Clinical Scholarship

Comparing Various Short-Form Geriatric Depression Scales Leads to the GDS-5/15

Sandra Kenney Weeks, Paul E. McGann, Teresa King Michaels, Brenda W.J.H. Penninx
Purpose: To compare three published short GDS scales and to identify a valid and reliable short-form alternative to the 15-item Geriatric Depression Scale. Design: Comparative validation study via retrospective chart review of 816 acute care patients in an 830-bed academic medical center in the USA in 2001. Methods: Data of the 15-item Geriatric Depression Scale, the Mini-Mental State Examination, and demographic data were extracted from medical records after patient discharge. Three scales: the DAth GDS-4, van Marwijk GDS-4, and Hoyl GDS-5, were compared to the 15item Geriatric Depression Scale. Results: The Hoyl 5-item version showed the highest sensitivity (97.9%). Concern for GDS-5 false positives when compared to the 15-item GDS (specificity 72.7%) led to re-ordering the 15 GDS items into a new two-tiered instrument, the GDS-5/15. In this study of 816 older adult inpatients, 60% were screened as not depressed using the first 5 items on the GDS5/15, leaving 40% for continued screening and completion of all 15 GDS items. Conclusions: A shorter screening tool might encourage more providers to add depression screening to routine health care visits. The GDS-5/15 is an alternative screening tool.

JOURNAL

OF

NURSING SCHOLARSHIP, 2003; 35:2, 133-137. 2003 SIGMA THETA TAU INTERNATIONAL.

[Key words: GDS, depression, older adults]

* Health care providers should identify depression so that patients may receive adequate treatment and enjoy better outcomes. A standardized rating scale such as the Geriatric Depression Scale (GDS) is one method providers use to screen for depression in the older population (Sheikh & Yesavage, 1986; Yesavage et al., 1983). Yesavage and colleagues introduced the GDS in 1983. This 30-item self-report instrument was designed to meet the need for a reliable screening test for depression in elderly populations that would be simple to administer and not require the time or skills of a trained interviewer (p. 45). The 30item GDS was subsequently reduced to 15 items (Sheikh & Yesavage, 1986).

inshaw called for research to identify older adults who are at high risk for loss of independence in daily activities or who are approaching increased dependence (2000, p. 119). One risk factor for loss of independence is undiagnosed, untreated depression. Depression is one of the most common illnesses present in older adults. Current estimates indicate that 15%-20% of older adults experience significant depressive symptoms (Beekman, Copeland, & Prince, 1999). Depression increases both health care use and costs (Badger, McNiece, & Gagan, 2000) and leads to functional decline and loss of independence (Espiritu et al., 2001). Depression often occurs concurrently with other diseases in hospitalized adults, with major depression in 10%-20% and minor depression in another 20%-30% of older in-patients (Onder et al., 2000). Evidence indicates that comorbid depression is often misdiagnosed, undertreated, or not diagnosed at all in critically ill patients (Whall & Hoes-Gurevich, 1999). Although safe and cost-effective treatments are available to treat depression, comorbid depression remains underdiagnosed and undertreated in inpatients and comorbidity adversely affects patient outcomes (Kaye, Morton, Bowcutt, & Maupin, 2000). Left untreated, depression can lead to increased morbidity and mortality among older adults (Penninx et al., 1999).

Sandra Kenney Weeks, RN, MSN, CRRN, Theta Sigma, Associate Director of Nursing, North Carolina Baptist Hospital; Paul E. McGann, SM, MD, FRCPC, Associate Professor of Geriatrics, Wake Forest University School of Medicine; Teresa King Michaels, RN, MSN, BC, Alpha Rho, Geriatric Clinical Nurse Specialist, North Carolina Baptist Hospital; Brenda W.J.H. Penninx, MPH, PhD, Associate Professor of Research, Wake Forest University School of Medicine; all at Wake Forest University Baptist Medical Center, Winston Salem, NC. Correspondence to Ms. Weeks, Wake Forest University Baptist Medical Center, Winston Salem, NC 27157. E-mail sweeks@wfubmc.edu Accepted for publication September 15, 2002. Journal of Nursing Scholarship Second Quarter 2003 133

