Research report
Three studies on self-report scales to detect bipolar disorder

https://doi.org/10.1016/j.jad.2010.07.012Get rights and content

Abstract

Background

This study investigated the usefulness of self-report scales for detecting bipolar disorder in several settings.

Methods

Study 1 developed a short form of the Hypomanic Personality Scale (the HPS-6) based on clinic/community and undergraduate samples. Study 2 used this scale for recruiting participants with bipolar disorder from the community. Study 3 administered the full-length Hypomanic Personality Scale, the Mood Disorder Questionnaire, and a short form of the General Behavior Inventory (the GBI-15) to an undergraduate sample. Each study featured a reference standard diagnostic interview.

Results

In Study 2, about half of those responding to the advertisement (based on the HPS-6 developed in Study 1) reported a history of at least one hypomanic episode on a telephone-based SCID. In Study 3, the most robust findings emerged for the GBI-15: about one-third of participants screening positive on that measure met criteria for bipolar disorder using the Structured Clinical Interview for the DSM-IV (SCID).

Limitations

Despite large sample sizes and stratified sampling, this study was limited by a low number of participants with bipolar I disorder.

Conclusions

These three studies produced mixed findings regarding the detection of bipolar disorder via self-report. The HPS-6 was reasonably successful in recruiting participants with a history of at least one manic or hypomanic episode into a study on bipolar disorder. The GBI-15 showed some promise as a screening tool in an undergraduate setting, but there is a need for more sensitive and specific scales. Discussion focuses on potential strategies for developing such scales.

Introduction

Bipolar disorder is associated with high health care costs, lost productivity, elevated medical comorbidity, hospitalization, and the risk of suicide (Andlin-Sobocki and Wittchen, 2005, Baldessarini and Tondo, 2003, Kupfer, 2005, Peele et al., 2003). The lifetime prevalence of bipolar I disorder is reported to be approximately 1% (Judd and Akiskal, 2003, Kessler et al., 1997, Kessler et al., 2005, Kessler et al., 1994, Narrow et al., 2002, Regier et al., 1993, Weissman et al., 1996). For bipolar II disorder, some authors argue that estimates of 1.1% (Merikangas et al., 2007) are artifactually low, due to poor sensitivity of diagnostic measures (Angst and Cassano, 2005). Estimates of the prevalence of cyclothymia range from 0.4% (American Psychiatric Association, 2000) up to 4% (Regeer et al., 2004).

Even though bipolar disorder affects many people, it is often misdiagnosed. Studies conducted in the United States and Europe indicate that 26 to 70% of people with bipolar disorder may either receive an incorrect diagnosis or go undiagnosed altogether. Proper diagnosis may not occur for a decade after bipolar symptoms emerge, especially for those with bipolar II disorder (Ghaemi et al., 2005, Lewis, 2004, Lish et al., 1994, Mantere et al., 2004). A significant percentage of those presenting with unipolar depression may actually suffer from some type of bipolar disorder (Dunner, 2003).

Several factors may contribute to this lack of recognition of the illness (Benazzi et al., 2004). Most people with bipolar disorder who seek treatment do so because of their depression rather than mania (Hirschfeld et al., 2005) and perceive their depression to be more troublesome than their mania (Calabrese et al., 2004). Some people with bipolar disorder lack insight into the presence or consequences of their manic symptoms. Hypomanic episodes, by definition, do not cause severe impairment or distress, so they are less likely to be defined as a problem by people who experience them. When people with bipolar disorder seek treatment for depression, their symptom profile may be virtually indistinguishable from that of unipolar depression (Cuellar et al., 2005). In addition, many clinicians do not screen for bipolar disorder, even when a client presents with depressive symptoms (Brickman et al., 2002).

Unfortunately, the consequences of misdiagnoses are serious. Misdiagnosis as unipolar depression can lead to antidepressant monotherapy, which may trigger mania or rapid cycling in up to 50% of people with bipolar disorder (Ghaemi et al., 2004).

One potential solution to the problem of misdiagnosis could be the use of brief self-report scales to alert providers to the possibility of bipolar disorder. This would allow clinicians to complete a more rigorous diagnostic assessment with patients who seem most likely to have a bipolar disorder. Several such scales have been developed, including the Hypomanic Personality Scale (HPS; Eckblad and Chapman, 1986), the Mood Disorder Questionnaire (MDQ; Hirschfeld et al., 2000), the Temperament Evaluation of Memphis, Pisa, Paris, and San Diego—Autoquestionnaire version (TEMPS-A; Akiskal et al., 2005), the Bipolar Spectrum Diagnostic Scale (BSDS; Ghaemi et al., 2005), the General Behavior Inventory (GBI; Depue et al., 1989), and the Hypomania Checklist (HCL; Angst et al., 2005).

