Early detection and intervention evaluation for people at risk of psychosis: multisite randomised controlled trial
BMJ 2012; 344 doi: https://doi.org/10.1136/bmj.e2233 (Published 05 April 2012) Cite this as: BMJ 2012;344:e2233
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
I follow with interest the responses to the work of Morrison et al (2012): Early detection and intervention evaluation for people at risk of psychosis: multisite randomised controlled trial. Despite the range of recruitment and analysis refinements and associated cautions suggested by respondents, the key issue in this paper stands. We need to reconsider whether young people meeting criteria are indeed at “ultra high risk of psychosis” and should be directed towards a treatment intervention.
In this study, as in the range of cited trials with smaller samples (1,2,3,4,5,6), identified symptoms for a large majority group did resolve in the follow up periods. In all studies other than in Amminger et al 2010, the transition to psychosis was not different across control and intervention, by the end of the follow up periods.
This gives us reason also to wonder whether the next research and intervention effort is best directed at refining assessment criteria or interventions with at risk or ultra high risk groups. There is a stronger case for focussing psychosis research on follow up studies, regarding longer term effectiveness and outcomes for young and adult populations, after provision of a range of high quality early interventions. From a population perspective and for the patients and families, there is more to be gained by extending research and intervention efforts beyond the initial two to three years from diagnosis, to minimise impairments and assist with fullest possible recovery.
References:
1. McGorry PD, Yung AR, Phillips LJ, Yuen HP, Francey S, Cosgrave EM, et al. Randomized controlled trial of interventions designed to reduce the risk of progression to first-episode psychosis in a clinical sample with subthreshold symptoms. Arch Gen Psychiatry 2002;59:921-8.
2. Yung AR, Phillips LJ, Nelson B, Francey SM, PanYuen H, Simmons MB, et al. Randomized controlled trial of interventions for young people at ultra high risk for psychosis: 6-month analysis. J Clin Psychiatry 2011;72:430-40.
3. McGlashan TH, Zipursky RB, Perkins D, Addington J, Miller T, Woods SW, et al. Randomized, double-blind trial of olanzapine versus placebo in patients prodromally symptomatic for psychosis. Am J Psychiatry 2006;163:790-9.
4. Addington J, Epstein I, Liu L, French P, Boydell KM, Zipursky RB. A randomized controlled trial of cognitive behavioral therapy for individuals at clinical high risk of psychosis. Schizoph Res 2011;125:54-61.
5. Morrison AP, French P, Walford L, Lewis SW, Kilcommons A, Green J, et al. Cognitive therapy for the prevention of psychosis in people at ultra-high risk: randomised controlled trial. Br J Psychiatry 2004;185:291-7.
6. Amminger GP, Schafer MR, Papageorgiou K, Klier CM, Cotton SM, Harrigan SM, et al. Long-chain omega-3 fatty acids for indicated prevention of psychotic disorders: a randomized, placebo-controlled trial. Arch Gen Psychiatry 2010;67:146-54.
Competing interests: No competing interests
Dear Sir/Madam
We read with interest Morrison et al. ‘Early detection and intervention evaluation for people at risk of psychosis: a multisite randomised controlled trial’ as well as the subsequent coverage in the Guardian http://www.guardian.co.uk/society/2012/apr/06/drugs-psychosis-schizophre...
The authors’ concluded that cognitive therapy did not significantly reduce transition to psychosis or symptom related distress. However, the most surprising finding of the study was the low transition rate (8%), which meant that the study was not sufficiently powered to detect a true difference between the interventions. The authors suggest several reasons for this: the exclusion of participants at risk of imminent transition; the sampling and recruitment strategy may have led to including participants who, due to their pathways into care and degree of help-seeking, may not have been ‘risk enriched’ [1-3]; those most at risk may not be willing to enter trials; the relatively short follow-up period; and that the ‘control’ condition (monitoring of mental state, with warm, empathic supportive listening) may itself have reduced transition rates. Whatever the reasons, the transition rate was much lower than that reported in the existing literature. A recent meta-analysis of 27 studies found that in a combined sample of about 2500 subjects, the transition rate was 18% at 6 months after onset of symptoms, 22% at 1 year, 29% at 2 years, and 32% at 3 years [4]. In agreement with the meta-analytic cohort findings, substantially higher 12 months transition rates (12.5% to 37.5%) were observed in the control conditions of earlier randomised controlled intervention trials in this population [5-10].
