Eva Ørnbøl

Diagnosing depression in primary care: a Rasch analysis of the major depression inventory

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaperJournal articleResearchpeer-review

OBJECTIVE: This study aims to assess the measurement properties of the Major Depression Inventory (MDI) in a clinical sample of primary care patients.

DESIGN: General practitioners (GPs) handed out the MDI to patients aged 18-65 years on clinical suspicion of depression.

SETTING: Thirty-seven general practices in the Central Denmark Region participated in the study.

PATIENTS: Data for 363 patients (65% females, mean age: 49.8 years, SD: 17.7) consulting their GP were included in the analysis.

MAIN OUTCOME MEASURES: The overall fit to the Rasch model, individual item and person fit, and adequacy of response categories were tested. Statistical tests for local dependency, unidimensionality, differential item functioning, and correct targeting of the scale were performed. The person separation reliability index was calculated. All analyses were performed using RUMM2030 software.

RESULTS: Items 9 and 10 demonstrated misfit to the Rasch model, and all items demonstrated disordered response categories. After modifying the original six-point to a five-point scoring system, ordered response categories were achieved for all 10 items. The MDI items seemed well targeted to the population approached. Model fit was also achieved for core symptoms of depression (items 1-3) and after dichotomization of items according to diagnostic procedure.

CONCLUSION: Despite some minor problems with its measurement structure, the MDI seems to be a valid instrument for identification of depression among adults in primary care. The results support screening for depression based on core symptoms and dichotomization of items according to diagnostic procedure. Key points The Major Depression Inventory (MDI) is widely used for screening, diagnosis and monitoring of depression in general practice. This study demonstrates misfit of items 9 and 10 to the Rasch model and a need to modify the scoring system The findings support screening for depression based on core symptoms and dichotomization of items according to diagnostic procedure. Minor problems with measurement structure should be addressed in future revisions of the MDI.

Original languageEnglish
JournalScandinavian Journal of Primary Health Care
Volume37
Issue1
Pages (from-to)105-112
Number of pages8
ISSN0281-3432
DOIs
Publication statusPublished - Mar 2019

    Research areas

  • Depression, Diagnosis, Mass Screening, Primary Health Care, Psychiatric Status Rating Scales, Psychometrics, MEASUREMENT MODEL, INDEX, VALIDITY, SEVERITY, Depression/diagnosis, Reproducibility of Results, Humans, Middle Aged, Male, Young Adult, Adolescent, Denmark, Mental Status Schedule, Adult, Female, Surveys and Questionnaires, Depressive Disorder, Major/diagnosis

See relations at Aarhus University Citationformats

ID: 143287655