Each of these scales has strengths. For example, they generally have high internal consistency, and each yields significantly higher scores in patient populations than among healthy controls. Several of them have demonstrated at least modest sensitivity and specificity. For instance, the MDQ had a sensitivity of .73 and a specificity of .90 in one study (Hirschfeld et al., 2000). More than three quarters of a sample of students scoring high on the HPS had a history of at least one hypomanic episode; none with low scores had a hypomanic history (Eckblad and Chapman, 1986). The GBI has shown a sensitivity of .78 and a specificity of greater than .98 in two studies (Depue et al., 1981, Klein et al., 1989). The HCL (32-item version) has more limited specificity in detecting bipolar disorder (Angst et al., 2005); however, this may reflect its goal of detecting hypomania, which is less easily distinguished from normal functioning. The TEMPS-A has been studied extensively in Europe, and was the subject of an issue of the Journal of Affective Disorders (Akiskal et al., 2005); however, sensitivity and specificity estimates for this measure have not yet been published.

Some measures are too long to be widely used as screeners. For example, although the GBI has been well-validated (Angst et al., 2005), its full 73-item version can take more than 20 min to complete. Shorter versions of the GBI have been developed (Lewinsohn et al., 2003, Meyer and Hautzinger, 2003), but have not been validated in other samples. In sum, successful detection of bipolar disorder via self-report scales faces difficult challenges. First, it is not clear that existing scales have adequate sensitivity to bipolar disorder for use in routine screening. Second, the length of some scales (e.g., the GBI) limits their utility for screening.

The studies reported here were designed to add to the literature on the detection of bipolar disorder by further testing measures that have shown promising sensitivity and specificity in previous research. Study 1 developed a brief version of the HPS and tested it across samples. The HPS was chosen because it has yet to be applied to clinical samples and may hold promise for detecting bipolar II disorder due to its attention to hypomanic traits. Study 2 used the small set of items derived from Study 1 in an advertising campaign to recruit research participants into a study on bipolar disorder. Study 3 involved administration of screening questionnaires (including the HPS-6 produced in Study 1 and tested in Study 2) to a large sample of undergraduates, yielding a sample enriched for vulnerability to bipolar disorder that was studied in detail. That study included the HPS due to its long-standing application to undergraduate populations, and the MDQ because it has been extensively studied in clinical settings. The only study to use the MDQ in an undergraduate population did not include a diagnostic interview to examine correspondence between the MDQ and bipolar diagnoses (Chandler et al., 2008). The complete GBI is too long to serve as a screening measure, but we included a shortened version, the GBI-15 (Meyer and Johnson, 2003) based on the existing psychometric evidence for the 73-item version.

Section snippets

Study 1

The goal of Study 1 was to develop a briefer version of the HPS that could be used to reliably screen for bipolar disorder. The original HPS has 48 true–false items and takes about 7 min to complete. The goal was to create a subset of items that could be completed in less than 2 min, which would be appropriate for advertisements or posters. Such a set of items then could be used to recruit potentially bipolar participants for studies through notices in papers or public transportation.1

Study 2

The goal of Study 2 was to determine whether advertising featuring items derived from Study 1 could be used to recruit research participants with bipolar disorder. Few studies have investigated the usefulness of self-report scales in the general population. What little research exists suggests that self-report screens are likely insensitive to bipolar disorder in the community (Hirschfeld et al., 2003).

Study 3

The first two studies focused on the usefulness of the HPS in detecting bipolar disorder. Few studies have directly compared the sensitivity and specificity of different self-report screening tools for bipolar disorder diagnoses. In one study, Youngstrom et al. (2004) found that adolescents with bipolar disorder scored .7 standard deviations higher on the GBI and the Youth Self-Report Form (Achenbach, 1991) than their non-bipolar counterparts. Sensitivity and specificity were not reported in

General discussion and future directions

This article presented results from three studies on the detection of bipolar disorder via brief self-report scales. In Study 1, the HPS was administered to an undergraduate and a clinic/community sample. This revealed a small subset of HPS items that correlated particularly well with bipolar diagnosis across samples, allowing creation of the briefer HPS-6. In Study 1, the HPS-6 had good sensitivity and modest specificity in both samples, although different cutoffs were required across samples.

Role of funding source

Funding for these studies was supported by grants from the National Institute of Mental Health (NIMH R01-075021), the National Science Foundation (BCS0544617), the Scottish Rite Schizophrenia Research Program, and University of Miami Seed Funding. None of these organizations played any further role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and/or in the decision to submit the paper for publication.

Conflict of interest

None of the authors have any conflicts of interest.

References (69)

  • J.D. Lish et al.

    The National Depressive and Manic-Depressive Association (DMDA) survey of bipolar members

    J. Affect. Disord.

    (1994)
  • D.M. Meads et al.

    Rasch analysis and item reduction of the Hypomanic Personality Scale

    Pers. Indiv. Differ.

    (2008)
  • T.D. Meyer et al.

    Screening for bipolar disorders using the Hypomanic Personality Scale

    J. Affect. Disord.

    (2003)
  • C.J. Miller et al.

    Sensitivity and specificity of the Mood Disorder Questionnaire for detecting bipolar disorder

    J. Affect. Disord.

    (2004)
  • J.P. Rice et al.

