The Clinical Utility of the Cornell Scale for Depression in Dementia as a Routine Assessment in Nursing Homes Yun-Hee Jeon, R.N., Ph.D., Zhicheng Li, B.H.Sc. (Hons), Lee-Fay Low, Ph.D., Lynn Chenoweth, R.N., Ph.D., Daniel O’Connor, M.D., Elizabeth Beattie, Ph.D., Zhixin Liu, Ph.D., Henry Brodaty, M.D., D.Sc.
Objective: To examine the clinical utility of the Cornell Scale for Depression in Dementia (CSDD) in nursing homes. Setting: 14 nursing homes in Sydney and Brisbane, Australia. Participants: 92 residents with a mean age of 85 years. Measurements: Consenting residents were assessed by care staff for depression using the CSDD as part of their routine assessment. Specialist clinicians conducted assessment of depression using the Semi-structured Clinical Diagnostic Interview for DSMIV-TR Axis I Disorders for residents without dementia or the Provisional Diagnostic Criteria for Depression in Alzheimer Disease for residents with dementia to establish expert clinical diagnoses of depression. The diagnostic performance of the staff completed CSDD was analyzed against expert diagnosis using receiver operating characteristic (ROC) curves. Results: The CSDD showed low diagnostic accuracy, with areas under the ROC curve being 0.69, 0.68 and 0.70 for the total sample, residents with dementia and residents without dementia, respectively. At the standard CSDD cutoff score, the sensitivity and specificity were 71% and 59% for the total sample, 69% and 57% for residents with dementia, and 75% and 61% for residents without dementia. The Youden index (for optimizing cut-points) suggested different depression cutoff scores for residents with and without dementia. Conclusion: When administered by nursing home staff the clinical utility of the CSDD is highly questionable in identifying depression. The complexity of the scale, the time required for collecting relevant information, and staff skills and knowledge of assessing depression in older people must be considered when using the CSDD in nursing homes. (Am J Geriatr Psychiatry 2015; 23:784e793) Key Words: Depression, nursing homes, assessment, Cornell Scale for Depression in Dementia, dementia
Received May 14, 2014; revised August 26, 2014; accepted August 29, 2014. From the Sydney Nursing School, the University of Sydney (Y-HJ, ZL), Sydney, Australia; the Dementia Collaborative Research Centre (L-FL, ZL, HB) and Centre for Healthy Brain Ageing (L-FL, LC, ZL, HB), University of New South Wales, Sydney, Australia; the Faculty of Health, University of Technology Sydney (LC), Sydney, Australia; the Southern Clinical School, Monash University (DOC), Clayton, Australia; and the Faculty of Health, School of Nursing, Queensland University of Technology (EB), Brisbane, Australia. Send correspondence and reprint requests to Yun-Hee Jeon, R.N., Ph.D., Sydney Nursing School, University of Sydney, NSW 2050, Sydney, Australia. e-mail: [email protected] Ó 2015 American Association for Geriatric Psychiatry http://dx.doi.org/10.1016/j.jagp.2014.08.013
784
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al.
D
epression is recognized as the most common psychiatric disorder in the elderly, leading to poor quality of life, limited activities of daily living, and an increased risk of medical comorbidity and suicide.1e3 The incidence and prevalence of depression has been consistently high in residential aged care homes (nursing homes, hereafter): ranging from 6% to 24% for major depression and 25%e40% for depressive symptoms or minor depression,4e7 (in particular among residents with dementia8,9). Approximately 52% of all permanent aged care residents in Australia are reported to have mild, moderate, or major depressive symptoms.10 Similarly, U.S. statistics report 54.4% of the residents in nursing homes are diagnosed with depression over the first year of stay.11 The treatment and management of depression in nursing homes remain a challenge.6,7,12 A fundamental issue is the lack of adequate depression assessment in nursing homes, where dementia is common but staff awareness of depression and its assessment are poor, and is associated with limited workforce capabilities and capacity.13 Since its introduction in late 1980s, the Cornell Scale for Depression in Dementia (CSDD) has been widely recommended as the preferred measure of depression for people with dementia or cognitive impairment. Its use has also been validated in people without dementia.14,15 It has good validity, including high sensitivity and specificity in detecting depression when compared against gold standards.14,16,17 Although the CSDD’s internal consistency is excellent, further work is necessary to improve inter-rater reliability.17,18 Notably, most studies testing the reliability and validity of the CSDD are based on ratings by specialist clinicians or trained researchers. It is not known how reliable the CSDD is when administered by nursing home staff, as there is no standardized process in place to assess the rater’s knowledge and skill base prior to administering the CSDD, a condition required by the instrument developers. A U.S. study conducted in long-term care settings found that a modified version (CSDD-M-LTCS) of the CSDD administered by direct care staff who were trained in the use of the CSDD failed to discriminate residents with depression and had low inter-rater reliability.19 In that study, the low accuracy rates may have been due to assessments relying solely on proxy (other care staff) interviews, rather than on multiple sources of information (e.g., including resident interviews and their medical records).
Am J Geriatr Psychiatry 23:8, August 2015
Few studies have systematically assessed nursing home staff’s use of the CSDD as routine assessment. Despite limited evidence concerning skilled use of the CSDD in the nursing home setting, the CSDD has been continually recommended as the tool of choice for assessing depression in older people with and without dementia. The Australian government adopted the CSDD as part of the mandatory Aged Care Funding Instrument (ACFI) in 2008, which is the means of allocating the government subsidy to aged care providers.20 The ACFI data provide an indication of the relative care needs of nursing home residents and the costs associated with managing such needs. The choice of the CSDD was based on the advice to Government that the tool does not rely solely on selfreporting by people with significant cognitive impairment, such as occurs in dementia, and the demonstrated validity of the CSDD for people with and without dementia.17 The inclusion of the CSDD as part of the ACFI had the additional purpose of raising awareness of the importance of depression assessment and management in residents, and of addressing the prevailing ad hoc practice of depression assessment in nursing home settings. Key issues concerning the appropriate use of the CSDD in nursing homes warrant consideration. These include the complex nature of the scale that requires specialist knowledge and skills and the need for adequate time to administer the scale. The CSDD has been identified as being “a complex instrument that requires specific training in its administration” (p. 37).21 Two Australian nursing home studies demonstrated that CSDD items 16e19, measuring suicidal ideation, self-deprecation, pessimism and mood-congruent delusions, were frequently omitted by staff as they found it difficult to evaluate ideational disturbance in residents unable to converse comprehensibly.6,22 A third study using the ACFI data found that the frequency of non-completion for these four items ranged from 41.5% to 43.5% when implemented by a trained research nurse and 50%e54.3% by nursing home staff.23 Furthermore, the authors found no correlation between the CSDD rating obtained by the trained research nurse and the ACFI depression domain score.23 It is not surprising, therefore, that 20 of the 98 submissions made to the first Australian national review of the ACFI raised specific concerns about the use of the CSDD—for example, its complexity in obtaining
785
Utility of the Cornell Scale for Depression in Dementia accurate information, scoring rules, the lengthy time taken to administer correctly, and its unpopularity among medical practitioners.20 Indeed, Australian medical practitioners often dismiss the results of the staff-administered CSDD, and they are rarely informed of the cases with high CSDD scores.24 The other major concern with use of the CSDD is the composition and skill sets of the nursing home workforce, comprising mainly nursing staff (registered nurses and enrolled nurses) and unlicensed care staff who have not been comprehensively trained in CSDD use. In Australia, as in many other countries, there is a shortage of aged care staff with specialist dementia and mental health care skills.12,24 This is a major concern, because over 75% of the residents in nursing homes are 80 years or older, over 50% have dementia, and almost 80% have been reported as having mental health conditions.25 Despite the complexity of care required for these residents, the proportion of registered nurses in Australian nursing homes is now less than 15%, and most of the direct care workforce consists of personal care assistants (62%).26 The proportion of allied health professionals in nursing homes is only 1.7%,26 with psychologists almost nonexistent. Similar trends are observed in the United States, where only 11.7% of full-time employees in nursing homes are registered nurses, and the rest are licensed practical or vocational nurses (22.9%) and nursing aides (65.4%).27 In summary, the lack of mental health services in aged care,28 skilled clinicians, and multidisciplinary approaches further challenges appropriate and timely assessment and management of depression in nursing homes, including the proper use of the CSDD as routine aged care practice. In view of the uncertainty as to whether adequate assessment and diagnosis of depression in nursing home residents occur when nursing home staff administer the CSDD, we aimed to examine the clinical utility of the nursing home staffecompleted CSDD for residents with and without dementia in Australia.
