Are there acceptable routine screening tests?

Instruments for routine case-finding in primary care settings must be of acceptable quality, brief and easy to use. Presumably the patients would complete the test while waiting to see their physician. Most self-administered tests have been designed for routine screening purposes and not as diagnostic aids. Some tests have been developed for settings in which the prevalence of depression is higher than in primary care. Because a lower prevalence decreases the positive predictive value, instruments may appear less effective when used in the primary care setting.

In reviewing the literature on self-administered instruments we applied the following criteria.

• Is there a sound conceptual basis on which the instrument was developed, and is this credible?

• Is it intended to detect solely depression or other emotional problems as well?

• Does it measure the severity of depression?

• Has it been tested across a broad spectrum of patients, especially those in primary care or in the community?

• Is it feasible and easy to apply in a clinical setting?

• Is there evidence to support its sensitivity to changes in clinical status?

• Is there evidence to support its reliability?

• Is there evidence that it is a valid measure for the early detection of depression?

A search of MEDLINE and the Science Citation Index identified six instruments that met those criteria sufficiently well to warrant a more in-depth review. A review by Wells provided additional information.

Beck Depression Inventory (BDI)

The BDI was designed specifically for use in clinical settings to identify depression as well as to measure its depth. The test has been used in many different situations with people suffering from various disabilities. The original 21-item version has been scaled down to 13 items. It is generally self-administered, takes less than 10 minutes to complete and asks people to describe their emotional state during the previous week. The total score allows classification into five levels, from normal mental status to severe depression.

The methodologic properties have been extensively assessed, and the performance characteristics are quite good. The BDI (Beck Depression Inventory) is sensitive to change over time with treatment. Reliability studies have shown a coefficient α of 0.73 to 0.92, and concurrent validity studies that compared the BDI with clinical ratings and other scales have shown correlation coefficients of 0.60 to 0.74.

Center for Epidemiologic Studies Depression Scale (CES-D)

This instrument was developed to measure depressive symptoms in the general population. It has been used in clinical settings as well. High scores reflect the distress that accompanies depression but are not diagnostic of depression. The 20-item scale is simple to understand and easy to administer.

The CES-D is sensitive to change in clinical status. Reliability studies have reported a coefficient α of 0.84 to 0.90. Validity studies, in both clinical and community populations, have shown coefficients of correlation with other measures, including structured psychiatric interviews, of 0.51 to 0.89. In studies involving depressed patients the sensitivity has been as high as 91% and the specificity 56%. However, in the general population the sensitivity was only 64%. This relates in part to the instrument’s ability to identify people with symptoms suggestive of psychologic distress rather than to detect clinical depression.

Zung Self-Assessment Depression Scale (SDS)

This brief, self-administered instrument measures the presence and the severity of depressive symptoms but is not a diagnostic tool. Although the SDS is popular there has been almost no empiric testing of the selected items. The heavy weighting of items addressing physical symptoms makes its use difficult among patients whose symptoms are caused by physical disorders.

The SDS (Zung Self-Assessment Depression Scale) is sensitive to change in clinical status, but its reliability has been inadequately evaluated. Validity studies comparing the SDS with other scales, as well as global ratings by psychiatrists, have revealed correlation coefficients of 0.40 to 0.80.

General Health Questionnaire (GHQ)

This instrument, designed specifically to detect psychiatric disorders in primary care, has a subscore for depression. It has been widely used in primary care settings. The original version had 60 questions, but shorter versions of 30, 28, 20 and 12 items have been developed. Areas assessed by the original version included depression and anxiety, social functioning, psychophysiologic symptoms, general health and vague aches and pains.

The internal consistency of the GHQ has been from 80% to 90% and the coefficients of correlation with global clinical assessments of psychopathology 0.55 to 0.83. The overall sensitivity has been about 68% and the specificity 81%.

Hopkins Symptom Checklist

The current 25-question version, shortened from the original 90-question version, measures only depression and anxiety and has been designed for use in primary care settings. The original version was intended for use in studies of psychotherapy and chemotherapy among psychiatric patients. The instrument is sensitive to changes in clinical status. Studies assessing its measurement properties have demonstrated an internal consistency as high as 95%, and coefficients of correlation with psychiatrists’ global ratings of severity of depression have been 0.70 to 0.77. The correlation has been shown to be lower for scores of severely depressed patients.

Mental Health Inventory (MHI)

The MHI was developed to measure mental health in the Rand Health Insurance Experiment. The original 38-question format has not been changed, and an extremely brief 5-question version is available. The internal consistency has been in the range of 96%. Although the MHI is not specific for depression, studies have shown that it does predict the future use of mental health services.

Hospital Anxiety and Depression Scale (HADS)

Designed for use in nonpsychiatric hospital clinics this brief scale identifies only depression and anxiety and measures their severity. It consists of seven items for each component. Only psychiatric symptoms are targetted; thus, symptoms that might be attributed to physical or emotional problems are avoided. The sensitivity to change has not been assessed. Internal consistency has been 30% to 60%; coefficients of correlation with other instruments, as well as clinical assessment, have been as high as 0.79 for the depression component. Although short and feasible, the HADS has not been extensively tested or widely used in North America.

Related posts:

1. Early detection of depression. Part 4 Special problems The early detection of depression is much more complex and challenging than the detection of other conditions seen in the primary care physician’s office. Perhaps the most significant reason for this is how the natural history of depression differs from that...
2. Early detection of depression. Part 1 Depression is common and can usually be treated effectively. However, in many cases physicians may not recognize depression, especially in its early stages. Because the burden of suffering can be high, including death through suicide, early recognition and treatment are worth while. Such...
3. Early detection of depression. Part 3 Effectiveness of early detection The critical question with any intervention for early detection and treatment is whether it does more good than harm as compared with the results of later diagnosis and intervention. The appropriate research tool for answering such a question is...
4. Managing Resistant Depression. Part 2 Definition Treatment-resistant depression is difficult to define. The term embraces many variables that can influence response to treatment. Treatment-resistant depression must be differentiated from a chronic depressive illness (i.e., present for many years before the onset of treatment), and from a recurrent depression...
5. Vitamin B12 Deficiency Linked to Depression in Older Women Brenda Penninx, Ph.D., and colleagues from The Johns Hopkins University, the University of Colorado in Denver and Vrije University in the Netherlands have found a positive correlation between vitamin B12 (cobalamin) deficiency and depression in physically disabled elderly women. Data from the Women’s...