Geriatric Depression Scale

Used worldwide, the GDS has been translated into many languages and validated in Argentina (Carrete et al., 2001), Brazil (Almeida & Almeida, 1999), China (Boey, 2000), Greece (Fountoulakis et al., 1999), India (Ganguli et al., 1999), Israel (Cwikel & Ritchie, 1989), Saudi Arabia (Al-Shammari & Al-Subaie, 1999), Spain (Fernandez-San Martin et al., 2002), Sweden (Gottfries, Noltorp, & Norgaard, 1997), Turkey (Ertan & Eker, 2000), the United Kingdom (Shah, Phongsathorn, Bielawski, & Katona, 1996), and the United States (Koenig, Meador, Cohen, & Blazer, 1988; Lesher & Berryhill, 1994; Norris, Gallagher, Wilson, & Winograd, 1987). Despite the availability of this valid and reliable screening tool, many older people are still not screened for depression during interactions with health care workers; thus many older people with undiagnosed depression remain untreated. The low screening rate may be because of the time required to ask an older person so many questions about depression. Also if many negatively stated questions result in sad comments by patients, and sad feelings in providers, providers might be reluctant to repeat the screening experience. Research in the United Kingdom (DAth, Katona, Mullan, Evans, & Katona, 1994), the Netherlands (van Marwijk, Wallace, de Bock, Kaptein, & Mulder, 1995), and the United States (Hoyl et al., 1999) has yielded shorter versions of the GDS with four or five questions extracted from the 15-item GDS tool (Sheikh & Yesavage, 1986). Hoyl and colleagues reported that their GDS-5 had superior performance characteristics to the DAth GDS-4 and van Marwijk GDS-4. They noted, however, that the focus of their research and the research introducing both GDS-4 short-form versions was limited to outpatient settings and they suggested replication in other settings. This study was designed to compare the DAth GDS-4, van Marwijk GDS-4, and Hoyl GDS-5 instruments to the Sheikh and Yesavage GDS in an inpatient setting. It began as a comparative analysis of four studies to determine if one of three short-form GDS tools could be a valid and reliable alternative to the 15-item GDS. It ended with re-ordering the 15 GDS questions into a new two-tiered alternative, the GDS-5/15.

total scores on the Mini-Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975), and demographic data from the medical records of 926 adults discharged from acute care units to the TCU during 1999 and 2000. Data were collected in 2001 and included age, sex, race, years of education, living location at the time of admission to acute care, and length of stay in the acute care unit before being discharged to the TCU. Of these 926 records, no GDS data were available for 110 adults. These 110 people did not differ from the 816 people with GDS information in age, sex, race, years of education, MMSE score, or length of stay in acute care before discharge to TCU (all p values greater than .10). Instruments The GDS (Sheikh & Yesavage, 1986) was used as the standard for this research. We used the widely accepted cutoff of more than 5 positive scores on the 15 GDS questions to indicate suggesting depression. In 1994, DAth and colleagues reported a short-form alternative consisting of 4 of the 15 GDS questions. In a sample of 194 outpatients in Great Britain and using a cut-off of 50% (two positive indicators out of four), the Cronbachs alpha coefficient was .55, sensitivity was .61, and specificity was .81. In 1995, van Marwijk and colleagues reported another GDS-4 consisting of 4 of the 15 GDS items. In a sample of 586 patients studied in nine general practices in the Netherlands, again using a cut-off of 50% (two positive indicators out of four), the Cronbachs alpha coefficient was .64, sensitivity .67, and specificity .66. In 1999, Hoyl and colleagues reported a 5-item GDS extracted from the 15-item GDS. They used this GDS-5 in screening 74 older veterans in outpatient clinics in California. Using a cut-off of 40% (two positive indicators out of five), the Cronbachs alpha coefficient was .80, sensitivity .97 and specificity .85. Data Analysis We made no attempt to analyze the merits of the widelyused 15-item Geriatric Depression Scale or the three shortform GDS tools published by DAth, van Marwijk, or Hoyl and their colleagues. We used the 15-item GDS answers found in patient records as a reference standard in a tool-to-tool comparison to determine which, if any, of the short forms indicated reliable, similar results compared to the 15-item GDS. Answers from the 15-item GDS tools were used to reconstruct the scores for the three short forms and then to compare the four as if the 816 subjects had taken all four GDS tests. Internal consistency for the Sheikh and Yesavage GDS (1986) and the three short-form versions was determined by calculating the Cronbachs alpha: .72 for the 15-item GDS, .61 for the DAth, .36 for the van Marwijk, and .51 for the Hoyl short forms. Sensitivity and specificity and receiver-operating characteristic (ROC) curves for each of the three short-item versions were determined using the 15-item GDS scale as the standard.

