Age and Ageing 2008; 37: 690695 doi: 10.

1093/ageing/afn193

The Author 2008. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Does depression predict adverse outcomes for older medical inpatients? A prospective cohort study of individuals screened for a trial
SARAH CULLUM1 , CHRIS METCALFE2 , CHRIS TODD3 , C AROL BRAYNE4
Division of Psychiatry, Cotham House, Cotham Hill, Bristol BS6 6JL, UK Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UK 3 School of Nursing, Midwifery and Social Work, University Place, University of Manchester, Oxford Road, Manchester M13 9PL, UK 4 Department of Public Health & Primary Care, Institute of Public Health, University Forvie Site, Robinson Way, Cambridge, CB2 2SR, UK
2 1

Address correspondence to: Sarah Cullum, Division of Psychiatry, Cotham House, Cotham Hill, Bristol BS6 6JL, UK. Email: sarah.cullum@awp.nhs.uk This work was performed at West Suffolk Hospital and Department of Public Health and Primary Care, University of Cambridge.

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

Abstract
Objective: to examine the relationship between depressive symptoms and hospital outcomes in an unselected consecutive sample of older medical inpatients. Design: a prospective cohort study of individuals screened for a trial. Setting: medical wards of UK district general hospital in rural East Anglia. Participants: six hundred and seventeen medical inpatients aged 65+ were randomly selected from consecutive admissions. Baseline measures: 15-item Geriatric Depression Scale (GDS-15), the Abbreviated Mental Test Score (AMTS) and the Cumulative Illness Rating ScaleGeriatric (CIRS-G). Main outcome measures: length of hospital stay; discharge to a community hospital (for rehabilitation), institutional care or usual place of residence; dying in hospital. Results: depressive symptoms are independently associated with an increased likelihood of inpatient death and transfer to a community hospital for rehabilitation, but are not associated with longer length of stay. Conclusions: research evaluating effectiveness of identication and treatment of depression in older medical inpatients should consider including inpatient death and use of rehabilitation services as potential outcomes.
Keywords: depression, aged, inpatients, cohort study, elderly

Introduction
Studies suggest that a substantial proportion of older hospital inpatients have depressive symptoms that may hinder their recovery. Adverse outcomes include delayed discharge [1, 2], greater risk of readmission [3], increased need for rehabilitation or institutional care [4] and higher mortality [49]. In order to attempt to reduce the impact of depression on patients health outcomes and healthcare costs, recent UK health policy encourages screening and treatment programmes for older people in mainstream healthcare settings [10, 11]. However, the research evidence cited above derives mainly from countries other than the UK and may not be generalisable to older people in UK hospitals. The aim of

this study is to examine the relationship between depressive symptoms and hospital outcomes in an unselected consecutive sample of older medical inpatients in a UK hospital, thus informing the selection of measurable outcomes that might be impacted by a policy of screening and treatment of depression in this population.

Methods
Participants

Acute admissions to the medical wards of a district general hospital aged 65 and over who had been in hospital between 3 and 7 days were eligible for the study. As

690

Adverse outcomes for older medical inpatients

there were insufcient resources to assess all eligible patients, 50% of consecutive admissions were randomly selected for the study using a list generated using block randomisation. Data were collected over a period of 15 months. Patients were excluded if they had severe dysphasia, severe deafness or were too physically unwell and/or confused to be interviewed; the remainder were approached for consent to take part in the study. Participants who screened positive for depression (22% of the total) were subsequently invited to a further diagnostic interview and to take part in a nested randomised controlled trial of liaison psychiatric nursing [12]. gories of each covariate. All analyses employed Stata version 9 (StataCorp, 2005, College Station, TX, USA).
Ethical issues

Informed consent was obtained from all participants. Participants who answered afrmatively to the screening question regarding suicidal ideas were immediately identied to ward staff with a recommendation for further assessment by a mental health professional. This study was approved by the West Suffolk Hospital LREC.

