A different approach toward screening for bipolar disorder: The prototype matching method

Department of Psychiatry and Human Behavior, Brown University School of Medicine, Providence, RI 02905, USA.
Comprehensive psychiatry (Impact Factor: 2.25). 07/2010; 51(4):340-6. DOI: 10.1016/j.comppsych.2009.09.004
Source: PubMed

ABSTRACT Most screening scales for psychiatric disorders consist of a series of questions about the signs and symptoms of the disorder of interest, and to determine whether a patient screens positive, the scores of the individual items are summed and the total score is compared with an empirically derived threshold. A problem with the score summation approach toward case identification on screening scales is that different studies may find that different thresholds are optimal for distinguishing cases from noncases. An alternative approach toward screening is the prototype matching approach, in which respondents are asked to indicate how well their clinical history matches the described prototype. In the present report from the Rhode Island Methods to Improve Diagnostic Assessment and Services project, we compared the symptom summation and prototype matching approaches toward screening for bipolar disorder in a large sample of psychiatric outpatients. Nine hundred sixty-one psychiatric outpatients were interviewed with the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition and completed the Bipolar Spectrum Disorders Scale (BSDS). The BSDS is a unique screening scale consisting of a prototypic description of bipolar disorder. The respondent checks off which items in the prototypic paragraph describes them and also answers a single multiple-choice question at the end of the paragraph asking how well the paragraph describes them. The results of a receiver operating curve analysis found that the score summation and prototype matching approaches toward screening on the BSDS performed equally well. These findings provide preliminary evidence that an alternative approach toward psychiatric screening, the prototype matching approach, is as effective as the traditional score summation method. This raises the intriguing possibility of developing a combined screening scale/educational instrument that can be formatted as a brochure and thus placed in clinicians' waiting rooms, thereby facilitating use of the measure.

10 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: The DSM-5 Personality and Personality Disorders (PDs) Work Group has recommended a reformulation of the PD section, one component of which is a replacement of specified operational criteria with a prototype matching dimensional rating system. The Work Group indicated that prototype ratings have been demonstrated to have good interrater reliability. No study was cited to support this statement, and a review of the reliability literature does not support this claim. The one study that directly compared the reliability of prototype and DSM-IV criteria counting approaches found the DSM-IV approach was much more reliably applied. The Work Group cited 2 studies supporting the validity of the prototype matching approach, one of which had significant methodological limitations and the other changed the a priori threshold on the PD prototype dimensional rating scale to categorize patients into PD positive and negative groups. The Work Group also cited 2 studies suggesting that prototype matching approaches are preferred by clinicians. Several studies have raised concerns about the adequacy of psychiatric diagnostic evaluations conducted in routine clinical practice thereby raising questions about the value of studies of clinicians' preferences in comparing different diagnostic practices. In conclusion, if the prototype matching dimensional approach described in the DSM-5 draft proposal is adopted, then it will have been adopted with essentially no empirical support demonstrating improved reliability or validity. In fact, there is evidence that reliability will be worse than the DSM-IV approach.
    Journal of personality disorders 04/2011; 25(2):206-21. DOI:10.1521/pedi.2011.25.2.206 · 3.08 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Screening for clinical depression and bipolar disorder remains controversial. Screening is usually based on finding discriminating symptoms, but not all tools perform equally well. Clinicians should be able to assess the clinical utility of screening tests as well as their accuracy and acceptability. Screening for depression using the Patient Health Questionnaire (PHQ) has been extensively examined. Four main versions of scoring the PHQ exist. The two-item PHQ2, the nine-item PHQ9, the PHQ DSM-IV algorithm, and the two-step PHQ2 then PHQ9. Recent results suggest that the PHQ9 is more accurate than the PHQ2, and that the algorithm scoring method is preferred to the linear cut-off score. The two-step procedure has promise, but it has not been adequately tested. Two screening questions may be a useful compromise in medical settings, as they take less than 2 min, but about a quarter of patients do not receive screening even when implemented systematically. Alternative customized questionnaires have been developed in medical settings such as the Depression Screening in Parkinson's Disease DESPAR and Neurological Disorders Depression Inventory for Epilepsy (NDDI-E). Screening for bipolar disorders is an even greater challenge than screening for unipolar depression. Screening in primary care and the community has low positive predictive value. Screening in high-risk samples, such as those with known depression is somewhat more successful, but not yet sufficiently accurate to be used alone. Screening for depression can bring added value to routine unassisted recognition, but only if followed by good-quality treatment. Screening for bipolar disorder is not yet sufficiently accurate to be used reliably in clinical practice.
    Current opinion in psychiatry 01/2012; 25(1):24-31. DOI:10.1097/YCO.0b013e32834de45b · 3.94 Impact Factor
  • Source
    Bipolar Disorders 02/2012; 14(1):1-5. DOI:10.1111/j.1399-5618.2011.00972.x · 4.97 Impact Factor
Show more