F E A Verhoeven1, L S M A Swaab2, I V E Carlier2, A M van Hemert2, F G Zitman2, H G Ruhé1, R A Schoevers1, E J Giltay3. 1. University Medical Center Groningen, Department of Psychiatry, The Netherlands. 2. Leiden University Medical Center, Department of Psychiatry, The Netherlands. 3. Leiden University Medical Center, Department of Psychiatry, The Netherlands. Electronic address: E.J.Giltay@lumc.nl.
Abstract
BACKGROUND: Standardized Diagnostic Interviews (SDIs) such as the Mini International Neuropsychiatric Interview (MINI) are widely used to systematically screen for psychiatric disorders in research. To support generalizability of results to clinical practice, we assessed agreement between the MINI and clinical diagnoses. METHODS: Agreement was assessed in a large, real life dataset (n = 7016) using concordance statistics such as sensitivity, specificity, efficiency and area under the curve (AUC). RESULTS: 41.5% of clinical diagnoses were mood disorders, 26.5% were anxiety disorders. Overall, we found moderate agreement between MINI and clinical diagnoses (median efficiency: 0.92, median AUC: 0.79). For mood disorders, the AUC for all participants showed a range between 0.55 and 0.81 (median: 0.73), and for anxiety disorders the AUC ranged from 0.78 to 0.88 (median: 0.83). The AUC showed better agreement for mood disorders in the single diagnosis group than in the total group (median 0.77 vs. 0.71). For anxiety disorders, the AUC for the single diagnosis group was comparable to the AUC of the total group (median: 0.81 vs. 0.83 respectively). Numbers of false positives were high for both mood and anxiety diagnoses, but less so in the single diagnosis group. LIMITATIONS: Time lag between MINI and clinical diagnosis, the availability of only the primary clinical diagnosis, and relatively high severity of the current sample are limitations of the current study. CONCLUSIONS: Agreement between MINI and clinical diagnoses was moderate at best, which partly reflects the difference between the different measures used in the current study.
BACKGROUND: Standardized Diagnostic Interviews (SDIs) such as the Mini International Neuropsychiatric Interview (MINI) are widely used to systematically screen for psychiatric disorders in research. To support generalizability of results to clinical practice, we assessed agreement between the MINI and clinical diagnoses. METHODS: Agreement was assessed in a large, real life dataset (n = 7016) using concordance statistics such as sensitivity, specificity, efficiency and area under the curve (AUC). RESULTS: 41.5% of clinical diagnoses were mood disorders, 26.5% were anxiety disorders. Overall, we found moderate agreement between MINI and clinical diagnoses (median efficiency: 0.92, median AUC: 0.79). For mood disorders, the AUC for all participants showed a range between 0.55 and 0.81 (median: 0.73), and for anxiety disorders the AUC ranged from 0.78 to 0.88 (median: 0.83). The AUC showed better agreement for mood disorders in the single diagnosis group than in the total group (median 0.77 vs. 0.71). For anxiety disorders, the AUC for the single diagnosis group was comparable to the AUC of the total group (median: 0.81 vs. 0.83 respectively). Numbers of false positives were high for both mood and anxiety diagnoses, but less so in the single diagnosis group. LIMITATIONS: Time lag between MINI and clinical diagnosis, the availability of only the primary clinical diagnosis, and relatively high severity of the current sample are limitations of the current study. CONCLUSIONS: Agreement between MINI and clinical diagnoses was moderate at best, which partly reflects the difference between the different measures used in the current study.
Authors: Leticia Coín-Aragüez; Francisco Javier Pavón; Alba Contreras; Adriana-Mariel Gentile; Said Lhamyani; Yolanda De Diego-Otero; Yolanda Casado; Wilfredo Oliva Olivera; Gabriel Olveira; Francisco J Tinahones; Lucía Pérez Costillas; Rajaa El Bekay Journal: Sci Rep Date: 2018-11-30 Impact factor: 4.379
Authors: Mike C Jentsch; Huibert Burger; Marjolein B M Meddens; Lian Beijers; Edwin R van den Heuvel; Marcus J M Meddens; Robert A Schoevers Journal: Int J Mol Sci Date: 2020-04-25 Impact factor: 5.923
Authors: Marie Asp; Daniel Lindqvist; Johan Fernström; Livia Ambrus; Eva Tuninger; Margareta Reis; Åsa Westrin Journal: PLoS One Date: 2020-01-02 Impact factor: 3.240
Authors: Marie-Paule V Austin; Victoria Mule; Dusan Hadzi-Pavlovic; Nicole Reilly Journal: Arch Womens Ment Health Date: 2021-08-05 Impact factor: 3.633