Abstract
BACKGROUND: The Major Depression Inventory (MDI) is widely used in Danish general practice as a screening tool to assess depression in symptomatic patients. Nevertheless, no validation studies of the MDI have been performed. The aim of this study was to validate the web-based version of the MDI against a fully structured telephone interview in a population selected on clinical suspicion of depression (ie, presence of two or three core symptoms of depression) in general practice.
MATERIALS AND METHODS: General practitioners (GPs) invited consecutive persons suspected of depression to complete the web-based MDI in a primary care setting. The validation was based on the Munich-Composite International Diagnostic Interview (M-CIDI) by phone. GPs in the 22 practices in our study included 132 persons suspected of depression. Depression was rated as yes/no according to the MDI and M-CIDI. Sensitivity, specificity, and positive predictive value of the International Classification of Diseases, Tenth Revision (ICD-10) algorithms of the MDI were examined.
RESULTS: According to the M-CIDI interview, 87.9% of the included population was depressed and 64.4% was severely depressed. According to the MDI scale, 59.1% of the population was depressed and 31.8% was severely depressed. The sensitivity of the MDI for depression was 62.1% (95% confidence interval [95% CI]: 52.6-70.9) and the specificity was 62.5% (95% CI: 35.4-84.8). The sensitivity for severe depression was 42.2% (95% CI: 30.6-52.4) and the specificity was 85.1% (95% CI: 71.7-93.8). The receiver operating curve showed an area under the curve of 0.66 (95% CI: 0.52-0.81) for any depression and of 0.72 (95% CI: 0.63-0.81) for severe depression.
CONCLUSION: The MDI is a conservative instrument for diagnosing ICD-10 depression in a clinical setting compared to the M-CIDI interview. Only a few false-positive diagnoses were identified when the MDI was used on clinical suspicion of depression.
Originalsprog | Engelsk |
---|---|
Tidsskrift | Clinical Epidemiology |
Vol/bind | 9 |
Sider (fra-til) | 355-365 |
Antal sider | 11 |
ISSN | 1179-1349 |
DOI | |
Status | Udgivet - 2017 |