falling short of the standards of others. Instead, they would be more likely to view their own behavior as reasonable and the behavior of others as unjustifiably malevolent.
We have also been asked to distinguish between severe social anxiety disorder and schizoid personality disorder. Schizoid individuals do not desire or enjoy close relationships. Socially anxious persons, in contrast, yearn for the social contact they fear. Occasionally, persons with social anxiety disorder will, during the course of treatment, make statements such as, “Maybe I am just not interested in making conversation with coworkers” or “I don’t care about dating”—even when these statements could not be further from the truth. In most instances, gentle probing will reveal that these statements are reflective of intense anxiety regarding confronting a feared social situation and the associated desire to avoid it, as opposed to a genuine lack of interest in other people.
Importance of the Diagnostic Evaluation
The time devoted to a thorough diagnostic evaluation is time well spent. Clients whose symptoms are better accounted for by a diagnosis other than social anxiety disorder may experience little improvement with this treatment and may experience, among other things, increased frustration, hopelessness, and financial hardship. In instances when misdiagnosed clients are included in a group treatment, the group process is likely to be disrupted as the client’s poor fit with the group becomes apparent.
A thorough diagnostic assessment is also important in determining whether social anxiety disorder is only one aspect of a more complex clinical picture. Comorbidity is common among persons with social (p. 153) anxiety disorder (Magee, Eaton, Wittchen, McGonagle, & Kessler, 1996; Schneier et al., 1992). We routinely treat clients with comorbid anxiety and mood disorders in both group and individual treatment. As noted previously, that comorbidity is associated with greater pretreatment and posttreatment impairment (Erwin et al., 2002), but it does not preclude significant therapeutic benefit. Comorbidity should suggest to you and the client that more extended treatment of social anxiety or supplemental treatment directed at the comorbid disorder may be required for the client to achieve optimal end-state functioning.
For clients highly impaired by disorders or life problems in addition to social anxiety, the clinician and the client should come to an agreement regarding what the initial focus of treatment should be. Severe depression, substance dependence, uncontrolled bipolar disorder, severe eating disorders, or acute life stressors (e.g., bereavement) are examples of problems that, in most cases, would need to be addressed prior to initiating this treatment program. In the case of individual therapy, treatments can be sequenced or integrated in a manner that best meets the needs of the client.
Give the client feedback regarding the diagnoses arrived at (or being considered) following the initial interview. Also discuss the importance of completing additional assessments to understand more fully the nature of the client’s social anxiety prior to starting treatment. We believe that an adequate pretreatment assessment battery also includes self-report and clinician-rated measures of social anxiety. A behavior test, in which the client confronts one or more fear-provoking social situations within the clinic setting, also provides important clinical information. Outside of research studies, we have recently omitted a separate behavioral test prior to starting treatment and used the first in-session exposure as a de facto assessment of social performance instead. (See more on the use of the first exposure as a behavior test in the discussion that follows.) It should be possible to complete this battery over the course of one or two sessions.
Even for cases in which social anxiety is the presenting complaint and obviously the primary problem, administration of validated self-report (p. 154) and clinician-rated instruments is important. In the context of an initial clinical interview, shame and evaluative fears can lead clients to give descriptions of their problems that do not fully reflect the severity or pervasiveness of their social fear and avoidance. Administration of psychometrically sound self-report and clinician-rated measures facilitates the assessment of fear and avoidance in a broad range of social situations within a relatively brief time frame. Furthermore, the availability of data regarding the typical performance of socially anxious clients and normal controls allows you to objectively evaluate the severity of a particular client’s symptoms relative to meaningful reference points. Last, pretreatment scores provide a baseline against which progress can be objectively assessed.
Recommended Assessment Battery
To facilitate the pretreatment assessment, we include copies of widely used measures of social anxiety, scoring information, and normative data in Appendix B. For more information on the measures presented, please see the work of Herbert et al. (2014). In a standard pretreatment assessment battery, we typically include the Social Interaction Anxiety Scale (SIAS) and Social Phobia Scale (SPS). These self-report instruments assess fear of interpersonal situations and fear of being scrutinized by others, respectively (Mattick & Clarke, 1998), and both have been shown to discriminate between persons with social anxiety disorder and those with other anxiety disorders (E. J. Brown et al., 1997; Heimberg, Mueller, Holt, Hope, & Liebowitz, 1992). To assess the fear of being judged unfavorably by others, we include the self-report Brief Fear of Negative Evaluation Scale (BFNE; Leary, 1983). Note that recent research suggested that for the BFNE (Rodebaugh et al., 2004, 2011; Weeks et al., 2005) and the SIAS (Rodebaugh, Woods, & Heimberg, 2007; Rodebaugh et al., 2011), the straightforwardly worded items may provide more valid measurement than those items that require reverse scoring; see the cited works for details.