Results
From 2112 eligible admissions over a period of 15 months, 1009 older medical inpatients were randomly selected into the study. Three hundred and eighteen did not meet the inclusion criteria for the screen, 42 refused consent and 31 were excluded after the screen (including one participant who expressed suicidal ideas), as reported in [20]. Data from one participant who had not been discharged from hospital at the end of the study were not included; thus data from 617 participants were used in the analyses. Of 617 participants, 181 (29%) reported a history of depression and 76 (12%) were already known to psychiatric Table 1. Length of stay in days by covariates
................................................................ All sample 617 19.48 12.0 5.0060.00 Age groups 6569 57 12.05 9.0 4.029.0 7074 104 15.04 9.0 4.046.0 7579 119 21.99 11.0 5.069.0 8084 150 21.65 15.0 5.064.0 8589 122 20.84 13.0 5.060.0 90+ 64 21.05 13.0 5.061.0 P for trend 0.001 Sex Female 363 20.46 12.0 5.064.0 Male 254 18.08 12.0 4.053.0 P for trend 0.20 CIRS-G 27 122 19.40 11.0 5.069.0 89 108 18.33 11.5 4.053.0 1011 132 20.27 12.0 5.059.0 1213 134 18.21 13.0 5.061.0 1421 121 21.12 14.0 4.047.0 P for trend 0.68 AMTS 6/7 94 25.26 15.5 5.080.0 8 104 18.82 13.0 5.046.0 9 184 20.72 5.060.0 12.5 10 235 16.49 10.0 4.047.0 P for trend 0.006 GDS-15 04 347 18.17 11.0 4.053.0 57 133 21.13 13.0 5.069.0 810 89 17.43 12.0 5.043.0 1115 48 28.19 16.5 5.095.0 P for trend 0.13
P-values adjusted for exposure to the trial intervention. n Mean Median 90% range

Materials
Participants were assessed using the 15-item Geriatric Depression Scale (GDS-15) [13], the Abbreviated Mental Test Score (AMTS) [14] and the Cumulative Illness Rating Scale Geriatric (CIRS-G) [15], a measure of physical co-morbidity. A higher score on the GDS-15 indicates that the patient has more depressive symptoms. The CIRS-G scores severity on 13 different organ systems ranged from 0 to 4; a higher score indicates a higher level of physical co-morbidity. The AMTS is a commonly used brief measure of cognitive function scored out of 10; a cut-off of 7 or 8 is suggested to discriminate between cognitive impairment and normality [16]. Participants with moderately severe impaired cognitive function (AMTS<6) were excluded, as their responses to the other rating scales were likely to be unreliable. All rating scales have been validated in older people [16 18]. Patients were tracked on the hospital information system to ascertain discharge date and destination or inpatient death.
Statistical analysis

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

The associations of length of stay with depressive symptoms and cognitive function were estimated using ordinary leastsquares regression. To accommodate the large positive skew in the distribution of length of stay, condence intervals were obtained using the bias-corrected and accelerated percentile bootstrap method (4999 resamplings), and corresponding P-values were calculated using a robust estimate of the standard error [19]. Logistic regression was used to compare the odds of the patient being discharged to a community hospital (for rehabilitation) or institutional care, or dying in hospital, rather than being discharged to their usual place of residence. Data from 62 participants who subsequently entered the trial and received the active treatment were included, but we allowed for exposure to the intervention in all the regression analyses. Consequently, all analyses were repeated three times with accumulating adjustment for exposure to the trial intervention, age and sex, physical co-morbidity and cognitive function. Adjustment was achieved by adding dummy variables to the regression equation, distinguishing the cate-

691

S. Cullum et al.
Table 2. Association of length of hospital stay with depressive symptoms
Crude (1) Mean difference 95% CI + Age, sex adjusted (2) Mean difference 95% CI Fully adjusted (3) Mean difference 95% CI

................................................................................................................ GDS-15 04 Baseline Baseline Baseline 57 2.96 1.37, 7.29 2.96 1.33, 7.25 2.65 1.61, 6.90 810 0.21 5.01, 5.43 0.31 4.84, 5.46 1.75 5.88, 2.37 1115 11.42 2.84, 25.67 11.06 3.48, 25.61 9.46 1.80, 20.73 P for trend 0.13 0.14 0.17
Adjusted for (1) exposure to the trial intervention, (2) plus age and sex, (3) plus co-morbidity and cognitive function.