It would be interesting to know how many of the cohort remitted from the at-risk mental state during their time in the study, and whether this was influenced by the two interventions [11]. The clinical significance of the results is difficult to interpret because the ordinal CAARMS intensity scales appear to have been weighted by the corresponding frequency ratings to create a 'severity scale' which was statistically analysed as though it were a continuous variable. The authors then commented that a 4 point reduction in this severity scale represented clinically significant change. However, spread across the four subscales examined this might actually reflect a modest clinical change.
The authors reported that of n=634 participants assessed for eligibility, n=346 were excluded and n=288 were randomised. Of those excluded, 156 either were taking antipsychotic medication or were found to be psychotic (24.6%) and n=110 were sub threshold for psychosis (17.35%). It is worth bearing in mind that many of those on anti-psychotics will not have had a psychotic disorder, but have been at-risk subjects with a relatively high risk of transition [12]. Hence, excluding this group may have contributed to the low transition rate. In the study by Morrison et al, 45.4% of those assessed met inclusion criteria. However, in an earlier publication from the same group that described the study design [13], the authors indicated that n=867 participants had been referred to the study, suggesting that n=233 were not assessed. This was reported as being due to a loss of contact or subjects lacking interest. However, this loss of potential participants is another potential source of bias, as these individuals may be those who are the most distressed or disadvantaged, and hence at greatest risk of psychosis. Without knowing more about the sampling procedures employed by the study, such as how the team tried to engage with those with whom they had lost contact, and demographic data comparing this subgroup to those included in the study, it is hard to know whether this may have also contributed to the low transition rate.
The authors suggest that the low transition rate that they identified raises questions about the validity of the At Risk Mental State. However, given how atypical the rate is relative to that in the literature, this may be premature. The concept of the At Risk Mental State has stimulated a body of new research that has significantly advanced our knowledge of the mechanisms underlying psychosis [14-22], and has led to the development of clinical services that permit the earlier detection and management of mental health problems [23, 24]. Research at this stage is a particularly powerful way of investigating the mechanisms underlying psychosis, as the same individual can be studied before and after the onset of illness, without the confounding effects of previous treatment or long-lasting disease-related effects [25]. Nevertheless, the data from this and other studies [26-29] suggest that the existing inclusion criteria, which are relatively recent and mainly based on positive psychotic symptoms [11, 30-35], could certainly be improved. In particular, it would be useful to include items relating to affective and negative psychotic symptoms, and self-perceived [36] and cognitive changes [37] or through the introduction of second step risk stratification [38, 39].
Dr Matthew Broome, University of Warwick
Dr Paolo Fusar-Poli, King’s College London
Dr Majella Byrne, King’s College London
Dr Andreas Bechdolf, University of Cologne
Dr Stefan Borgwardt, University Hospital Basel
Dr Stephan Ruhrmann, University of Cologne
Professor Tyrone Cannon, University of California, Los Angele
Professor Pat McGorry, University of Melbourne
Professor Anita Riecher-Rössler, University Hospital Basel
Professor Jean Addington, University of Calgary
Professor Philip McGuire, King’s College London
1. Van Os, J. and P. Delespaul, Toward a world consensus on prevention of schizophrenia. Dialogues in Clinical Neuroscience, 2005. 7(1): p. 53-67.
2. van Os, J. and R.J. Linscott, Introduction: The Extended Psychosis Phenotype—Relationship With Schizophrenia and With Ultrahigh Risk Status for Psychosis. Schizophrenia Bulletin, 2012. 38(2): p. 227-230.
3. Broome, M.R., et al., What causes the onset of psychosis? Schizophrenia Research, 2005. 79(1): p. 23-34.
4. Fusar-Poli, P., et al., Predicting Psychosis: Meta-analysis of Transition Outcomes in Individuals at High Clinical Risk. Arch Gen Psychiatry, 2012. 69(3): p. 220-229.
5. Amminger, G.P., et al., Long-chain omega-3 fatty acids for indicated prevention of psychotic disorders: a randomized, placebo-controlled trial. Archives of General Psychiatry, 2010. 67(2): p. 146-54.
