Computer diagnosis in psychiatry: a Bayes approach

R. Hirschfeld, R. L. Spitzer, R. G. Miller

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

A model is presented based on applied probability theory (Bayes method) which has been tested successfully in other medical fields but has only been proposed without sufficient testing in psychiatry. The logic of Bayes method is described, detailing the samples used to develop and test the model, comparing the results of the model with the performance of a well established different type of computer model (DIAGNO), and the implications of this are discussed. Bayes method requires estimates of the relative frequencies of relevant symptoms in specified diseases (e.g., the relative incidence of disturbed reality testing in paranoid schizophrenia as compared with that in alcoholism). In this study estimates of these frequencies are derived from a sample of patients and nonpatients interviewed by New York psychiatrists using Spitzer's Current and Past Psychopathology Scales (CAPPS). The formal mathematical procedure (Bayes method) which translates this information into predicted diagnoses is briefly described. The model is tested on a subset of this sample, and then on 3 completely separate samples: an inpatient group from Columbia and the Institute of Living, a group of women in a maternity clinic who were selected by a screening questionnaire for schizophrenia, and a mixed group of Italian inpatients and outpatients interviewed by Italian psychiatrists. The CAPPS records are processed by both Bayes method and DIAGNO, and the results compared. The agreement among clinician and computer varies between 40 and 70% for Bayes method, and between 45 and 55% for DIAGNO. The reasons for this difference are discussed. Finally a comparison of the advantages and disadvantages of the respective methods is presented.

Original languageEnglish (US)
Pages (from-to)399-407
Number of pages9
JournalJournal of Nervous and Mental Disease
Volume158
Issue number6
StatePublished - 1974

Fingerprint

Psychiatry
Psychopathology
Inpatients
Reality Testing
Probability Theory
Paranoid Schizophrenia
Computer Simulation
Alcoholism
Schizophrenia
Outpatients
Incidence

ASJC Scopus subject areas

  • Psychiatry and Mental health

Cite this

Hirschfeld, R., Spitzer, R. L., & Miller, R. G. (1974). Computer diagnosis in psychiatry: a Bayes approach. Journal of Nervous and Mental Disease, 158(6), 399-407.

Computer diagnosis in psychiatry : a Bayes approach. / Hirschfeld, R.; Spitzer, R. L.; Miller, R. G.

In: Journal of Nervous and Mental Disease, Vol. 158, No. 6, 1974, p. 399-407.

Research output: Contribution to journalArticle

Hirschfeld, R, Spitzer, RL & Miller, RG 1974, 'Computer diagnosis in psychiatry: a Bayes approach', Journal of Nervous and Mental Disease, vol. 158, no. 6, pp. 399-407.
Hirschfeld R, Spitzer RL, Miller RG. Computer diagnosis in psychiatry: a Bayes approach. Journal of Nervous and Mental Disease. 1974;158(6):399-407.
Hirschfeld, R. ; Spitzer, R. L. ; Miller, R. G. / Computer diagnosis in psychiatry : a Bayes approach. In: Journal of Nervous and Mental Disease. 1974 ; Vol. 158, No. 6. pp. 399-407.
@article{fec27dac3fa64bca90750d00df4ceb10,
title = "Computer diagnosis in psychiatry: a Bayes approach",
abstract = "A model is presented based on applied probability theory (Bayes method) which has been tested successfully in other medical fields but has only been proposed without sufficient testing in psychiatry. The logic of Bayes method is described, detailing the samples used to develop and test the model, comparing the results of the model with the performance of a well established different type of computer model (DIAGNO), and the implications of this are discussed. Bayes method requires estimates of the relative frequencies of relevant symptoms in specified diseases (e.g., the relative incidence of disturbed reality testing in paranoid schizophrenia as compared with that in alcoholism). In this study estimates of these frequencies are derived from a sample of patients and nonpatients interviewed by New York psychiatrists using Spitzer's Current and Past Psychopathology Scales (CAPPS). The formal mathematical procedure (Bayes method) which translates this information into predicted diagnoses is briefly described. The model is tested on a subset of this sample, and then on 3 completely separate samples: an inpatient group from Columbia and the Institute of Living, a group of women in a maternity clinic who were selected by a screening questionnaire for schizophrenia, and a mixed group of Italian inpatients and outpatients interviewed by Italian psychiatrists. The CAPPS records are processed by both Bayes method and DIAGNO, and the results compared. The agreement among clinician and computer varies between 40 and 70{\%} for Bayes method, and between 45 and 55{\%} for DIAGNO. The reasons for this difference are discussed. Finally a comparison of the advantages and disadvantages of the respective methods is presented.",
author = "R. Hirschfeld and Spitzer, {R. L.} and Miller, {R. G.}",
year = "1974",
language = "English (US)",
volume = "158",
pages = "399--407",
journal = "Journal of Nervous and Mental Disease",
issn = "0022-3018",
publisher = "Lippincott Williams and Wilkins",
number = "6",

