Mark Zimmerman, MD

Borderline personality disorder (BPD) is associated with impaired psychosocial functioning,^1-4 reduced health-related quality of life,⁵ high utilization of services,^6,7 and excess mortality.^8-10 Although BPD occurs in up to 40% of psychiatric inpatients¹¹ and 10% of outpatients,¹² it is underrecognized.^13-15 Often, patients with BPD do not receive an accurate diagnosis until ≥10 years after initially seeking treatment.¹⁶ The treatment and clinical implications of failing to recognize BPD include overprescribing medication and underutilizing empirically effective psychotherapies.¹⁴

This review summarizes studies of the underdiagnosis of BPD in routine clinical practice, describes which patients should be screened, and reviews alternative approaches to screening.

Underrecognition of BPD

The Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project is an ongoing clinical research study involving the integration of research assessment methods into routine clinical practice.¹⁷ In an early report from the MIDAS project, BPD diagnoses derived from structured and unstructured clinical interviews were compared between 2 groups of psychiatric outpatients in the same practice.¹⁵ Individuals in the structured interview cohort were 35 times more often diagnosed with BPD than individuals evaluated with an unstructured clinical interview. Importantly, when the information from the structured interview was presented to the clinicians, BPD was more likely to be diagnosed clinically.

Other studies^13,16 also found that the rate of diagnosing BPD was higher when the diagnosis was based on a semi-structured diagnostic interview compared with an unstructured clinical interview, and that clinicians were reluctant to diagnose BPD during their routine intake diagnostic evaluation.

Clinicians, however, do not use semi-structured interviews in their practice, and they also do not tend to diagnose personality disorders (PDs) based on direct questioning, as they typically would when assessing a symptom-based disorder such as depression or anxiety. Rather, clinicians report that they rely on longitudinal observations to diagnose PDs.¹⁸ However, the results from the MIDAS project were inconsistent with clinicians’ reports. When clinicians were presented with the results of the semi-structured interview, they usually would diagnose BPD, even though it was the initial evaluation. If clinicians actually relied on longitudinal observations and considered information based on the direct question approach of research interviews to be irrelevant or invalid, then the results from the semi-structured interview should not have influenced the rate at which they diagnosed BPD. This suggests that the primary issue in diagnosing PDs is not the need for longi­tudinal observation but rather the need for more information, and that there is a role for screening questionnaires.

One potential criticism of studies demonstrating underrecognition of BPD in clinical practice is that patients typically were interviewed when they presented for treatment, when most were depressed or anxious. The possible pathologizing effects of psychiatric state on personality have been known for years.¹⁹ However, a large body of literature examining the treatment, prognostic, familial, and biological correlates of PDs supports the validity of diagnosing PDs in this manner. Moreover, from a clinical perspective, the sooner a clinician is aware of the presence of BPD, the more likely this information can be used for treatment planning.

Who should be screened for BPD?

BPD is underrecognized and underdiagnosed because patients with BPD often also have comorbid mood, anxiety, or substance use disorders.^20,21 The symptoms associated with these disorders are typically the chief concern of patients with undiagnosed BPD who present for treatment. Patients with BPD rarely present for an intake evaluation and state that they are struggling with abandonment fears, chronic feelings of emptiness, or an identity disturbance. If patients identified these problems as their chief concerns, BPD would be easier to recognize.

Although several studies have documented the frequency of BPD in patients with a specific psychiatric diagnosis such as major depressive disorder (MDD) or attention-deficit/hyperactivity disorder,^22-26 the MIDAS project examined the frequency of BPD in patients with various diagnoses and evaluated which disorders were associated with a significantly increased rate of BPD.²⁷ The highest rate of BPD was found in patients with bipolar disorder. Approximately 25% of patients with bipolar II disorder and one-third of those with bipolar I disorder were diagnosed with BPD; these rates were significantly higher than the rate of BPD in patients without these disorders (Table 1²⁷). The rate of BPD was second highest in patients with a principal diagnosis of posttraumatic stress disorder (PTSD) and MDD; however, the rate of BPD in these patients was not significantly elevated compared with patients who did not have these principal diagnoses. Three disorders were associated with a significantly lower rate of BPD: adjustment disorder, dysthymic disorder, and generalized anxiety disorder.

It would be easy to recommend screening for BPD in all psychiatric patients. However, that is not feasible or practical. In making screening recommendations, absolute risk should be considered more important than relative risk. Clinicians should screen for BPD in patients presenting to a general psychiatric outpatient practice with a principal diagnosis of MDD, bipolar disorder, PTSD, or panic disorder with agoraphobia. That is, I recommend screening for BPD in patients with a principal diagnosis in which the prevalence of BPD is ≥10% (Table 1²⁷).

 

 

A brief review of screening statistics

Screening tests for most psychiatric disorders are based on multi-item scales in which a total score is computed from a sum of item scores, and a cutoff point is established to determine who does and does not screen positive on the test. However, sensitivity, specificity, and positive and negative predictive values are not invariant properties of a screening test with a continuous score distribution. Rather, the performance statistics of a scale can be altered by changing the threshold score to distinguish cases from non-cases. When the screening threshold is lowered, sensitivity increases and specificity decreases.

For screening, a broad net needs to be cast so that all (or almost all) cases are included. Therefore, the cutoff score should be set low to prioritize the sensitivity of the instrument. A screening scale also should have high negative predictive value so that the clinician can be confident that patients who screen negative on the test do not have the disorder.

Screening questionnaires for BPD

Several questionnaires have been developed to screen for PDs (Table 2^28-35). Some screen for each of the DSM PDs,^28,36-42 and some screen more broadly for the presence or absence of any PD.^29,43,44 The most commonly studied self-report scale for BPD is the McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD),³⁰ a 10-item self-report scale derived from a subset of questions from the BPD module of a semi-structured diagnostic interview.

The initial validation study³⁰ found that the optimal cutoff score was 7, which resulted in a sensitivity of 81% and specificity of 89%. Three studies have evaluated the scale in adolescents and young adults,^45-47 and 3 studies examined the scale in adult outpatients.^48-50 Across all 6 studies, at the optimal cutoff scores determined in each study, the sensitivity of the MSI-BPD ranged from 68% to 94% (mean, 80%) and the specificity ranged from 66% to 80% (mean, 72%).

Problems with screening questionnaires. Although screening scales have been developed for many psychiatric disorders, they have not been widely used in mental health settings. In a previous commentary, I argued that the conceptual justification for using self-report screening scales for single disorders in psychiatric settings is questionable.⁵¹ Another problem with screening scales is their potential misuse as case-finding instruments. In the literature on bipolar disorder screening, several researchers misconstrued a positive screen to indicate caseness.⁵¹ Although this is not a problem with the screening measures or the selection of a cutoff score, caution must be taken to not confuse screening with diagnosis.⁵²

 

 

Screening for BPD as part of your diagnostic interview

An alternative approach to using self-administered questionnaires for screening is for clinicians to include questions in their evaluation as part of a psychiatric review of systems. When conducting a diagnostic interview, clinicians typically screen for disorders that are comorbid with the principal diagnosis by asking about the comorbid disorders’ necessary features or “gate criteria.” For example, in a patient with a principal diagnosis of MDD, the clinician would inquire about the presence of panic attacks, excessive worry, or substance use to screen for the presence of panic disorder, generalized anxiety disorder, or a substance use disorder. In contrast, for polythetically defined disorders such as BPD, there is no single gate criterion, because the disorder is diagnosed based on the presence of at least 5 of 9 criteria and no single one of these criteria is required to be present to establish the diagnosis.

As part of the MIDAS project, the psycho­metric properties of the BPD criteria were examined to determine if it was possible to identify 1 or 2 criteria that could serve as gate criteria to screen for the disorder. If the sensitivity of 1 criterion or a combination of 2 BPD criteria was sufficiently high (ie, >90%), then the assessment of this criterion (or these criteria) could be included in a psychiatric review of systems, thus potentially improving the detection of BPD. Researchers hypothesized that affective instability, considered first by Linehan⁵³ and later by other theorists⁵⁴ to be of central importance to the clinical manifestations of BPD, could function as a gate criterion. In the sample of 3,674 psychiatric outpatients who were evaluated with a semi-structured interview, the sensitivity of the affective instability criterion was 92.8%, and the negative predictive value of the criterion was 99%.