6. Bechdolf, A., et al., Preventing progression to first-episode psychosis in early initial prodromal states. British Journal of Psychiatry, 2012. 200(1): p. 22-9.
7. McGlashan, T.H., et al., Randomized, double-blind trial of olanzapine versus placebo in patients prodromally symptomatic for psychosis. American Journal of Psychiatry, 2006. 163(5): p. 790-9.
8. McGorry, P.D., et al., Randomized controlled trial of interventions designed to reduce the risk of progression to first-episode psychosis in a clinical sample with subthreshold symptoms. Arch Gen Psychiatry, 2002. 59(10): p. 921-8.
9. Morrison, A.P., et al., Cognitive therapy for the prevention of psychosis in people at ultra-high risk: randomised controlled trial. Br J Psychiatry, 2004. 185: p. 291-7.
10. Addington, J., et al., A randomized controlled trial of cognitive behavioral therapy for individuals at clinical high risk of psychosis. Schizophrenia Research, 2011. 125(1): p. 54-61.
11. Simon, A.E., et al., Ultra high-risk state for psychosis and non-transition: A systematic review. Schizophrenia Research, 2011. 132(1): p. 8-17.
12. Cornblatt, B.A., et al., Can antidepressants be used to treat the schizophrenia prodrome? Results of a prospective, naturalistic treatment study of adolescents. Journal of Clinical Psychiatry, 2007. 68(4): p. 546-57.
13. Morrison, A.P., et al., Early detection and intervention evaluation for people at high-risk of psychosis-2 (EDIE-2): trial rationale, design and baseline characteristics. Early Intervention in Psychiatry, 2011. 5(1): p. 24-32.
14. Shaikh, M., et al., Reduced mismatch negativity predates the onset of psychosis. Schizophrenia Research, 2012. 134(1): p. 42-48.
15. Allen, P., et al., Transition to Psychosis Associated With Prefrontal and Subcortical Dysfunction in Ultra High-Risk Individuals. Schizophrenia Bulletin, 2012.
16. Fusar-Poli, P., et al., Abnormal prefrontal activation directly related to pre-synaptic striatal dopamine dysfunction in people at clinical high risk for psychosis. Mol Psychiatry, 2011. 16(1): p. 67-75.
17. Fusar-Poli, P., et al., Prefrontal Function at Presentation Directly Related to Clinical Outcome in People at Ultrahigh Risk of Psychosis. Schizophrenia Bulletin, 2011. 37(1): p. 189-198.
18. Broome, M.R., et al., Neural correlates of visuospatial working memory in the ‘at-risk mental state’. Psychological Medicine, 2010. 40(12): p. 1987-1999.
19. Howes, O.D., et al., Elevated Striatal Dopamine Function Linked to Prodromal Signs of Schizophrenia. Archives of General Psychiatry, 2009. 66(1): p. 13-20.
20. Broome, M.R., et al., Neural correlates of executive function and working memory in the 'at-risk mental state'. The British Journal of Psychiatry, 2009. 194(1): p. 25-33.
21. Broome, M.R., et al., Delusional ideation, manic symptomatology and working memory in a cohort at clinical high-risk for psychosis: A longitudinal study. European Psychiatry, 2012. 27(4): p. 258-263.
22. Carletti, F., et al., Alterations in White Matter Evident Before the Onset of Psychosis. Schizophrenia Bulletin, 2012.
23. Valmaggia, L.R., et al., Economic impact of early intervention in people at high risk of psychosis. Psychological Medicine, 2009. 39(10): p. 1617-1626.
24. Broome, M.R., et al., Outreach and support in south London (OASIS): implementation of a clinical service for prodromal psychosis and the at risk mental state. European Psychiatry: the Journal of the Association of European Psychiatrists, 2005. 20(5-6): p. 372-8.
25. Fusar-Poli, P., S. Borgwardt, and P.K. McGuire, eds. Vulnerability to psychosis: from psychopathology to neurosciences. . Maudsley Monographs2011, Psychology Press, Routledge: London.
26. Miller, T.J., et al., Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: predictive validity, interrater reliability, and training to reliability.[see comment][erratum appears in Schizophr Bull. 2004;30(2):following 217]. Schizophrenia Bulletin, 2003. 29(4): p. 703-15.
