The excellent article by Chan & Sireling ^{Reference Chan and Sireling1} about the recent increase in public awareness of bipolar disorder mirrors our own experience in research and practice, and highlights important issues for health services.

This article is very timely because there is ongoing debate about the extent to which bipolar disorder may be over- or underdiagnosed. ^{Reference Zimmerman2,Reference Smith and Ghaemi3} Both over- and underdiagnosis occur and are problematic. Some people may be inappropriately labelled, whereas others who would benefit from the diagnosis are missed. Optimal treatment of depression is different in bipolar and unipolar disorders. This is one of many examples in psychiatry where making an early and correct diagnosis is highly likely to have a very direct and important effect on the quality of care offered to, and quality of life experienced by, a patient. ^{Reference Craddock, Antebi, Attenburrow, Bailey, Carson and Cowen4}

Chan & Sireling highlight new cases of bipolar disorder from the primary care setting. Preliminary data from our ongoing studies of primary care patients with depression suggest that bipolar (i.e. manic/hypomanic) features are relatively common in this group (unpublished data; available from the authors on request). In our wider research in individuals with both bipolar and unipolar mood disorders, we have found that those with a diagnosis of recurrent unipolar depression who have a history of mild manic symptoms tend to respond less well to antidepressants. ^{Reference Smith, Forty, Russell, Caesar, Walters and Cooper5}

Inevitably, increasing awareness of any illness has the potential to lead to overdiagnosis and this could cause problems for the patient as well as for services. Thus, a balance must always be struck between the need to increase awareness appropriately among patients, public and clinicians, while not causing a tsunami of uncritical overdiagnosis and self-labelling. As psychiatrists we must ensure we are pragmatic and put the patient's well-being at the centre of decision-making. This will require us to have knowledge of the developing evidence base, make a comprehensive diagnosis based on a detailed lifetime history of both depressed and manic mood (including asking an informant), and have an awareness of the boundaries of clinically relevant symptomatology.