Methods
The 15-item GDS was administered to each patient by one of two registered nurses who evaluated patients on the day of discharge from acute care to the 49-bed subacute transitional care unit (TCU) of an 830-bed academic medical center in North Carolina, USA. This tool is part of the standard clinical assessment for discharge to TCU. On the assumption that some patients might be cognitively impaired, nurses were instructed to use their clinical judgment when a patient appeared confused about the questions and to record answers or inability to answer as indicated by the patient. After institutional review board approval, we obtained responses to the 15 GDS questions (Sheikh & Yesavage, 1986),
134 Second Quarter 2003 Journal of Nursing Scholarship

Geriatric Depression Scale

Sensitivity indicates the percentage of the people screened as suggesting depressed as defined by a score of more than 5 on the 15-item GDS scale who also were screened as suggesting depressed on the short-item version (score of 2 or more on the 4- or 5-item versions). Specificity indicates the percentage of people who were screened as not depressed on the 15-item GDS scale (score of 5 or less) who also were screened as not depressed on the short-item version (score of 1 or less). Finally, data were stratified by cognitive status to examine whether internal consistency of the GDS scales and the sensitivity and specificity of the short-item GDS scales were different between cognitively intact (MMSE>24) and cognitively impaired people (MMSE<24).

Findings
Characteristics of the sample (N=816) are shown in Table 1. A high percentage of patients (36.9%) showed evidence of cognitive impairment on the day of discharge from acute care to TCU.

Figure 1. Receiver-operating characteristic (ROC) curves for the 4-item (DAth), 4-item (van Marwijk), and the 5-item (Hoyl) GDS versions. Of the 816 patients in the sample, 500 were not cognitively impaired (>24 on MMSE), 293 were cognitively impaired (MMSE <24), and 23 were missing MMSE scores. The mean MMSE score for those with cognitive impairment was 18.6 compared to 27.2 for those without cognitive impairment. Table 3 shows results dichotomized by cognitive status and indicates that significantly more cognitively impaired people were screened as suggesting depressed (15-item 24.6% and 5-item 45.4%) compared to those not cognitively impaired (15-item 12.2% and 5-item 34.8%), p<.001. Table 3 also shows sensitivity and specificity of the three short tools compared to the GDS in relation to cognitive impairment. The GDS scale characteristics are not different for cognitively impaired and cognitively intact people.
Table 3. Comparison of Results in Relation to Cognitive Impairment (N =793)
Not cognitively impaired (N =500) GDS version 15-item GDS 5-item Hoyl GDS 4-item van Marwijk GDS 4-item D'Ath GDS Suggest depressed 12.2% 34.8% 27.2% 7.2% Sensitivity 96.7% 82.0% 52.5% Specificity 73.8% 80.4% 99.1%

Table 1. Sample Characteristics (N =816)

Age, mean years (SD) Female African American <9 years of education Cognitively impaired (MMSE<24) Living at home before acute care admission Mean days of acute care prior to GDS (SD) 74.3 (12.0) 62.7% 17.3% 22.8% 36.9% 97.7% 18.2 (13.4)

Table 2 shows the percentage of people suggesting depressed and the sensitivity and specificity of each of the three short forms compared to the 15-item GDS. The Hoyl GDS-5 instrument had the highest sensitivity (97.9%) with a specificity of 72.7%, lowest of the three short-form alternatives.