For clinicians looking to keep the item count to a minimum, a number of abbreviated versions of the SIAS and SPS have been developed (Carleton et al., 2009; Fergus, Valentiner, McGrath, Gier-Lonsway, & Kim, 2012; Kupper & Denollet, 2012; (p. 155) Peters, Sunderland, Andrews, Rapee, & Mattick, 2012) and may present valid, briefer alternatives. All have reasonable psychometric characteristics and correlate well with the original scales. The six-item versions of the SIAS and SPS developed by Fergus et al. (2012) and Peters et al. (2012) have been directly compared to each other and to the original scales by Le Blanc et al. (2014); although they differ on some psychometric indices, there is little to recommend one set over the other. The 14-item Social Interaction Phobia Scale, which takes items from both the SIAS and SPS, has been the most extensively validated of the various short forms. It has one advantage over the other short forms in that it has clinically validated cutoff scores that differentiate individuals with social anxiety disorder from those with no disorder (Carleton et al., 2009) or with generalized anxiety disorder (Menatti et al., 2015). See the source articles for additional information on these various short-form options.
Finally, we administer the Liebowitz Social Anxiety Scale (LSAS; Liebowitz, 1987), which is a clinician-administered instrument designed to assess fear and avoidance of social interaction and performance situations. We have conducted considerable research on the psychometric characteristics of the LSAS and demonstrated that it is highly correlated with other measures of social anxiety, able to discriminate between persons with social anxiety disorder and persons with generalized anxiety disorder, and responsive to the effects of both cognitive–behavioral and pharmacological treatment (Fresco et al., 2001; Heimberg et al., 1998, 1999; Heimberg & Holaway, 2007; Mennin et al., 2002). A self-report version of the LSAS is also available. The LSAS-SR (Fresco et al., 2001) correlates quite highly with the clinician-administered version and demonstrates virtually identical psychometric characteristics (e.g., Rytwinski et al., 2009).
Additional Assessment Measures
Other instruments are available that were designed specifically for use in the assessment of social anxiety disorder. Probably the most commonly used self-report instruments not included in this Therapist Guide are the SPIN and the Social Phobia and Anxiety Inventory (SPAI; Turner, Beidel, Dancu, & Stanley, 1989). The Brief Social Phobia Scale (BSPS; (p. 156) Davidson et al., 1991) is probably the next most frequently used clinician-administered instrument, following the LSAS.
A relatively new self-report measure that may be worth serious consideration is the Subtle Avoidance Frequency Examination (SAFE), which measures the extent to which the client engages in a range of safety behaviors and other forms of subtle avoidance (Cuming et al., 2009). In the original study, three factors emerged from the SAFE, which appeared to reflect active safety behaviors, subtle restriction of behavior, and behaviors aimed at avoiding or concealing physical symptoms. The SAFE demonstrated strong internal consistency, good construct validity, and the ability to discriminate between clinical and nonclinical participants (Rowa et al., 2015). In addition, the SAFE was responsive to the effects of treatment (Rapee, Gaston, & Abbott, 2009). Although further evaluation is required, the SAFE may prove to be a useful measure both to identify avoidance behaviors that may be important to target in exposure-based treatment and as an additional means to assess treatment outcome. In our study of the relative efficacy of cognitive–behavioral therapy (CBT) and mindfulness-based stress reduction for individuals with social anxiety disorder, the SAFE was one of the few measures to discriminate between the two treatments, demonstrating greater improvement among individuals in the CBT groups (Goldin et al., 2016). A copy of the SAFE is included in Appendix B.
Supplemental Assessment Measures
In addition to social anxiety instruments, we routinely administer measures that tap other constructs. We often administer the Beck Depression Inventory—Second Edition (BDI-II), which assesses symptoms of depression (A. T. Beck et al., 1996). An alternative to the BDI-II that is being used with increasing frequency in primary care settings is the nine-item Patient Health Questionnaire (PHQ-9; Kroenke, Spitzer, & Williams, 2001). Disability measures such as the Liebowitz Self-Rated Disability Scale (Schneier et al., 1994) or the Sheehan Disability Scale (Hambrick, Turk, Heimberg, Schneier, & Liebowitz, 2004; Sheehan, 1983) are useful in assessing various levels of functional impairment across a number of domains. Perhaps the best standard by which to judge the impact of a treatment is its effect on the client’s overall sense of (p. 157) well-being. Therefore, we also administer the Quality of Life Inventory, a self-report measure that assesses the extent to which clients perceive themselves as satisfied in the areas of their lives that they deem important to their happiness (Frisch, 1994; Frisch, Cornell, Villanueva, & Retzlaff, 1992).
