What is early detection of psychosis?
Early detection refers to the correct identification of individuals who are at high risk of developing schizophrenia, with an emphasis on the development of frank psychosis. Several assessment tools have been constructed to effectively identify such individuals. Sensitivity of an assessment tool refers to the proportion of people who develop psychosis that were previously identified by the assessment tool as being at high risk. Specificity is the proportion of people who do not develop psychosis that were previously identified as not being at high risk. Assessment tools therefore aim to have both high sensitivity and high specificity.
Generally, there are two approaches that dictate the characteristics used as markers for detection. The ultra-high risk approach focuses on a triad of at-risk mental states defined as:
1. Having a family history of psychosis plus non-specific symptoms and a recent decline in functioning
2. Showing recent onset of attenuated psychotic symptoms with a decline in functioning
3. Symptoms must be brief, intermittent and limited
The other approach is based on Huber’s Basic Symptoms, which focuses on a detailed way of describing subjective disturbances, and may be an earlier indicator of risk than the first approach.
What is the evidence for early detection of psychosis?
Moderate to high quality evidence finds the mean rate of transition to psychosis in those assessed as being at clinical high risk for psychosis is around 16% by 2 years and 29% by 3 years. In people assessed as being at clinical high risk of obsessive-compulsive disorder are at higher risk of psychosis than people assessed as being at clinical high risk of bipolar disorder, which in turn has higher risk of psychosis than people assessed as being at high risk of depression. However, the rate of transition to psychosis are only one third the rate of transition to non-psychotic disorders in people at assessed as being at clinical high risk for non-psychotic disorders.
In children and adolescents assessed as being at clinical high risk of psychosis, transition rates were between 17% and 20% by 1 year follow-up and between 7% and 21% by 2 year follow-up. 36% of children and adolescents recovered from their clinical high risk status by 6-year follow-up, and 40% continued to meet clinical high risk criteria without transition to psychosis.
Studies with older samples reported higher transition rates than studies with younger samples, and more recent publications reported lower transition rates than older publications. Studies using the basic symptoms approach reported higher transition rates than studies using the ultra-high risk approach. Studies of people receiving psychosocial treatments (e.g. cognitive behavioral therapy) reported lower transition rates than studies of people receiving standard care (e.g. case management). Studies of people on antipsychotics also reported lower transition rates than studies of people not on antipsychotics.
Moderate to high quality evidence suggests instruments based on ultra-high risk criteria have good sensitivity and moderate specificity. Moderate to low quality evidence also suggests the BSABS scale, based on basic symptoms approach, has good sensitivity and moderate specificity. This indicates validated instuments are generally good at correctly identifying individuals who do develop psychosis, but not as good at identifying individuals who do not develop psychosis.
Moderate quality evidence suggests the model with the best predictive value (86%) for transition to psychosis was a clinical model including odd beliefs, marked impairment in role functioning, blunted affect, auditory hallucinations, and anhedonia/asociality. A biological model using grey matter volume, and a neurocognitive model using IQ, verbal memory, executive functioning, attention, processing speed, and speech perception, both had positive predictive values of 83%. An environmental model with a positive predictive value of 63% involved urbanicity, social-sexual aspects, and social-personal adjustment. The best combination model had a positive predictive value of 82% and involved disorganised communication, suspiciousness, verbal memory deficit, and decline in social functioning.
Green - Topic summary is available.
Orange - Topic summary is being compiled.
Red - Topic summary has no current systematic review available.