The At-Risk Mental State for Prediction of Early Psychosis

The at-risk mental state (ARMS) for psychosis is a clinical construct used to identify help‑seeking individuals who have a substantially elevated short‑to-medium‑term risk of developing a first episode of psychosis, but who have not yet crossed the threshold into full psychotic disorder. It is closely related to, and often used interchangeably with, the ultra‑high risk (UHR) or clinical high risk (CHR) concepts used in early psychosis services.
 
Most ARMS/UHR frameworks (e.g. CAARMS, SIPS/SOPS) converge on three main inclusion pathways:
Attenuated psychotic symptoms (APS):

Subthreshold positive symptoms (e.g. suspiciousness, unusual thought content, perceptual disturbances) that are below psychotic intensity or frequency, but clearly abnormal and distressing/impairing; typically present in the recent period (e.g. past year).
 
Brief, limited, intermittent psychotic symptoms (BLIPS/BIPS):
Fully psychotic‑level symptoms (e.g. clear hallucinations or delusions) that emerge but resolve spontaneously within a short duration (often defined as less than 1 week) without sustained antipsychotic treatment and recur intermittently.
 
Genetic risk plus functional decline:

Either a family history of psychotic disorder in a first‑degree relative or a schizotypal personality disorder, combined with a significant recent decline in psychosocial or occupational functioning.
 
Individuals meeting any of these pathways, and who are help‑seeking, are considered to have an ARMS and are candidates for specialized monitoring and indicated preventive interventions in early psychosis services.
 
ARMS criteria aim to enrich for imminent risk of psychosis compared with the general help‑seeking or community population. Transition rates in ARMS cohorts are typically:
Approximately 10–15% within the first year.
Around 20–30% by 2–3 years of follow‑up.
 
Thus, the construct is prognostically meaningful, but most ARMS individuals do not transition within these time frames, highlighting both the utility and the limitation of ARMS as a prediction tool. Within the ARMS population, certain baseline features consistently increase transition risk and are used to refine prediction beyond the categorical ARMS label:
 
Greater severity/frequency of subthreshold positive symptoms (unusual thought content, suspiciousness, perceptual abnormalities, disorganized speech).
 
Longer duration of attenuated/BLIPS symptoms before presentation.
 
Poor social and role functioning and functional decline.
 
Negative symptoms and cognitive problems (e.g. poor attention, subtle cognitive deficits).
 
Older age within the typical ARMS range (often late vs early adolescence), which may index longer duration of prodromal phenomena.
 
These variables underpin multivariate risk calculators that attempt individualized risk estimates within the ARMS group, rather than treating ARMS as a homogeneous category. Beyond the canonical ARMS criteria, several related constructs and markers have been investigated to improve early prediction:
 
Basic symptoms:
Very subtle, self‑experienced disturbances in thinking, perception, and self‑experience that are subjectively recognized as abnormal, sometimes preceding APS/BLIPS; certain basic‑symptom profiles are associated with higher long‑term risk of schizophrenia.
 
Neurocognitive and neurobiological markers:

Impairments in tasks such as verbal learning, social cognition, and electrophysiological indices (e.g. mismatch negativity) have shown added predictive value in some cohorts.

 
Structural and functional neuroimaging–based pattern‑recognition approaches have reached proof‑of‑concept accuracy for predicting transition in ARMS samples, although these are not yet routine clinical tools. Overall, the ARMS concept is a pragmatic clinical gateway: it identifies a small, enriched group at substantially elevated risk, within which more fine‑grained clinical, cognitive, and biological predictors can be used to estimate individual transition risk and to guide indicated preventive interventions.
 
 
References
Bonnett, L.J., Hunt, A., Flores, A., Smith, C. T., Varese, F., et al. (2025, February 27). Clinical Prediction Model for Transition to Psychosis in Individuals Meeting At Risk Mental State Criteria. Schizophrenia. 
 
McGorry, P.D., Hartmann, J.A., Spooner, R. & Nelson, B. (2018, May 24). Beyond the “At-Risk Mental State” Concept: Transitioning Transdiagnostic Psychiatry.World Psychiatry.
 
Radez, J., Waite, F., Izon, E. & Johns, L. (2023, January 11). Identifying Individuals At Risk of Developing Psychosis: A Systematic Review of the Literature in Primary Care Services. Early Intervention in Psychiatry.
 
Tavares, V., Vassos, E., Marquand, A., Stone, J., Valli, I., et al. (2023, January 19). Prediction of Transition to Psychosis From an At-Risk Mental State Using Structural Neuroimaging, Genetic, and Environmental Data. Frontiers in Psychiatry.
 
Thompson, A., Marwaha, S. & Broome, M.R. (2018, January 2). At-Risk Mental State for Psychosis: Identification in Current Treatment Approaches. BJPsych Advances. Cambridge University Press.
 
Zarogianni, E., Storkey, A.J., Borgwardt, S., Smieskova, R., Studerus, E., Riecher-Rossler, A. & Lawrie, S.M. (2019, December). Individualized Prediction of Psychosis in Subjects with an At-Risk Mental State. Schizophrenia Research.
 
Yung, A.R., Wood, S.J., Malla, A., Nelson, B., McGorry, P. & Shah, J. (2019, October 28). The Reality of At-Risk Mental State Services: A Response to Recent Criticisms. Psychological Medicine.