Table 3. Hospital outcome by GDS-15 score

Discharge destination Community hospital

................................................................................................................ 04 290 (83.5%) 8 (2.5%) 16 (4.5%) 14 (4.0%) 19 (5.5%) 57 101 (76.0%) 0 (0.0%) 4 (3.0%) 8 (6.0%) 20 (15.0%) 810 61 (68.5%) 2 (2.5%) 6 (6.5%) 9 (10.0%) 11 (12.5%) 1115 30 (62.5%) 1 (2.0%) 6 (12.5%) 5 (10.5%) 6 (12.5%) Total 482 (78.0%) 11 (2.0%) 32 (5.0%) 36 (6.0%) 56 (9%) Odds of hospital outcome by GDS-15 score Crude (1) OR 04 57 810 1115 P for trend 04 57 810 1115 P for trend 04 57 810 1115 P for trend Baseline 1.64 4.20 5.53 0.001 Baseline 0.72 1.42 2.76 0.258 Baseline 3.02 2.41 2.53 0.012 Age and sex adjusted (2) 95% CI OR 95% CI Community hospital (n = 36) 0.67, 4.03 1.66, 10.61 1.69, 18.06 Baseline 1.66 4.64 6.23 0.67, 4.14 1.78, 12.08 1.79, 21.73 Fully adjusted (3) OR Baseline 1.66 3.35 3.98 0.004 Baseline 0.70 1.71 4.05 0.032 Baseline 2.67 2.11 2.80 0.017 95% CI

GDS-15 score

Usual residence

Care home

Other hospital

Death

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

0.66, 4.19 1.29, 8.73 1.22, 12.97

0.001 Other hospital (n = 64) 0.23, 2.20 0.44, 4.59 0.78, 9.73 Baseline 0.69 1.35 3.02 Death (n = 56) 1.55, 5.89 0.95, 6.09 0.77, 8.37 Baseline 3.19 2.70 2.48 0.010 0.247 0.22, 2.13 0.41, 4.49 0.84, 10.92

0.22, 2.17 0.60, 4.85 1.34, 12.25

1.62, 6.29 1.06, 6.87 0.71, 8.66

1.33, 5.39 0.91, 4.90 0.99, 7.95

Each destination in turn is compared with discharge to usual place of residence, adjusted for (1) exposure to the trial intervention, (2) plus age and sex, (3) plus co-morbidity and cognitive function

services. The high depression score correlated with a lower cognition score (Spearmans r = 0.10, P = 0.013) and a higher co-morbidity score (r = 0.26, P<0.001), whereas there was no evidence of association between cognition and co-morbidity scores (r = 0.04, P = 0.38).
Length of stay

nitive function and higher score on the depression rating GDS-15 but was largely unaffected by gender. Table 2 shows that, after adjustment for potential confounding factors, depressive symptoms were not associated with longer length of stay (P for trend = 0.17).
Hospital outcomes

Table 1 shows that median length of stay in hospital increased with age, greater physical co-morbidity, poorer cog-

Table 3 shows the hospital outcome for inpatients in terms of discharge destination and mortality by a GDS-15 score on admission.

692

Adverse outcomes for older medical inpatients

Rehabilitation

Of the 617 participants, 36 (5.8%) were transferred to a community hospital for rehabilitation prior to discharge. Table 3 shows the odds of being transferred to a community hospital (compared with discharge to usual place of residence) by depression rating score on admission. The adjusted odds were higher in participants with a higher GDS-15 depression rating score (P for trend = 0.004).
Institutional care

Only 11 of the 617 (1.8%) participants were discharged from hospital to institutional care (residential or nursing home), which precludes an investigation of the association with depressive symptoms.
Inpatient mortality

both the eligibility and inclusion criteria. Patients who were discharged before 3 days or were too ill/confused to participate in the study might have had different outcomes to those who took part. Nonetheless, the ndings are generalisable to those older medical inpatients who are most likely to be screened for depression in a real world NHS setting, i.e. those who stay in hospital for more than 3 days and are t enough to complete a screening questionnaire. It is important to note that this study examined the effect of depressive symptoms rather than the clinical syndrome of depressive disorder although there is a strong association between the GDS score and clinical diagnosis. An associated prevalence study included one-third of the participants in the present study, with diagnostic interviews allowing the prevalence of ICD-10 depressive disorder to be estimated as 17.7% (95% CI 12.922.5) [20].

Fifty-six of the 617 (9.1%) participants died in hospital. Table 3 shows the odds of death in hospital compared with discharge to usual place of residence by depression rating score on admission. A higher GDS-15 score on admission was associated with a higher risk of inpatient death, even after adjustment for potential confounding factors such as age, physical co-morbidity and cognitive function (P = 0.017).