These results from the MIDAS project were consistent with the results of other, smaller studies that found that >90% of patients with BPD report affective instability, and it was the most frequent BPD criterion.^55-62 The largest of these studies, the multisite Collaborative Longitudinal Investigation of Personality Study (CLPS), found that sensitivity of affective instability was 94%, which was higher than the sensitivity of the other BPD criteria.⁶² Moreover, the CLPS examined the sensitivity of the BPD criteria assessed at baseline in relation to a diagnosis of BPD that was made 2 years later.⁶³ Affective instability had a 90% sensitivity and 95% negative predictive value in predicting a future diagnosis of BPD. Both of these figures were the highest of the BPD criteria. Other studies have found a negative predictive value >95%.^55,58-61 Therefore,a clinician can be highly confident in ruling out a diagnosis of BPD in patients who do not report affective instability. Table 3 lists questions used to assess affective instability in semi-structured interviews.

Identifying a single BPD criterion that is present in the vast majority of patients diagnosed with BPD will allow clinicians to follow their usual clinical practice when conducting a psychiatric review of systems and inquire about the gate criteria of various disorders. Several studies have found that >90% of patients with BPD report affective instability. However, this does not mean that the diagnosis of BPD can be abbreviated to an assessment of the presence or absence of affective instability. Many patients who screen positive will not have BPD when a more definitive diagnostic evaluation is conducted. In the case of BPD, the more costly definitive diagnostic procedure simply entails inquiry of the other diagnostic criteria.

Borderline personality disorder (BPD) is associated with impaired psychosocial functioning,^1-4 reduced health-related quality of life,⁵ high utilization of services,^6,7 and excess mortality.^8-10 Although BPD occurs in up to 40% of psychiatric inpatients¹¹ and 10% of outpatients,¹² it is underrecognized.^13-15 Often, patients with BPD do not receive an accurate diagnosis until ≥10 years after initially seeking treatment.¹⁶ The treatment and clinical implications of failing to recognize BPD include overprescribing medication and underutilizing empirically effective psychotherapies.¹⁴

This review summarizes studies of the underdiagnosis of BPD in routine clinical practice, describes which patients should be screened, and reviews alternative approaches to screening.

Underrecognition of BPD

The Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project is an ongoing clinical research study involving the integration of research assessment methods into routine clinical practice.¹⁷ In an early report from the MIDAS project, BPD diagnoses derived from structured and unstructured clinical interviews were compared between 2 groups of psychiatric outpatients in the same practice.¹⁵ Individuals in the structured interview cohort were 35 times more often diagnosed with BPD than individuals evaluated with an unstructured clinical interview. Importantly, when the information from the structured interview was presented to the clinicians, BPD was more likely to be diagnosed clinically.

Other studies^13,16 also found that the rate of diagnosing BPD was higher when the diagnosis was based on a semi-structured diagnostic interview compared with an unstructured clinical interview, and that clinicians were reluctant to diagnose BPD during their routine intake diagnostic evaluation.

Clinicians, however, do not use semi-structured interviews in their practice, and they also do not tend to diagnose personality disorders (PDs) based on direct questioning, as they typically would when assessing a symptom-based disorder such as depression or anxiety. Rather, clinicians report that they rely on longitudinal observations to diagnose PDs.¹⁸ However, the results from the MIDAS project were inconsistent with clinicians’ reports. When clinicians were presented with the results of the semi-structured interview, they usually would diagnose BPD, even though it was the initial evaluation. If clinicians actually relied on longitudinal observations and considered information based on the direct question approach of research interviews to be irrelevant or invalid, then the results from the semi-structured interview should not have influenced the rate at which they diagnosed BPD. This suggests that the primary issue in diagnosing PDs is not the need for longi­tudinal observation but rather the need for more information, and that there is a role for screening questionnaires.

One potential criticism of studies demonstrating underrecognition of BPD in clinical practice is that patients typically were interviewed when they presented for treatment, when most were depressed or anxious. The possible pathologizing effects of psychiatric state on personality have been known for years.¹⁹ However, a large body of literature examining the treatment, prognostic, familial, and biological correlates of PDs supports the validity of diagnosing PDs in this manner. Moreover, from a clinical perspective, the sooner a clinician is aware of the presence of BPD, the more likely this information can be used for treatment planning.

Who should be screened for BPD?

BPD is underrecognized and underdiagnosed because patients with BPD often also have comorbid mood, anxiety, or substance use disorders.^20,21 The symptoms associated with these disorders are typically the chief concern of patients with undiagnosed BPD who present for treatment. Patients with BPD rarely present for an intake evaluation and state that they are struggling with abandonment fears, chronic feelings of emptiness, or an identity disturbance. If patients identified these problems as their chief concerns, BPD would be easier to recognize.

Although several studies have documented the frequency of BPD in patients with a specific psychiatric diagnosis such as major depressive disorder (MDD) or attention-deficit/hyperactivity disorder,^22-26 the MIDAS project examined the frequency of BPD in patients with various diagnoses and evaluated which disorders were associated with a significantly increased rate of BPD.²⁷ The highest rate of BPD was found in patients with bipolar disorder. Approximately 25% of patients with bipolar II disorder and one-third of those with bipolar I disorder were diagnosed with BPD; these rates were significantly higher than the rate of BPD in patients without these disorders (Table 1²⁷). The rate of BPD was second highest in patients with a principal diagnosis of posttraumatic stress disorder (PTSD) and MDD; however, the rate of BPD in these patients was not significantly elevated compared with patients who did not have these principal diagnoses. Three disorders were associated with a significantly lower rate of BPD: adjustment disorder, dysthymic disorder, and generalized anxiety disorder.

It would be easy to recommend screening for BPD in all psychiatric patients. However, that is not feasible or practical. In making screening recommendations, absolute risk should be considered more important than relative risk. Clinicians should screen for BPD in patients presenting to a general psychiatric outpatient practice with a principal diagnosis of MDD, bipolar disorder, PTSD, or panic disorder with agoraphobia. That is, I recommend screening for BPD in patients with a principal diagnosis in which the prevalence of BPD is ≥10% (Table 1²⁷).

 

 

A brief review of screening statistics

Screening tests for most psychiatric disorders are based on multi-item scales in which a total score is computed from a sum of item scores, and a cutoff point is established to determine who does and does not screen positive on the test. However, sensitivity, specificity, and positive and negative predictive values are not invariant properties of a screening test with a continuous score distribution. Rather, the performance statistics of a scale can be altered by changing the threshold score to distinguish cases from non-cases. When the screening threshold is lowered, sensitivity increases and specificity decreases.

For screening, a broad net needs to be cast so that all (or almost all) cases are included. Therefore, the cutoff score should be set low to prioritize the sensitivity of the instrument. A screening scale also should have high negative predictive value so that the clinician can be confident that patients who screen negative on the test do not have the disorder.

Screening questionnaires for BPD

Several questionnaires have been developed to screen for PDs (Table 2^28-35). Some screen for each of the DSM PDs,^28,36-42 and some screen more broadly for the presence or absence of any PD.^29,43,44 The most commonly studied self-report scale for BPD is the McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD),³⁰ a 10-item self-report scale derived from a subset of questions from the BPD module of a semi-structured diagnostic interview.

The initial validation study³⁰ found that the optimal cutoff score was 7, which resulted in a sensitivity of 81% and specificity of 89%. Three studies have evaluated the scale in adolescents and young adults,^45-47 and 3 studies examined the scale in adult outpatients.^48-50 Across all 6 studies, at the optimal cutoff scores determined in each study, the sensitivity of the MSI-BPD ranged from 68% to 94% (mean, 80%) and the specificity ranged from 66% to 80% (mean, 72%).

Problems with screening questionnaires. Although screening scales have been developed for many psychiatric disorders, they have not been widely used in mental health settings. In a previous commentary, I argued that the conceptual justification for using self-report screening scales for single disorders in psychiatric settings is questionable.⁵¹ Another problem with screening scales is their potential misuse as case-finding instruments. In the literature on bipolar disorder screening, several researchers misconstrued a positive screen to indicate caseness.⁵¹ Although this is not a problem with the screening measures or the selection of a cutoff score, caution must be taken to not confuse screening with diagnosis.⁵²

 

 

Screening for BPD as part of your diagnostic interview

An alternative approach to using self-administered questionnaires for screening is for clinicians to include questions in their evaluation as part of a psychiatric review of systems. When conducting a diagnostic interview, clinicians typically screen for disorders that are comorbid with the principal diagnosis by asking about the comorbid disorders’ necessary features or “gate criteria.” For example, in a patient with a principal diagnosis of MDD, the clinician would inquire about the presence of panic attacks, excessive worry, or substance use to screen for the presence of panic disorder, generalized anxiety disorder, or a substance use disorder. In contrast, for polythetically defined disorders such as BPD, there is no single gate criterion, because the disorder is diagnosed based on the presence of at least 5 of 9 criteria and no single one of these criteria is required to be present to establish the diagnosis.

