Screening for Depression in Older Adults: Recommended Instruments and Considerations for Community-Based Practice
Depression is a treatable medical condition and one of the most frequent causes of emotional distress in older adults. Depressive symptoms are not a normal part of aging, and a recent population-based survey documents the prevalence of current depressive symptoms (defined as an 8-item Patient Health Questionnaire [PHQ-8] score of 10 or higher) as 4.9% among community-dwelling older adults 65 years of age and older.1 Depression in older adults complicates chronic conditions (eg, cardiac disease, diabetes) and reduces quality of life and functional abilities.2 Older adults with depression are less capable of performing self-care behaviors necessary for managing chronic conditions and participating in disease-prevention activities.3,4 Depression negatively affects overall healthcare by affecting patient-provider relationships, decreasing access to care, and increasing costs of care.5 Older adults are also at the greatest risk for suicide as compared with other age cohorts,6 making the detection and treatment of depression even more vital.7 Despite its importance, depression continues to be underrecognized in this population.8 In this article, we review depression screening instruments determined to be appropriate for older adults, and that are reliable and feasible for use by community-based clinicians.
Clinical and Community Recommendations for Depression Screening
Several national organizations have recognized depression as a significant public health issue for the growing number of older adults.9,10 The U.S. Preventive Services Task Force (USPSTF) currently recommends depression screening for adults (including older adults) in clinical practices that have systems in place to assure accurate diagnosis, effective treatment, and follow-up.10 The Guide to Community Preventive Services recommends both home- and clinic-based depression care management for older adults.9 These evidence-based interventions emanate from the depression care management model11 and require use of a validated depression measure to both screen for depression and to monitor the outcomes of the treatment. As noted by the USPSTF, because there is little evidence that screening by itself will lead to effective treatment, screening needs to be done in a setting capable of providing the necessary diagnostic evaluation and treatment.10 Establishing a screening program includes selecting a depression screening instrument that is appropriate for the target older adult population. Screening instruments signify the risk or likelihood of having depression and by themselves do not constitute a diagnosis. The diagnosis and care of older patients with depression requires clinical evaluation by someone trained in evaluating both the criteria symptoms and other medical problems associated with depressive symptoms. It is critical for clinicians who care for older adults to be familiar with appropriate depression screening measures and practical considerations for establishing a depression screening program.
Depression Screening Instruments for Older Adults: Review and Recommendations
Recommendations throughout this article were generated regarding the effectiveness and feasibility of several depression screening instruments that could be used by community-based providers. These recommendations were based on findings from a systematic literature review and evaluation of the review by a 14-member expert panel of researchers and community-based practitioners.12 Detailed study methods are described elsewhere.12,13 The expert panel considered several criteria when making recommendations for appropriate screening instruments for community-based settings. From the systematic literature review, the panel considered the effectiveness of the instruments as reflected in screening performance statistics (eg, sensitivity, specificity, predictive values). As noted in prior reviews,8,14 it also considered the following issues as critical areas to examine relevant to community-based practice: the setting and other work to be performed (eg, whether time administering the tool is prohibitive); population (eg, some ethnic groups do not identify with the questions about life satisfaction); method of administration (in-person or telephone); and purpose for screening (eg, initial screening for depression or assessing treatment outcomes). Similarly, the range of ages of the population might guide the choice, as a clinical practice that wants to screen both younger and older patients might not find geriatric-specific instruments as useful. Using these key dimensions, panel members assessed the strengths and limitations of a set of depression screening measures using a rating process. They assessed 28 screening instruments from 53 articles published between 1981 and 2005.
Based on the ratings of the panel, six instruments were recommended and are described below. It is important to note that several single-item screens were also considered but eliminated. They were not recommended because of limited screening effectiveness. Additionally, they were determined not to be appropriate outcome measures for any of the recommended evidence-based clinical interventions. It is also worth noting that the PHQ-9 was not included in this review because of inadequate data at the time of the review to evaluate the effectiveness of the PHQ-9 in community-based older adults.15 In that study, authors reported an area under the curve (AUC) of 0.96 for subjects over the age of 60 years but did not report sensitivity or specificity data for these participants. Several recent studies suggest that the PHQ-9 is an effective tool for depression screening and monitoring because of its usefulness in diagnosing major depressive disorder, based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition-Text Revision (DSM-IV-TR) criteria, and for quantifying severity.16,17 A list of all reviewed instruments and their respective citations are available from the study authors (Lesley Steinman, firstname.lastname@example.org).