METHODS Participants Participants were recruited from 14 nursing homes with or without dementia specific care in Sydney, New South Wales, and Brisbane, Queensland, Australia. All
786
participating facilities were fully accredited by the Aged Care Accreditation Agency, with a mix of private-for-profit and not-for-profit organizations. Residents were considered to be eligible if they were due to be assessed for depression using the CSDD, usually within 4e6 weeks after being admitted to the facility or 12 months after the last assessment. A research nurse from each city liaised with the manager (or a nominated person) from the participating facility to identify eligible residents. Written informed consent was obtained from 92 residents. For those residents who were unable to provide direct consent, proxy consent was obtained from their close family member or guardian. Measures The CSDD29 contains 19 items measuring five domains of depressionemood, behavioral disturbance, physical signs, cyclic functions, and ideational disturbance. Each item is rated on a scale of 0e2 (absent, mild or intermittent, and severe) based on two semi-structured interviews (one with the resident and one with a staff informant), resident’s clinical records, and observations. Scores above 10 indicate a probable major depression and scores above 18 indicate a definite major depression. Scores below 6 are associated with absence of significant depressive symptoms. In the current study, a cutoff score of greater than 10 was used for classifying residents with depression. The Semi-structured Clinical Diagnostic Interview for DSM-IV-TR Axis I Disorders (DSM-IV, hereafter)30 was used to establish a gold standard diagnosis of depression for residents without dementia. For residents with dementia we used the Provisional Diagnostic Criteria for Depression in Alzheimer Disease (PDCdAD).31 The PDCdAD is equivalent to the DSM-IV in many ways except that it takes account of depressive features within the course of dementia per se. These include symptoms of irritability and social isolation/withdrawal that are not featured in the DSM-IV. Meanwhile, concentration problems, listed as a symptom of depression in the DSM-IV, are omitted in the PDCdAD as they are often associated with dementia. Classification of dementia was determined using the standardized Mini-Mental State Examination (MMSE)32 and the Global Deterioration Rating Scale
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al. (GDRS).33 Based on their clinical experience, we deemed that residents’ medical records were too unreliable to provide a basis for the diagnosis of dementia. The MMSE is the most widely used measure for testing cognitive functions such as arithmetic, orientation, and memory. Missing values on the MMSE due to sensory impairment were prorated using the mean of the remaining items. The GDRS provides an overview of the stages of dementia, consisting of seven stages, with stages 4e7 indicating the loss of cognitive and other functions experienced by persons with dementia. To meet the criteria for dementia in this study, residents were required to have scores of 4 and above on the GDRS and of 23 and below on the MMSE. Procedures The study was approved by the University of Sydney Human Research and Ethics Committee (HREC Ref No: 12898). Letters of support were received from all participating nursing homes and approved by the HREC. Following approvals, consented residents were assessed by manager-approved nursing home staff using the CSDD, as part of the mandatory ACFI assessment protocol. Following training by an experienced psychogeriatrician (HB), two specialist clinician raters (one each in Sydney and Brisbane) established a high level of agreement in depression diagnoses (k ¼ 1.0). Within ten days of the CSDD assessment, the specialist clinicians who were “blinded” to the CSDD results reviewed the resident’s medical chart and conducted a diagnostic interview with the resident and a staff informant. The MMSE and the GDRS were administrated following the interview to determine the presence of dementia. The DSM-IV and the PDCdAD criteria were used to establish expert diagnosis of depression for residents without and with dementia, respectively. Data Analysis Sample size estimation was calculated based on the expected area under the receiver operating characteristic (ROC) curve using MedCalc version 12.7.7. The ROC curve is the plot that displays the trade-off between sensitivity and specificity at all possible cutoff points. The diagnostic accuracy of the CSDD was calculated as the area under the ROC curve. The area under the curve (AUC) is a useful measure for
Am J Geriatr Psychiatry 23:8, August 2015
the overall performance as it represents the average value of sensitivity for all possible values of specificity. An AUC value of 0.5 indicates 50e50 chances for the test to correctly distinguish the diseased cases from normal cases. A perfectly accurate test would have an AUC value of 1.0, reflecting its ability to correctly identify all positive cases. AUCs of 0.9 and above, 0.7e0.9, and below 0.7 are considered to have high, moderate, and low accuracy, respectively.34 The CSDD usually has an average AUC of 0.80 when administered using the standard procedure by trained mental health clinicians.35,36 We expected that the CSDD administered by nursing home staff to have a lower AUC and a value of 0.75 was assumed for sample estimation. With the prevalence of depression in the nursing home setting set at 15% based on the DSM-IV diagnostic criteria,37 93 participants were expected to provide sufficient statistical power (80%) for the comparison of the AUC with the null hypothesis value 0.5. All analyses were conducted using SPSS version 20 except those for ROC curves. Residents with and without depression were compared on demographic and other characteristics. Categorical variables were analyzed using the Pearson c2 test. Continuous variables were tested for normality using the Shapiro-Wilk test. For variables that were not normally distributed, the Mann-Whitney U test was used as the nonparametric alternative to the independent-samples t test. Missing items on the CSDD (“a”e”unable to evaluate”) were replaced using the lowest possible score of “0”, a practice recommended by Leontjevas and colleagues.35 MedCalc version 12.7.7 was used for the ROC curve analysis. The Youden index (J)38 is commonly used to determine the optimal cutoff score. It optimizes the instrument’s differentiating ability when equal weight is assigned to sensitivity and specificity. Analyses using this test were conducted for all residents, then separately for residents with and without dementia.
