Accuracy of the family history method in affective illness. Comparison with direct interviews in family studies.

We interviewed available spouses and first-degree relatives of 140 bipolar and unipolar probands for current and past psychopathology, and assessed interrater reliability. Diagnoses based on direct interviews of relatives were compared with those based on reports of the probands and of all other interviewed family members. Probands underestimated the prevalence of affective illness and other psychiatric disorders in their relatives, and overestimated the age of onset of illness in their ill relatives. Probands reported more accurately about illness in their spouses and parents than in their siblings and children, but accuracy reached acceptable levels for spouses only. Diagnoses on relatives derived by combining reports of all other interviewed family members, including the proband, were slightly more accurate than those based on the proband's reports alone. Good accuracy was obtained only for reports about spouses.