Am J Psychiatry 1992; 149:1225-1227
Copyright © 1992 by American Psychiatric Association
Sensitivity of psychiatric diagnosis based on the best estimate procedure
TA Kosten and BJ Rounsaville
Department of Psychiatry, Yale University School of Medicine, New Haven, Conn. 06519.
OBJECTIVE: A "best estimate" diagnosis is one made by expert clinicians on
the basis of diagnostic information from direct interview conducted by
another clinician plus information from medical records and from reports of
family members. The authors address the question of whether the best
estimate procedure can enhance the classification of psychiatric diagnoses
of subjects who are interviewed directly. METHOD: Four hundred seventy-five
subjects were interviewed directly: 201 opiate-addicted probands who sought
treatment from a university-based clinic and 274 of their spouses and/or
first-degree relatives. Subjects were interviewed by trained clinical
assessors using the Schedule for Affective Disorders and Schizophrenia and
classified according to Research Diagnostic Criteria. Two psychologists
independently diagnosed the same subjects by applying the best estimate
procedure. Lifetime rates of major and minor depressive disorder,
antisocial personality, alcoholism, and drug abuse were calculated. The
rates of diagnoses made on the basis of direct interviews alone were
compared with the rates of diagnoses made according to the best estimate
procedure. RESULTS: Higher rates of diagnoses of all four disorders were
made when the best estimate procedure was applied than when direct
interview alone was used; the best estimate procedure also resulted in a
minimal rate of false positives. CONCLUSIONS: The higher rate of diagnoses
based on the best estimate procedure may represent an enhancement in the
accuracy of psychiatric diagnoses or an increase in erroneous diagnoses.
The authors consider the second possibility less likely.