}

TY - JOUR

T1 - Computer diagnosis in psychiatry

T2 - a Bayes approach

AU - Hirschfeld, R.

AU - Spitzer, R. L.

AU - Miller, R. G.

PY - 1974

Y1 - 1974

N2 - A model is presented based on applied probability theory (Bayes method) which has been tested successfully in other medical fields but has only been proposed without sufficient testing in psychiatry. The logic of Bayes method is described, detailing the samples used to develop and test the model, comparing the results of the model with the performance of a well established different type of computer model (DIAGNO), and the implications of this are discussed. Bayes method requires estimates of the relative frequencies of relevant symptoms in specified diseases (e.g., the relative incidence of disturbed reality testing in paranoid schizophrenia as compared with that in alcoholism). In this study estimates of these frequencies are derived from a sample of patients and nonpatients interviewed by New York psychiatrists using Spitzer's Current and Past Psychopathology Scales (CAPPS). The formal mathematical procedure (Bayes method) which translates this information into predicted diagnoses is briefly described. The model is tested on a subset of this sample, and then on 3 completely separate samples: an inpatient group from Columbia and the Institute of Living, a group of women in a maternity clinic who were selected by a screening questionnaire for schizophrenia, and a mixed group of Italian inpatients and outpatients interviewed by Italian psychiatrists. The CAPPS records are processed by both Bayes method and DIAGNO, and the results compared. The agreement among clinician and computer varies between 40 and 70% for Bayes method, and between 45 and 55% for DIAGNO. The reasons for this difference are discussed. Finally a comparison of the advantages and disadvantages of the respective methods is presented.

AB - A model is presented based on applied probability theory (Bayes method) which has been tested successfully in other medical fields but has only been proposed without sufficient testing in psychiatry. The logic of Bayes method is described, detailing the samples used to develop and test the model, comparing the results of the model with the performance of a well established different type of computer model (DIAGNO), and the implications of this are discussed. Bayes method requires estimates of the relative frequencies of relevant symptoms in specified diseases (e.g., the relative incidence of disturbed reality testing in paranoid schizophrenia as compared with that in alcoholism). In this study estimates of these frequencies are derived from a sample of patients and nonpatients interviewed by New York psychiatrists using Spitzer's Current and Past Psychopathology Scales (CAPPS). The formal mathematical procedure (Bayes method) which translates this information into predicted diagnoses is briefly described. The model is tested on a subset of this sample, and then on 3 completely separate samples: an inpatient group from Columbia and the Institute of Living, a group of women in a maternity clinic who were selected by a screening questionnaire for schizophrenia, and a mixed group of Italian inpatients and outpatients interviewed by Italian psychiatrists. The CAPPS records are processed by both Bayes method and DIAGNO, and the results compared. The agreement among clinician and computer varies between 40 and 70% for Bayes method, and between 45 and 55% for DIAGNO. The reasons for this difference are discussed. Finally a comparison of the advantages and disadvantages of the respective methods is presented.

UR - http://www.scopus.com/inward/record.url?scp=0016258514&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0016258514&partnerID=8YFLogxK

M3 - Article

C2 - 4601526

AN - SCOPUS:0016258514

VL - 158

SP - 399

EP - 407

JO - Journal of Nervous and Mental Disease

JF - Journal of Nervous and Mental Disease

SN - 0022-3018

IS - 6

ER -