RESULTS Characteristics of the Study Participants Demographic and other characteristics are shown in Table 1. The average age of the 92 residents was 84.72
787
Utility of the Cornell Scale for Depression in Dementia
TABLE 1. Demographic and Other Characteristics of the Study Population
Characteristics Age (years), median Length of residency (months), median Sex (female), N (%) Born in Australia, N (%) English ability, N (%) Not well Well Very well Facility location, N (%) Sydney Brisbane Room arrangement, N (%) One-bed room Shared room Unstated Dementia diagnosis in the chart, N (%) Yes No Unstated Mini-Mental Status Examination, median Global Deterioration Rating Scale, N (%) No cognitive decline Very mild cognitive decline Mild cognitive decline Moderate cognitive decline Moderately severe cognitive decline Severe cognitive decline Very severe cognitive decline Diagnosis of depression, N (%)a a
Total (N [ 92)
Residents with Depression (N [ 17)
Residents without Depression (N [ 75)
Residents with Dementia (N [ 46)
Residents without Dementia (N [ 46)
87.00 3.89 52 (56.5) 60 (65.2)
80.00 1.74 12 (70.6) 10 (58.8)
87.00 6.47 40 (53.3) 50 (66.7)
87.00 3.60 26 (56.5) 28 (60.9)
87.00 3.93 26 (56.5) 32 (69.6)
5 (5.5) 28 (30.4) 59 (64.1)
2 (11.8) 8 (47.1) 7 (41.2)
3 (4.0) 20 (26.7) 52 (69.3)
5 (5.5) 18 (39.1) 23 (50.0)
0 10 (21.7) 36 (78.3)
58 (63.0) 34 (37.0)
12 (70.6) 5 (29.4)
46 (61.3) 29 (38.7)
33 (71.7) 13 (28.3)
25 (54.3) 21 (45.7)
53 (57.6) 33 (35.9) 6 (6.5)
7 (41.2) 9 (52.9) 1 (1.1)
46 (61.3) 24 (32.0) 5 (6.7)
21 (45.7) 23 (50.0) 2 (4.3)
32 (69.6) 10 (21.7) 4 (8.7)
42 (45.7) 32 (34.8) 18 (19.6) 21.50
13 (76.5) 2 (11.8) 2 (11.8) 10.38
29 (38.7) 30 (40.0) 16 (21.3) 22.22
34 (73.9) 7 (15.2) 5 (10.9) 11.89
8 (17.4) 25 (54.3) 13 (28.3) 26.00
0 0 0 (28.3) (30.4) (32.6) (8.7) (28.3)
4 (8.7) 14 (30.4) 28 (60.9) 0 0 0 0 4 (8.7)
4 14 28 13 14 15 4 17
(4.3) (15.2) (30.4) (14.1) (15.2) (16.3) (4.3) (18.5)
1 3 3 3 6 1
0 (5.9) (17.6) (17.6) (17.6) (35.3) (5.9) 17
4 13 25 10 11 9 3
(5.3) (17.3) (33.3) (13.3) (14.7) (12.0) (4.0) e
13 14 15 4 13
Diagnosis of depression is based on DSM-IV-TR/PDCdAD.
(SD: 9.85) years. Half of the residents met the criteria for dementia. Seventeen residents (18.5%) were diagnosed with depression according to the specialist clinicians’ assessments, either by the DSM-IV or the PDCdAD. The proportion of residents diagnosed with depression was higher among residents with dementia than those without dementia. Residents with depression scored significantly lower on the MMSE than those without depression (U ¼ 342.5, z ¼ 2.97, p 10 69.23 (38.6e90.9) >6 92.31 (64.0e99.8) Residents without dementia (N ¼ 45) >10 75.00 (19.4e99.4) >12 75.00 (19.4e99.4)
Notes: The standard error of the area under the curve (AUC) was calculated using the method of Hanley and McNeil39; the p value was generated by a z test that compares the AUC to the null hypothesis AUC ¼ 0.5.
Our findings corroborate the findings of a similar study by Watson et al:,19 in which the AUC of care staff completed CSDD was 0.66 and did not reach statistical significance against the DSM-IV diagnosis, even though the researchers modified the CSDD to make it more accessible to non-clinicians and provided specific CSDD training for the care staff. This suggests that the problem may not be exclusively attributed to the lack of staff knowledge and skills. The CSDD is designed to derive a rating based on a comprehensive set of information from the resident, informant (caregiver/staff), clinical records, and observations. Yet only 27% of the 92 assessments in the present study were based on the combination of resident interview and staff informant interview. The rest of the assessments were based on only one source of information—the resident (interview and/ or observation) or the staff informant. It is possible that when discrepancies occur between the resident’s self-report and care staff’s observations, the care staff do not have the clinical expertise to integrate the information provided by the resident to make an informed judgment on the CSDD score.24 Furthermore, frequent missing values identified in items 16e19 in the present study confirm that those items are more difficult to score in the nursing home context, possibly because they are “too confrontational” for staff,24 or too difficult to evaluate in residents with cognitive impairment.6,22 Making an informed decision on CSDD scores in these circumstances requires extensive training and clinical experience in managing people with depression. A recent Australian study has also shown that the lack of adherence to the standard CSDD administration protocol and frequent missing values in the use of the
Am J Geriatr Psychiatry 23:8, August 2015
CSDD in the nursing home setting is often associated with insufficient staff resource and time constraints.24 Our research on routine assessment practices in nursing homes also indicates there are inadequate and inappropriate assessments of behavioral and psychological disturbances in residents with dementia, which are highly prevalent in this care environment. Our research suggests that these inadequate care practices are multifactorial and extend beyond the lack of staff training.13,23 Our findings indicate that achieving a high level of accuracy in detecting depression using the CSDD may be difficult in the time-pressured and poorly resourced nursing home context, where issues of understaffing, undertraining, high staff turnover, and irregular schedules are common.19,41 However, the study findings need to be interpreted with caution due to the study limitation of sample size, which was slightly lower than the required number of 93. For the sub-group analyses (i.e., those with dementia and those without dementia), the sample sizes were unlikely to be adequate for statistical significance. The generalizability of the findings needs to be established in a larger sample.
CONCLUSION The routine use of the CSDD in the nursing home setting by care staff does not appear to be a valid method of assessing depression in aged care residents with and without dementia. The complexity of the CSDD scale, the time required for collecting relevant information, and the specialist skills and knowledge of depression in older people, must be considered when using the CSDD in nursing homes.
791
Utility of the Cornell Scale for Depression in Dementia In the absence of any other depression assessment tool that can be easily implemented by nursing home staff, a continued effort needs to be made to develop better methods for assessing depression in older residents with or without dementia. This study has been funded by the Dementia Collaborative Research CentreeAssessment and Better Care,
University of New South Wales, as part of an Australian Government Initiative. Additional funding was provided by Queensland University of Technology to support data collection from Brisbane. The authors express appreciation to the aged care providers for their participation, registered nurses Eesa Witt and Judy McCrow for data collection, and psychologists Kim Burns and Leander Mitchell for conducting the clinical assessments for our participants.