As part of the MIDAS project, the psycho­metric properties of the BPD criteria were examined to determine if it was possible to identify 1 or 2 criteria that could serve as gate criteria to screen for the disorder. If the sensitivity of 1 criterion or a combination of 2 BPD criteria was sufficiently high (ie, >90%), then the assessment of this criterion (or these criteria) could be included in a psychiatric review of systems, thus potentially improving the detection of BPD. Researchers hypothesized that affective instability, considered first by Linehan⁵³ and later by other theorists⁵⁴ to be of central importance to the clinical manifestations of BPD, could function as a gate criterion. In the sample of 3,674 psychiatric outpatients who were evaluated with a semi-structured interview, the sensitivity of the affective instability criterion was 92.8%, and the negative predictive value of the criterion was 99%.

These results from the MIDAS project were consistent with the results of other, smaller studies that found that >90% of patients with BPD report affective instability, and it was the most frequent BPD criterion.^55-62 The largest of these studies, the multisite Collaborative Longitudinal Investigation of Personality Study (CLPS), found that sensitivity of affective instability was 94%, which was higher than the sensitivity of the other BPD criteria.⁶² Moreover, the CLPS examined the sensitivity of the BPD criteria assessed at baseline in relation to a diagnosis of BPD that was made 2 years later.⁶³ Affective instability had a 90% sensitivity and 95% negative predictive value in predicting a future diagnosis of BPD. Both of these figures were the highest of the BPD criteria. Other studies have found a negative predictive value >95%.^55,58-61 Therefore,a clinician can be highly confident in ruling out a diagnosis of BPD in patients who do not report affective instability. Table 3 lists questions used to assess affective instability in semi-structured interviews.

Identifying a single BPD criterion that is present in the vast majority of patients diagnosed with BPD will allow clinicians to follow their usual clinical practice when conducting a psychiatric review of systems and inquire about the gate criteria of various disorders. Several studies have found that >90% of patients with BPD report affective instability. However, this does not mean that the diagnosis of BPD can be abbreviated to an assessment of the presence or absence of affective instability. Many patients who screen positive will not have BPD when a more definitive diagnostic evaluation is conducted. In the case of BPD, the more costly definitive diagnostic procedure simply entails inquiry of the other diagnostic criteria.

Borderline personality disorder (BPD) is associated with impaired psychosocial functioning,^1-4 reduced health-related quality of life,⁵ high utilization of services,^6,7 and excess mortality.^8-10 Although BPD occurs in up to 40% of psychiatric inpatients¹¹ and 10% of outpatients,¹² it is underrecognized.^13-15 Often, patients with BPD do not receive an accurate diagnosis until ≥10 years after initially seeking treatment.¹⁶ The treatment and clinical implications of failing to recognize BPD include overprescribing medication and underutilizing empirically effective psychotherapies.¹⁴

This review summarizes studies of the underdiagnosis of BPD in routine clinical practice, describes which patients should be screened, and reviews alternative approaches to screening.

Underrecognition of BPD

The Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project is an ongoing clinical research study involving the integration of research assessment methods into routine clinical practice.¹⁷ In an early report from the MIDAS project, BPD diagnoses derived from structured and unstructured clinical interviews were compared between 2 groups of psychiatric outpatients in the same practice.¹⁵ Individuals in the structured interview cohort were 35 times more often diagnosed with BPD than individuals evaluated with an unstructured clinical interview. Importantly, when the information from the structured interview was presented to the clinicians, BPD was more likely to be diagnosed clinically.

Other studies^13,16 also found that the rate of diagnosing BPD was higher when the diagnosis was based on a semi-structured diagnostic interview compared with an unstructured clinical interview, and that clinicians were reluctant to diagnose BPD during their routine intake diagnostic evaluation.

Clinicians, however, do not use semi-structured interviews in their practice, and they also do not tend to diagnose personality disorders (PDs) based on direct questioning, as they typically would when assessing a symptom-based disorder such as depression or anxiety. Rather, clinicians report that they rely on longitudinal observations to diagnose PDs.¹⁸ However, the results from the MIDAS project were inconsistent with clinicians’ reports. When clinicians were presented with the results of the semi-structured interview, they usually would diagnose BPD, even though it was the initial evaluation. If clinicians actually relied on longitudinal observations and considered information based on the direct question approach of research interviews to be irrelevant or invalid, then the results from the semi-structured interview should not have influenced the rate at which they diagnosed BPD. This suggests that the primary issue in diagnosing PDs is not the need for longi­tudinal observation but rather the need for more information, and that there is a role for screening questionnaires.

One potential criticism of studies demonstrating underrecognition of BPD in clinical practice is that patients typically were interviewed when they presented for treatment, when most were depressed or anxious. The possible pathologizing effects of psychiatric state on personality have been known for years.¹⁹ However, a large body of literature examining the treatment, prognostic, familial, and biological correlates of PDs supports the validity of diagnosing PDs in this manner. Moreover, from a clinical perspective, the sooner a clinician is aware of the presence of BPD, the more likely this information can be used for treatment planning.

Who should be screened for BPD?

BPD is underrecognized and underdiagnosed because patients with BPD often also have comorbid mood, anxiety, or substance use disorders.^20,21 The symptoms associated with these disorders are typically the chief concern of patients with undiagnosed BPD who present for treatment. Patients with BPD rarely present for an intake evaluation and state that they are struggling with abandonment fears, chronic feelings of emptiness, or an identity disturbance. If patients identified these problems as their chief concerns, BPD would be easier to recognize.

Although several studies have documented the frequency of BPD in patients with a specific psychiatric diagnosis such as major depressive disorder (MDD) or attention-deficit/hyperactivity disorder,^22-26 the MIDAS project examined the frequency of BPD in patients with various diagnoses and evaluated which disorders were associated with a significantly increased rate of BPD.²⁷ The highest rate of BPD was found in patients with bipolar disorder. Approximately 25% of patients with bipolar II disorder and one-third of those with bipolar I disorder were diagnosed with BPD; these rates were significantly higher than the rate of BPD in patients without these disorders (Table 1²⁷). The rate of BPD was second highest in patients with a principal diagnosis of posttraumatic stress disorder (PTSD) and MDD; however, the rate of BPD in these patients was not significantly elevated compared with patients who did not have these principal diagnoses. Three disorders were associated with a significantly lower rate of BPD: adjustment disorder, dysthymic disorder, and generalized anxiety disorder.

It would be easy to recommend screening for BPD in all psychiatric patients. However, that is not feasible or practical. In making screening recommendations, absolute risk should be considered more important than relative risk. Clinicians should screen for BPD in patients presenting to a general psychiatric outpatient practice with a principal diagnosis of MDD, bipolar disorder, PTSD, or panic disorder with agoraphobia. That is, I recommend screening for BPD in patients with a principal diagnosis in which the prevalence of BPD is ≥10% (Table 1²⁷).

 

 

A brief review of screening statistics

Screening tests for most psychiatric disorders are based on multi-item scales in which a total score is computed from a sum of item scores, and a cutoff point is established to determine who does and does not screen positive on the test. However, sensitivity, specificity, and positive and negative predictive values are not invariant properties of a screening test with a continuous score distribution. Rather, the performance statistics of a scale can be altered by changing the threshold score to distinguish cases from non-cases. When the screening threshold is lowered, sensitivity increases and specificity decreases.

For screening, a broad net needs to be cast so that all (or almost all) cases are included. Therefore, the cutoff score should be set low to prioritize the sensitivity of the instrument. A screening scale also should have high negative predictive value so that the clinician can be confident that patients who screen negative on the test do not have the disorder.

Screening questionnaires for BPD

Several questionnaires have been developed to screen for PDs (Table 2^28-35). Some screen for each of the DSM PDs,^28,36-42 and some screen more broadly for the presence or absence of any PD.^29,43,44 The most commonly studied self-report scale for BPD is the McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD),³⁰ a 10-item self-report scale derived from a subset of questions from the BPD module of a semi-structured diagnostic interview.

The initial validation study³⁰ found that the optimal cutoff score was 7, which resulted in a sensitivity of 81% and specificity of 89%. Three studies have evaluated the scale in adolescents and young adults,^45-47 and 3 studies examined the scale in adult outpatients.^48-50 Across all 6 studies, at the optimal cutoff scores determined in each study, the sensitivity of the MSI-BPD ranged from 68% to 94% (mean, 80%) and the specificity ranged from 66% to 80% (mean, 72%).