Recommended Screening Instruments for Community Practice
The six instruments that were recommended include the 20-item Center for Epidemiologic Studies Depression Scale (CES-D),18-29 the CES-D-10 (a 10-item version),19,27,30 the 30-item Geriatric Depression Scale (GDS),19,23,27,29,31-37 the GDS-15 (a 15-item version),19,27,35-53 the two-item PHQ (PHQ-2),19,27,54 and the 20-item Zung Self-Rating Depression Scale (ZSDS)32,34,50,55 (see the Table for a summary of the characteristics of these screening instruments). The overall recommendations are first described followed by examination of the instruments along two major dimensions: the effectiveness of the screening instrument and considerations for clinical practice (ie, time required to screen, diagnosing major depression, screening in specific racial and ethnic groups, responsiveness to change over time, screening cognitively impaired older adults).
The public health significance of depression in older adults suggests that many settings that traditionally have not focused on depression need to be engaged in this effort. The panel avoided recommending a single instrument because they recognized the importance of selecting an instrument for use that was appropriate and feasible for the circumstances of a particular clinical practice. The CES-D-10 and the GDS-15 were strongly recommended for use to screen for depression in community-based older adults. The PHQ-2 was also strongly recommended as a screening instrument by most of the expert panel, given its brevity and assessment of depression-criteria symptoms, but it was viewed as being of mixed effectiveness for overall use as a screening and treatment-outcome measure. The brevity of the shorter instruments makes them particularly attractive to providers who address depression as one of several other social and healthcare concerns. The CES-D, GDS, and ZSDS were recommended as well, with less strength because of the additional length of these longer forms. However, these longer versions may be more appropriate for monitoring and assessing treatment response and outcomes.
Effectiveness of Recommended Screening Instruments
The effectiveness of the screening instruments was evaluated based on statistics to assess the accuracy of the instrument. Sensitivity and specificity, as well as AUC, are the most widely used statistics utilized to describe a test. The screening statistics for most of the recommended instruments were similar, suggesting comparable effectiveness as case-finding instruments. The sensitivity of the recommended instruments, or how often the instrument detects a case of depression when the patient truly has depression, ranged from 60% to 100%. The specificity, or how often the instrument accurately indicates the absence of depression when the patient is not depressed, varied from 64% to 87%. The AUC, a measure of how accurate a screening test is at assessing the absence or presence of depression, was reported for four instruments, ranging from 0.5 to 0.94. AUC values range from ≤ 0.5 (no discriminatory ability) to 1.0, with values between 0.7 and 0.9 being in the acceptable-to-excellent range.56 While the screening statistics were fairly similar between instruments, the slight variations found may relate to differences in depression diagnoses, subjects, and settings.
Considerations for Community Practice
In addition to accuracy, several other factors should be considered in determining the use of depression screening instruments. Characteristics to consider include how much time is required to complete the instrument, how closely the items of the instrument correspond to the criteria for major depression, whether there is information on use of the instrument among specific racial and/or ethnic groups, responsiveness of the instrument to change over time, and the appropriateness of use for older adults with cognitive impairment.
Time Required to Screen
The GDS-15 and CES-D-10 take approximately 3-5 minutes to complete. The ZSDS takes approximately 10 minutes because of the number of items. Many providers as well as patients have noted that the “yes” and “no” format of the GDS-15 and CES-D-10 is easier to use and follow when the questions are delivered orally. In contrast, some older adults may find choosing a simple “yes” or “no” answer challenging because their experiences are perceived as more complex and not easily dichotomized. The four-item response choices for the ZSDS and PHQ-2 allow for varied levels of responses, but they may be challenging to patients who are cognitively impaired and have trouble remembering the four response options when given orally. Several options that can assist when using these instruments include using “cue cards” with written response options for the ZSDS, and using the phrase “most of the days” when asking GDS or CES-D-10 items if patients struggle with the “yes” or “no” response options.
