[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Views 1,191
Citations 0
March 11, 2020

Clinical High Risk for Psychosis—Not Seeing the Trees for the Wood

Author Affiliations
  • 1Orygen, Parkville, Victoria, Australia
  • 2Centre for Youth Mental Health, The University of Melbourne, Parkville, Victoria, Australia
JAMA Psychiatry. Published online March 11, 2020. doi:10.1001/jamapsychiatry.2019.4635

“It's not the tools that you have faith in: tools are just tools. They work, or they don't work. It's people you have faith in or not.”

Steve Jobs

Three decades ago, the schizophrenia field finally began to challenge the intrinsic pessimism that had inhibited preventive approaches for a century. Early detection and specialized early treatment models for first-episode psychosis have since become the global standard of care, producing better outcomes that “bend the curve” of the early course of illness1 and have opened the door to the prevention or delay of the first episode of psychosis. The development of operational criteria (the “ultra” or “clinical” high risk [CHR] criteria) for identifying what we originally termed the at-risk mental state meant that an even earlier stage of illness could be identified prospectively and studied for its heuristic and therapeutic potential.

In their umbrella review of 42 meta-analyses, Fusar-Poli et al claim that the innovative trigger of the CHR concept “represents the most established preventive approach in clinical psychiatry.”2 This is a big statement worth serious scrutiny, particularly because the field has attracted its genuine critics as well as “merchants of doubt.”3 How should this scrutiny be carried out? The technology of evidence-based medicine has provided us with tools, which include Cochrane reviews and various species of meta-analysis, all of which aim for and claim objectivity and accuracy. These tools are of undoubted value, but they also have blind spots, can be subject to error and bias, are increasingly subject to inexpert use and overuse, and require clarity of formulation and interpretation. If this is lacking, there is risk of spurious precision and a misreading of the evidence. To what extent can the true meaning of data be obscured by boilerplate technology? The limits of a helicopter view are that one cannot see the trees for the wood. Key insights and findings can be hidden from view, and individuals who have not themselves been involved in the culture and conduct of the clinical research may formulate the questions crudely or unfairly, interpret the data incorrectly, and discount or overlook valid or inconvenient findings from individual studies that could be critical for further progress. These errors may go in both directions. However, we are concerned in particular that this could undermine confidence in offering intervention to people in urgent need of care and conceal the pathway to an even more substantial paradigm shift in intervening transdiagnostically in the early stages of mental disorders.4

Morbidity and Prediction

The data from this umbrella review show that help-seeking patients with this phenotype are characterized by increased levels of premorbid risk factors and are highly symptomatic with substantial comorbidity, substance use, suicidal behavior, and functional impairment. They not only demonstrate a clear-cut need for care, despite being subthreshold for first-episode psychosis, but also a range of neurobiologic disturbances, including structural brain changes, neurocognitive impairment, and blood biomarker changes. Some authors have alleged that this cohort of help-seeking patients are within the normative range and to offer care might be harmful through labeling or overtreatment.5 The Fusar-Poli et al review2 clearly validates their morbidity and risk. The question then arises of how to find and engage the people with a genuine need for care enriched for risk of psychosis and other potentially poor outcomes and how to intervene safely to reduce these risks and improve outcomes. Another challenge is to engage more than a minority of such patients3 for whom solutions have been developed.4

The meta-analytic approach, especially at the umbrella level, fails to fully capture the cutting edge of knowledge and clinical utility of prediction, particularly with regard to timing and sequence of interventions. More creative and potent predictive strategies, such as machine learning, sequential enrichment, risk stratification, joint modeling, and risk calculators, are not sufficiently captured, and much meaning and relevance is lost.

To engage and detect those genuinely at risk and in need of care requires a population health perspective combined with novel cultures of care and dynamic sequential risk enrichment strategies operating in transdiagnostic clinical settings. Static approaches anchored to baseline data are not fit for purpose. Staged care intervention designs based on treatment response are also consistent with this dynamic approach.6 A broader range of initial clinical phenotypes is optimal in which a range of diagnostic and functional outcomes can be studied.4


There have been more than 20 randomized clinical trials of interventions for CHR patients. A number of studies were underpowered; however, meta-analyses of the first wave of studies were able to overcome this problem and were positive, demonstrating a risk reduction of 50% and an average number needed to treat of 9.7 More recently, network meta-analyses have aimed to determine whether any 1 intervention is superior to others, including a range of control conditions. These studies showed, as with many other mental and physical disorders, there are several interventions that have the same effect. The authors tend to interpret these data to suggest that, rather than all interventions being equally effective, intervention is ineffective. However, the control condition in most of these studies against which specific interventions are tested is an active psychosocial intervention, and the symptomatic and functional outcomes for both arms are generally equally improved in a highly morbid group. For ethical reasons, there is rarely a truly inactive control or monitoring group, reflecting the level of clinical need in this cohort. The reason for the falling transition rate in CHR intervention studies over the last decade may not only be owing to sampling issues but also that all patients, especially in research cohorts, are being provided with effective interventions and being treated early in the course of symptoms.