Problems with screening questionnaires. Although screening scales have been developed for many psychiatric disorders, they have not been widely used in mental health settings. In a previous commentary, I argued that the conceptual justification for using self-report screening scales for single disorders in psychiatric settings is questionable.⁵¹ Another problem with screening scales is their potential misuse as case-finding instruments. In the literature on bipolar disorder screening, several researchers misconstrued a positive screen to indicate caseness.⁵¹ Although this is not a problem with the screening measures or the selection of a cutoff score, caution must be taken to not confuse screening with diagnosis.⁵²

 

 

Screening for BPD as part of your diagnostic interview

An alternative approach to using self-administered questionnaires for screening is for clinicians to include questions in their evaluation as part of a psychiatric review of systems. When conducting a diagnostic interview, clinicians typically screen for disorders that are comorbid with the principal diagnosis by asking about the comorbid disorders’ necessary features or “gate criteria.” For example, in a patient with a principal diagnosis of MDD, the clinician would inquire about the presence of panic attacks, excessive worry, or substance use to screen for the presence of panic disorder, generalized anxiety disorder, or a substance use disorder. In contrast, for polythetically defined disorders such as BPD, there is no single gate criterion, because the disorder is diagnosed based on the presence of at least 5 of 9 criteria and no single one of these criteria is required to be present to establish the diagnosis.

As part of the MIDAS project, the psycho­metric properties of the BPD criteria were examined to determine if it was possible to identify 1 or 2 criteria that could serve as gate criteria to screen for the disorder. If the sensitivity of 1 criterion or a combination of 2 BPD criteria was sufficiently high (ie, >90%), then the assessment of this criterion (or these criteria) could be included in a psychiatric review of systems, thus potentially improving the detection of BPD. Researchers hypothesized that affective instability, considered first by Linehan⁵³ and later by other theorists⁵⁴ to be of central importance to the clinical manifestations of BPD, could function as a gate criterion. In the sample of 3,674 psychiatric outpatients who were evaluated with a semi-structured interview, the sensitivity of the affective instability criterion was 92.8%, and the negative predictive value of the criterion was 99%.

These results from the MIDAS project were consistent with the results of other, smaller studies that found that >90% of patients with BPD report affective instability, and it was the most frequent BPD criterion.^55-62 The largest of these studies, the multisite Collaborative Longitudinal Investigation of Personality Study (CLPS), found that sensitivity of affective instability was 94%, which was higher than the sensitivity of the other BPD criteria.⁶² Moreover, the CLPS examined the sensitivity of the BPD criteria assessed at baseline in relation to a diagnosis of BPD that was made 2 years later.⁶³ Affective instability had a 90% sensitivity and 95% negative predictive value in predicting a future diagnosis of BPD. Both of these figures were the highest of the BPD criteria. Other studies have found a negative predictive value >95%.^55,58-61 Therefore,a clinician can be highly confident in ruling out a diagnosis of BPD in patients who do not report affective instability. Table 3 lists questions used to assess affective instability in semi-structured interviews.

Identifying a single BPD criterion that is present in the vast majority of patients diagnosed with BPD will allow clinicians to follow their usual clinical practice when conducting a psychiatric review of systems and inquire about the gate criteria of various disorders. Several studies have found that >90% of patients with BPD report affective instability. However, this does not mean that the diagnosis of BPD can be abbreviated to an assessment of the presence or absence of affective instability. Many patients who screen positive will not have BPD when a more definitive diagnostic evaluation is conducted. In the case of BPD, the more costly definitive diagnostic procedure simply entails inquiry of the other diagnostic criteria.

Since the appearance of social anxiety disorder (SAD) in the DSM-III in 1980, research on its prevalence, characteristics, and treatment have grown (Box 1^1,2). In addition to the name, the definition of SAD has changed over the years; as a result, its prevalence has increased in recent cohort studies. This has led to debate over whether the experience of shyness is being over-pathologized, or whether SAD has been underdiagnosed in earlier decades. Those who argue that shyness is being over-pathologized note that it is a normal human experience that has evolutionary functions (eg, preventing engagement in harmful social relationships³). Others argue that a high degree of shyness is not beneficial in terms of evolution because it causes the individual to be shunned, so to speak, by society.⁴

Why worry about ‘over-pathologizing’?

The medicalization of shyness might be a reflection of Western societal values of assertiveness and gregariousness; other societies that value modesty and reticence do not over-pathologize shyness.⁵ It is important not to assume that someone who is shy necessarily has a “pathologic” level of social anxiety, especially because some people who are shy view that condition as a positive quality, much like sensitivity and conscientiousness.⁵

The broader issue of what constitutes a mental disorder arises in this debate. A “disorder” is a socially constructed label that describes a set of symptoms occurring together and its associated behaviors, not a real entity with etiological homogeneity.⁶ Labeling emotional problems “disordered” assumes that happiness is the natural homeostatic state, and distressing emotional states are abnormal and need to be changed.⁷ A diagnostic label can help improve communication and understand maladaptive behaviors; if that label is reified, however, it can lead to assumptions that the etiology, course, and treatment response are known. Proponents of the diagnostic psychiatric nomenclature have acknowledged the dangers of over-pathologizing normal experiences of living (such as fear) by way of diagnostic labeling.⁸

Determining when shyness becomes a clinically significant problem—what we call SAD here—demands a delicate distinction that has important implications for treatment. On one hand, if shyness is over-pathologized, persons who neither desire nor need treatment might be subjected to unnecessary and costly intervention. On the other hand, if SAD is underdiagnosed, some persons will not receive treatment that might be beneficial to them.

In this article, we review the similarities and differences between shyness and SAD, and provide recommendations for determining when shyness becomes a more clinically significant problem. We also highlight the importance of this distinction as it pertains to management, and provide suggestions for treatment approaches.

SAD: Definition, prevalence

SAD is defined as a significant fear of embarrassment or humiliation in social or performance-based situations, to a point at which the affected person often avoids these situations or endures them only with a high level of distress⁹ (Table 1, and Box 2). SAD can be distinguished from other anxiety disorders based on the source and content of the fear (ie, the source being social interaction or performance situations, and the content being a fear that one will show a behavior that will cause embarrassment). SAD also should be distinguished from autism spectrum disorders, in which persons have limited social communication capabilities and inadequate age-appropriate social relationships.

SAD is most highly comorbid with mood and anxiety disorders, with rates of at least 30% in clinical samples.¹⁰ The disorder also is highly comorbid with avoidant personality disorder—to a point at which it is argued that they are one and the same disorder.¹¹ 

As with other psychiatric disorders, anxiety must cause significant impairment or distress. What constitutes significant impairment or distress is subjective, and the arbitrary nature of this criterion can influence estimates of the prevalence of SAD. For example, prevalence ranges as widely as 1.9% to 20.4% when different cut-offs are used for distress ratings and the number of impaired domains.¹²

The prevalence of SAD varies from 1 epidemiological study to another (ie, the Epidemiological Catchment Area [ECA] Study and the National Comorbidity Survey [NCS])—in part, a consequence of the differing definitions of significant impairment or distress. The ECA study assessed the clinical significance of each symptom in anxiety disorders; the NCS assessed overall clinical significance of the disorder. When the clinical significance criterion was applied at the symptom level to the NCS dataset (as was done in the ECA study), 1-year prevalence decreased by 50% (from 7.4% to 3.7%).¹³ The manner in which significant impairment or distress is defined (ie, conservatively or liberally) impacts whether social anxiety symptoms are classified as disordered or non-disordered.   

 

Shyness: Definition, prevalence

Shyness often refers to 1) anxiety, inhibition, reticence, or a combination of these findings, in social and interpersonal situations, and 2) a fear of negative evaluation by others.¹⁴ It is a normal facet of personality that combines the experience of social anxiety and inhibited behavior,¹⁵ and also has been described as a stable temperament.¹⁶ Shyness is common; in the NCS study,¹⁷ 26% of women and 19% of men characterized themselves as “very shy”; in the NCS Adolescent study,¹⁸ nearly 50% of adolescents self-identified as shy. 

Persons who are shy tend to self-report greater social anxiety and embarrassment in social situations than non-shy persons do; they also might experience greater autonomic reactivity—especially blushing—in social or performance situations.¹⁵ Furthermore, shy persons are more likely to have axis I comorbidity and traits of introversion and neuroticism, compared with non-shy persons.¹⁴

Research suggests that temperament and behavioral inhibition are risk factors for mood and anxiety disorders, and appear to have a particularly strong relationship with SAD.¹⁹ A recent prospective study showed that shyness tends to increase steeply in toddlerhood, then stabilizes in childhood. Shyness in childhood—but not toddlerhood—is predictive of anxiety, depression, and poorer social skills in adolescence.²⁰

A qualitative, or just quantitative, difference?

It is clear that SAD and shyness share several features—including anxiety and embarrassment—in social interactions. This raises a question: Are SAD and shyness distinct qualitatively, or do they represent points along a continuum, with SAD being an extreme form of shyness?