References 1. Bogner HR, Morales KH, Reynolds CF, et al: Course of depression and mortality among older primary care patients. Am J Geriatr Psychiatry 2012; 20:895e903 2. Brown PJ, Roose SP, Fieo R, et al:Frailty and depression in older adults: a high-risk clinical population [Epub ahead of print Aug 20, 2013]. Am J Geriatr Psychiatry 3. Alexopoulos G, Kelly RJ: Research advances in geriatric depression. World Psychiatry 2009; 8:140e149 4. Ames D: Depressive disorders among elderly people in longterm institutional care. Aust NZ J Psychiatry 1993; 27:379e391 5. Whelan PJ, Gaughran F, Walwyn R, et al: ‘Do you feel that your life is empty?’ The clinical utility of a one-off question for detecting depression in elderly care home residents. Age Ageing 2008; 37:475e478 6. Snowdon J: Depression in nursing homes. Int Psychogeriatr 2010; 22:1143e1148 7. McSweeney K, O’Connor D: Depression among newly admitted Australian nursing home residents. Int Psychogeriatr 2008; 20: 724e737 8. Debruyne H, Van Buggenhout M, Le Bastard N, et al: Is the geriatric depression scale a reliable screening tool for depressive symptoms in elderly patients with cognitive impairment? Int J Geriatr Psychiatry 2009; 24:556e562 9. Lyketsos CG, Olin J: Depression in Alzheimer’s disease: overview and treatment. Biol Psychiatry 2002; 52:243e252 10. Australian Institute of Health and Welfare: Depression in residential aged care 2008e2012. Vol AGE 73. Canberra, Australian Institute of Health and Welfare, 2013 11. Hoover DR, Siegel M, Lucas J, et al: Depression in the first year of stay for elderly long-term nursing home residents in the USA. Int Psychogeriatr 2010; 22:1161e1171 12. Snowdon J, Fleming R: Recognising depression in residential facilities: an Australian challenge. Int J Geriatr Psychiatry 2008; 23: 295e300 13. Jeon Y-H, Govett J, Low L-F, et al: Care planning practices for behavioural and psychological symptoms of dementia (BPSD) in residential aged care: a pilot of an education toolkit informed by the Aged Care Funding Instrument (ACFI). Contemp Nurse 2013; 44:156e169 14. Kørner A, Lauritzen L, Abelskov K, et al: The Geriatric Depression Scale and the Cornell Scale for Depression in Dementia. A validity study. Nord J Psychiatry 2006; 60:360e364 15. Alexopoulos GS, Abrams RC, Young RC, et al: Use of the Cornell scale in nondemented patients. J Am Geriatr Soc 1988; 36:230e236 16. Pachana N, Helmes E, Byrne G, et al: Screening for mental disorders in residential aged care facilities. Int Psychogeriatr 2010; 22:1107e1120 17. Sansoni J, Marosszeky N, Jeon Y-H, et al: Final Report: Dementia Outcomes Measurement Suite Project. Wollongong, NSW, Centre for Health Service Development, University of Wollongong, 2007
792
18. Koopmans R, Zuidema S, Leontjevas R, et al: Comprehensive assessment of depression and behavioral problems in long-term care. Int Psychogeriatr 2010; 22:1054e1062 19. Watson LC, Zimmerman S, Cohen LW, et al: Practical depression screening in residential care/assisted living: five methods compared with gold standard diagnoses. Am J Geriatr Psychiatry 2009; 17:556e564 20. Australian Government Department of Health and Ageing: The Review of the Aged Care Funding Instrument. Canberra, Australian Government Department of Health and Ageing, 2011 21. Sansoni J, Marosszeky N, Fleming G, et al: Selecting Tools for ACAT Assessment: A Report for the Aged Care Assessment Program (ACAP) Expert Clinical Reference Group. Canberra, Centre for Health Service Development, University of Wollongong, 2010 22. Snowdon J, Rosengren D, Daniel D, et al: Australia’s use of the Cornell scale to screen for depression in nursing homes. Australas J Ageing 2011; 30:33e36 23. Jeon Y-H, Govett J, Low L, et al: Assessment of behavioural and psychological symptoms of dementia in the era of the aged care funding instrument. Internet J Psychiatry 2014; 3 24. Davison TE, Snowdon J, Castle N, et al: An evaluation of a national program to implement the Cornell Scale for Depression in Dementia into routine practice in aged care facilities. Int Psychogeriatr 2012; 24:631e641 25. Australian Institute of Health and Welfare: Residential aged care and aged care packages in the community 2011e12. (online). Available at: http://www.aihw.gov.au/aged-care/residential-andcommunity-2011-12/. Accessed December 10, 2013 26. King D, Mavromaras K, Wei Z, et al: The Aged Care Workforce, 2012. Canberra, Australian Government Department of Health and Ageing, 2012 27. Harris-Kojetin L, Sengupta M, Park-Lee E, et al: Long-term care services in the United States: 2013 overview. (online). Available at: http://www.cdc.gov/nchs/nsltcp/nsltcp_products.htm. Accessed April 1, 2014 28. Snowdon J, Ames D, Chiu E, et al: A survey of psychiatric services for elderly people in Australia. Aust NZ J Psychiatry 1995; 29: 207e214 29. Alexopoulos G, Abrams R, Young R, et al: Cornell scale for depression in dementia. Biol Psychiatry 1988; 23:271e284 30. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSMIV TR). Washington, DC, American Psychiatric Association, 2000 31. Olin J, Katz I, Meyers B, et al: Provisional diagnostic criteria for depression of Alzheimer disease. Am J Geriatr Psychiatry 2002; 10:125e128 32. Folstein M, Folstein S, McHugh P: ‘‘Mini-Mental-State’’: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189e198
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al. 33. Reisberg B, Ferris S, de Leon M, et al: The global deterioration scale for assessment of primary degenerative dementia. Am J Psychiatry 1982; 139:1136e1139 34. Fischer J, Bachmann L, Jaeschke R: A readers’ guide to the interpretation of diagnostic test properties: clinical example of sepsis. Intensive Care Med 2003; 29:1043e1051 35. Leontjevas R, Gerritsen DL, Vernooij-Dassen M, et al: Comparative validation of proxy-based Montgomery- Asberg Depression Rating Scale and Cornell Scale for Depression in Dementia in nursing home residents with dementia. Am J Geriatr Psychiatry 2012; 20:985e993 36. Barca ML, Engedal K, Selbæk G: A reliability and validity study of the Cornell Scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord 2010; 29:438e447
Am J Geriatr Psychiatry 23:8, August 2015
37. Watson LC, Pignone M: Screening accuracy for late-life depression in primary care: a systematic review. J Fam Pract 2003; 52:956e964 38. Youden WJ: Index for rating diagnostic tests. Cancer 1950; 3: 32e35 39. Hanley JA, McNeil BJ: The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29e36 40. McCabe MP, Davison TE, Mellor D, et al: Depression among older people with cognitive impairment: prevalence and detection. Int J Geriatr Psychiatry 2006; 21:633e644 41. Maas M, Specht J, Buckwalter K, et al: Nursing home staffing and training recommendations for promoting older adults’ quality of care and life: part 1. Deficits in the quality of care due to understaffing and undertraining. Res Gerontol Nurs 2008; 1:123e152
793
Objective: To examine the clinical utility of the Cornell Scale for Depression in Dementia (CSDD) in nursing homes. Setting: 14 nursing homes in Sydney and Brisbane, Australia. Participants: 92 residents with a mean age of 85 years. Measurements: Consenting residents were assessed by care staff for depression using the CSDD as part of their routine assessment. Specialist clinicians conducted assessment of depression using the Semi-structured Clinical Diagnostic Interview for DSMIV-TR Axis I Disorders for residents without dementia or the Provisional Diagnostic Criteria for Depression in Alzheimer Disease for residents with dementia to establish expert clinical diagnoses of depression. The diagnostic performance of the staff completed CSDD was analyzed against expert diagnosis using receiver operating characteristic (ROC) curves. Results: The CSDD showed low diagnostic accuracy, with areas under the ROC curve being 0.69, 0.68 and 0.70 for the total sample, residents with dementia and residents without dementia, respectively. At the standard CSDD cutoff score, the sensitivity and specificity were 71% and 59% for the total sample, 69% and 57% for residents with dementia, and 75% and 61% for residents without dementia. The Youden index (for optimizing cut-points) suggested different depression cutoff scores for residents with and without dementia. Conclusion: When administered by nursing home staff the clinical utility of the CSDD is highly questionable in identifying depression. The complexity of the scale, the time required for collecting relevant information, and staff skills and knowledge of assessing depression in older people must be considered when using the CSDD in nursing homes. (Am J Geriatr Psychiatry 2015; 23:784e793) Key Words: Depression, nursing homes, assessment, Cornell Scale for Depression in Dementia, dementia