Table 2. Comparison of Results With Use of Four Short-Form Scales (N =816)

GDS version 15-item GDS 5-item Hoyl GDS 4-item van Marwijk GDS 4-item D'Ath GDS Suggest depressed 17.2% 39.5% 31.6% 9.1% Sensitivity 97.9% 81.6% 48.2% Specificity 72.7% 78.8% 99.1%

Cognitively impaired (N =293) GDS version Suggest depressed Sensitivity 15-item GDS 24.6% 5-item Hoyl GDS 45.4% 98.6% 4-item van Marwijk GDS 38.9% 81.9% 4-item D'Ath GDS 11.6% 44.4%

Specificity 71.9% 75.1% 99.1%

The receiver operating characteristic (ROC) curves were calculated for the three short GDS versions. The GDS version with the largest area under the ROC curve indicates better test effectiveness (balancing sensitivity and specificity). The 5-item Hoyl GDS version had the highest area under the curve (.95), compared to .86 for the 4-item DAth and .85 for the 4item van Marwijk GDS versions.

Note. *p<.05; **p<.01; ***p<.001 when comparing cognitively impaired with not cognitively impaired people.

Discussion
This study was conducted to determine which, if any, of three tools developed by DAth and colleagues (1994), van Marwijk and colleagues (1995) and Hoyl and colleagues
Journal of Nursing Scholarship Second Quarter 2003 135

Geratric Depression Scale

(1999) could serve as a valid and reliable short-form alternative to the 15-item GDS (Sheikh & Yesavage, 1986) in an acute care inpatient setting. Of the tools studied, the best combination of test characteristics was associated with the Hoyl GDS-5. However, 22.3% false positives were indicated with the GDS-5. Based on these results, we re-ordered the 15 GDS questions and adopted the Hoyl GDS-5 as the first part of a two-tiered GDS-5/15 geriatric depression scale used to screen older adults (see Table 4). If the patient scores 0 or 1 on the GDS-5 (60% of our sample of 816 inpatients), the patient is classified as not depressed and no further questions are asked. If the patient scores 2 or more on the GDS-5 (40% of our sample), the screener continues asking the remaining 10 questions and classifies the patient as suggesting depressed or not depressed according to the GDS guidelines. Those classified as suggesting depressed on the full GDS-5/15 (17.2% of the total sample of 816 patients) receive further clinical investigation for symptoms of depression.
Table 4. The GDS-5/15 Geriatric Depression Scale
1. Are you basically satisfied with your life?1,2,3 2. Do you often get bored?3 3. Do you often feel helpless?3 4. Do you prefer to stay home rather than going out and doing new things?2,3 5. Do you feel pretty worthless the way you are now?3 Yes Yes* Yes* Yes* Yes* No* No No3 No No

and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994). These forms are designed to screen for depressive symptoms to determine the need for further assessment and they cannot be generalized to a medical diagnosis of depression. This study has several unique strengths. The results have the potential for worldwide application to practice. The sample size of 816 older adults is large for a geriatric research study and larger than the samples used in developing any of the three short-form alternatives tested here. This study is the first known comparison of the DAth GDS-4, van Marwijk GDS-4, and Hoyl GDS-5 in an inpatient hospital setting. Results led to the construction of an innovative two-tiered tool, the GDS-5/15. This tool has the potential to save valuable time for nurses and physicians who use the GDS in their practice. The shorter time required to ask 5 versus 15 questions might allow providers to screen more people for depression. If more providers screen for depression, more people may be identified as at risk for depression and referred for further clinical evaluation. This study also indicated a high proportion of older patients with evidence of cognitive impairment on day of discharge from acute care to TCU.

