Premium
Using discriminant function analysis with clinical, demographic and historical variables to diagnose schizophrenia
Author(s) -
Newmark Charles S.,
Jones M. Tazewell,
McKee Daphne C.,
Simpson Mariett,
Boren Richard B.,
Prillaman Kathy
Publication year - 1980
Publication title -
british journal of medical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.102
H-Index - 62
eISSN - 2044-8341
pISSN - 0007-1129
DOI - 10.1111/j.2044-8341.1980.tb02564.x
Subject(s) - generalizability theory , psychology , schizophrenia (object oriented programming) , discriminant function analysis , psychopathology , interview , sample (material) , clinical psychology , psychiatry , developmental psychology , statistics , chemistry , chromatography , political science , law , mathematics
An effort was made to formalize the critical clinical, demographic and historical features of schizophrenia through the development of a reliable assessement questionnaire. The criterion diagnosis was established via traditional hospital procedures and was substantiated by psychometric measures. Raters with comparable diagnostic experience were familiarized with each of the clinical, demographic and historical features before evaluating the presence or absence in a sample of both psychiatric in‐patients and out‐patients. A structured, standardized interview was used to reduce inconsistency and oversight due to variability in interviewing techniques and coverage of psychopathology. Using the coefficients of the discriminative functions of five major symptoms (loose associations, autism, loss of ego boundaries, emotional blunting, delusions) and two demographic and historical features (family history of mental illness, poor social relations) correctly classified 98 per cent of the total sample of 253 patients. A subsequent investigation demonstrated the utility and generalizability of the proposed system. The assets and liabilities of using such an approach are discussed.