Continuum hypothesis. Support for the continuum hypothesis includes evidence that SAD and shyness share several features, including autonomic arousal, deficits in social skills (eg, aversion of gaze, difficulty initiating and maintaining conversation), avoidance of social situations, and fear of negative evaluation.^21,22 In addition, both shyness and SAD are highly heritable,²³ and mothers of shy children have a significantly higher rate of SAD than non-shy children do.²⁴ No familial or genetic studies have compared heritability and familial aggregation in shyness and SAD. 

According to the continuum hypothesis, if SAD is an extreme form of shyness, all (or nearly all) persons who have a diagnosis of SAD also would be characterized as shy. However, only approximately one-half of such persons report having been shy in childhood.¹⁷ Less than one-quarter of shy persons meet criteria for SAD.^14,18 Because many persons who are shy do not meet criteria for SAD, and many who have SAD were not considered shy earlier in life, it has been suggested that this supports a qualitative distinction. 

Qualitative distinctiveness. Despite having similarities, several features distinguish the experience of SAD from that of shyness. Compared with shyness, a SAD diagnosis is associated with:

• greater comorbidity
• greater severity of avoidance and impairment
• poorer quality of life.^18,21,25

Studies that compared SAD, shyness without SAD, and non-shyness have shown that the shyness without SAD group more closely resembles the non-shy group than the SAD group—particularly with regard to impairment, presence of substance use, and other behavioral problems.^18,25

Given the evidence, experts have concluded that shyness and a SAD diagnosis are overlapping yet different constructs that encapsulate qualitative and quantitative differences.²⁵ There is a spectrum of shyness that ranges from a normative level to a higher level that overlaps the experience of SAD, but the 2 states represent different constructs.²⁵

Guidance for making an assessment. Because of similarities in anxiety, embarrassment, and other symptoms in social situations, the best way to determine whether shyness crosses the line into a clinically significant problem is to assess the severity of the anxiety and associated degree of impairment and distress. More severe anxiety paired with distress about having anxiety and significant impairment in multiple areas of functioning might indicate more problematic social anxiety—a diagnosis of SAD—not just “normal” shyness. 

It is important to take into account the environmental and cultural context of a patient’s distress and impairment because these features might fall within a normal range, given immediate circumstances (such as speaking in front of a large audience when one is not normally called on to do so, to a degree that does not interfere with general social functioning⁶).

What is considered a normative range depends on the developmental stage:

• Among children, a greater level of shyness might be considered more normative when it manifests during developmental stages in which separation anxiety appears.
• Among adolescents, a greater level of shyness might be considered normative especially during early adolescence (when social relationships become more important), and during times of transition (ie, entering high school).
• In adulthood, a greater level of normative shyness or social anxiety might be present during a major life change (eg, beginning to date again after the loss of a lengthy marriage or romantic relationship).

 

Assessment tools

Assessment tools can help you differentiate normal shyness from SAD. Several empirically-validated rating scales exist, including clinician-rated and self-report scales.

Liebowitz Social Anxiety Scale²⁶ rates the severity of fear and avoidance in a variety of social interaction and performance-based situations. However, it was developed primarily as a clinician-rated scale and might be more burdensome to complete in practice. In addition, it does not provide cut-offs to indicate when more clinically significant anxiety might be likely.

Clinically Useful Social Anxiety Disorder Outcome Scale (CUSADOS)²⁷ and Mini-Social Phobia Inventory (Mini-SPIN)²⁸ are brief self-report scales that provide cut-offs to suggest further assessment is warranted. A cut-off score of 16 on the CUSADOS suggests the presence of SAD with 73% diagnostic efficiency.

One disadvantage to relying on a rating scale alone is the narrow focus on symptoms. Given that shyness and SAD share similar symptoms, it is necessary to assess the degree of impairment related to these symptoms to determine whether the problem is clinically significant. The overly narrow focus on symptoms utilized by the biomedical approach has been criticized for contributing to the medicalization of normal shyness.⁵ 

Diagnostic interviews, such as the Structured Clinical Interview for DSM-IV Axis I Disorders²⁹ include sections on SAD that assess avoidance and impairment/distress associated with anxiety. Because these interviews may increase the time burden during an office visit, there are several general questions outside of a structured interview that you can ask, such as: “Has this anxiety interfered with your ability to initiate or maintain friendships? If so, how?” (Table 2). Persons with clinically significant social anxiety, rather than shyness, tend to report greater effects on their relationships and on work or school performance, as well as greater distress about having that anxiety.   

Treatment approaches based on distinctions

Exercise care in making the distinction between normal shyness and dysfunctional and impairing levels of anxiety characteristic of SAD, because persons who display normal shyness but who are overdiagnosed might feel stigmatized by a diagnostic label.⁵ Also, overpathologizing shyness takes what is a social problem out of context, and could promote treatment strategies that might not be helpful or effective.³⁰

Unnecessary diagnosis might lead to unnecessary treatment, such as prescribing an antidepressant or benzodiazepine. Avoiding such a situation is important, because of the side effects associated with medication and the potential for dependence and withdrawal effects with benzodiazepines.

Persons who exhibit normal shyness do not require medical treatment and, often, do not want it. However, some people may be interested in improving their ability to function in social interactions. Self-help approaches or brief psychotherapy (eg, cognitive-behavioral therapy [CBT]) should be the first step—and might be all that is necessary. 

The opposite side of the problem.  Under-recognition of clinically significant social anxiety can lead to under-treatment, which is common even in patients with a SAD diagnosis.³¹ Treatment options include CBT, medication, and CBT combined with medication (Table 3):

• several studies have demonstrated the short- and long-term efficacy of CBT alone for SAD
• medication alone has been efficacious in the short-term, but less efficacious than CBT in the long-term
• combined treatment also has been shown to be more efficacious than CBT or medication alone in the short-term
• there is evidence to suggest that CBT alone is more efficacious in the long-term compared with combined treatment.^a

CBT is recommended as an appropriate first-line option, especially for mild and moderate SAD; it is the preferred initial treatment option of the United Kingdom’s National Institute for Health and Care Excellence (NICE). For more severe presentations (such as the presence of comorbidity) or when a patient did not respond to an adequate course of CBT, combined treatment might be an option—the goal being to taper the medication and continue CBT as a longer-term treatment. Research has shown that continuing CBT while discontinuing medication helps prevent relapse.^32,33

Appropriate pharmacotherapy options include selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors.³⁴ Increasingly, benzodiazepines are considered less desirable; they are not recommended for routine use in SAD in the NICE guidelines. Those guidelines call for continuing pharmacotherapy for 6 months when a patient responds to treatment within 3 months, then discontinuing medication with the aid of CBT.

Bottom Line

The severity of anxiety and associated impairment and distress are the main variables that differentiate normal shyness and clinically significant social anxiety. Taking care not to over-pathologize normal shyness and common social anxiety concerns or underdiagnose severe, impairing social anxiety disorder has important implications for treatment—and for whether a patient needs treatment at all.

 

Related Resources

National Institute for Health and Care Excellence. Social anxiety disorder: recognition, assessment, and treatment of social anxiety disorder. http://guidance.nice.org.uk/cg159.

• Hofmann SG, DiBartolo PM, eds. Social anxiety: clinical, developmental, and social perspectives, 2nd ed. London, United Kingdom: Academic Press; 2010.

• The Shyness Institute. www.shyness.com.

Drug Brand Names

Alprazolam • Xanax        Clonazepam • Klonopin        Fluoxetine • Prozac

Fluvoxamine • Luvox      Paroxetine • Paxil                Phenelzine • Nardil

Sertraline • Zoloft          Venlafaxine • Effexor 

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Featured Audio
Kristy L. Dalrymple, PhD, discusses, treating social anxiety disorder. Dr. Dalrymple is Staff Psychologist, Department of Psychiatry, Rhode Island Hospital, and Assistant Professor of Psychiatry and Human Behavior, Alpert Medical School of Brown University, Providence, Rhode Island.

Since the appearance of social anxiety disorder (SAD) in the DSM-III in 1980, research on its prevalence, characteristics, and treatment have grown (Box 1^1,2). In addition to the name, the definition of SAD has changed over the years; as a result, its prevalence has increased in recent cohort studies. This has led to debate over whether the experience of shyness is being over-pathologized, or whether SAD has been underdiagnosed in earlier decades. Those who argue that shyness is being over-pathologized note that it is a normal human experience that has evolutionary functions (eg, preventing engagement in harmful social relationships³). Others argue that a high degree of shyness is not beneficial in terms of evolution because it causes the individual to be shunned, so to speak, by society.⁴

Why worry about ‘over-pathologizing’?

