Review of Hamilton Depression Rating Scale

Abstract

Hamilton Depression Rating Scale which is observer-rating scale is used to assess severity and type of depression. The aim of this paper is to review and evaluate psychometric properties of HDRS. Six articles which used HDRS are analyzed in terms of reliability and validity evidence of HDRS.

Hamilton Depression Rating Scale is a standard scale used for measurement of the severity of depression (Hamilton, 1960). In fact Hamilton stated that this scale has been designed for using on patients who already have diagnosis of affective disorder of depressive type (1960).HDRS is administered by evaluators, mostly clinicians, through interview. Although this scale was initially administered through unstructured interview to obtain information from patients, modified version of the test (Structured Interview Guide for the HDRS) was developed by BW Williams to standardize the administration of the scale. Original version of HDRS includes 17 variables measured either on 5-point or 3-point scale, and total score of HDRS with 17 variables can range from 0 to 53.Moreover, on 0-4 point scale, scores that are greater than 2 and on 0-2 point scale, scores that are greater than 1 indicate the presence of that symptom (Hamdi, Amin, Abou-Saleh, 1997). In other words, increasing the total score indicates increasing the severity of depression. Some of the items on the scale are defined by categories which have increasing intensity, whereas others are planned as a number of equal-valued items (Hamilton, 1960). These 17 variables are designed to assess conditions respectively: depressed mood, guilt, suicide, insomnia initial, insomnia middle, insomnia delayed, work and interests, retardation, agitation, anxiety psychic, anxiety somatic, somatic gastrointestinal, somatic general, genital, hypochondriasis, insight and loss of weight. In fact , The HDRS has 4 additional items (diurnal variation, derealization, paranoid symptoms, obsessional symptoms) but these 4 items are excluded from quantitative measurement according to Hamilton, because, diurnal variation defines the type of depression and other three items are not seen frequently in the population (1960).

The purposes of the scale are to assess severity of depression and to measure change in depressive symptoms because of therapy or antidepressant trials (Hammond, 1998; Akdemir et al., 2001). The scale was devised to be used on adults and its administration takes about 25 minutes.

HDRS is a multi-factor test. Hamilton (1960) examined the factor saturations for the scale, and he found four main 4 factors: retarded depression, agitated depression, anxiety reaction and last factor being untitled. Other studies on factor structure of HDRS support multi-dimensional structure of the scale but numbers of the factors change from 4 to 7.

There are several empirical articles in literature which evaluated HDRS in terms of reliability and validity. I want to show their results under two main topics: reliability analyses and validity analyses.

Reliability Analyses

It is possible to estimate reliability through four simple methods: test-retest reliability, alternate forms reliability, split-half reliability and internal consistency reliability. The estimating reliability for this scale was usually examined through internal consistency reliability, test-retest reliability and split half reliability. Moreover, it is important to test inter-rater reliability for such tests assessed by clinician-administered rating scales. For this reason, inter rater reliability for HDRS were examined in several studies.

Internal Consistency Reliability

Gençöz, Gençöz and Soykan (2007) used 17 items of HDRS to assess depression in 45 endstage renal disease (ESRD) patients undergoing hemodialysis in Turkey.They administered the scale twice, at baseline and next month, and they examined via Cronbach's alpha. The result of the study shows that internal consistency reliability of the test is .83 at baseline assessment and .81 at follow-up one. (Gençöz et al., 2007). HDRS has high coefficient alpha when used on ESRD patients undergoing HD treatment, so it is reliable to use for this population.

Hammond (1998) used HDRS with 17 items to test reliability of the scale on 100 elderly physically ill patients. Hammond found initial coefficient alpha was 0.46, and then he excluded 11 items from the scale to adapt the elderly physically ill population (1998) because according to Hammond (1998), some conditions in the HDRS are seen not as depressive symptoms but normal among those people such as loss of libido. Hence, Hammond used HDRS with 6 items that are middle and late insomnia, agitation, physic and somatic anxiety and hypochondriasis (1998). This six-item scale revealed 0.60 of coefficient alpha. Therefore, HDRS with 17 items showed low reliability among this population.

Hamdi, Amin and Abou-Saleh used both HDRS-21 and HDRS-17 on 100 depressed patients residing in the UAE to evaluate reliability of the scale (1998). They found that Cronbach's alpha is .51, so reliability of the HDRS-21 is low reliability when used on depressed patients who live in the UAE (Hamdi et. Al., 1998).Moreover, they conducted factor analysis to determine whether the scale homogeneous construct or not. The results of their study showed that HDRS has contained seven factors, so it assesses heterogeneous aspects of depression (Hamdi et al., 1998). They examined reliability Croanbach's alpha respectively for each factor (except last one): .65, .52, .49, .52, .47 and .39. (Hamdi et al., 1998). Because low reliability alpha levels of especially last two factors, HDRS would mostly be invalid to measure "somatization components" and "psychotic symptoms".

Akdemir et al. used Turkish version of the Hamilton Depression Rating Scale on 94 patients who suffer from depressive mood disorder to test reliability and validity of the scale (2001). They found that Cronbach's alpha is .75, so Turkish version of HDRS-17 shows moderate internal consistency reliability.

Kunugi et al. (2013) used Japanese IVR-HDRS-17 items on 60 Japanese patients who suffer from major depressive episode or dysthymia to test reliability and validity of IVR- HDRS program. In this study, they used two psychologists, one clinician and interactive voice response technology (IVR) to administer HDRS-17. They investigated the internal consistency reliability for clinician HDRS, psychologists HDRS and IVR-HDRS (day 1 and day 2) assessment, and Cronbach's alphas were obtained by evaluators is respectively: .77, .79,