Diagnosing Major Depression
Regardless of the length of the screening instrument, clinicians are still faced with the problem that screening instruments do not provide a diagnosis by themselves. For a clinician whose scope of practice includes making a diagnosis of major depression, none of the scales that were highly recommended fully covers each of the nine criteria defined for major depression in the DSM-IV-TR. The GDS-15 indirectly addresses five of the nine items, namely anhedonia, sad mood, worthlessness, energy, and concentration. Thus, a clinician interested in making a diagnosis will need to inquire about the nine criteria symptoms in addition to the 15 GDS items. Similarly, the CES-D-10 covers only two to three of the criteria symptoms (depressed mood and sleep change). The 20 items of the ZSDS at least partially address each of the DSM-IV-TR criteria symptoms. The PHQ-2 directly addresses the two “cardinal” symptoms for major depression, which are depressed mood and anhedonia, one of which must be present for a diagnosis. It is also notable that the PHQ-9 includes the items from the PHQ-2 and each of the remaining seven symptoms of major depression. When a qualified clinician performs it, few additional questions are necessary to establish the presence of the DSM-IV-TR symptoms.
Screening in Specific Racial and Ethnic Populations
The major focus in this area is on the translation of these measures to other languages. The GDS has been translated into Chinese, Spanish, Japanese, and Dutch; the GDS-15 has been translated into Chinese and Spanish. A study of African-Caribbean residents of the United Kingdom yielded acceptable sensitivity and specificity45; however, the instrument was found to have lower sensitivity when used with African Americans in the United States.57 The CES-D-10 has been translated into Spanish for Spanish-speaking populations in the United States and Spain and was found to have good accuracy. In translating the CES-D-10 into Chinese, the authors deviated from the “yes” or “no” response format but reported that the instrument performed adequately. The ZSDS has been translated into Spanish and Chinese languages, and these have shown good sensitivity and specificity. Of course, translation into other languages is only one step in determining the cultural appropriateness of the instruments, as other issues such as cultural beliefs regarding mental illness, mood, and stigma associated with mental illness must also be addressed. Determining the validity of depression screening instruments in other cultural groups is complex and there is growing awareness that symptom recognition, meaning, and report are expressed differently by older persons of different cultures.58
Responsiveness to Change Over Time
Ideally, a screening instrument would be used to identify older adults with depression, monitor change in depression status over time, and be responsive to effective treatment interventions over time. Very brief instruments such as the PHQ-2 may be too restricted in range to accurately reflect change over time with or without treatment being given. The GDS-15 has been found to be acceptably responsive to change over time, as has the ZSDS, which was used in many early studies examining the effectiveness of antidepressants. The CES-D-10 has less published information on responsiveness to change. Given that the range of scores in the CES-D-10 can vary between 0 and 20, comparable to the range for the GDS-15, it may also be responsive to change over time; however, this requires further testing.
Screening Older Adults Who Are Cognitively Impaired
Cognitive impairment complicates screening and treatment of older adults. Some experts advise that when conducting depression screening for older adults living in institutional settings such as nursing homes, clinicians should avoid using self-report instruments in persons with more than mild cognitive impairment.59 Most studies examined in this review excluded persons with cognitive impairment, often screening them out if they scored less than 24 on the Folstein Mini-Mental State Examination (MMSE). One study, using the GDS, found no difference in the screening characteristics (eg, sensitivity) of persons who scored between 24 to 28 versus those who scored greater than 28 on the MMSE.46
Barriers and Opportunities for Screening in Community-Based Settings
Even once an appropriate depression screening instrument is selected, challenges in implementing the screening activities and response to the results of the screening need to be considered. These include identifying and training staff to conduct screening, finding the most opportune time and settings for screening, and overcoming the barriers encountered when trying to link older adults with depression to treatment.
Given the prevalence of depression in older adults, providers who serve as safety net providers for high-risk older adults are likely to encounter significant numbers of individuals with depression. Thus, these are naturally occurring opportunities to screen for depression in home and community-based care settings responsible for comprehensive assessments to guide care planning. Similarly, home healthcare providers, senior housing providers, and adult day health programs are potentially high-yield places to establish depression screening and evidence-based treatment programs.
These different settings have staff of varying backgrounds, and so the issue of training staff to administer depression screening instruments often needs to be addressed. The screening instruments described and recommended above were all designed and validated as self-report instruments. Thus, for some settings, minimal training is required if the older adults being targeted have adequate vision, reading, writing, and cognitive abilities to complete the written forms. This facilitates having the instruments completed while the patient is waiting to be seen by a staff member.
