Investigation of a High Risk Systems Approach to the Prediction of Psychosis and Bipolar Disorder

Psychotic and bipolar disorders are associated with huge individual suffering and societal cost. Prevention is a key goal in treating these disorders and requires detecting individuals before illness onset, ideally before adulthood. However, current approaches reach only a small proportion of individuals at high risk. In the first study of this PhD thesis, we reviewed the existing evidence on the ability of the symptom-based ‘At Risk Mental State’ approach to detect children and adolescents at high risk of psychosis. The findings suggested that this approach has limited predictive ability and reaches only a small proportion of high-risk individuals. This highlights the need for new, additional approaches to psychosis risk detection strategies.

Building on this review, we proposed and investigated a “systems” based approach to psychosis and bipolar disorder prediction. Instead of only focusing on individuals presenting with (potentially prodromal) symptoms, we suggest evaluating individual’s risk of psychotic and bipolar disorders by focusing on their contacts with different healthcare or other systems. Given that psychotic and bipolar disorders are often preceded by various problems in health and development, we hypothesised that there would be substantial concentration of risk among individuals in early life health care and education systems related to these problems. 

We examined three specific systems that we hypothesised would be associated with enriched, early life risk for psychosis and bipolar disorders: 1) Presenting to a hospital for self-harm, 2) Attending Child and Adolescent Mental Health Services (CAMHS) and 3) Non-Normative School Progression (relative to same-aged peers). 

Using nationwide register data on all Finns born in 1987 (N =59 476) recorded from birth to age 28 years, we assessed the absolute risk of later psychotic and bipolar disorders and the predictive capacity associated with these systems. We found that:

• Among individuals who presented to hospital for self-harm, 18% were subsequently diagnosed with psychosis or bipolar disorder by age 28 years. 4% of all psychosis and bipolar disorder cases were preceded by presenting to a hospital for self-harm. 
• Among individuals who presented to CAMHS, 12% were subsequently diagnosed with psychosis or bipolar disorder by age 28 years. Among individuals who were admitted to inpatient CAMHS, 17% were subsequently diagnosed with psychosis or bipolar disorder by age 28 years. 45% of all psychosis and bipolar disorder cases were preceded by a CAMHS contact, and 20% of all psychosis and bipolar disorder cases occurred in individuals who had a preceding inpatient CAMHS admission.
• Among individuals with non-normative school progression at age 16 (relative to same aged peers), 4% were subsequently diagnosed with psychosis or bipolar disorder by age 28 years. 14% of psychosis and bipolar disorder cases between ages 17-28 years occurred in individuals who had non-normative school progression at age 16 years.

Our findings suggest that there is significant potential for risk systems to identify individuals at risk of psychosis and bipolar disorder during adolescence. Given that 45% of psychosis and bipolar disorders diagnosed by age 28 were in individuals who had attended CAMHS, this points to CAMHS in particular as a very important risk system for psychosis prediction and prevention. Future studies are needed for: international replication of our findings; identification of additional risk markers for psychosis and bipolar disorder within CAMHS that allow stratification into high-, medium-, and low-risk groups; intervention trials that aim to reduce risk of developing psychosis and bipolar disorder.