A Six-Item Screener Outperforms the MDQ for Bipolar Depression in Adolescents

Misdiagnosing an emerging bipolar disorder as unipolar depression is one of the more consequential errors in adolescent psychiatry, and the field's workhorse self-report — the 13-item MDQ — was never built for teenagers. This study tests the Rapid Mood Screener, a six-item instrument originally designed to flag bipolarity inside depressed adults seen in primary and general care, in the exact population where the question bites hardest: adolescents already in a depressive episode.

The result is practically interesting because the RMS-C did not merely "work." It outperformed two longer, better-established screeners on the same patients, at a fraction of the response burden.

What the data shows

The discrimination figures are the headline. Against a structured diagnostic reference, the RMS-C produced an AUC of 0.81 for detecting any bipolar disorder, climbing to 0.82 for BD-II and 0.83 for the BD-II-with-anxious-distress subgroup — the presentations that look most like ordinary depression and are most often missed. In the same dataset the MDQ and HCL-32 trailed on these indices, which matters because both are longer instruments that ask adolescents to recall and self-rate hypomanic features they may never have labelled as abnormal.

Six items with a cutoff of 4 is a low-friction screen. The two-factor solution (CFI = 0.92) and test-retest ICCs of 0.71–0.79 across two months indicate the tool is measuring something stable, not mood-state noise. The authors frame it as a rapid first-pass filter, not a diagnostic substitute — a positive screen routes the adolescent toward a fuller clinical interview rather than toward a label.

How to use it in clinic

Treat the RMS-C as a triage gate, not a verdict. The natural insertion point is any adolescent presenting with a depressive episode, especially one with early onset, atypical features, mixed or anxious distress, irritability, or a family history of bipolarity — exactly the profile where starting an antidepressant without mood-stabiliser cover carries the most risk. A score at or above the cutoff of 4 should trigger a structured assessment of past hypomania, episode course, and functional change before any prescribing decision is finalised.

Two cautions for direct use. First, this validation is in a Chinese adolescent sample, so the specific cutoff should be read as a working anchor rather than a transportable constant until local data confirm it. Second, a six-item screen trades depth for speed: it will over-call in highly anxious or irritable presentations, which is acceptable for a gate whose job is to lower the threshold for a closer look, not to confirm. Documented alongside the interview, a positive RMS plus corroborating history is a defensible reason to slow down on antidepressant monotherapy.