Duration of untreated psychosis: definition, measurement and association with outcome

Introduction The course of schizophrenia and other psychotic disorders shows substantial individual variation, suggesting that perhaps something can be done to improve outcome (Wiersma et al., 1998). However, most of the established predictors of outcome, such as gender, age of onset and premorbid adjustment (Harrigan, McGorry & Krstev, 2003), cannot be easily changed. In this respect, duration of untreated psychosis (DUP) is exceptional, because it is potentially modifiable. This raises the exciting possibility that outcome could be improved through early detection programmes aimed at reducing DUP. In this chapter, we will systematically assess the strength and robustness of the association between DUP and outcome; however, as an essential preliminary, we will first consider the difficulties of defining and assessing DUP. Variability of findings in duration of untreated psychosis: what reason? There is much potential for measurement error in a concept as rarefied as DUP, and it is not reassuring that there is such a wide variability in estimates of DUP across studies (Norman & Malla, 2001; Perkins et al., 2005). This variability might be attributed to heterogeneity amongst different psychotic disorders (Keshavan & Schooler, 1992) or to differences between societies and health services. However, it could equally be accounted for by measurement error. We will consider some of the most likely sources of measurement error below. In summary, these are (1) difficulties in defining the onset and offset of untreated psychosis, (2) the problems of retrospective assessment, (3) discrepancies between the accounts of patients and carers, (4) sample bias, and (5) failure to use standardized assessment instruments.