Family history research diagnostic criteria (FH-RDC)35 were assessed by having affected participants complete the FH-RDC interview36, the best validated instrument available, to report on psychotic illness in relatives. Detailed study of the reliability, sensitivity and specificity of the FH-RDC versus personal interview of first-degree family members in the Roscommon Family Study37 showed that between informants, the kappa±SE was 0.49±0.02; between FH-RDC and best-estimate diagnoses, the kappa±SE was 0.43±0.02. Validated against the best-estimate diagnosis (which included personal interview and hospital records in most cases), the sensitivity was 0.37 and the specificity was 0.996. The FH-RDC criteria38 and instrument36 thus miss true illness in relatives but almost never produce a false positive.