Using Behavioral Observation in Clinical Settings
An assessment of the client’s behavior in social situations is an important part of the pretreatment assessment. Behavioral observation provides objective information about how anxious the client becomes during social interaction, performance, or observational situations. It also provides an objective index of the quality of the client’s social behavior and helps the therapist to estimate how likely the client’s efforts are to be met with a positive response in the real world. This objective information cannot be replaced by self-report given that research has shown that socially anxious individuals describe their social behavior as more inadequate (e.g., Norton & Hope, 2001; Rapee & Lim, 1992; Stopa & Clark, 1993) and their anxiety as more obvious (e.g., Alden & Wallace, 1995) than do independent judges. In previous editions of this Therapist Guide, we recommended a formal, separate behavioral test to gather this information. However, we now believe that a separate behavior test prior to starting treatment is only necessary if there are questions about whether significant social skills deficits are present or behavioral data are needed to confirm the diagnosis. A detailed description of behavioral tests can be found in the work of Heimberg and Becker (2002).
Good clinicians begin behavioral assessment with the first interactions with the client. Socially anxious clients are often anxious in interactions with the therapist or office staff, and the astute clinician can observe how much this anxiety interferes with client performance. Significant social skills deficits that are beyond impaired social performance due to anxiety are usually also apparent in these interactions. If the client’s social performance is reasonably adequate or improves as he or she becomes more accustomed to the therapist, we typically just use the first in-session exposure as the baseline behavioral assessment. As with a formal behavior test, you should be looking for the following information:
■ Overall level of social skill and quality of performance
■ Overall visibility of anxiety
■ Visibility of specific anxiety symptoms that may be a focus of treatment, such as a hand tremor
■ Concordance between the client’s self-report of social skill/quality of performance or visibility of anxiety symptoms and the impressions of the therapist
■ Propensity of the client to avoid all or part of the situation or escape
You will also have self-reported subjective anxiety and cognitions before, during, and after the exposure to consider with the behavioral data. You can view all of the measures listed qualitatively, or you can easily think of them on a scale (e.g., 1–5 on social skills with 1 being inadequate performance, 5 being excellent performance). Making notes about the behavioral data or using a quantitative rating would allow you to track client progress without the inevitable biases of memory.
We find behavioral data extremely useful in case conceptualization and treatment planning. As noted previously, socially anxious individuals often believe they have deficient social skills. Most often this represents a biased perception of their own behavior or their anxiety is interfering with performance. However, occasionally true social skills deficits are an important part of the clinical picture and a more formal behavioral test should be conducted. As described by Meier and Hope (1998), behavioral tests can help tease apart this difference. If a client can perform the requisite social skills during a behavior test, then the client’s deficit is logically one in performance, not lack of skills. Therefore, if in the first behavioral test the client demonstrates reasonable social skills, the assessment can stop, even if the client is convinced that his or her skills are lacking. However, if performance is noticeably poor, then you can conduct additional behavioral tests that evoke less anxiety. If the requisite skills are present in the less threatening behavioral test, then performance quality should improve when the client experiences less anxiety. As a last resort for highly anxious individuals, we often have the client give a verbal account of what one might say or how one might act in a given social situation. If, at a minimum, the client can state what skilled behavior would look like, then we proceed with the exposure-based treatment in this Therapist Guide without additional (p. 159) skills training. However, if one cannot construct a behavioral test or elicit a verbal report that yields evidence of social skills, then the treatment plan should probably include some attention to the training of social skills.
When judging social skills or the visibility of anxiety symptoms, it is important for you to consider the standards by which you are making that judgment. There are many very successful people who do not have perfect social skills. The standard of what is acceptable varies widely across cultures and situations. One does not have to have the oratory skills of Presidents John F. Kennedy, Ronald Reagan, or Barack Obama to make the report at a weekly staff meeting of 10 people. Verbal or nonverbal behaviors that will be viewed very negatively by many people or a clear lack of awareness of what is generally considered socially appropriate may require skills training. However, subtle concerns may be within the broad range of normal behavior or behaviors that will change in response to the natural contingencies of social exchanges once the client stops avoiding and starts entering more social situations. Finally, you should consider whether safety behaviors may be masquerading as poor skills. For example, a client may make poor eye contact when he is anxious because it helps reduce anxiety, but he knows he should make better eye contact and he does so when not anxious.