27. Miller, T.J., et al., Prospective diagnosis of the initial prodrome for schizophrenia based on the Structured Interview for Prodromal Syndromes: preliminary evidence of interrater reliability and predictive validity. Am J Psychiatry, 2002. 159(5): p. 863-5.
28. Yung, A.R., et al., Psychotic-Like Experiences in Nonpsychotic Help-Seekers: Associations With Distress, Depression, and Disability. Schizophr Bull, 2006. 32(2): p. 352-359.
29. Yung, A.R., et al., Psychosis prediction: 12-month follow up of a high-risk ("prodromal") group. Schizophr Res, 2003. 60(1): p. 21-32.
30. Broome, M. and P. Fusar-Poli, Philosophical Issues in the Prodromal Phase of Psychosis. Current Pharmaceutical Design, 2012. 18(4): p. 596-605.
31. Fusar-Poli, P. and A.R. Yung, Should attenuated psychosis syndrome be included in DSM-5? Lancet, 2012. 379(9816): p. 591-2.
32. Kaymaz, N., et al., Do subthreshold psychotic experiences predict clinical outcomes in unselected non-help-seeking population-based samples? A systematic review and meta-analysis, enriched with new results. Psychological Medicine. FirstView: p. 1-15.
33. Kelleher, I., J.A. Jenner, and M. Cannon, Psychotic symptoms in the general population – an evolutionary perspective. The British Journal of Psychiatry, 2010. 197(3): p. 167-169.
34. Yung, A.R., et al., Declining transition rate in Ultra High Risk (prodromal) services: dilution or reduction of risk? Schizophenia Bulletin, 2007. 33: p. 673-681.
35. Kelleher, I., et al., Clinicopathological significance of psychotic experiences in non-psychotic young people: evidence from four population-based studies. The British Journal of Psychiatry, 2012.
36. Schultze-Lutter F, et al., Schizophrenia Proneness Instrument - Adult version (SPI-A)2007, Rome: Giovanni Fioriti.
37. Fusar-Poli P, et al., Cognitive functioning in prodromal psychosis: A meta-analysis. . Archives of General Psychiatry, in press.
38. Ruhrmann, S., et al., Prediction of psychosis in adolescents and young adults at high risk: results from the prospective European prediction of psychosis study. Archives of General Psychiatry, 2010. 67(3): p. 241-51.
39. Riecher-Rossler, A., et al., Efficacy of using cognitive status in predicting psychosis: a 7-year follow-up. Biological Psychiatry, 2009. 66(11): p. 1023-30.
Competing interests: No competing interests
Dear Editor,
In their well-designed study on the possibilities to prevent the onset of psychotic disorders in people at ultrahigh risk (the EDIE-2 trial), Morrison and colleagues (2012) did not find a significant effect for cognitive behavioural therapy (CBT) compared with monitoring. However, this could very well be explained by a lack of statistical power and the low transition rates in the control group. Power had been calculated on the basis of estimated transition rates of 15% in the experimental and 30% in the control condition, but the rates proved to be only 6.9% versus 9%. In their discussion, the authors suggest that the ultrahigh risk strategy should be revised, because of a lack of predictive power.
However, the major lesson from the failure of this study to test the hypothesis that CBT can reduce the risk of transition lies in sampling and statistics. As Sackett (1995) showed some time ago, the maximum predictive capacity of any “test” (and the UHR criteria are a test in this sense) is dependent on the base rate of true positives in the sample. The latter must exceed around 20% before any test, no matter how good, can perform efficiently. Earlier studies of the positive predictive value (PPV) of the UHR criteria and clinical trials were conducted in samples, which were sufficiently enriched that the threshold of 20% was passed sufficiently (McGorry et al, 2002; Yung et al, 2004; Morrison et al, 2004). More recent studies including the EDIE 2 trial appear to have been conducted in samples more dilute (less enriched) for genuine risk.
