Open AccessImprovement of the quality of the DSM‐IV diagnostic definitions through computerized algorithms using macros in SPSS and SAS language
Author(s) -
Granero Rosario,
Doménech Josém.,
Ezpeleta Lourdes
Publication year - 2000
Publication title -
international journal of methods in psychiatric research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.275
H-Index - 73
eISSN - 1557-0657
pISSN - 1049-8931
DOI - 10.1002/mpr.88
Subject(s) - computer science , medical diagnosis , standardization , macro , natural language processing , algorithm , elaboration , programming language , artificial intelligence , machine learning , medicine , philosophy , pathology , humanities , operating system
This work addresses methodological and conceptual issues related to the translation of the DSM‐IV diagnoses from verbal language into formal language through computerized algorithms, with the aim of guaranteeing its quality. A great number of biases can affect this process, so the main difficulties connected with each phase are outlined, proposals to standardize the process are advanced, and practical solutions to avoid errors in the formal diagnostic definitions are presented. The steps followed are: (a) classification of the disorders in seven different groups depending on the diagnostic structure they have in the taxonomy; (b) the creation of macros in SPSS and SAS languages that express formally the structures/groups identified; (c) elaboration of 162 macro calls, specifying formally all the particular diagnostic conditions for each DSM‐IV disorder; and (d) checking the correct functioning of the formal definitions proposed in a test data file (that has also been created in this work). Moreover, because the standardization of the process that creates the diagnoses in programming language requires the homogenization of the variable names and the codification formats, we have produced a proposal compatible with the verbal language for identifying, through letters and numbers, all the DSM‐IV criteria. The main contribution of this work consists of facilitating computerized and universal algorithms for obtaining automatically any DSM‐IV diagnosis in three categories (present, absent and not evaluable due to the lack of information) starting from a vector that includes all the criteria for that disorder. This study will also contribute to endowing the classification system most used in psychopathology (DSM‐IV) with greater methodological rigour. Copyright © 2000 Whurr Publishers Ltd.