Research News

Uddin, Md. Zahir 

Supervised by Professor Dr. M. Kamruzzaman Mozumder & Professor Dr. Graham E. Powell

Abstract:

Psychological support for patients with a mental health condition is a relatively new addition in Bangladesh. There is an immense need for contextually sensitive instruments for screening and measuring psychological problems. The Depression Scale developed two decades ago has been in productive use by professionals, but there is an emerging need for an updated scale. The present study was therefore designed to revise the Depression Scale. A ground up approach was taken for the revision to ensure the new version became a contextually sensitive modern tool for assessing depression. Multi phased mixed method design was used for the present research. Two exploratory studies using a qualitative approach were carried out at the first phase which was followed by a quantitative study. A phenomenological study was done in the first phase to understand the essence of depression in the specific Bangladesh context, where a series of in-depth interviews and focus group discussions were conducted with patients, caregivers, and professionals across four divisions of Bangladesh. Rich descriptions on the symptoms of depression generated from the phenomenological exploration contributed in the development of contextually relevant items in the later part of this research. Additionally, a small qualitative exploration was carried out to understand the expectations of the stakeholders about scale for assessing depression. Psychiatric patients and mental health professionals were interviewed and some of their suggestions were later incorporated into devising the items and designing the scale. A pool of 282 items was developed, items generated from the phenomenological study of depression, clinical notes from mental health professionals, the existing Depression Scale developed by the current researcher and a review of the literature on depression. Items for the experimental tryout of the revised scale were selected through four stages of evaluation and revision by mental health experts. The 24 items identified in IV this rigorous evaluation went through some minor linguistic revision in the field testing. The final selection of items was made through item analysis of the 5th draft of the scale. All the 24 initially selected items passed the stringent dual criterion, namely ability to discriminate depressed patients from healthy adult (F = 35.21 to 359.15, p < .01) and corrected item-total correlation (r = .37 to .81, p < .01). The finalized 24-item depression scale underwent a series of testing for validity and reliability. The findings demonstrated high reliability of the newly developed scale in terms of both internal consistency (Cronbach‟s alpha = .96) and stability (test-retest correlation over two weeks gap (r = .87 p < .01). Content validity of the scale was established though the process of its development. Criterion related validity of the newly developed depression scale was established with the Bengali version of the widely recognized BDI-II (r = .94, p < .01) and diagnosis of depression as an external criterion (the scale could differentiate between depressed and healthy respondents; F =182.63, p < .01). Convergent validity was estimated for the new depression scale by looking at correlations with related constructs (with hopelessness, r =.76, p < .01; on the Bengali version of the Beck Hopelessness Scale and r = .59, p < .01 on the subjective rating; with stress, r =.75, p < .01 on the Bengali version of Perceived Stress Scale and r =.78, p < .01 on subjective rating; with well-being r = -.77, p < .01 on the Bengali version of WHO Five Well-Being Index; with intelligence r = -.37, p < .01 on the matrix reasoning sub-test of Wechsler Abbreviated Scale of Intelligence; with symptoms of depression like fatigue, r = .65, p < .01 on subjective rating; with desire to commit suicide, r = .64, p < .01 on subjective rating; with reduction of interest, r = .61, p < .01 on subjective rating; with low mood, r = .83, p < .01 on subjective rating; with level of confidence. r = -.60, p < .01 with subjective rating; and with impairment of functioning, r = .32, p < .05 and .42, p < .01 on V subjective ratings). The scale did not correlate with subjective rating of occupational functioning and family relationship. For the new depression scale, two types of norms were estimated: severity and screening norm. Four levels of severity namely mild, moderate, severe and profound, based on percentile points were considered for severity norm. Sensitivity and specificity were calculated for different cut off scores of the scale. The optimal cut off score with best combination of sensitivity (89%) and specificity (88%) was found to be at scale score of 25. However, researchers can use a different cutoff value if they need to use the tool with higher sensitivity or specificity depending on their purpose. Based on the sensitivity and specificity calculation, the scale has good diagnostic performance as reflected in the Receiver Operator Characteristic (ROC) curve. The area under the ROC curve for the new depression scale was .96, which is considered an outstanding diagnostic performance for a psychological scale. The newly developed depression scale was constructed following proper steps of scale construction and it has good reliability and validity. The initial normative analysis demonstrated high sensitivity and specificity. The scale is now ready to use in clinical and research settings.

Description:

This dissertation submitted to the Department of Clinical Psychology, University of Dhaka in partial fulfilment of requirements of degree of Doctor of Philosophy in Clinical Psychology.