Abstract
In a commentary in this journal, Dr Amos has criticized the TIPS study (Amos, 2012). We are grateful for the opportunity to comment upon his criticism and to clarify what we regard as the main results.
Dr Amos attacks our interpretation of results. He states that ‘even accepting their results at face value it seems perverse that the authors valorize the small difference in negative symptoms favoring the early detection group while ignoring the large difference in functional outcome favoring the no early detection group’. He suggests that we systematically overstate findings supporting early detection and similarly understate findings pointing in the opposite direction. In our opinion, he demonstrates the very tendency he assigns to his opponents: ‘selective use of research to support a position rather than weigh the evidence’. We have published several critical papers (Larsen et al., 2001; Friis et al., 2003, 2004; Rossberg et al., 2010) demonstrating our commitment to a balanced discussion of potential benefits and shortcomings of early intervention. We are convinced that such an open-minded, unbiased position is the best way to help our patients. For sure, we started the project genuinely hopeful about the possibilities of early intervention, but we have always kept an open eye for potential limitations, believing firmly that this is the only way to secure an enduring position for early intervention. Over the years, we have seen that uncritical endorsement of a method fosters an uncritical rejection of equal enthusiasm.
We agree that care is recommended concerning conclusions based on abstracts. But as Dr Amos himself seems to jump to conclusions from abstracts; we do not think that he follows his own advice. Furthermore, he leaves doubt as to how well he has read the abstracts upon which he comments. He criticizes Professor Yung (2012) for incorrectly referring to an abstract by McGlashan et al. in Schizophrenia Bulletin 37 (suppl 1) 273−274, entitled ‘Early detection and intervention (EDI) in first episode psychosis: can it increase the chance for full recovery? TIPS 10 year findings’. He claims that the correct reference is Friis et al. on page 98 of the same issue. In fact, the reference given by Professor Yung is correct. It concludes that the early intervention group had double the percentage of recovered patients at 10 years. These results have been published in the American Journal of Psychiatry (Hegelstad et al., 2012).
Dr Amos is right about there being an abstract from the TIPS study by Friis et al. on page 98. However, the abstract has quite a different title, one clearly showing that we do not sweep problems under the carpet: ‘Early detection and intervention (EDI) in first episode psychosis: can it reduce the risk for poor outcome? TIPS 10 year findings’ (italics added). In this abstract, we focus on the flip side of the coin, i.e. for a considerable group of patients we were not able to demonstrate long-term benefits of early intervention. A thorough presentation of these results has recently been submitted for publication.
Let us end by condensing what we consider to be the primary lessons from the TIPS study so far.
The combined use of information campaigns and early detection teams made it possible to significantly reduce Duration of Untreated Psychosis (DUP) in the EDI area. DUP was significantly shorter, both compared to a previous period in the same area and compared to an Ordinary Intervention (OI) area in the same period (Melle et al., 2004).
Because of the shorter DUP, EDI patients had lower levels of positive and negative symptoms (Melle et al., 2004) and of suicidal behaviour (Melle et al., 2006) than OI patients. While it has never been measured to what degree people with untreated psychosis commit suicide, the fact is that psychosis clearly increases such a risk and that EDI can reduce this risk and prevent pre-treatment suicide.
In both study areas, the level of positive symptoms dropped rapidly after the start of treatment. Consequently the difference between groups in positive symptoms had disappeared by 3 months. But the difference in negative symptoms persisted throughout the 2-year intervention period (Melle et al., 2008).
At 5 years, the difference in negative symptoms was narrowed (Larsen et al., 2011), and at 10 years it was gone on the group level. However, the 10-year outcome was complex. EDI patients had twice the rate of recovery (31 vs. 15%), while the rate of non-remission was disappointingly high in both areas (48 vs. 52%) (Hegelstad et al., 2012).
The OI area lost more patients at the 10-year follow-up, and the lost 10-year OI patients tended to have higher levels of negative symptoms at 5 year than those not lost, while the EDI area tended to lose patients with a lower symptom level at 5 years.
The selective loss of patients probably implies that we underestimated the EDI advantage concerning remission and recovery at 10 years.
In the EDI patients, the total duration of the first psychotic episode was only half that of OI patients. Even if this difference was not related to better long-term outcome for the majority of patients, we argue that EDI met the ethical imperative of reducing more rapidly patient suffering from psychosis.
To sum up: EDI has clearly demonstrated advantages over OI, but EDI is not magic. Finding better treatment methods for the most severely ill remains a huge challenge.
See Commentary by Amos, 2012, 46(9): 903–904