The 2019 Cochrane review8 reinforced the impression that there is no evidence that intervention is effective. The review characterized studies that actually used the best possible methodology as low quality without fully accessing the relevant material essential to make an accurate judgment regarding quality. The misleading and harmful headline message from these meta-analyses and reviews is that patients do not benefit at all from intervention.9


The conventional tools of evidence-based medicine have a major role in guiding the field of psychiatry. However, they have limitations and are still subject to error and interpretation bias. We believe these orthodox technologies need modernization and better regulation because how they are deployed and interpreted can yield misleading results, which are accorded automatic respect, owing to the brand under which they appear.

In another 2019 umbrella review on CHR interventions,10 boldly titled “Preventive Treatments for Psychosis: Umbrella Review (Just the Evidence),” the authors go beyond the evidence into sociological speculation. They characterize the evolution of the CHR field as having moved from a “peak of inflated expectations” into a “trough of disillusionment.” They foreshadow a “slope of enlightenment.” While no doubt in need of refinement and more fine-grained understanding, we view the evidence in a much more positive light and celebrate the tangible progress that this clinical research paradigm has yielded thanks to the efforts of skilled and dedicated research teams globally.

The CHR field has been genuinely heuristic, not only showing it is possible to predict and delay the onset of fully-fledged psychosis but also to demonstrate that neurobiologic change is actively occurring during the perionset stages of emerging mental disorders. We have learned that the CHR phenotype poses transdiagnostic risk for a range of syndromes that emerge heterotypically and comorbidly as well as homotypically. This knowledge has spawned a new diagnostic model, namely clinical staging, which is of potential value not only for research strategy but also for clinical trials and clinical care. As we move to parse the mix of heterogeneity through the next phase of neurobiological research and deploy novel and emerging biological and psychosocial therapies in a more targeted and sequential way, there is no need for disappointment as we search for further knowledge.

Back to top
Article Information

Corresponding Author: Patrick D. McGorry, MD, PhD, Orygen, 35 Poplar Rd (Locked Bag 10), Parkville, VIC 3052, Australia (pat.mcgorry@orygen.org.au).

Published Online: March 11, 2020. doi:10.1001/jamapsychiatry.2019.4635

Conflict of Interest Disclosures: None reported.

Correll  CU, Galling  B, Pawar  A,  et al.  Comparison of early intervention services vs treatment as usual for early-phase psychosis: a systematic review, meta-analysis, and meta-regression.  JAMA Psychiatry. 2018;75(6):555-565. doi:10.1001/jamapsychiatry.2018.0623PubMedGoogle ScholarCrossref
Fusar-Poli  P, Salazar de Pablo  G, Correll  CU,  et al.  Prevention of psychosis: advances in detection, prognosis, and intervention [published online March 11, 2020].  JAMA Psychiatry. doi:10.1001/jamapsychiatry.2019.4779.Google Scholar
Ajnakina  O, David  AS, Murray  RM.  ‘At risk mental state’ clinics for psychosis - an idea whose time has come - and gone!  Psychol Med. 2018;49(4):529-534. doi:10.1017/S0033291718003859PubMedGoogle ScholarCrossref
McGorry  PD, Hartmann  JA, Spooner  R, Nelson  B.  Beyond the “at risk mental state” concept: transitioning to transdiagnostic psychiatry.  World Psychiatry. 2018;17(2):133-142. doi:10.1002/wps.20514PubMedGoogle ScholarCrossref
Moritz  S, Gawęda  Ł, Heinz  A, Gallinat  J.  Four reasons why early detection centers for psychosis should be renamed and their treatment targets reconsidered: we should not catastrophize a future we can neither reliably predict nor change.  Psychol Med. 2019;49(13):2134-2140. doi:10.1017/S0033291719001740PubMedGoogle ScholarCrossref
Nelson  B, Amminger  GP, Yuen  HP,  et al.  Staged Treatment in Early Psychosis: a sequential multiple assignment randomised trial of interventions for ultra high risk of psychosis patients.  Early Interv Psychiatry. 2018;12(3):292-306. doi:10.1111/eip.12459PubMedGoogle ScholarCrossref
van der Gaag  M, Smit  F, Bechdolf  A,  et al.  Preventing a first episode of psychosis: meta-analysis of randomized controlled prevention trials of 12 month and longer-term follow-ups.  Schizophr Res. 2013;149(1-3):56-62. doi:10.1016/j.schres.2013.07.004PubMedGoogle ScholarCrossref
Bosnjak Kuharic  D, Kekin  I, Hew  J, Rojnic Kuzman  M, Puljak  L. Interventions for prodromal stage of psychosis. Cochrane Database of Systematic Reviews. 2019(11). doi:10.1002/14651858.CD012236.pub
Nelson  B, Amminger  GP, Bechdolf  A,  et al.  Evidence for preventive treatments in young patients at clinical high risk of psychosis: the need for context [published online December 19, 2019].  Lancet Psychiatry. doi:10.1016/S2215-0366(19)30513-9Google Scholar
Fusar-Poli  P, Davies  C, Solmi  M,  et al.  Preventive treatments for psychosis: umbrella review (just the evidence).  Front Psychiatry. 2019;10:764. doi:10.3389/fpsyt.2019.00764Google ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words