In the Dutch EDIE trial (EDIE-NL) two recruitment strategies could be compared with each other. The traditional referral strategy that was also used in EDIE-2, resulted in a sub sample that is very comparable to the EDIE-2 sample: The subjects were young (19 versus 20 years of age), predominantly male (63% versus 85%), mostly single and living with parents and with a low transition rate (7.5% in 12 months versus 8% in 24 months). In another region, however, an alternative strategy to identify at-risk people was used. Here all consecutive help-seeking people in secondary mental health care were screened with the Prodromal Questionnaire (PQ; Loewy et al, 2005) and then, in the second step, diagnosed them with the instrument to identify at-risk people (the CAARMS; Yung et al, 2005). This sub sample was older (25.5 years), predominantly female (60%), and more often living with a partner. The transition rate in this group was considerably higher (22.5% in 12 months). The screened sample was also much more depressed and anxious, reported lower quality of life and were much more distressed by their subclinical symptoms at baseline (Rietdijk et al, 2012).
These results fit well with the model of enrichment proposed by van Os and Linscott (2012). This model describes that depending at which point along the extended psychosis phenotype the UHR subjects are selected, the subsequent transition rate varies from 7% to 20%. We suggest using the decline in social functioning criterion strictly (a score of ≤55 on the SOFAS, a much-used instrument to assess social functioning, or a drop of 30% within the last few months) and to add a threshold level of sustained distress (together indicating a need for care) as well as active help-seeking for an axis 1 or 2 disorder. In this way the sample should be sufficiently enriched with psychosis-proneness such that the number of false positive subjects is sufficiently low to allow clinical trials to be conducted with sufficient power to detect clinically important differences between treatment options. A two-stage strategy with help-seeking people in secondary mental health is essential i.e. such patients must be screened and enriched using for instance the PQ-16 (Ising et al, 2012) and then the ultrahigh can be detected with the CAARMS.
It is quite understandable that the results of EDIE-2 were at first glance disappointing to many in the field. However, this should not lead to the conclusion that identifying people at risk for developing a psychotic disorder is not possible. Identifying people at high-risk is possible if the right techniques are used. And more importantly: if groups with high incidence rates are identified, prevention of the onset of the first episode may still be possible, despite the negative effects of the EDIE 2 trial. Of course, further refinement and wider re-engineering of the framework to capture early clinical phenotypes must also continue (McGorry et al, 2006; 2010).
Mark van der Gaag, Professor of Clinical Psychology, VU University, Amsterdam, The Netherlands, m.vander.gaag@vu.nl
Patrick McGorry, Professor of Youth Mental Health, Orygen Youth Health Research Centre, University of Melbourne, Australia, pmcgorry@unimelb.edu.au
Pim Cuijpers, Professor of Clinical Psychology, VU University, Amsterdam, The Netherlands, p.cuijpers@vu.nl
References:
Ising, H. K., Veling, W., Loewy, R. L., Rietveld, M. W., Rietdijk, J., Dragt, S., et al. (2012). The validity of the 16-item version of the prodromal questionnaire (PQ-16) to screen for ultra high risk of developing psychosis in the general help-seeking population. Schizophrenia Bulletin, doi:10.1093/schbul/sbs068
Loewy, R. L., Bearden, C. E., Johnson, J. K., Raine, A., & Cannon, T. D. (2005). The prodromal questionnaire (PQ): Preliminary validation of a self-report screening measure for prodromal and psychotic syndromes. Schizophrenia Research, 77(2-3), 141-9.
McGorry, P. D., Yung, A. R., Phillips, L. J., Yuen, H. P., Francey, S., Cosgrave, E. M., et al. (2002). Randomized controlled trial of interventions designed to reduce the risk of progression to first-episode psychosis in a clinical sample with subthreshold symptoms. Archives of General Psychiatry, 59(10), 921-8.
McGorry, P. D., Hickie, I. B., Yung, A. R., Pantelis, C., & Jackson, H. J. (2006). Clinical staging of psychiatric disorders: A heuristic framework for choosing earlier, safer and more effective interventions. The Australian and New Zealand Journal of Psychiatry, 40(8), 616-22.
McGorry, P. D. (2010). Risk syndromes, clinical staging and DSM V: New diagnostic infrastructure for early intervention in psychiatry. Schizophrenia Research, 120, 49-53.
Morrison, A. P., French, P., Walford, L., Lewis, S. W., Kilcommons, A., Green, J., et al. (2004). Cognitive therapy for the prevention of psychosis in people at ultra-high risk: Randomised controlled trial. The British Journal of Psychiatry : The Journal of Mental Science, 185, 291-7.