Development of the Treatment Contract
For clients with complex clinical presentations, the pretreatment assessment battery provides more data regarding whether the client’s symptoms are best accounted for by the diagnosis of social anxiety disorder and whether social anxiety should be the initial focus of treatment. Our preference is for clients beginning this program to make it the primary focus for an agreed-on period of time, such as 12–16 weeks, even if significant co-primary or secondary diagnoses, such as depression, are present. Otherwise, at particularly challenging points during treatment—such as the initiation of exposure exercises—it is easy to fall into the trap of switching the focus of therapy to other issues. In such instances, avoidance behavior is reinforced. Furthermore, we believe, based on clinical experience and research, that clients who do not receive a sufficient dose of exposure are unlikely to improve. Importantly, we do not want clients (p. 160) who have read the Client Workbook but completed few or no in-session or in vivo exposures to incorrectly conclude that this program did not work for them.
After the course of treatment initially contracted with the client, the assessment battery delivered at pretreatment should be repeated (i.e., self-report questionnaires, clinician-administered instruments). The posttreatment assessment informs decisions regarding whether the client is ready to discontinue treatment, whether more sessions focusing on social anxiety are necessary, or whether supplemental or alternative treatments need to be considered.
At the end of 16 sessions, it is unlikely that all of a client’s social fears will have been eliminated. In fact, elimination of social anxiety should not be the criterion for ending treatment. Instead, if a client stops avoiding key social situations, experiences a meaningful reduction of anxiety in a few areas, and believes that she can use the skills gained in therapy to continue to work independently, then the client is probably ready to stop treatment. Such a client is likely to continue to make progress in the coming months. Follow-up appointments are recommended to monitor the client’s clinical status (e.g., 1 month and 6 months posttreatment).
In other cases, it may be apparent that treatment is working, but the client’s anxiety and avoidance are still too severe and pervasive for the client to be ready to continue alone. This phenomenon may also be observed among some clients who are depressed or have other significant comorbidities. In these cases, continued treatment is recommended. The data obtained from the most recent assessment should be used to guide decisions about the length of the new treatment contract and the domains that will be targeted during that time. The new treatment contract, as with the original one, should be for 16 sessions or less, at which point another posttreatment assessment should occur. For some clients, tapering of sessions over the course of the new treatment contract may be appropriate (e.g., meet weekly for 2 months, biweekly for 2 months, monthly for 2 months).
(p. 161) Last, in some cases the assessment will reveal that treatment has resulted in little or no improvement. In some cases, you may decide that the client’s symptoms are better accounted for by another diagnosis, and you should recommend appropriate alternative treatment. In other cases, social anxiety disorder may indeed be the primary problem, but factors can be identified that have contributed to a poor outcome. Modifying these factors would become a priority. For example, environmental factors that promote continued avoidance behavior might be present for some clients (e.g., a husband who discourages his wife’s increasing independence; an adult child whose social anxiety symptoms allow him to remain cared for financially by a parent). Therapy that addresses these environmental factors would be recommended before attempting additional treatment for social anxiety. In another example, overwhelming anxiety during exposures may be identified as resulting in poor homework compliance that has, in turn, contributed to a poor outcome. Supplemental treatments such as applied relaxation (Öst, 1987) or concomitant medication may make exposures more easily tolerated by these clients. Similarly, clients whose anxiety is partially maintained by the negative reactions that they elicit from others during exposures due to poor social skills may benefit from social skills training.
Culturally Sensitive Assessment
Multicultural competency is an essential feature of contemporary clinical assessment and practice. Although a full discussion of that competency is beyond the scope of this Therapist Guide, there are a couple of issues that are especially important when working with individuals who present with social anxiety disorder. First, there are wide cultural variations in appropriate social interactions. You must evaluate the social anxiety within the client’s cultural context. For example, gaze aversion that might be seen as a symptom of social anxiety in European American culture is culturally appropriate among certain Native American tribes (De Coteau, Anderson, & Hope, 2006). Care must be taken not to pathologize a culturally appropriate behavior or normalize a behavior that is unacceptable within the person’s culture. Second, heterocentrism, the assumption that heterosexuality is normal or better (Herek, 2009), can impact assessment of social anxiety disorder. Be careful not to assume (p. 162) a client is heterosexual during the assessment process, especially when asking about dating anxiety. Until the client specifically indicates the gender of the person he or she would prefer to date, keep your language neutral or ask specifically. A socially anxious individual who is also a sexual minority may be too unassertive to correct a therapist who assumes that he or she is heterosexual, leading to potential withdrawal from treatment or an awkward secret that likely disrupts the therapeutic process. It should be noted that several of the standard self-report instruments for social anxiety assume heterosexuality. We have ongoing research that recommends more inclusive wording options (Shulman & Hope, 2016; Weiss, Hope & Capozzoli, 2013).