Comparison with previous studies

Length of stay

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

Discussion
The main ndings of this study are that a higher depression rating on admission to hospital is associated with increased odds of inpatient death and increased odds of being transferred to a community hospital for rehabilitation independent of age, sex and physical status. Depressive symptoms were not associated with longer length of hospital stay. Most of the screen positive participants were invited to take part in a nested trial of liaison psychiatric nursing. The active intervention in the trial was assessment and implementation of a management plan by a liaison psychiatric nurse compared to standard care. It is possible that the active intervention may have biased our estimate of the association between depressive symptoms and adverse outcomes, but the trial intervention had little effect and moreover was adjusted for in the analyses. A Hawthorne effect, due to all trial participants being interviewed again at follow-up, may also have biased the ndings, as trial participants had higher depression ratings. However, this is unlikely as most of the interviews took place after discharge.
Strengths and weaknesses

Studies carried out in the USA [21] and Italy [22, 23] support the ndings of our study that, after adjustment for potential confounding factors, length of stay is not associated with depressive symptoms. However, a Swiss study [24] found that depressed mood (GDS-15>6) was associated with inappropriate hospital days (P = 0.001), indicating that, although depressive symptoms may not increase total length of stay, they may increase inappropriate use of inpatient hospital resources.
Other hospital outcomes

Very few studies have considered whether depressed mood affects the need for rehabilitative or long-term institutional care, but our study found that depressive symptoms were associated with transfer to a community hospital for rehabilitation. In Switzerland, Bula et al. [4] similarly found that depressive symptoms were associated with an increased risk of nursing home placement and increased average costs of both acute and rehabilitation services in elderly medical inpatients. However, after adjustment for demographic factors and physical status, only the association for rehabilitation services remained. This is consistent with the ndings in our study.
Mortality

The main strength of the study was its large sample size, which was greater than that in any other identied study that has measured similar outcomes. The large sample size, plus the measurement of and adjustment for potential confounding factors, permits a more condent interpretation of the ndings. The main weakness of the study was that the ndings could only be generalised to the 45% of admissions that met

A number of studies, with follow-ups ranging from 90 days to 9 years, have reported an association between depressive symptoms and longer-term mortality [49]. However, very few studies have examined the effect of depressive symptoms on inpatient mortality. One study [1] reported that mortality in older medical inpatients was signicantly higher among depressed compared with nondepressed controls (6 vs 0 deaths, P = 0.03), but the analyses were not adjusted for potential confounding factors associated with depressive symptoms such as physical co-morbidity. Another study [25], carried out in a population of younger medical inpatients (mean age

693

S. Cullum et al.
50 years), found that, after adjustment for age, number, type and severity of patients illnesses, the diagnosis of major depressive disorder did predict mortality, whereas demographic variables and medical diagnoses did not. These ndings are supported by the results of our study that inpatient mortality is independently associated with severity of depressive symptoms.
Does identication and treatment of depression make a difference?

tively improves (or does not improve) health outcomes will inuence the development of healthcare policy. Researchers carrying out these studies should consider including inpatient death and use of rehabilitation services as potential outcomes.

Key points
r Depression is common in older medical inpatients, but

We can only speculate about the causal pathways underlying these ndings. Does depression reect an underlying diminution in the bodys ability to recover from serious physical illness, thus increasing the likelihood of inpatient death and/or need for rehabilitation services? Or do hospital staff misdiagnose depression as slower recovery from physical illness and thus recommend rehabilitation when they should be recommending treatment for the depressive illness instead? Frailty is used loosely to describe a range of conditions in older people, including general debility and cognitive impairment associated with increased economic costs [26]. It is generally asserted that a combination of biomedical and psychosocial factors inuences frail peoples physiological state to the extent that its function is reduced, and subsequent exposure to further minor environmental stresses may lead to dependence. The ndings of this study suggest that depressive symptoms may be an important indicator of frailty. If depressive symptoms are a marker for frailty, then the question is: will treatment of depression have any impact on health outcomesmental or physical? Studies in primary care demonstrate that treatment of depression improves mental health, physical health and quality of life [2729], but the evidence for the effectiveness of treatment of depression in general hospital inpatients is less clear [12, 3032]. Two recent RCTs [12, 33] carried out in general hospital settings (in Canada and the UK) were unable to demonstrate that systematic detection and multidisciplinary care of depression was more benecial than usual care for elderly medical inpatients, but both of these studies were underpowered. Another larger UK study [34] evaluated the effectiveness of a nurseled mental health liaison service for older medical inpatients and found that subjects randomised to the intervention arm had signicantly lower GDS scores at follow-up than those receiving usual care.

its recognition is poor and treatment often inadequate, despite research suggesting that depression may be independently associated with adverse hospital outcomes. r The aim of this study is to examine the relationship between depressive symptoms and hospital outcomes in older medical inpatients. r Our study found that depressive symptoms are independently associated with an increased likelihood of inpatient death and transfer to a community hospital for rehabilitation, but are not associated with longer length of stay. r Further research is required to establish whether screening, assessment and treatment of depression identied in this population will reduce the risk of these adverse outcomes.

