Table 2

Description of three hypothetical cases, from real-world scenarios, illustrating the impact of different pretest probabilities of psychosis onset related to potential sampling bias on estimated post-test probabilities of psychosis onset after CHR assessment, computed using the LRs indicated in figure 222

Case 1Case 2Case 3
PopulationGeneral populationGenetic high-risk populationMental health service users
SettingSchoolRelative's self-support groupCHR service
GenderMaleFemaleMale
Age21 years19 years20 years
EthnicityBlack CaribbeanWhiteWhite
Family history of psychosisNoMother diagnosed with schizophreniaNo
Estimated 3-year risk of psychosis of population ≈ 3-year pretest probability0.27% (assuming an annual incidence of schizophrenia in the general population of Black Caribbean migrants of 0.091%30)4% (assuming a 3-year risk of psychosis in offspring of mothers with schizophrenia as estimated from the Kaplan-Meier, figure 2 from31)15% (assuming a 3-year risk for psychosis of 15% in help-seeking samples referred to CHR services8)
AssessmentInterview (CHR assessment)Interview (CHR assessment)Interview (CHR assessment)
3-year post-test probability in case of positive result (LR+ 1.82)220.49%7.05%26.00%
3-year post-test probability in case of negative result (LR− 0.09)220.02%0.38%1.56%
  • CHR, clinical high risk; LR, likelihood ratio.