Abstract
Professor Yung’s recent article (2012) neatly illustrates a technique common in the early intervention literature in its selective use of research to support a position rather than weigh the evidence. Professor Yung relies upon the Treatment and Illness Perceptions Survey (TIPS) study (McGlashan et al., 2011, incorrectly referenced in Yung, 2012) to propose that reducing the duration of untreated psychosis is associated with improved outcomes for up to 10 years. This group examined the effects of implementing a public health campaign raising awareness of psychosis and access to help, with associated assessment resources. They compared outcomes in two regions with and two regions without the campaign and resources. The sceptic could legitimately wonder why Professor Yung refers to a congress abstract rather than the peer-reviewed research published by the group, such as Larsen et al. (2011). In any case, a critical view of this abstract might note a different abstract by the same authors at the same congress. Friis et al. (2011) reported that while the early detection group were slightly less affected by negative symptoms at 10-year follow-up (no effect size is given), they also had twice the rate of non-recovered patients with no capacity for independent living (21% vs 11%); that is, this paper appears to show that early detection slightly reduces measured psychopathology and doubles the rate of severe functional deficit.
If the interested reader was interested enough to read the peer-reviewed work of the TIPS group, they might also note that while the effort to reduce duration of untreated psychosis was certainly successful, at the 5-year follow-up patients from the early detection groups had spent on average 1.5 times as many days in hospital as those from the no early detection group (mean weeks in hospital 45.3 vs 30.8 over 5 years in Larsen et al., 2011). It can be argued that Larsen et al. (2011) rely on a flawed statistical analysis of their results to imply a significant difference on negative symptoms at 5 years (Amos, 2012). Even accepting their results at face value, it seems perverse that the authors valorise the small difference in negative symptoms favouring the early detection group while ignoring the large difference in functional outcome favouring the no early detection group.
The selective use of evidence becomes apparent when examining Professor Yung’s claim that specialised services for first-episode psychosis have been shown to be superior to generic services at up to 5 years, based in part on the OPUS trial. This group examined the effects of implementing specialist services for patients with first-episode psychosis including smaller case loads, cognitive behavioural therapy for psychosis, and family support, comparing regions with and without this model of care. Bertelsen et al. (2008) conclude that at 5 years there were no differences in symptomatology, but that specialist services were associated with less time in hospital (149 vs 193 days) and with a lower proportion in supported housing (4% vs 10%).
There is a neat symmetry here: reporting the TIPS trial Professor Yung notes the small difference in negative symptoms, while ignoring the lack of difference in positive symptoms, and the evidence of harm in accommodation and days in hospital; reporting the OPUS trial Professor Yung ignores the lack of difference in symptoms, but reports the advantage in accommodation and days in hospital. Professor Yung does not appear interested in particular outcomes, such as psychotic symptoms, days in hospital, or supported housing, but only in positive outcomes favouring early intervention treatments. This approach seems closer to advocacy than scientific enquiry, and conclusions based on such selective review of the literature should be treated with significant scepticism.
See Debate by Yung, 2012, 46(1): 7–9