Morrison, . P., French, P., Stewart, S. L. K., Birchwood, M., Fowler, D., Gumley, . I., et al. (2012). Early detection and intervention evaluation for people at risk of psychosis: Multisite randomised controlled trial. BMJ (Clinical Research Ed.), 344(apr05 1), e2233.
Rietdijk, J., Klaassen, R., Ising, H., Dragt, S., Nieman, D. H., van de Kamp, J., et al. (2012). Detection of people at risk of developing a first psychosis: Comparison of two recruitment strategies. Acta Psychiatrica Scandinavica, doi:10.1111/j.1600-0447.2012.01839.x.
Sacket, D. L. (1995). Clinical epidemiology: A basic science for clinical medicine (2nd ed.). Boston, Masschusets: Little, Browne and Company.
van Os, J. & Linscott, R. J. (2012). Introduction: The extended psychosis phenotype--relationship with schizophrenia and with ultrahigh risk status for psychosis. Schizophrenia Bulletin, 38(2), 227-30.
Yung, A. R., Phillips, L. J., Yuen, H. P., & McGorry, P. D. (2004). Risk factors for psychosis in an ultra high-risk group: Psychopathology and clinical features. Schizophrenia Research, 67(2-3), 131-42.
Yung, A. R., Yuen, H. P., McGorry, P. D., Phillips, L. J., Kelly, D., Dell'Olio, M., et al. (2005). Mapping the onset of psychosis: The comprehensive assessment of at-risk mental states. The Australian and New Zealand Journal of Psychiatry, 39(11-12), 964-971.
Competing interests: No competing interests
I appreciate the hard work of Professor Morrison and colleagues in developing their study design and data interpretation (1). This is the largest trial to date looking at a population at risk for psychosis. And it is good when solid study methodology and a good sample size lead to conclusions that contradict prior pessimistic formulations. In this case it would be great news to report that the conversion rate to psychosis in at-risk population is lower than what’s been previously reported. The 8% conversion rate of Morrison et al. study is much better that rates of up to 38% conversion in untreated at risk samples (2). But is this truly the case?
Unfortunately to confidently draw this conclusion one would need much better clarity with regards to the subjects' antipsychotic medication history. Instead we are summarily told that 14 subjects in the cognitive therapy arm and 13 subjects in the monitoring arm were prescribed antipsychotics. However we do not know if these were the subjects that eventually converted; if this were the case, this would indeed greatly strengthen the conclusion that perhaps the rate of conversion might have been overestimated in prior studies. However if it turns out that the antipsychotics were prescribed for a completely different group of patients (than the patients who ended up converting) this would indicate that the antipsychotics were in fact the causa proxima for the relatively low rate of conversion. If this were to be the case the logical conclusion would be that we are only looking at a conversion rate lowered by the antipsychotic protective effect.
The authors appropriately address the potentially confounding effect of ”active monitoring” which might have resulted in more of an active intervention than just plain monitoring. I actually find the term “monitoring” a bit misleading. As in fact we are looking at “an enhancement over routine care” including non-judgmental, empathic listening and psychosocial services, factors that have all been reported to be therapeutically effective. To make matters even more complicated the active group were offered CBT on top of active monitoring, rather than as a parallel intervention. As a result, we are looking at controls who have received a lot of services, or active subjects who have received not only a lot of services but also additional therapy.
Then the fact that the subjects in this study did better than prior samples might simply be an effect of an active intervention across study groups. Not a true lower conversion rate but a treatment effect.
Finally the study design and findings raise questions about which was the active intervention. And the evidence seems to indicate that we are looking at an "active monitoring" effect that plateaus and could not be further improved by the addition of CBT.
In a way Professor Morrison’s study leads to a somewhat paradoxical conclusion: in lack of a clear additional gain, the recommendation would be against CBT (the "active" intervention) and in favor of active monitoring (the "control" intervention), which turns out to be the true active intervention, for a population at risk for psychosis.
Adrian Preda, Health Sciences Associate Professor of Psychiatry, apreda@uci.edu
References
1. Morrison AP, French P, Stewart SLK, Birchwood M, Fowler D, Gumley AI, Jones PB, Bentall RP, Lewis SW, Murray GK, Patterson P, Brunet K, Conroy J, Parker S, Reilly T, Byrne R, Davies LM, Dunn G. Early detection and intervention evaluation for people at risk of psychosis: multisite randomised controlled trial. BMJ 2012; 344: e2233 doi: 10.1136/bmj.e2233.