Received May 14, 2014; revised August 26, 2014; accepted August 29, 2014. From the Sydney Nursing School, the University of Sydney (Y-HJ, ZL), Sydney, Australia; the Dementia Collaborative Research Centre (L-FL, ZL, HB) and Centre for Healthy Brain Ageing (L-FL, LC, ZL, HB), University of New South Wales, Sydney, Australia; the Faculty of Health, University of Technology Sydney (LC), Sydney, Australia; the Southern Clinical School, Monash University (DOC), Clayton, Australia; and the Faculty of Health, School of Nursing, Queensland University of Technology (EB), Brisbane, Australia. Send correspondence and reprint requests to Yun-Hee Jeon, R.N., Ph.D., Sydney Nursing School, University of Sydney, NSW 2050, Sydney, Australia. e-mail: [email protected] Ó 2015 American Association for Geriatric Psychiatry http://dx.doi.org/10.1016/j.jagp.2014.08.013
784
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al.
D
epression is recognized as the most common psychiatric disorder in the elderly, leading to poor quality of life, limited activities of daily living, and an increased risk of medical comorbidity and suicide.1e3 The incidence and prevalence of depression has been consistently high in residential aged care homes (nursing homes, hereafter): ranging from 6% to 24% for major depression and 25%e40% for depressive symptoms or minor depression,4e7 (in particular among residents with dementia8,9). Approximately 52% of all permanent aged care residents in Australia are reported to have mild, moderate, or major depressive symptoms.10 Similarly, U.S. statistics report 54.4% of the residents in nursing homes are diagnosed with depression over the first year of stay.11 The treatment and management of depression in nursing homes remain a challenge.6,7,12 A fundamental issue is the lack of adequate depression assessment in nursing homes, where dementia is common but staff awareness of depression and its assessment are poor, and is associated with limited workforce capabilities and capacity.13 Since its introduction in late 1980s, the Cornell Scale for Depression in Dementia (CSDD) has been widely recommended as the preferred measure of depression for people with dementia or cognitive impairment. Its use has also been validated in people without dementia.14,15 It has good validity, including high sensitivity and specificity in detecting depression when compared against gold standards.14,16,17 Although the CSDD’s internal consistency is excellent, further work is necessary to improve inter-rater reliability.17,18 Notably, most studies testing the reliability and validity of the CSDD are based on ratings by specialist clinicians or trained researchers. It is not known how reliable the CSDD is when administered by nursing home staff, as there is no standardized process in place to assess the rater’s knowledge and skill base prior to administering the CSDD, a condition required by the instrument developers. A U.S. study conducted in long-term care settings found that a modified version (CSDD-M-LTCS) of the CSDD administered by direct care staff who were trained in the use of the CSDD failed to discriminate residents with depression and had low inter-rater reliability.19 In that study, the low accuracy rates may have been due to assessments relying solely on proxy (other care staff) interviews, rather than on multiple sources of information (e.g., including resident interviews and their medical records).
Am J Geriatr Psychiatry 23:8, August 2015
Few studies have systematically assessed nursing home staff’s use of the CSDD as routine assessment. Despite limited evidence concerning skilled use of the CSDD in the nursing home setting, the CSDD has been continually recommended as the tool of choice for assessing depression in older people with and without dementia. The Australian government adopted the CSDD as part of the mandatory Aged Care Funding Instrument (ACFI) in 2008, which is the means of allocating the government subsidy to aged care providers.20 The ACFI data provide an indication of the relative care needs of nursing home residents and the costs associated with managing such needs. The choice of the CSDD was based on the advice to Government that the tool does not rely solely on selfreporting by people with significant cognitive impairment, such as occurs in dementia, and the demonstrated validity of the CSDD for people with and without dementia.17 The inclusion of the CSDD as part of the ACFI had the additional purpose of raising awareness of the importance of depression assessment and management in residents, and of addressing the prevailing ad hoc practice of depression assessment in nursing home settings. Key issues concerning the appropriate use of the CSDD in nursing homes warrant consideration. These include the complex nature of the scale that requires specialist knowledge and skills and the need for adequate time to administer the scale. The CSDD has been identified as being “a complex instrument that requires specific training in its administration” (p. 37).21 Two Australian nursing home studies demonstrated that CSDD items 16e19, measuring suicidal ideation, self-deprecation, pessimism and mood-congruent delusions, were frequently omitted by staff as they found it difficult to evaluate ideational disturbance in residents unable to converse comprehensibly.6,22 A third study using the ACFI data found that the frequency of non-completion for these four items ranged from 41.5% to 43.5% when implemented by a trained research nurse and 50%e54.3% by nursing home staff.23 Furthermore, the authors found no correlation between the CSDD rating obtained by the trained research nurse and the ACFI depression domain score.23 It is not surprising, therefore, that 20 of the 98 submissions made to the first Australian national review of the ACFI raised specific concerns about the use of the CSDD—for example, its complexity in obtaining
785
Utility of the Cornell Scale for Depression in Dementia accurate information, scoring rules, the lengthy time taken to administer correctly, and its unpopularity among medical practitioners.20 Indeed, Australian medical practitioners often dismiss the results of the staff-administered CSDD, and they are rarely informed of the cases with high CSDD scores.24 The other major concern with use of the CSDD is the composition and skill sets of the nursing home workforce, comprising mainly nursing staff (registered nurses and enrolled nurses) and unlicensed care staff who have not been comprehensively trained in CSDD use. In Australia, as in many other countries, there is a shortage of aged care staff with specialist dementia and mental health care skills.12,24 This is a major concern, because over 75% of the residents in nursing homes are 80 years or older, over 50% have dementia, and almost 80% have been reported as having mental health conditions.25 Despite the complexity of care required for these residents, the proportion of registered nurses in Australian nursing homes is now less than 15%, and most of the direct care workforce consists of personal care assistants (62%).26 The proportion of allied health professionals in nursing homes is only 1.7%,26 with psychologists almost nonexistent. Similar trends are observed in the United States, where only 11.7% of full-time employees in nursing homes are registered nurses, and the rest are licensed practical or vocational nurses (22.9%) and nursing aides (65.4%).27 In summary, the lack of mental health services in aged care,28 skilled clinicians, and multidisciplinary approaches further challenges appropriate and timely assessment and management of depression in nursing homes, including the proper use of the CSDD as routine aged care practice. In view of the uncertainty as to whether adequate assessment and diagnosis of depression in nursing home residents occur when nursing home staff administer the CSDD, we aimed to examine the clinical utility of the nursing home staffecompleted CSDD for residents with and without dementia in Australia.
