Developing and Internally Validating a Prognostic Model (P Risk) to Improve the Prediction of Psychosis in a Primary Care Population Using Electronic Health Records: The MAPPED Study

Background: An accurate risk prediction algorithm could improve psychosis outcomes by reducing duration of untreated psychosis. The objective was to develop and validate a risk prediction model for psychosis, for use by family doctors, using linked electronic health records. Methods: A prospective prediction study. Records from family practices were used between 1/1/2010 to 31/12/2017 of 300,000 patients who had consulted their family doctor for any nonpsychotic mental health problem. Records were selected from Clinical Practice Research Datalink Gold, a routine database of UK family doctor records linked to Hospital Episode Statistics, a routine database of UK secondary care records.  Each patient had 5 to 8 years of follow up data. Study predictors were consultations, diagnoses and/or prescribed medications, during the study period or historically, for 13 nonpsychotic mental health problems and behaviours, age, gender, number of mental health consultations, social deprivation, geographical location, and ethnicity. The outcome was time to an ICD10 psychosis diagnosis. Findings: 830 individual diagnoses of psychosis were made. Patients were from 216 family practices, mean age was 45.3 years and 43.5% were male. Median follow-up was 6.5 years (IQR 5.6, 7.8). Overall 8-year incidence of psychosis was 45.8 (95% CI 42.8, 49.0)/100,000 person years at risk. A risk prediction model including age, sex, ethnicity, social deprivation, consultations for suicidal behaviour, depression/anxiety and substance abuse, a history of consultations for suicidal behaviour, smoking history and substance abuse and prescribed medications for depression/anxiety/PTSD/OCD and total number of consultations resulted in good discrimination (Harrell’s C=0.77 after internal validation). Identifying patients with predicted risk exceeding 1% over 6 years had sensitivity of 69% and specificity of 71%. Interpretation: An accurate prediction model for risk of psychosis developed from electronic health records could be used to facilitate early detection of psychosis by family doctors. Funding: NIHR, School for Primary Care Research. Declaration of Interest: None of the authors have any declarations of interest to declare. Ethical Approval: Approval was obtained from the CPRD’s Independent Scientific Advisory Committee.