    The stability of diagnosis with an application to bipolar II disorder

    J. Psychiatr. Res.

    (1986)
  • E.A. Youngstrom et al.

    Comparing the diagnostic accuracy of six potential screening instruments for bipolar disorder in youths aged 5 to 17 years

    J. Am. Acad. Child Adolesc. Psychiatry

    (2004)
  • T.M. Achenbach

    Manual for the Youth Self-Report Form and 1991 Profile

    (1991)
  • American Psychiatric Association

    Diagnostic and Statistical Manual of Mental Disorders

    Text Revision

    (2000)
  • P. Andlin-Sobocki et al.

    Cost of affective disorders in Europe

    Eur. J. Neurol.

    (2005)
  • N.C. Andreasen et al.

    Reliability of lifetime diagnosis: a multicenter collaborative perspective

    Arch. Gen. Psychiatry

    (1981)
  • J. Angst et al.

    The mood spectrum: improving the diagnosis of bipolar disorder

    Bipolar Disord.

    (2005)
  • AstraZeneca

    Is it more than just depression?

    Newsweek, CXLVII

    (2006)
  • R.J. Baldessarini et al.

    Suicide risk and treatments for patients with bipolar disorder

    JAMA

    (2003)
  • R.P. Bentall et al.

    Emotional stroop performance and the manic defense

    Br. J. Clin. Psychol.

    (1990)
  • A. Brickman et al.

    Screening for bipolar disorder [Letter to the editor]

    Psychiatr. Serv.

    (2002)
  • J.S. Cacciola et al.

    Comparability of telephone and in-person Structured Clinical Interview for DSM-III-R (SCID) diagnoses

    Assessment

    (1999)
  • J.R. Calabrese et al.

    Impact of depressive symptoms compared with manic symptoms in bipolar disorder: results of a U.S. community-based sample

    J. Clin. Psychiatry

    (2004)
  • R.A. Chandler et al.

    A new US–UK diagnostic project: mood elevation and depression in first-year undergraduates at Oxford and Stanford universities

    Acta. Psych. Scand.

    (2008)
  • W. Coryell et al.

    Long-terms stability of polarity distinctions in the affective disorders

    Am. J. Psychiatry

    (1995)
  • R.A. Depue et al.

    A behavioral paradigm for identifying persons at risk for bipolar depressive disorder: a conceptual framework and five validation studies

    J. Abnorm. Psychol.

    (1981)
  • R.A. Depue et al.

    General behavior inventory identification of unipolar and bipolar affective conditions in a nonclinical university population

    J. Abnorm. Psychol.

    (1989)
  • D.L. Dunner

    Clinical consequences of under-recognized bipolar spectrum disorder

    Bipolar Disord.

    (2003)
  • M. Eckblad et al.

    Development and validation of a scale for hypomanic personality

    J. Abnorm. Psychol.

    (1986)
  • J. Endicott et al.

    A diagnostic interview: the schedule for affective disorders and schizophrenia

    Arch. Gen. Psychiatry

    (1978)
  • Cited by (21)

    • Behavioral activation system (BAS) dysregulation and bipolar spectrum psychopathology in daily life: An online-diary study

      2020, Psychiatry Research
      Citation Excerpt :

      Following previous research (Kwapil et al., 2011; Walsh et al., 2013), a considerable number of high-risk participants with subclinical manifestations were included with reference to the HPS. As noted earlier, the HPS has been extensively tested to possess reliability, convergent, and predictive validity as a measure of BD vulnerability (Kwapil et al., 2000, 2011; Meyer and Hautzinger, 2003b; Miller et al., 2011). In addition, the use of high-risk samples has been justified in the existing literature of the BAS model of BD due to strategic advantages such as avoiding confounding variables (e.g., medication, illness course), which might interfere with the natural BAS response (Johnson et al., 2012).

    • A latent profile analysis of schizotypal dimensions: Associations with psychopathology and personality

      2017, Psychiatry Research
      Citation Excerpt :

      The MDQ is also used as a screening tool for mania, for which categorical determination is made based on Criterion 1 and 2 only. This has improved sensitivity and specificity in some studies (Gervasoni et al., 2009; Miller et al., 2011). In our study, we used a Spanish-translated version that has been validated in both patients and non-clinical young adults (Fonseca-Pedrero et al., 2016c; Sanchez-Moreno et al., 2008).

    • The link between hypomania risk and creativity: The role of heightened behavioral activation system (BAS) sensitivity

      2017, Journal of Affective Disorders
      Citation Excerpt :

      The 48 items pertain to episodic fluctuation in affect, behavior and energy level. The reliability and validity of the HPS were widely tested (Kwapil et al., 2000, 2011; Meyer and Hautzinger, 2003; Miller et al., 2011; Johnson et al., 2015a). In fact, the HPS is one of the most frequently used measures for assessing personality variant vulnerability of BD (Bowins, 2008).

    View all citing articles on Scopus
    View full text