Context Depressive disorders are highly prevalent in the general population,
but recognition and accurate diagnosis are made difficult by the lack of a
simple confirmatory test.
Objective To review the accuracy and precision of depression questionnaires and
the clinical examination for diagnosing clinical depression.
Data Sources We searched the English-language literature from 1970 through July 2000
using MEDLINE, a specialized registry of depression trials, and bibliographies
of selected articles.
Study Selection Case-finding studies were included if they used depression questionnaires
with easy to average literacy requirements, evaluated at least 100 primary
care patients, and compared questionnaire results with accepted diagnostic
criteria for major depression. Eleven questionnaires, ranging in length from
1 to 30 questions, were assessed in 28 published studies. Reliability studies
for the clinical examination required criterion-based diagnoses made by at
least 2 clinicians who interviewed the patient or reviewed a taped examination.
Fourteen studies evaluated interrater reliability.
Data Extraction Pairs of authors independently reviewed articles. For case-finding studies,
quality assessment addressed sample size and whether patients were selected
consecutively or randomly, the criterion standard was administered and interpreted
independently of and blind to the results of the case-finding instrument,
and the proportion of persons receiving the criterion standard assessment
was less than or more than 50% of those approached for criterion standard
assessment. For reliability studies, quality assessment addressed whether
key patient characteristics were described, the interviewers collected clinical
history independently, and diagnoses were made blinded to other clinicians'
Data Synthesis In case-finding studies, average questionnaire administration times
ranged from less than 1 minute to 5 minutes. The median likelihood ratio positive
for major depression was 3.3 (range, 2.3-12.2) and the median likelihood ratio
negative was 0.19 (range, 0.14-0.35). No significant differences between questionnaires
were found. For mental health care professionals using a semistructured interview,
agreement was substantial to almost perfect for major depression (κ
= 0.64-0.93). Nonstandardized interviews yielded somewhat lower agreement
(κ = 0.55-0.74). A single study showed that primary care clinicians
using a semistructured interview have high agreement with mental health care
professionals (κ = 0.71).
Conclusions Multiple, practical questionnaires with reasonable performance characteristics
are available to help clinicians identify and diagnose patients with major
depression. Diagnostic confirmation by mental health care professionals using
a clinical interview or by primary care physicians using a semistructured
interview can be made with high reliability.