METHODS Participants Participants were recruited from 14 nursing homes with or without dementia specific care in Sydney, New South Wales, and Brisbane, Queensland, Australia. All
786
participating facilities were fully accredited by the Aged Care Accreditation Agency, with a mix of private-for-profit and not-for-profit organizations. Residents were considered to be eligible if they were due to be assessed for depression using the CSDD, usually within 4e6 weeks after being admitted to the facility or 12 months after the last assessment. A research nurse from each city liaised with the manager (or a nominated person) from the participating facility to identify eligible residents. Written informed consent was obtained from 92 residents. For those residents who were unable to provide direct consent, proxy consent was obtained from their close family member or guardian. Measures The CSDD29 contains 19 items measuring five domains of depressionemood, behavioral disturbance, physical signs, cyclic functions, and ideational disturbance. Each item is rated on a scale of 0e2 (absent, mild or intermittent, and severe) based on two semi-structured interviews (one with the resident and one with a staff informant), resident’s clinical records, and observations. Scores above 10 indicate a probable major depression and scores above 18 indicate a definite major depression. Scores below 6 are associated with absence of significant depressive symptoms. In the current study, a cutoff score of greater than 10 was used for classifying residents with depression. The Semi-structured Clinical Diagnostic Interview for DSM-IV-TR Axis I Disorders (DSM-IV, hereafter)30 was used to establish a gold standard diagnosis of depression for residents without dementia. For residents with dementia we used the Provisional Diagnostic Criteria for Depression in Alzheimer Disease (PDCdAD).31 The PDCdAD is equivalent to the DSM-IV in many ways except that it takes account of depressive features within the course of dementia per se. These include symptoms of irritability and social isolation/withdrawal that are not featured in the DSM-IV. Meanwhile, concentration problems, listed as a symptom of depression in the DSM-IV, are omitted in the PDCdAD as they are often associated with dementia. Classification of dementia was determined using the standardized Mini-Mental State Examination (MMSE)32 and the Global Deterioration Rating Scale
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al. (GDRS).33 Based on their clinical experience, we deemed that residents’ medical records were too unreliable to provide a basis for the diagnosis of dementia. The MMSE is the most widely used measure for testing cognitive functions such as arithmetic, orientation, and memory. Missing values on the MMSE due to sensory impairment were prorated using the mean of the remaining items. The GDRS provides an overview of the stages of dementia, consisting of seven stages, with stages 4e7 indicating the loss of cognitive and other functions experienced by persons with dementia. To meet the criteria for dementia in this study, residents were required to have scores of 4 and above on the GDRS and of 23 and below on the MMSE. Procedures The study was approved by the University of Sydney Human Research and Ethics Committee (HREC Ref No: 12898). Letters of support were received from all participating nursing homes and approved by the HREC. Following approvals, consented residents were assessed by manager-approved nursing home staff using the CSDD, as part of the mandatory ACFI assessment protocol. Following training by an experienced psychogeriatrician (HB), two specialist clinician raters (one each in Sydney and Brisbane) established a high level of agreement in depression diagnoses (k ¼ 1.0). Within ten days of the CSDD assessment, the specialist clinicians who were “blinded” to the CSDD results reviewed the resident’s medical chart and conducted a diagnostic interview with the resident and a staff informant. The MMSE and the GDRS were administrated following the interview to determine the presence of dementia. The DSM-IV and the PDCdAD criteria were used to establish expert diagnosis of depression for residents without and with dementia, respectively. Data Analysis Sample size estimation was calculated based on the expected area under the receiver operating characteristic (ROC) curve using MedCalc version 12.7.7. The ROC curve is the plot that displays the trade-off between sensitivity and specificity at all possible cutoff points. The diagnostic accuracy of the CSDD was calculated as the area under the ROC curve. The area under the curve (AUC) is a useful measure for
Am J Geriatr Psychiatry 23:8, August 2015
the overall performance as it represents the average value of sensitivity for all possible values of specificity. An AUC value of 0.5 indicates 50e50 chances for the test to correctly distinguish the diseased cases from normal cases. A perfectly accurate test would have an AUC value of 1.0, reflecting its ability to correctly identify all positive cases. AUCs of 0.9 and above, 0.7e0.9, and below 0.7 are considered to have high, moderate, and low accuracy, respectively.34 The CSDD usually has an average AUC of 0.80 when administered using the standard procedure by trained mental health clinicians.35,36 We expected that the CSDD administered by nursing home staff to have a lower AUC and a value of 0.75 was assumed for sample estimation. With the prevalence of depression in the nursing home setting set at 15% based on the DSM-IV diagnostic criteria,37 93 participants were expected to provide sufficient statistical power (80%) for the comparison of the AUC with the null hypothesis value 0.5. All analyses were conducted using SPSS version 20 except those for ROC curves. Residents with and without depression were compared on demographic and other characteristics. Categorical variables were analyzed using the Pearson c2 test. Continuous variables were tested for normality using the Shapiro-Wilk test. For variables that were not normally distributed, the Mann-Whitney U test was used as the nonparametric alternative to the independent-samples t test. Missing items on the CSDD (“a”e”unable to evaluate”) were replaced using the lowest possible score of “0”, a practice recommended by Leontjevas and colleagues.35 MedCalc version 12.7.7 was used for the ROC curve analysis. The Youden index (J)38 is commonly used to determine the optimal cutoff score. It optimizes the instrument’s differentiating ability when equal weight is assigned to sensitivity and specificity. Analyses using this test were conducted for all residents, then separately for residents with and without dementia.
