ABSTRACT
Objective: To determine the accuracy of depression screening instruments for older adults in primary care.
Study Design: Systematic review
Data Sources: MEDLINE, PsycINFO (search dates 1966 to January 2002), and the Cochrane database on depression, anxiety and neurosis. We also searched the second Guide to Clinical Preventive Services, the 1993 Agency for Health Care Policy and Research Clinical Practice Guideline on Depression, and recent systematic reviews. Hand-checking of bibliographies and extensive peer review were also used to identify potential articles.
Outcomes Measured: A predefined search strategy targeted only studies of adults aged 65 years or older in primary care or community settings, including long-term care. Articles were included in this review if they reported original data and tested depression screening instruments against a criterion standard, yielding sensitivity and specificity.
Results: Eighteen articles met criteria and are included in this review, representing 9 different screening instruments. The most commonly evaluated were the Geriatric Depression Scale (30-and 15-item versions), the Center for Epidemiologic Studies Depression Scale, and the SelfCARE(D). Differences in the performance of these 3 instruments were minimal; sensitivities ranged from 74% to 100% and specificities ranged from 53% to 98%.
Conclusions: Accurate and feasible screening instruments are available for detecting late-life depression in primary care. More research is needed to determine the accuracy of depression screening instruments for demented individuals, and for those with subthreshold depressive disorders.
When depression is detected and treated in older patients, not only do symptoms subside, but behavior, cognitive functioning, and overall quality of life improve.1 We conducted a systematic review to determine the accuracy of instruments for detecting unrecognized late-life depression in the primary care setting. Several instruments are comparable in sensitivity and specificity, though the 15-item Geriatric Depression Scale is particularly useful in the primary care setting.
Search methods
As a part of a broader review for the US Preventive Services Task Force and the Research Triangle Institute–University of North Carolina at Chapel Hill Evidence-Based Practice Center, we prepared a strategy to identify articles relevant to the accuracy of depression screening instruments for older adults in the primary care setting. We searched for articles in MEDLINE, PsycINFO (search dates 1966 to January 2002), and the Cochrane database on depression, anxiety, and neurosis. We also searched the second Guide to Clinical Preventive Services,2 the 1993 Agency for Health Care Policy and Research (AHCPR) Clinical Practice Guideline on Depression, and recent systematic reviews.3 We also hand-checked bibliographies and used extensive peer review to identify potential articles.
We used the search terms depression, depressive disorder, mass screening, sensitivity and specificity, reproducibility of results, primary health care, ambulatory care, family practice, and the names of common screening and diagnostic instruments used to detect depression. Our search was limited to English-language texts and to ages greater than 65 years.
Inclusion and exclusion criteria
For inclusion, articles must have reported on depression screening in a primary care population of adults aged greater than 65 years. They must have used a criterion standard as comparison and provided information on diagnostic accuracy (usually sensitivity and specificity). Studies performed in the community and in long-term care settings, but not in psychiatric facilities or clinics, were included.
We excluded studies that extracted briefer instruments from the parent version retrospectively; for example, if an investigator evaluated a 5-item version of the Geriatric Depression Scale (GDS), he or she must have defined the specific questions prior to administering the instrument, rather than extracting the 5 items based on posthoc analyses.
The criterion standards must have been commonly accepted, structured or semistructured diagnostic interviews or independent evaluations performed by psychiatrists based on Diagnostic and Statistic Manual of Mental Disorders, revised 3rd or 4th editions (DSM-IIIR, DSM-IV), International Classification of Diseases, 10th ed (ICD-10), or Research Diagnostic Criteria. Our selection criteria are consistent with recognized standards for reviewing diagnostic tests, specifically in eliminating spectrum bias and requiring a criterion standard.4
Review standards
Both authors independently reviewed the abstracts and full articles generated from the searches. Discrepancies about eligibility were resolved by consensus after review of the entire article. For each included study, we extracted information about the screening instrument, the criterion standard, sensitivity and specificity, average age of participants, their dementia status, and the study setting. To further estimate accuracy, we calculated 95% confidence intervals around each measure of sensitivity and specificity. Multiple screening instruments precluded a meaningful meta-analysis of these results.
Results
Our initial search strategy yielded 1325 potential articles, 1269 of which could be eliminated by title review. Of the 56 articles remaining, 38 were eliminated after identifying exclusion criteria in the abstract or the manuscript: 17 because there was no criterion standard, 7 because the setting was not appropriate, 8 because the population was not geriatric, and 6 with varying methodologic exclusions. Eighteen articles met our inclusion criteria and specifically examined the performance of depression screening instruments for older adults in primary care ( Table 1 ).