Implementación informática y validación de la prism 5 (psychiatric research interview for substance and mental disorders for DSM 5)

Abstract

Given the need to facilitate the detection and diagnosis of psychiatric pathology in substance users in healthcare and research settings, we have carried out the computer implementation and validation of the diagnostic instrument for dual pathology “Psychiatric Research Interview for Substance and Mental Disorders” according to DSM 5 criteria (PRISM 5). A cross-sectional study was carried out with 197 patients who consulted or were admitted for substance addiction in the Parc de Salut Mar healthcare setting or in one of the collaborating centres. It was carried out in 3 phases: 1) Adaptation of the PRISM instrument to the DSM 5 criteria; 2) Computerisation in Spanish of the PRISM 5; 3) Validation of the PRISM 5 instrument, using the gold standard “Longitudinal assessment, done by Expert diagnosticians, using All Data that are available” (LEAD), and Data analysis. After updating the PRISM to DSM 5 criteria, patients were recruited. Diagnoses were obtained by administration of the computerised version of the PRISM 5 and by assessment by medical experts using the LEAD method. For the validation of the PRISM 5 interview, Cohen’s kappa coefficient (k) was used by analysing the concordance between the diagnoses obtained through the PRISM 5 interview and those obtained by the LEAD method. In this way, the Reliability of PRISM 5 was measured. In addition, Validity was measured, with Sensitivity (S) and Specificity (S), Safety, with Positive Predictive Value (PPV) and Negative Predictive Value (NPV), and the Positive Likelihood Ratio (PLR) and Negative Likelihood Ratio (NLR), statistics that are not conditioned by prevalence as occurs with the PPV and NPV, were also obtained. In results, PRISM 5 obtained a k≥0.7 in Reliability in most of the Substance Use Disorders, in current Primary Major Depressive Disorder, in the group of any past Primary Psychotic Spectrum Disorder, in current Panic Disorder and in Antisocial Personality Disorder. In Validity PRISM 5 has obtained excellent results, being better in Specificity than in Sensitivity; in Confidence it has presented very good results in PPV, being excellent in NPV, although we must be cautious in the interpretation of these two statistics (PPV and NPV) since they depend on prevalence; and in PSVR PRISM 5 has presented excellent results being more modest in NPV.