Conclusions
Using the two-tiered GDS-5/15 will save time for nurses, physicians, and patients by reducing by 66% (from 15 to 5) the questions asked of many patients being screened. A shorter GDS with similar results shows caring for older adults by saving their time and energy and relieving them from having to answer all 15 questions, many of them negatively stated. Finally, reducing the number of questions from 15 to 5, and asking only the final 10 questions if indicated, might enable more physicians and nurses to include depression screening in their routine assessments. Additional providers screening for depression might identify additional older adults at high risk for depression and lead them to clinical evaluation and treatment that might help prevent depression-related loss of independence.
References Al-Shammari, S.A., & Al-Subaie, A. (1999). Prevalence and correlates of depression among Saudi elderly. International Journal of Geriatric Psychiatry, 14(9), 739-747. Almeida, O.P., & Almeida, S.A. (1999). Short versions of the Geriatric Depression Scale: A study of their validity for the diagnosis of a major depressive episode according to ICD-10 and DSM-IV. International Journal of Geriatric Psychiatry, 14, 858-865. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Badger, T.A., McNiece, C., & Gagan, M.J. (2000). Depression, service need, and use in vulnerable populations. Archives of Psychiatric Nursing, 14(4), 173-182. Beekman, A.T.F., Copeland, J.R.M., & Prince, M.J. (1999). Review of community prevalence of depression in later life. British Journal of Psychiatry, 174, 307-311. Boey, K.W. (2000). The use of GDS-15 among the older adults in Beijing. Clinical Gerontologist, 21(2), 49-60.

Score GDS-5= ________ Score of 2 or more on GDS-5? Please continue with the remaining 10 questions: 6. Have you dropped many of your activities and interests?2 7. Do you feel that your life is empty?1 8. Are you in good spirits most of the time? 9. Are you afraid that something bad is going to happen to you?1 10. Do you feel happy most of the time?1,2 11. Do you feel you have more problems with memory than most? 12. Do you think it is wonderful to be alive now? 13. Do you feel full of energy? 14. Do you feel your situation is hopeless? 15. Do you think that most people are better off than you are? Yes* Yes* Yes Yes* Yes Yes* Yes Yes Yes* Yes* No No No* No No* No No* No* No No

Score GDS-15= ________ Circle each answer. Each answer indicated by * counts as 1 point.

Note. 1 =Included on 4-Item, D'Ath. 2 =Included on 4-Item, van Marwijk. 3 =Included on 5-Item, Hoyl. GDS-5 score of 2 or more indicates possible depression (Hoyl et al., 1999); ask remaining 10 questions. GDS-15 score of 5-9 indicates possible depression; scores above 9 usually indicate depression (Sheikh & Yesavage, 1986).

No attempt was made to correlate screening for depression with medical diagnosis, initiation of treatment, or effects of treatment on functional outcomes. Neither the GDS-5 nor the GDS-5/15 can be used as a substitute for a clinical diagnosis of depression based on criteria provided by the Diagnostic
136 Second Quarter 2003 Journal of Nursing Scholarship

