There is a marked paucity of high quality trials for both pharmacological and psychological treatments of BP II. This stands in stark contrast to the substantial prevalence data, resulting in an evidence vacuum. As depression is the dominant clinical issue, treatment of BP II is essentially the acute and maintenance treatment of depression. In the absence of adequate data, the clinician is faced with a choice of extrapolating from either the unipolar depression or bipolar depression databases, acknowledging that even the latter is threadbare at present. In practice, the pharmacological choice is the balance in the algorithm between the role of accepted mood stabilisers and antidepressants. A core component of the decision-making process is the weighting of risk and benefit. In this context, a dominant issue is the potential for antidepressant therapy to induce rapid cycling and mixed states.

Complicating the issue of the assessment of switching and cycling is the methodology of such assessment. In general, prospective designs are likely to give far higher rates of event detection than retrospective ones. The availability of tools specifically designed to detect switching and cycling is essential, as methodologies relying on spontaneous reports of adverse events are likely to report lower rates than studies using structured tools for the purpose. SSRI-induced sexual dysfunction is a good example of a situation where, based on spontaneous reports, initial assessments of prevalence were negligible, while later studies utilising specific detection tools in prospective designs consistently reported substantial rates.