RESULTS Characteristics of the Study Participants Demographic and other characteristics are shown in Table 1. The average age of the 92 residents was 84.72
787
Utility of the Cornell Scale for Depression in Dementia
TABLE 1. Demographic and Other Characteristics of the Study Population
Characteristics Age (years), median Length of residency (months), median Sex (female), N (%) Born in Australia, N (%) English ability, N (%) Not well Well Very well Facility location, N (%) Sydney Brisbane Room arrangement, N (%) One-bed room Shared room Unstated Dementia diagnosis in the chart, N (%) Yes No Unstated Mini-Mental Status Examination, median Global Deterioration Rating Scale, N (%) No cognitive decline Very mild cognitive decline Mild cognitive decline Moderate cognitive decline Moderately severe cognitive decline Severe cognitive decline Very severe cognitive decline Diagnosis of depression, N (%)a a
Total (N [ 92)
Residents with Depression (N [ 17)
Residents without Depression (N [ 75)
Residents with Dementia (N [ 46)
Residents without Dementia (N [ 46)
87.00 3.89 52 (56.5) 60 (65.2)
80.00 1.74 12 (70.6) 10 (58.8)
87.00 6.47 40 (53.3) 50 (66.7)
87.00 3.60 26 (56.5) 28 (60.9)
87.00 3.93 26 (56.5) 32 (69.6)
5 (5.5) 28 (30.4) 59 (64.1)
2 (11.8) 8 (47.1) 7 (41.2)
3 (4.0) 20 (26.7) 52 (69.3)
5 (5.5) 18 (39.1) 23 (50.0)
0 10 (21.7) 36 (78.3)
58 (63.0) 34 (37.0)
12 (70.6) 5 (29.4)
46 (61.3) 29 (38.7)
33 (71.7) 13 (28.3)
25 (54.3) 21 (45.7)
53 (57.6) 33 (35.9) 6 (6.5)
7 (41.2) 9 (52.9) 1 (1.1)
46 (61.3) 24 (32.0) 5 (6.7)
21 (45.7) 23 (50.0) 2 (4.3)
32 (69.6) 10 (21.7) 4 (8.7)
42 (45.7) 32 (34.8) 18 (19.6) 21.50
13 (76.5) 2 (11.8) 2 (11.8) 10.38
29 (38.7) 30 (40.0) 16 (21.3) 22.22
34 (73.9) 7 (15.2) 5 (10.9) 11.89
8 (17.4) 25 (54.3) 13 (28.3) 26.00
0 0 0 (28.3) (30.4) (32.6) (8.7) (28.3)
4 (8.7) 14 (30.4) 28 (60.9) 0 0 0 0 4 (8.7)
4 14 28 13 14 15 4 17
(4.3) (15.2) (30.4) (14.1) (15.2) (16.3) (4.3) (18.5)
1 3 3 3 6 1
0 (5.9) (17.6) (17.6) (17.6) (35.3) (5.9) 17
4 13 25 10 11 9 3
(5.3) (17.3) (33.3) (13.3) (14.7) (12.0) (4.0) e
13 14 15 4 13
Diagnosis of depression is based on DSM-IV-TR/PDCdAD.
(SD: 9.85) years. Half of the residents met the criteria for dementia. Seventeen residents (18.5%) were diagnosed with depression according to the specialist clinicians’ assessments, either by the DSM-IV or the PDCdAD. The proportion of residents diagnosed with depression was higher among residents with dementia than those without dementia. Residents with depression scored significantly lower on the MMSE than those without depression (U ¼ 342.5, z ¼ 2.97, p 10 69.23 (38.6e90.9) >6 92.31 (64.0e99.8) Residents without dementia (N ¼ 45) >10 75.00 (19.4e99.4) >12 75.00 (19.4e99.4)
Notes: The standard error of the area under the curve (AUC) was calculated using the method of Hanley and McNeil39; the p value was generated by a z test that compares the AUC to the null hypothesis AUC ¼ 0.5.
Our findings corroborate the findings of a similar study by Watson et al:,19 in which the AUC of care staff completed CSDD was 0.66 and did not reach statistical significance against the DSM-IV diagnosis, even though the researchers modified the CSDD to make it more accessible to non-clinicians and provided specific CSDD training for the care staff. This suggests that the problem may not be exclusively attributed to the lack of staff knowledge and skills. The CSDD is designed to derive a rating based on a comprehensive set of information from the resident, informant (caregiver/staff), clinical records, and observations. Yet only 27% of the 92 assessments in the present study were based on the combination of resident interview and staff informant interview. The rest of the assessments were based on only one source of information—the resident (interview and/ or observation) or the staff informant. It is possible that when discrepancies occur between the resident’s self-report and care staff’s observations, the care staff do not have the clinical expertise to integrate the information provided by the resident to make an informed judgment on the CSDD score.24 Furthermore, frequent missing values identified in items 16e19 in the present study confirm that those items are more difficult to score in the nursing home context, possibly because they are “too confrontational” for staff,24 or too difficult to evaluate in residents with cognitive impairment.6,22 Making an informed decision on CSDD scores in these circumstances requires extensive training and clinical experience in managing people with depression. A recent Australian study has also shown that the lack of adherence to the standard CSDD administration protocol and frequent missing values in the use of the
Am J Geriatr Psychiatry 23:8, August 2015
CSDD in the nursing home setting is often associated with insufficient staff resource and time constraints.24 Our research on routine assessment practices in nursing homes also indicates there are inadequate and inappropriate assessments of behavioral and psychological disturbances in residents with dementia, which are highly prevalent in this care environment. Our research suggests that these inadequate care practices are multifactorial and extend beyond the lack of staff training.13,23 Our findings indicate that achieving a high level of accuracy in detecting depression using the CSDD may be difficult in the time-pressured and poorly resourced nursing home context, where issues of understaffing, undertraining, high staff turnover, and irregular schedules are common.19,41 However, the study findings need to be interpreted with caution due to the study limitation of sample size, which was slightly lower than the required number of 93. For the sub-group analyses (i.e., those with dementia and those without dementia), the sample sizes were unlikely to be adequate for statistical significance. The generalizability of the findings needs to be established in a larger sample.
CONCLUSION The routine use of the CSDD in the nursing home setting by care staff does not appear to be a valid method of assessing depression in aged care residents with and without dementia. The complexity of the CSDD scale, the time required for collecting relevant information, and the specialist skills and knowledge of depression in older people, must be considered when using the CSDD in nursing homes.
791
Utility of the Cornell Scale for Depression in Dementia In the absence of any other depression assessment tool that can be easily implemented by nursing home staff, a continued effort needs to be made to develop better methods for assessing depression in older residents with or without dementia. This study has been funded by the Dementia Collaborative Research CentreeAssessment and Better Care,
University of New South Wales, as part of an Australian Government Initiative. Additional funding was provided by Queensland University of Technology to support data collection from Brisbane. The authors express appreciation to the aged care providers for their participation, registered nurses Eesa Witt and Judy McCrow for data collection, and psychologists Kim Burns and Leander Mitchell for conducting the clinical assessments for our participants.