Geratric Depression Scale Carrete, P., Augustovski, F., Gimpel, N., Fernandez, S., Di Paolo, R., Schaffer, I., et al. (2001). Validation of a telephone-administered geriatric depression scale in a Hispanic elderly population. Journal of General Internal Medicine, 16(7), 446-450. Cwikel, J., & Ritchie, K. (1989). Screening for depression among the elderly in Israel: An assessment of the Short Geriatric Depression Scale (S-GDS). Israel Journal of Medical Sciences, 25, 131-137. DAth, P., Katona, P., Mullan, E., Evans, S., & Katona, C. (1994). Screening detection and management of depression in elderly primary care attenders. I: The acceptability and performance of the 15-item Geriatric Depression Scale (GDS15) and the development of short versions. Family Practice, 11(3), 260-266. Ertan, T., & Eker, E. (2000). Reliability, validity, and factor structure of the Geriatric Depression Scale in Turkish elderly: Are there different factor structures for different cultures? International Psychogeriatrics, 12(2), 163-172. Espiritu, D.A., Rashid, H., Mast, B.T., Fitzgerald, J., Steinberg, J., & Lichtenberg, P.A. (2001). Depression, cognitive impairment and function in Alzheimers disease. International Journal of Geriatric Psychiatry, 16(11), 1098-1103. Fernandez-San Martin, M.I., Andrade, C., Molina, J., Munoz, P.E., Carretero, B., Rodriguez, M., et al. (2002). Validation of the Spanish version of the geriatric depression scale (GDS) in primary care. International Journal of Geriatric Psychiatry, 17(3), 279287. Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975). Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-198. Fountoulakis, K.N., Tsolaki, M., Iacovides, A., Yesavage, J., OHara, R., Kazis, A., et al. (1999). The validation of the short form of the Geriatric Depression Scale (GDS) in Greece. Aging (Milano), 11(6), 367-372. Ganguli, M., Dube, S., Johnston, J.M., Pandav, R., Chandra, V., & Dodge, H.H. (1999). Depressive symptoms, cognitive impairment and functional impairment in a rural elderly population in India: A Hindi version of the geriatric depression scale (GDS-H). International Journal of Geriatric Psychiatry, 14(10), 807-820. Gottfries, C.G., Noltorp, S., & Norgaard, N. (1997). Experience with a Swedish version of the Geriatric Depression Scale in primary care centres. International Journal of Geriatric Psychiatry, 12(10), 10291034. Hinshaw, A.S. (2000). Nursing knowledge for the 21st century: Opportunities and challenges. Journal of Nursing Scholarship, 32, 117-123. Hoyl, M.T., Alessi, C.A., Harker, J.O., Josephson, K.R., Pietruszka, F.M., Koelfgen, M., et al. (1999). Development and testing of a fiveitem version of the Geriatric Depression Scale. Journal of the American Geriatrics Society, 47(7), 873-878. Kaye, J., Morton, J., Bowcutt, M., & Maupin, D. (2000). Depression: The forgotten diagnosis among hospitalized adults. Journal of Neuroscience Nursing, 32(1), 9-16. Koenig, H.G., Meador, K., Cohen, H.J., & Blazer, D. (1988). Self-rated depression scales and screening for major depression in the older hospitalized patient with medical illness. Journal of the American Geriatrics Society, 36(8), 699-706. Lesher, E.L., & Berryhill, J.S. (1994). Validation of the Geriatric Depression Scaleshort form among inpatients. Journal of Clinical Psychology, 50, 256-260. Norris, J.T., Gallagher, D., Wilson, A., & Winograd, C.H. (1987). Assessment of depression in geriatric medical outpatients: The validity of two screening measures. Journal of the American Geriatrics Society, 35(11), 989-995. Onder, G., Penninx, B.W.J.H., Landi, F., Atkinson, H., Cesari, M., Bernabei, R., et al. (2003). Depression and adverse drug reactions among hospitalized older adults. Archives of Internal Medicine, 163(3), 301-305. Penninx, B.W.J.H., Geerlings, S.W., Deeg, D.J.H., van Eijk, J.Th.M., van Tilburg, W., & Beekman, A.T.F. (1999). Minor and major depression and the risk of death in older persons. Archives of General Psychiatry, 56, 889-896. Shah, A., Phongsathorn, V., Bielawski, C., & Katona, C. (1996). Screening for depression among geriatric inpatients with short versions of the Geriatric Depression Scale. International Journal of Geriatric Psychiatry, 11, 915-918. Sheikh, J.L., & Yesavage, J.A. (1986). Geriatric depression scale (GDS): Recent evidence and development of a shorter version. In T.L. Brink (Ed.), Clinical gerontology: A guide to assessment and intervention. New York: Hawthorne Press. van Marwijk, H.W.J., Wallace, P., de Bock, G.H., Kaptein, A.A., & Mulder, J.D. (1995). Evaluation of the feasibility, reliability and diagnostic value of shortened versions of the geriatric depression scale. British Journal of General Practice, 45(393), 195-199. Whall, A.L., & Hoes-Gurevich, M.L. (1999). Missed depression in elderly individuals. Journal of Gerontological Nursing, 25(6), 44-46. Yesavage, J.A., Brink, T.L., Rose, T.L., Lum, O., Huang, V., Adey, M., et al. (1983). Development and validation of a geriatric depression screening scale: A preliminary report. Journal of Psychiatric Research, 17(1), 37-49.

Journal of Nursing Scholarship

Second Quarter 2003

137