Table 1

Measures of prognostic accuracy that can be calculated from a 2×2 table showing the cross-classification of a CHR assessment and subsequent gold standard assessment of psychosis

Population seeking help at CHR services
Psychotic disorder as defined with the gold standard ICD/DSM
Developing psychosisNot developing psychosis
CHR assessment outcomeAt Risk (CHR+)CHR+ T (true positive)CHR+ NT (false positive, type I error)
Not At Risk (CHR−)CHR− T (false negative, type II error)CHR− NT (true negative)
Pretest probability (x-axis in figure 2)The probability of developing psychosis before the CHR assessment, also called the prevalence of the disease: (true positive+false negative)/(true positive+false postive+false negative+true negative)
SensitivityProportion of patients who will develop the psychosis who will have a positive CHR assessment result: true positive/(true positive+false negative}
SpecificityThe proportion of patients without the development of psychosis who will have a negative CHR assessment result: true negative/(true negative+false positive)
Positive Predictive Value (PPV)The proportion of patients with a positive CHR assessment result who actually have developed psychosis: true positive/(true positive+false positive)
Negative predictive value (NPV)The proportion of patients with a negative CHR assessment result who do not have actually developed psychosis: true negative/(true negative+false negative)
AccuracyThe proportion of true results among the total number of CHR patients assessed: (true positive+true negative)/(true positive+false positive+true negative+false negative)
Positive likelihood ratio
LR+ (legend to figure 2)
The probability of a CHR patient who develop psychosis testing positive at the CHR assessment divided by the probability of a person who does not develop the psychosis testing positive at the CHR assessment: sensitivity/(1-specificity)
Negative likelihood ratio LR− (legend to figure 2)The probability of a patient who develop psychosis testing negative at the CHR assessment divided by the probability of a patient who does not develop psychosis testing negative at the CHR assessment: (1-sensitivity)/specificity
Post-test positive probability (red line in figure 2)The individual probability of developing the disease given a positive prognostic test result: Pretest probability×Sensitivity/((Pretest probability×Sensitivity)+(1−Pretest probability)×(1−Specificity))
Post-test negative probability (green line in figure 2)The individual probability of developing the disease given a negative prognostic test result: (1−Pretest probability)×Specificity/((1−Pretest probability)×Specificity)+Pretest probability×(1−Sensitivity))
  • APS, attenuated psychosis symptoms; BIPS, brief intermittent psychotic symptoms; BLIPS, brief limited intermittent psychotic symptoms; BS, basic symptoms. Adapted from;8 CHR, clinical high risk (UHR and/or BS); DSM, Diagnostic and Statistical Manual; GRFD, genetic risk and deterioration syndrome; ICD, International Classification of Diseases; UHR, ultra high risk (including BLIPS/BIPS and/or APS and/or GRFD).