2. McGlashan TH, Zipursky RB, Perkins D, Addington J, Miller T, Woods SW, Hawkins KA, Hoffman RE, Preda A, Epstein I, Addington D, Lindborg S, Trzaskoma Q, Tohen M, Breier A.Am J Psychiatry. Randomized, double-blind trial of olanzapine versus placebo in patients prodromally symptomatic for psychosis.2006 May;163(5):790-9.
Competing interests: No competing interests
When Professor Morrison and colleagues did the power calculation for their randomised trial they estimated that without active treatment 30% of patients with at risk mental states would go on to develop schizophrenia or a related condition. They based this figure on rates of transition to psychosis in previous experimental and observational investigations. However, half-way through the study the researchers had to amend their primary outcomes to incorporate frequency, severity and distress from symptoms. Only 9% of their control subjects and 7% of those who received cognitive therapy developed a psychotic illness (1).
Early intervention practitioners have been warned about this lowering of the transition rate for more than a decade (2,3). Richard Warner even provided worked examples to illustrate why this would happen in his commentary on a paper by two of the investigators (4). It has occurred because, as we all know, the positive predictive value of a diagnostic test depends on the prevalence of the condition in the population to which it is applied (5). Initial studies were conducted in patients who had been referred by consultant psychiatrists and general practitioners to specialised clinics for a putative schizophrenic prodrome. It turned out that as many as 30%-70% of these patients were, indeed, developing a psychotic illness (6). Clinical and research tools such as the Comprehensive Assessment of At-Risk Mental States predicted with a fair degree of accuracy which patients would become seriously ill. Inspired by this success early intervention for psychosis teams urged general practitioners to lower their referral threshold. They have also encouraged multidisciplinary mental health clinicians, primary care teams, counsellors, ministers, priests, teachers, probation workers, housing officers, and worried families to refer troubled young people directly to their clinics. Not surprisingly, true pre-psychotic states are infrequent amongst these referrals. Hence, assessments of at risk mental states will have a low positive predictive value. The brilliant BMJ paper by Mathers and Hodgkin in 1989 which clearly explains clinical and policy implications of the positive predictive value should be required reading for all early intervention practitioners (5).
Well-meaning claims about the primary prevention of schizophrenia have misled patients, families, journalists and politicians and this has led to a distortion of health care priorities in many countries. However, I have recently gained the impression that the epidemiological and clinical errors that underpin the early intervention for psychosis movement are gradually dawning on its advocates. Perhaps the spectre of “psychosis risk syndrome” or “attenuated psychotic symptoms syndrome” in the next edition of the Diagnostic and Statistical Manual of Mental Disorders has focussed their minds. They are, at last, voicing concerns about the dangers of over-diagnosis and unnecessary treatment with powerful psychological therapies and antipsychotic medicines. Hopefully Morrison and colleagues’ ambitious, important and methodologically admirable study - combined with proper understanding of positive predictive value – will be the start of rational policy making and sensible clinical care for people in the early stages of major mental disorders.
References
1. Morrison AP, French P, Stewart SLK, Birchwood M, Fowler D, Gumley AI, Jones PB, Bentall RP, Lewis SW, Murray GK, Patterson P, Brunet K, Conroy J, Parker S, Reilly T, Byrne R, Davies LM, Dunn G. Early detection and intervention evaluation for people at risk of psychosis: multisite randomised controlled trial. BMJ 2012; 344: e2233 doi: 10.1136/bmj.e2233.
2. Warner R. Limitations of the Bonn Scale for the Assessment of Basic Symptoms as a screening measure. Arch Gen Psychiatry 2002; 59: 470-1.
3. Pelosi AJ, Birchwood M. Is early intervention for psychosis a waste of valuable resources? British Journal of Psychiatry 2003; 182: 196-8.
4. Warner R. Fact versus fantasy: A reply to Bentall & Morrison. Journal of Mental Health 2003; 12: 351-7.
5. Mathers N, Hodgkin P. The Gatekeeper and the Wizard: a fairy tale. BMJ 1989; 298: 172-4.
6. Fusar-Poli P, Bonoldi I, Yung AR, Borgwardt S, Kempton MJ, Valmaggia L, Barale F, Caverzasi , McGuire P. Predicting psychosis. Meta-analysis of transition outcomes in individuals at high clinical risk. Arch Gen Psychiatry 2012; 69: 220-9.
