The Axial Diagnostic Criteria for Depression. Development, construct and predictive validity and reliability

Abstract

Introduction. The authors have developed a new axial diagnostic criterion for depression (ADCD) made up of seven items: mood, motivation/interest, impulse/drive, liking/pleasure, daily job, energy and different quality. They have aimed to examine its predictive validity and reliability, psychometric properties and constructive validity. There are few studies that have examined the psychometric properties of other diagnostic criteria for depression currently in use.

Material and methods. A total of 111 psychiatric outpatients who attended an out-patient clinic consecutively were interviewed. Sixty met the ICD-10 criteria for depressive episode and 51 formed a part of the control group: non-depressed psychiatric outpatients. For the interview, the authors used a brief self-administered questionnaire (IDASD) in which the patients indicated how they felt. Each item had a Visual Analogue Scale so that the subjects could quantify their answers.

Results. Four or more items are needed for the ADCD to correctly diagnose depression. At least two of these should belong to a group of three items that were extracted using a discriminant function (mood, energy and different quality).

The ADCD constructed in this way has a 0.93 sensitivity and 0.82 specificity, with a kappa reliability of 0.76 and a proportion of total cases correctly classified ranging from 88% to 93%. Specificity reaches up to 0.92 when the control group is formed exclusively by symptom-free psychiatric outpatients.

A factor analysis reveals that the ADCD is a one-dimensional model that has good construct validity (0.69). It also has good alpha reliability (alpha = 0.92), elevated consistency of the two halves of the test (R = 0.91) and a high test-retest correlation (r = 0.67).

The ADCD diagnostic agreement between two psychiatrists who use the IDASD as a data source is very high (kappa: 1.00).

Conclusions. The ADCD/IDASD system offers a valid and reliable procedure to diagnose depression. It also has an excellent internal architecture, good construct validity and internal consistency. These data are much more than what can be said about other more used diagnostic criteria, which lack this information.