The medicalization of shyness might be a reflection of Western societal values of assertiveness and gregariousness; other societies that value modesty and reticence do not over-pathologize shyness.⁵ It is important not to assume that someone who is shy necessarily has a “pathologic” level of social anxiety, especially because some people who are shy view that condition as a positive quality, much like sensitivity and conscientiousness.⁵

The broader issue of what constitutes a mental disorder arises in this debate. A “disorder” is a socially constructed label that describes a set of symptoms occurring together and its associated behaviors, not a real entity with etiological homogeneity.⁶ Labeling emotional problems “disordered” assumes that happiness is the natural homeostatic state, and distressing emotional states are abnormal and need to be changed.⁷ A diagnostic label can help improve communication and understand maladaptive behaviors; if that label is reified, however, it can lead to assumptions that the etiology, course, and treatment response are known. Proponents of the diagnostic psychiatric nomenclature have acknowledged the dangers of over-pathologizing normal experiences of living (such as fear) by way of diagnostic labeling.⁸

Determining when shyness becomes a clinically significant problem—what we call SAD here—demands a delicate distinction that has important implications for treatment. On one hand, if shyness is over-pathologized, persons who neither desire nor need treatment might be subjected to unnecessary and costly intervention. On the other hand, if SAD is underdiagnosed, some persons will not receive treatment that might be beneficial to them.

In this article, we review the similarities and differences between shyness and SAD, and provide recommendations for determining when shyness becomes a more clinically significant problem. We also highlight the importance of this distinction as it pertains to management, and provide suggestions for treatment approaches.

SAD: Definition, prevalence

SAD is defined as a significant fear of embarrassment or humiliation in social or performance-based situations, to a point at which the affected person often avoids these situations or endures them only with a high level of distress⁹ (Table 1, and Box 2). SAD can be distinguished from other anxiety disorders based on the source and content of the fear (ie, the source being social interaction or performance situations, and the content being a fear that one will show a behavior that will cause embarrassment). SAD also should be distinguished from autism spectrum disorders, in which persons have limited social communication capabilities and inadequate age-appropriate social relationships.

SAD is most highly comorbid with mood and anxiety disorders, with rates of at least 30% in clinical samples.¹⁰ The disorder also is highly comorbid with avoidant personality disorder—to a point at which it is argued that they are one and the same disorder.¹¹ 

As with other psychiatric disorders, anxiety must cause significant impairment or distress. What constitutes significant impairment or distress is subjective, and the arbitrary nature of this criterion can influence estimates of the prevalence of SAD. For example, prevalence ranges as widely as 1.9% to 20.4% when different cut-offs are used for distress ratings and the number of impaired domains.¹²

The prevalence of SAD varies from 1 epidemiological study to another (ie, the Epidemiological Catchment Area [ECA] Study and the National Comorbidity Survey [NCS])—in part, a consequence of the differing definitions of significant impairment or distress. The ECA study assessed the clinical significance of each symptom in anxiety disorders; the NCS assessed overall clinical significance of the disorder. When the clinical significance criterion was applied at the symptom level to the NCS dataset (as was done in the ECA study), 1-year prevalence decreased by 50% (from 7.4% to 3.7%).¹³ The manner in which significant impairment or distress is defined (ie, conservatively or liberally) impacts whether social anxiety symptoms are classified as disordered or non-disordered.   

 

Shyness: Definition, prevalence

Shyness often refers to 1) anxiety, inhibition, reticence, or a combination of these findings, in social and interpersonal situations, and 2) a fear of negative evaluation by others.¹⁴ It is a normal facet of personality that combines the experience of social anxiety and inhibited behavior,¹⁵ and also has been described as a stable temperament.¹⁶ Shyness is common; in the NCS study,¹⁷ 26% of women and 19% of men characterized themselves as “very shy”; in the NCS Adolescent study,¹⁸ nearly 50% of adolescents self-identified as shy. 

Persons who are shy tend to self-report greater social anxiety and embarrassment in social situations than non-shy persons do; they also might experience greater autonomic reactivity—especially blushing—in social or performance situations.¹⁵ Furthermore, shy persons are more likely to have axis I comorbidity and traits of introversion and neuroticism, compared with non-shy persons.¹⁴

Research suggests that temperament and behavioral inhibition are risk factors for mood and anxiety disorders, and appear to have a particularly strong relationship with SAD.¹⁹ A recent prospective study showed that shyness tends to increase steeply in toddlerhood, then stabilizes in childhood. Shyness in childhood—but not toddlerhood—is predictive of anxiety, depression, and poorer social skills in adolescence.²⁰

A qualitative, or just quantitative, difference?

It is clear that SAD and shyness share several features—including anxiety and embarrassment—in social interactions. This raises a question: Are SAD and shyness distinct qualitatively, or do they represent points along a continuum, with SAD being an extreme form of shyness?

Continuum hypothesis. Support for the continuum hypothesis includes evidence that SAD and shyness share several features, including autonomic arousal, deficits in social skills (eg, aversion of gaze, difficulty initiating and maintaining conversation), avoidance of social situations, and fear of negative evaluation.^21,22 In addition, both shyness and SAD are highly heritable,²³ and mothers of shy children have a significantly higher rate of SAD than non-shy children do.²⁴ No familial or genetic studies have compared heritability and familial aggregation in shyness and SAD. 

According to the continuum hypothesis, if SAD is an extreme form of shyness, all (or nearly all) persons who have a diagnosis of SAD also would be characterized as shy. However, only approximately one-half of such persons report having been shy in childhood.¹⁷ Less than one-quarter of shy persons meet criteria for SAD.^14,18 Because many persons who are shy do not meet criteria for SAD, and many who have SAD were not considered shy earlier in life, it has been suggested that this supports a qualitative distinction. 

Qualitative distinctiveness. Despite having similarities, several features distinguish the experience of SAD from that of shyness. Compared with shyness, a SAD diagnosis is associated with:

• greater comorbidity
• greater severity of avoidance and impairment
• poorer quality of life.^18,21,25

Studies that compared SAD, shyness without SAD, and non-shyness have shown that the shyness without SAD group more closely resembles the non-shy group than the SAD group—particularly with regard to impairment, presence of substance use, and other behavioral problems.^18,25

Given the evidence, experts have concluded that shyness and a SAD diagnosis are overlapping yet different constructs that encapsulate qualitative and quantitative differences.²⁵ There is a spectrum of shyness that ranges from a normative level to a higher level that overlaps the experience of SAD, but the 2 states represent different constructs.²⁵

Guidance for making an assessment. Because of similarities in anxiety, embarrassment, and other symptoms in social situations, the best way to determine whether shyness crosses the line into a clinically significant problem is to assess the severity of the anxiety and associated degree of impairment and distress. More severe anxiety paired with distress about having anxiety and significant impairment in multiple areas of functioning might indicate more problematic social anxiety—a diagnosis of SAD—not just “normal” shyness. 

It is important to take into account the environmental and cultural context of a patient’s distress and impairment because these features might fall within a normal range, given immediate circumstances (such as speaking in front of a large audience when one is not normally called on to do so, to a degree that does not interfere with general social functioning⁶).

What is considered a normative range depends on the developmental stage:

• Among children, a greater level of shyness might be considered more normative when it manifests during developmental stages in which separation anxiety appears.
• Among adolescents, a greater level of shyness might be considered normative especially during early adolescence (when social relationships become more important), and during times of transition (ie, entering high school).
• In adulthood, a greater level of normative shyness or social anxiety might be present during a major life change (eg, beginning to date again after the loss of a lengthy marriage or romantic relationship).

 

Assessment tools

Assessment tools can help you differentiate normal shyness from SAD. Several empirically-validated rating scales exist, including clinician-rated and self-report scales.

Liebowitz Social Anxiety Scale²⁶ rates the severity of fear and avoidance in a variety of social interaction and performance-based situations. However, it was developed primarily as a clinician-rated scale and might be more burdensome to complete in practice. In addition, it does not provide cut-offs to indicate when more clinically significant anxiety might be likely.

Clinically Useful Social Anxiety Disorder Outcome Scale (CUSADOS)²⁷ and Mini-Social Phobia Inventory (Mini-SPIN)²⁸ are brief self-report scales that provide cut-offs to suggest further assessment is warranted. A cut-off score of 16 on the CUSADOS suggests the presence of SAD with 73% diagnostic efficiency.