References 1. Bogner HR, Morales KH, Reynolds CF, et al: Course of depression and mortality among older primary care patients. Am J Geriatr Psychiatry 2012; 20:895e903 2. Brown PJ, Roose SP, Fieo R, et al:Frailty and depression in older adults: a high-risk clinical population [Epub ahead of print Aug 20, 2013]. Am J Geriatr Psychiatry 3. Alexopoulos G, Kelly RJ: Research advances in geriatric depression. World Psychiatry 2009; 8:140e149 4. Ames D: Depressive disorders among elderly people in longterm institutional care. Aust NZ J Psychiatry 1993; 27:379e391 5. Whelan PJ, Gaughran F, Walwyn R, et al: ‘Do you feel that your life is empty?’ The clinical utility of a one-off question for detecting depression in elderly care home residents. Age Ageing 2008; 37:475e478 6. Snowdon J: Depression in nursing homes. Int Psychogeriatr 2010; 22:1143e1148 7. McSweeney K, O’Connor D: Depression among newly admitted Australian nursing home residents. Int Psychogeriatr 2008; 20: 724e737 8. Debruyne H, Van Buggenhout M, Le Bastard N, et al: Is the geriatric depression scale a reliable screening tool for depressive symptoms in elderly patients with cognitive impairment? Int J Geriatr Psychiatry 2009; 24:556e562 9. Lyketsos CG, Olin J: Depression in Alzheimer’s disease: overview and treatment. Biol Psychiatry 2002; 52:243e252 10. Australian Institute of Health and Welfare: Depression in residential aged care 2008e2012. Vol AGE 73. Canberra, Australian Institute of Health and Welfare, 2013 11. Hoover DR, Siegel M, Lucas J, et al: Depression in the first year of stay for elderly long-term nursing home residents in the USA. Int Psychogeriatr 2010; 22:1161e1171 12. Snowdon J, Fleming R: Recognising depression in residential facilities: an Australian challenge. Int J Geriatr Psychiatry 2008; 23: 295e300 13. Jeon Y-H, Govett J, Low L-F, et al: Care planning practices for behavioural and psychological symptoms of dementia (BPSD) in residential aged care: a pilot of an education toolkit informed by the Aged Care Funding Instrument (ACFI). Contemp Nurse 2013; 44:156e169 14. Kørner A, Lauritzen L, Abelskov K, et al: The Geriatric Depression Scale and the Cornell Scale for Depression in Dementia. A validity study. Nord J Psychiatry 2006; 60:360e364 15. Alexopoulos GS, Abrams RC, Young RC, et al: Use of the Cornell scale in nondemented patients. J Am Geriatr Soc 1988; 36:230e236 16. Pachana N, Helmes E, Byrne G, et al: Screening for mental disorders in residential aged care facilities. Int Psychogeriatr 2010; 22:1107e1120 17. Sansoni J, Marosszeky N, Jeon Y-H, et al: Final Report: Dementia Outcomes Measurement Suite Project. Wollongong, NSW, Centre for Health Service Development, University of Wollongong, 2007
792
18. Koopmans R, Zuidema S, Leontjevas R, et al: Comprehensive assessment of depression and behavioral problems in long-term care. Int Psychogeriatr 2010; 22:1054e1062 19. Watson LC, Zimmerman S, Cohen LW, et al: Practical depression screening in residential care/assisted living: five methods compared with gold standard diagnoses. Am J Geriatr Psychiatry 2009; 17:556e564 20. Australian Government Department of Health and Ageing: The Review of the Aged Care Funding Instrument. Canberra, Australian Government Department of Health and Ageing, 2011 21. Sansoni J, Marosszeky N, Fleming G, et al: Selecting Tools for ACAT Assessment: A Report for the Aged Care Assessment Program (ACAP) Expert Clinical Reference Group. Canberra, Centre for Health Service Development, University of Wollongong, 2010 22. Snowdon J, Rosengren D, Daniel D, et al: Australia’s use of the Cornell scale to screen for depression in nursing homes. Australas J Ageing 2011; 30:33e36 23. Jeon Y-H, Govett J, Low L, et al: Assessment of behavioural and psychological symptoms of dementia in the era of the aged care funding instrument. Internet J Psychiatry 2014; 3 24. Davison TE, Snowdon J, Castle N, et al: An evaluation of a national program to implement the Cornell Scale for Depression in Dementia into routine practice in aged care facilities. Int Psychogeriatr 2012; 24:631e641 25. Australian Institute of Health and Welfare: Residential aged care and aged care packages in the community 2011e12. (online). Available at: http://www.aihw.gov.au/aged-care/residential-andcommunity-2011-12/. Accessed December 10, 2013 26. King D, Mavromaras K, Wei Z, et al: The Aged Care Workforce, 2012. Canberra, Australian Government Department of Health and Ageing, 2012 27. Harris-Kojetin L, Sengupta M, Park-Lee E, et al: Long-term care services in the United States: 2013 overview. (online). Available at: http://www.cdc.gov/nchs/nsltcp/nsltcp_products.htm. Accessed April 1, 2014 28. Snowdon J, Ames D, Chiu E, et al: A survey of psychiatric services for elderly people in Australia. Aust NZ J Psychiatry 1995; 29: 207e214 29. Alexopoulos G, Abrams R, Young R, et al: Cornell scale for depression in dementia. Biol Psychiatry 1988; 23:271e284 30. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSMIV TR). Washington, DC, American Psychiatric Association, 2000 31. Olin J, Katz I, Meyers B, et al: Provisional diagnostic criteria for depression of Alzheimer disease. Am J Geriatr Psychiatry 2002; 10:125e128 32. Folstein M, Folstein S, McHugh P: ‘‘Mini-Mental-State’’: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189e198
Am J Geriatr Psychiatry 23:8, August 2015
Jeon et al. 33. Reisberg B, Ferris S, de Leon M, et al: The global deterioration scale for assessment of primary degenerative dementia. Am J Psychiatry 1982; 139:1136e1139 34. Fischer J, Bachmann L, Jaeschke R: A readers’ guide to the interpretation of diagnostic test properties: clinical example of sepsis. Intensive Care Med 2003; 29:1043e1051 35. Leontjevas R, Gerritsen DL, Vernooij-Dassen M, et al: Comparative validation of proxy-based Montgomery- Asberg Depression Rating Scale and Cornell Scale for Depression in Dementia in nursing home residents with dementia. Am J Geriatr Psychiatry 2012; 20:985e993 36. Barca ML, Engedal K, Selbæk G: A reliability and validity study of the Cornell Scale among elderly inpatients, using various clinical criteria. Dement Geriatr Cogn Disord 2010; 29:438e447
Am J Geriatr Psychiatry 23:8, August 2015
37. Watson LC, Pignone M: Screening accuracy for late-life depression in primary care: a systematic review. J Fam Pract 2003; 52:956e964 38. Youden WJ: Index for rating diagnostic tests. Cancer 1950; 3: 32e35 39. Hanley JA, McNeil BJ: The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29e36 40. McCabe MP, Davison TE, Mellor D, et al: Depression among older people with cognitive impairment: prevalence and detection. Int J Geriatr Psychiatry 2006; 21:633e644 41. Maas M, Specht J, Buckwalter K, et al: Nursing home staffing and training recommendations for promoting older adults’ quality of care and life: part 1. Deficits in the quality of care due to understaffing and undertraining. Res Gerontol Nurs 2008; 1:123e152
793