To the extent that the screens are sometimes incorporated into more comprehensive assessments being performed by staff, it is quite feasible to have the questions read aloud by an interviewer. In these circumstances, staff such as case managers, social workers, and home health nurses can be trained relatively quickly to administer these screening instruments. Training can be performed in group settings, and, typically with use of role play or video presentation of the screening form, staff can proceed relatively quickly to using the instruments. Reconvening the group with an experienced clinician or trainer to answer questions and clarify issues that emerge only after staff has had time using the instrument can often be helpful. Training tends to focus on helping staff acquire skills involved in doing standardized administration of valid tools, which may differ from their typical assessment involving informal inquiry or observation to identify a patient’s needs or level of functioning. Staff members need to learn verbal strategies for respecting a patient’s need to “tell his/her story,” while still being able to obtain structured answers. Training approaches have been successfully implemented that teach staff how to help patients provide a definite choice when they are unsure of the precise answer, and to help staff learn to ask the questions and record the answers with minimal bias from their own perceptions about the patient’s depression status.60,61 Although it is true that some patients and staff are in denial regarding depression and mental illness, in most instances, the judgment of the patient prevails.
Healthcare providers also need to decide when to perform depression screening in the course of existing patient assessments. Typically, keeping the questions grouped together facilitates scoring. It is also relatively easy to orient the patient to the shift from other topics to the depression items by informing him/her that, “Now I’ll be asking some questions about your mood.” The shorter versions of the GDS and CES-D were strongly recommended, largely out of sensitivity to the fact that these questions often must be integrated into lengthy overall assessments and create another area that healthcare providers must respond to when significant levels of depression are found.
Consideration needs to be given to follow-up on repeated screenings, as often patients’ depression status will change over time. We are unaware of any evidence-based guidelines for frequency of screening older adults whose initial screens suggest no clinically relevant depression. In programs where the healthcare provider’s mission and priority will always be higher for other domains (eg, assuring safe housing or providing access to urgent in-home or emergency medical care), it may be more appropriate to routinely perform depression screening during a subsequent visit (vs the first contact) so that the staff member responsible for acting upon the results will not have too many competing demands.
Finally, to the extent that some screening instruments ask specifically about suicidal ideation, the healthcare provider must have a protocol for how to respond. Often, referring the patient to the existing emergency mental health crisis system is an option consistent with the community healthcare provider’s standard of care if a patient or family member had volunteered the information; this approach would also be appropriate in the event of elicitation of this information from depression screening. The evidence-based programs for treatment of depression also have protocols that can be referenced and used by healthcare providers performing screening.62
Several depression screening instruments had sufficient data from community settings to be recommended by an expert panel as appropriate for use by clinicians caring for community-based older adults. In keeping with the recommendations of the USPSTF, screening is of clear benefit only in settings that have the means to offer accurate diagnosis, effective treatment, and monitoring over time. Clinicians working in primary care clinics as well as home-based service-delivery organizations should have access to these programs. It is not possible to recommend any one instrument. This approach is consistent with other earlier reviews.19 Specifically, clinicians need to consider the population to be screened and the individual characteristics of the instruments in terms of time required, ease of administration, cultural appropriateness, responsiveness to change over time, and the ability of older adults with cognitive impairment to complete the instrument when deciding which instrument to use.
The authors report no relevant financial relationships.
Acknowledgments This research was funded by the Centers for Disease Control and Prevention (CDC) Healthy Aging Program through the CDC Prevention Research Program, Special Interest Project grant (U48 DP0000050) to the University of Washington Health Promotion Research Center. This material is based upon work supported in part by the Houston Center for Quality of Care & Utilization Studies, Health Services Research and Development Service, Office of Research and Development, Department of Veterans Affairs (HFP90-020). The views expressed are those of the authors and not necessarily those of the Department of Veterans Affairs. The authors wish to thank the Late Life Depression Special Interest Project expert panelists, and would like to acknowledge Lynda Anderson, PhD, Rosemarie Kobau, MPH, Jim Laditka, PhD, Jason Lang, MPH, MS, Kahla Mehta, MD, and Jaya Rao, MD, MPH, for their contributions to the expert panel. Special thanks to Yuki Durham, Barb Williams, and Antonia Bennett for their assistance with the literature search and data abstraction for this study.
Dr. Snowden is Associate Professor, Department of Psychiatry and Behavioral Sciences, University of Washington, and Medical Director, Geriatric Psychiatry Services Program, Harborview Medical Center, Seattle, WA; Ms. Steinman is Research Coordinator, Health Promotion Research Center, University of Washington; Dr. Frederick is Geriatric Psychiatrist, Virginia Mason Medical Center, Seattle, WA; and Ms. Wilson is Assistant Professor, Baylor College of Medicine, and Assistant Director for Program Development, Huffington Center on Aging, Houston, TX.