One disadvantage to relying on a rating scale alone is the narrow focus on symptoms. Given that shyness and SAD share similar symptoms, it is necessary to assess the degree of impairment related to these symptoms to determine whether the problem is clinically significant. The overly narrow focus on symptoms utilized by the biomedical approach has been criticized for contributing to the medicalization of normal shyness.⁵ 

Diagnostic interviews, such as the Structured Clinical Interview for DSM-IV Axis I Disorders²⁹ include sections on SAD that assess avoidance and impairment/distress associated with anxiety. Because these interviews may increase the time burden during an office visit, there are several general questions outside of a structured interview that you can ask, such as: “Has this anxiety interfered with your ability to initiate or maintain friendships? If so, how?” (Table 2). Persons with clinically significant social anxiety, rather than shyness, tend to report greater effects on their relationships and on work or school performance, as well as greater distress about having that anxiety.   

Treatment approaches based on distinctions

Exercise care in making the distinction between normal shyness and dysfunctional and impairing levels of anxiety characteristic of SAD, because persons who display normal shyness but who are overdiagnosed might feel stigmatized by a diagnostic label.⁵ Also, overpathologizing shyness takes what is a social problem out of context, and could promote treatment strategies that might not be helpful or effective.³⁰

Unnecessary diagnosis might lead to unnecessary treatment, such as prescribing an antidepressant or benzodiazepine. Avoiding such a situation is important, because of the side effects associated with medication and the potential for dependence and withdrawal effects with benzodiazepines.

Persons who exhibit normal shyness do not require medical treatment and, often, do not want it. However, some people may be interested in improving their ability to function in social interactions. Self-help approaches or brief psychotherapy (eg, cognitive-behavioral therapy [CBT]) should be the first step—and might be all that is necessary. 

The opposite side of the problem.  Under-recognition of clinically significant social anxiety can lead to under-treatment, which is common even in patients with a SAD diagnosis.³¹ Treatment options include CBT, medication, and CBT combined with medication (Table 3):

• several studies have demonstrated the short- and long-term efficacy of CBT alone for SAD
• medication alone has been efficacious in the short-term, but less efficacious than CBT in the long-term
• combined treatment also has been shown to be more efficacious than CBT or medication alone in the short-term
• there is evidence to suggest that CBT alone is more efficacious in the long-term compared with combined treatment.^a

CBT is recommended as an appropriate first-line option, especially for mild and moderate SAD; it is the preferred initial treatment option of the United Kingdom’s National Institute for Health and Care Excellence (NICE). For more severe presentations (such as the presence of comorbidity) or when a patient did not respond to an adequate course of CBT, combined treatment might be an option—the goal being to taper the medication and continue CBT as a longer-term treatment. Research has shown that continuing CBT while discontinuing medication helps prevent relapse.^32,33

Appropriate pharmacotherapy options include selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors.³⁴ Increasingly, benzodiazepines are considered less desirable; they are not recommended for routine use in SAD in the NICE guidelines. Those guidelines call for continuing pharmacotherapy for 6 months when a patient responds to treatment within 3 months, then discontinuing medication with the aid of CBT.

Bottom Line

The severity of anxiety and associated impairment and distress are the main variables that differentiate normal shyness and clinically significant social anxiety. Taking care not to over-pathologize normal shyness and common social anxiety concerns or underdiagnose severe, impairing social anxiety disorder has important implications for treatment—and for whether a patient needs treatment at all.

 

Related Resources

National Institute for Health and Care Excellence. Social anxiety disorder: recognition, assessment, and treatment of social anxiety disorder. http://guidance.nice.org.uk/cg159.

• Hofmann SG, DiBartolo PM, eds. Social anxiety: clinical, developmental, and social perspectives, 2nd ed. London, United Kingdom: Academic Press; 2010.

• The Shyness Institute. www.shyness.com.

Drug Brand Names

Alprazolam • Xanax        Clonazepam • Klonopin        Fluoxetine • Prozac

Fluvoxamine • Luvox      Paroxetine • Paxil                Phenelzine • Nardil

Sertraline • Zoloft          Venlafaxine • Effexor 

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Featured Audio
Kristy L. Dalrymple, PhD, discusses, treating social anxiety disorder. Dr. Dalrymple is Staff Psychologist, Department of Psychiatry, Rhode Island Hospital, and Assistant Professor of Psychiatry and Human Behavior, Alpert Medical School of Brown University, Providence, Rhode Island.

Since the appearance of social anxiety disorder (SAD) in the DSM-III in 1980, research on its prevalence, characteristics, and treatment have grown (Box 1^1,2). In addition to the name, the definition of SAD has changed over the years; as a result, its prevalence has increased in recent cohort studies. This has led to debate over whether the experience of shyness is being over-pathologized, or whether SAD has been underdiagnosed in earlier decades. Those who argue that shyness is being over-pathologized note that it is a normal human experience that has evolutionary functions (eg, preventing engagement in harmful social relationships³). Others argue that a high degree of shyness is not beneficial in terms of evolution because it causes the individual to be shunned, so to speak, by society.⁴

Why worry about ‘over-pathologizing’?

The medicalization of shyness might be a reflection of Western societal values of assertiveness and gregariousness; other societies that value modesty and reticence do not over-pathologize shyness.⁵ It is important not to assume that someone who is shy necessarily has a “pathologic” level of social anxiety, especially because some people who are shy view that condition as a positive quality, much like sensitivity and conscientiousness.⁵

The broader issue of what constitutes a mental disorder arises in this debate. A “disorder” is a socially constructed label that describes a set of symptoms occurring together and its associated behaviors, not a real entity with etiological homogeneity.⁶ Labeling emotional problems “disordered” assumes that happiness is the natural homeostatic state, and distressing emotional states are abnormal and need to be changed.⁷ A diagnostic label can help improve communication and understand maladaptive behaviors; if that label is reified, however, it can lead to assumptions that the etiology, course, and treatment response are known. Proponents of the diagnostic psychiatric nomenclature have acknowledged the dangers of over-pathologizing normal experiences of living (such as fear) by way of diagnostic labeling.⁸

Determining when shyness becomes a clinically significant problem—what we call SAD here—demands a delicate distinction that has important implications for treatment. On one hand, if shyness is over-pathologized, persons who neither desire nor need treatment might be subjected to unnecessary and costly intervention. On the other hand, if SAD is underdiagnosed, some persons will not receive treatment that might be beneficial to them.

In this article, we review the similarities and differences between shyness and SAD, and provide recommendations for determining when shyness becomes a more clinically significant problem. We also highlight the importance of this distinction as it pertains to management, and provide suggestions for treatment approaches.

SAD: Definition, prevalence

SAD is defined as a significant fear of embarrassment or humiliation in social or performance-based situations, to a point at which the affected person often avoids these situations or endures them only with a high level of distress⁹ (Table 1, and Box 2). SAD can be distinguished from other anxiety disorders based on the source and content of the fear (ie, the source being social interaction or performance situations, and the content being a fear that one will show a behavior that will cause embarrassment). SAD also should be distinguished from autism spectrum disorders, in which persons have limited social communication capabilities and inadequate age-appropriate social relationships.

SAD is most highly comorbid with mood and anxiety disorders, with rates of at least 30% in clinical samples.¹⁰ The disorder also is highly comorbid with avoidant personality disorder—to a point at which it is argued that they are one and the same disorder.¹¹ 

As with other psychiatric disorders, anxiety must cause significant impairment or distress. What constitutes significant impairment or distress is subjective, and the arbitrary nature of this criterion can influence estimates of the prevalence of SAD. For example, prevalence ranges as widely as 1.9% to 20.4% when different cut-offs are used for distress ratings and the number of impaired domains.¹²

The prevalence of SAD varies from 1 epidemiological study to another (ie, the Epidemiological Catchment Area [ECA] Study and the National Comorbidity Survey [NCS])—in part, a consequence of the differing definitions of significant impairment or distress. The ECA study assessed the clinical significance of each symptom in anxiety disorders; the NCS assessed overall clinical significance of the disorder. When the clinical significance criterion was applied at the symptom level to the NCS dataset (as was done in the ECA study), 1-year prevalence decreased by 50% (from 7.4% to 3.7%).¹³ The manner in which significant impairment or distress is defined (ie, conservatively or liberally) impacts whether social anxiety symptoms are classified as disordered or non-disordered.   

 

Shyness: Definition, prevalence

Shyness often refers to 1) anxiety, inhibition, reticence, or a combination of these findings, in social and interpersonal situations, and 2) a fear of negative evaluation by others.¹⁴ It is a normal facet of personality that combines the experience of social anxiety and inhibited behavior,¹⁵ and also has been described as a stable temperament.¹⁶ Shyness is common; in the NCS study,¹⁷ 26% of women and 19% of men characterized themselves as “very shy”; in the NCS Adolescent study,¹⁸ nearly 50% of adolescents self-identified as shy. 