Competing interests: No competing interests
Re: Early detection and intervention evaluation for people at risk of psychosis: multisite randomised controlled trial
We agree with Broom et al that the transition rate of below 10% at one year is a surprising finding. It potentially has very positive implications for young people with the type of mental state that has been considered as putting them at ultra-high risk for transition to a psychotic syndrome; this reassuring finding indicates that we can tell them that this is relatively unlikely to happen. Broom and colleagues concur with our discussion about possible selection biases that may have accounted for a lower than expected one-year prevalence of transition, but we point out that the rate in published reports has been decreasing inexorably over the past decade and should not be summarised as a cross-sectional estimate. This is likely due to the case-mix of the different studies changing over time and the fact that many of those included in early trials may already have had a psychosis syndrome; our double assessment at baseline excluded this, much like a careful history in clinical practice. Indeed, the Melbourne group has previously published thoughtfully on the decreasing trend in transition rates (Yung et al 2007). Similarly, the meta-analysis of transition rates that they cite was always likely to conclude that the rates are higher than the low rates being currently reported, since it will inevitably be skewed by the inclusion of the preliminary studies that found the highest rates.
Van der Gaag and colleagues, Preda and, last, Pelosi, are similarly concerned by our transitions over the follow-up period, pointing out that the positive predictive value of a test (and they mean the double CAARMS assessment, here) is dependent on the prevalence of the disorder in the population: low prevalence, low predictive value. We were reassured when we revisited their citations in support of this well known fact (Sackett, 1995 and the wonderfully iconoclastic piece by Mathers & Hodgkin, 1989). The fact is that we, like the authors of many of the responses to our paper, had the prior hypothesis that there would be many more transitions in the group at supposedly ultra-high risk group included in our trial. We, like them, were wrong, but this contributes to new knowledge already emerging over the life cycle of the trial (Yung et al 2007). The empirical finding that the prevalence of transition is lower than would have been expected a decade ago when research into ultra-high risk mental states began in earnest may be an uncomfortable fact for clinical researchers, but it may be of great solace to those using their services.
Hamilton is concerned that our data suggest that these young people may be needlessly treated. However, despite the fact that our study suggests they are much less at risk of developing psychosis than initially thought, it is important not to trivialise the difficulties that this population are experiencing, and to recognise the genuine distress, disability and help-seeking behaviour that is also evident from our data regarding symptoms and functioning. We do not recommend abandoning the at-risk and preventative concepts, but rather seek to refine the definition and enrich the sampling strategies (by inclusion of affective dimensions or using other enrichment strategies such as those mentioned by van der Gaag and Reitdijk et al. (2012)). It would also be sensible to widen the outcomes of interest so that they are transdiagnostic, since transition to a diagnosis such as schizophrenia that has questionable reliability and validity (see Bentall, 2003) may not be the most sensible outcome, and these young people are likely to be at risk of many potential mental health problems, not just psychosis. As we argue in our discussion, provision of treatment should be based on the evidence about which interventions work for which presenting problems, and aim to minimise the potential exposure to adverse effects.
References
Bentall, R. P. (2003). Madness explained: psychosis and human nature. Penguin Books Ltd: London, England.
Mathers N, Hodgkin P. The Gatekeeper and the Wizard: a fairy tale. BMJ 1989; 298: 172-4.
Rietdijk, J., Klaassen, R., Ising, H., Dragt, S., Nieman, D. H., van de Kamp, J., et al. (2012). Detection of people at risk of developing a first psychosis: Comparison of two recruitment strategies. Acta Psychiatrica Scandinavica, doi:10.1111/j.1600-0447.2012.01839.x.
Sacket, D. L. (1995). Clinical epidemiology: A basic science for clinical medicine (2nd ed.). Boston, Massachusetts: Little, Browne and Company.
Yung, A.R., et al., Declining transition rate in Ultra High Risk (prodromal) services: dilution or reduction of risk? Schizophenia Bulletin, 2007. 33: p. 673-681.
Competing interests: No competing interests