Three years ago a provocative editorial in the British Journal of Psychiatry argued that we should rethink the place of bipolar disorder in psychiatry. Rather than schizophrenia, bipolar disorder should be the ‘dominant paradigm – the heartland condition’ (Reference Goodwin and GeddesGoodwin 2007). One of the arguments was that the course of bipolar disorder ‘allows a much more meaningful distinction’ between acute and long-term care. Perhaps so, but for many clinicians the waters of clinical practice have become muddied over the past three years. Even though, as Goodwin Geddes noted, ‘we are not embarrassed to tell patients they have bipolar disorder’, who would have foreseen that the condition, endorsed by celebrities, would acquire a cachet, coming to be seen as ‘a favourable and even desirable diagnosis’ – a diagnosis associated with creativity and higher social status which, given the choice, one would ‘keep’ (Reference Chan and SirelingChan 2010)? Who would have guessed that one day The Psychiatrist's front cover would bear the legend ‘I want to be bipolar’? Perhaps for clinicians the crunch is whether bipolar disorder should be defined narrowly as a severe mental illness or not. Might it better be conceived of as a heterogeneous group of disorders including ‘softer’ elements – the bipolar spectrum? The implications are enormous, given that the estimated prevalence for bipolar-spectrum illness is a staggering 8.3% of the population.

Saunders & Goodwin (pp. 318–328) consider this and a number of key clinical issues in light of the evidence. What is the mean age at onset? (They note the ‘considerable downward shift’ reported in studies over the past two decades.) Is early diagnosis possible? A third of individuals wait 10 years for an accurate diagnosis. Is early intervention possible? Given the important link between suicide and the early stages of the illness, harm reduction might be possible (Reference Jamison, Hawton and HawtonJamison 2005). What is the minimum duration of a hypomanic episode? Some argue that it should be less than the 4 days required by DSM–IV. This has crucial significance in establishing the breadth of the spectrum. How should the clinician respond to subsyndromal symptoms? These are common and, although depressive symptoms cause impairment, subsyndromal hypomanic symptoms actually enhance functioning. And lastly comorbidity, almost the sine qua non of bipolar disorder, with huge prevalences of alcohol misuse and personality disorder in particular.