Persons who are shy tend to self-report greater social anxiety and embarrassment in social situations than non-shy persons do; they also might experience greater autonomic reactivity—especially blushing—in social or performance situations.¹⁵ Furthermore, shy persons are more likely to have axis I comorbidity and traits of introversion and neuroticism, compared with non-shy persons.¹⁴

Research suggests that temperament and behavioral inhibition are risk factors for mood and anxiety disorders, and appear to have a particularly strong relationship with SAD.¹⁹ A recent prospective study showed that shyness tends to increase steeply in toddlerhood, then stabilizes in childhood. Shyness in childhood—but not toddlerhood—is predictive of anxiety, depression, and poorer social skills in adolescence.²⁰

A qualitative, or just quantitative, difference?

It is clear that SAD and shyness share several features—including anxiety and embarrassment—in social interactions. This raises a question: Are SAD and shyness distinct qualitatively, or do they represent points along a continuum, with SAD being an extreme form of shyness?

Continuum hypothesis. Support for the continuum hypothesis includes evidence that SAD and shyness share several features, including autonomic arousal, deficits in social skills (eg, aversion of gaze, difficulty initiating and maintaining conversation), avoidance of social situations, and fear of negative evaluation.^21,22 In addition, both shyness and SAD are highly heritable,²³ and mothers of shy children have a significantly higher rate of SAD than non-shy children do.²⁴ No familial or genetic studies have compared heritability and familial aggregation in shyness and SAD. 

According to the continuum hypothesis, if SAD is an extreme form of shyness, all (or nearly all) persons who have a diagnosis of SAD also would be characterized as shy. However, only approximately one-half of such persons report having been shy in childhood.¹⁷ Less than one-quarter of shy persons meet criteria for SAD.^14,18 Because many persons who are shy do not meet criteria for SAD, and many who have SAD were not considered shy earlier in life, it has been suggested that this supports a qualitative distinction. 

Qualitative distinctiveness. Despite having similarities, several features distinguish the experience of SAD from that of shyness. Compared with shyness, a SAD diagnosis is associated with:

• greater comorbidity
• greater severity of avoidance and impairment
• poorer quality of life.^18,21,25

Studies that compared SAD, shyness without SAD, and non-shyness have shown that the shyness without SAD group more closely resembles the non-shy group than the SAD group—particularly with regard to impairment, presence of substance use, and other behavioral problems.^18,25

Given the evidence, experts have concluded that shyness and a SAD diagnosis are overlapping yet different constructs that encapsulate qualitative and quantitative differences.²⁵ There is a spectrum of shyness that ranges from a normative level to a higher level that overlaps the experience of SAD, but the 2 states represent different constructs.²⁵

Guidance for making an assessment. Because of similarities in anxiety, embarrassment, and other symptoms in social situations, the best way to determine whether shyness crosses the line into a clinically significant problem is to assess the severity of the anxiety and associated degree of impairment and distress. More severe anxiety paired with distress about having anxiety and significant impairment in multiple areas of functioning might indicate more problematic social anxiety—a diagnosis of SAD—not just “normal” shyness. 

It is important to take into account the environmental and cultural context of a patient’s distress and impairment because these features might fall within a normal range, given immediate circumstances (such as speaking in front of a large audience when one is not normally called on to do so, to a degree that does not interfere with general social functioning⁶).

What is considered a normative range depends on the developmental stage:

• Among children, a greater level of shyness might be considered more normative when it manifests during developmental stages in which separation anxiety appears.
• Among adolescents, a greater level of shyness might be considered normative especially during early adolescence (when social relationships become more important), and during times of transition (ie, entering high school).
• In adulthood, a greater level of normative shyness or social anxiety might be present during a major life change (eg, beginning to date again after the loss of a lengthy marriage or romantic relationship).

 

Assessment tools

Assessment tools can help you differentiate normal shyness from SAD. Several empirically-validated rating scales exist, including clinician-rated and self-report scales.

Liebowitz Social Anxiety Scale²⁶ rates the severity of fear and avoidance in a variety of social interaction and performance-based situations. However, it was developed primarily as a clinician-rated scale and might be more burdensome to complete in practice. In addition, it does not provide cut-offs to indicate when more clinically significant anxiety might be likely.

Clinically Useful Social Anxiety Disorder Outcome Scale (CUSADOS)²⁷ and Mini-Social Phobia Inventory (Mini-SPIN)²⁸ are brief self-report scales that provide cut-offs to suggest further assessment is warranted. A cut-off score of 16 on the CUSADOS suggests the presence of SAD with 73% diagnostic efficiency.

One disadvantage to relying on a rating scale alone is the narrow focus on symptoms. Given that shyness and SAD share similar symptoms, it is necessary to assess the degree of impairment related to these symptoms to determine whether the problem is clinically significant. The overly narrow focus on symptoms utilized by the biomedical approach has been criticized for contributing to the medicalization of normal shyness.⁵ 

Diagnostic interviews, such as the Structured Clinical Interview for DSM-IV Axis I Disorders²⁹ include sections on SAD that assess avoidance and impairment/distress associated with anxiety. Because these interviews may increase the time burden during an office visit, there are several general questions outside of a structured interview that you can ask, such as: “Has this anxiety interfered with your ability to initiate or maintain friendships? If so, how?” (Table 2). Persons with clinically significant social anxiety, rather than shyness, tend to report greater effects on their relationships and on work or school performance, as well as greater distress about having that anxiety.   

Treatment approaches based on distinctions

Exercise care in making the distinction between normal shyness and dysfunctional and impairing levels of anxiety characteristic of SAD, because persons who display normal shyness but who are overdiagnosed might feel stigmatized by a diagnostic label.⁵ Also, overpathologizing shyness takes what is a social problem out of context, and could promote treatment strategies that might not be helpful or effective.³⁰

Unnecessary diagnosis might lead to unnecessary treatment, such as prescribing an antidepressant or benzodiazepine. Avoiding such a situation is important, because of the side effects associated with medication and the potential for dependence and withdrawal effects with benzodiazepines.

Persons who exhibit normal shyness do not require medical treatment and, often, do not want it. However, some people may be interested in improving their ability to function in social interactions. Self-help approaches or brief psychotherapy (eg, cognitive-behavioral therapy [CBT]) should be the first step—and might be all that is necessary. 

The opposite side of the problem.  Under-recognition of clinically significant social anxiety can lead to under-treatment, which is common even in patients with a SAD diagnosis.³¹ Treatment options include CBT, medication, and CBT combined with medication (Table 3):

• several studies have demonstrated the short- and long-term efficacy of CBT alone for SAD
• medication alone has been efficacious in the short-term, but less efficacious than CBT in the long-term
• combined treatment also has been shown to be more efficacious than CBT or medication alone in the short-term
• there is evidence to suggest that CBT alone is more efficacious in the long-term compared with combined treatment.^a

CBT is recommended as an appropriate first-line option, especially for mild and moderate SAD; it is the preferred initial treatment option of the United Kingdom’s National Institute for Health and Care Excellence (NICE). For more severe presentations (such as the presence of comorbidity) or when a patient did not respond to an adequate course of CBT, combined treatment might be an option—the goal being to taper the medication and continue CBT as a longer-term treatment. Research has shown that continuing CBT while discontinuing medication helps prevent relapse.^32,33

Appropriate pharmacotherapy options include selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors.³⁴ Increasingly, benzodiazepines are considered less desirable; they are not recommended for routine use in SAD in the NICE guidelines. Those guidelines call for continuing pharmacotherapy for 6 months when a patient responds to treatment within 3 months, then discontinuing medication with the aid of CBT.

Bottom Line

The severity of anxiety and associated impairment and distress are the main variables that differentiate normal shyness and clinically significant social anxiety. Taking care not to over-pathologize normal shyness and common social anxiety concerns or underdiagnose severe, impairing social anxiety disorder has important implications for treatment—and for whether a patient needs treatment at all.

 

Related Resources

National Institute for Health and Care Excellence. Social anxiety disorder: recognition, assessment, and treatment of social anxiety disorder. http://guidance.nice.org.uk/cg159.

• Hofmann SG, DiBartolo PM, eds. Social anxiety: clinical, developmental, and social perspectives, 2nd ed. London, United Kingdom: Academic Press; 2010.

• The Shyness Institute. www.shyness.com.

Drug Brand Names

Alprazolam • Xanax        Clonazepam • Klonopin        Fluoxetine • Prozac

Fluvoxamine • Luvox      Paroxetine • Paxil                Phenelzine • Nardil

Sertraline • Zoloft          Venlafaxine • Effexor 

Disclosure
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Featured Audio
Kristy L. Dalrymple, PhD, discusses, treating social anxiety disorder. Dr. Dalrymple is Staff Psychologist, Department of Psychiatry, Rhode Island Hospital, and Assistant Professor of Psychiatry and Human Behavior, Alpert Medical School of Brown University, Providence, Rhode Island.