I was pleased to see that you included the BSDS on your PsychEducation.org website. I'd like to clarify a bit, though, about the "history" of this scale, as originally developed by me. (It was subsequently revised slightly and then "field-tested" by Dr Ghaemi's group, with lead investigator Chris Miller). Although I more or less understand the statistical basis for the "cut off" of 13--optimal sensitivity and specificity--statistics has never been my strong suit!--I never conceived of the scale as producing an "either-or" result for clinicians. In my own, "home-made" scoring system--before the Miller et al APA presentation, etc.--I used the following break down:

 

19 or higher=  bipolar spectrum disorder highly likely

11-18         =  moderate probability of bipolar spectrum disorder.

6-10           =  low probability of bipolar spectrum disorder

<6              =  bipolar spectrum disorder very unlikely

 

This was based solely on my own clinical experience after seeing many, many "refractory depressed" patients who turned out to be bipolar (based on family history, hypomanic periods, subsequent response to mood stabilizers, etc.). But, I still believe this gradation of probability most accurately reflects what the test "means"...By the way, I originally called this the "Bipolar Spectrum Story", because of its narrative quality. We somewhat reluctantly settled on the term "diagnostic scale" after considering many options. But the scale is not "diagnostic" of bipolar disorder or of anything. It is really just a screening instrument that gives the clinician a sort of "yellow light" or "red light", requiring the usual comprehensive diagnostic assessment.