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

Acknowledgements
Maggie Bray (research assistant); Barbara Lee (research nurse); Linda Sharples, Lu Gao and Tony Johnson (medical statisticians); Prof. Ann Louise Kinmonth (sponsorship of MRC Fellowship and advice on study design).

Conicts of interest
None.

Funding
This work was supported by a MRC Health Services Research Training Fellowship (G106/950) and a NHS Executive Eastern Research and Development Project Grant (RCC33041). Financial sponsors played no role in the design, execution, analysis and interpretation of data or writing of the study.

References Conclusions
This study has shown that depressive symptoms are associated with an increased likelihood of inpatient death and transfer to a community hospital for rehabilitation, but are not associated with longer length of stay. Currently, we have no robust research evidence regarding the impact and costeffectiveness of screening, assessment and treatment of depression identied in this population. There is clearly a need for further research in the area, as evidence that identication and treatment of depression in older medical inpatients effec1. Koenig HG, Shelp F, Goli V, Cohen HJ, Blazer DG. Survival and health care utilization in elderly medical inpatients with major depression. J Am Geriatr Soc 1989; 37: 599606. 2. Holmes J, House A. Psychiatric illness predicts poor outcome after surgery for hip fracture: a prospective cohort study. Psychol Med 2000; 30: 9219. 3. Koenig HG, Kuchibhatla M. Use of health services by medically ill depressed elderly patients after hospital discharge. Am J Geriatr Psychiatry 1999; 7: 4856. 4. Bula CJ, Wietlisbach V, Burnand B, Yersin B. Depressive symptoms as a predictor of 6-month outcomes and services

694

Adverse outcomes for older medical inpatients

utilization in elderly medical inpatients. Arch Intern Med 2001; 161: 260915. Covinsky KE, Kahana E, Chin MH, Palmer RM, Fortinsky RH, Landefeld CS. Depressive symptoms and 3-year mortality in older hospitalized medical patients. Ann Intern Med 1999; 130: 5639. Inouye SK, Peduzzi PN, Robison JT, Hughes JS, Horwitz RI, Concato J. Importance of functional measures in predicting mortality among older hospitalized patients. JAMA 1998; 279: 118793. Arfken CL, Lichtenberg PA, Tancer ME. Cognitive impairment and depression predict mortality in medically ill older adults. J Gerontol A Biol Sci Med Sci 1999; 54: M152 M156. Ganzini L, Smith DM, Fenn DS, Lee MA. Depression and mortality in medically ill older adults. J Am Geriatr Soc 1997; 45: 30712. Koenig HG, George LK, Larson DB, McCullough ME, Branch PS, Kuchibhatla M. Depressive symptoms and nine-year survival of 1,001 male veterans hospitalized with medical illness. Am J Geriatr Psychiatry 1999; 7: 12431. National Institute for Clinical Excellence. Depression: Management of Depression in Primary and Secondary Care. Clinical Guideline 2004, 23. Department of Health, Care Services Improvement Partnership. Everybodys Business. Integrated Mental Health Service for Older Adults: A Service Development Guide. London: Department of Health, 2005. Cullum S, Tucker S, Todd C, Brayne C. Effectiveness of liaison psychiatric nursing in older medical inpatients with depression: a randomised controlled trial. Age Ageing 2007; 36: 43642. Sheikh J, YesavageJ. Geriatric Depression Scale (GDS). Recent evidence and development of a shorter version. Clin Gerontol 1986; 5: 16573. Hodkinson M. Evaluation of a mental test score for assessment of mental impairment in the elderly. Age Ageing 1972; 1: 2338. Miller MD, Paradis CF, Houck PR et al. Rating chronic medical illness burden in geropsychiatric practice and research: application of the cumulative illness rating scale. Psych Res 1992; 41: 23748. Qureshi KN, Hodkinson HM. Evaluation of a ten-question mental test in the institutionalized elderly. Age Ageing 1974; 3: 1527. McDowell I, Newell C. Measuring Health: A Guide to Rating Scales and Questionnaires. Oxford: Oxford University Press, 1996. Harboun M, Ankri J. Comorbidity indexes: review of the literature and application to studies of elderly population. Rev Epidemiol Sante Publique 2001; 49: 28798. Kirkwood BR, Sterne JAC. Essential Medical Statistics. Oxford: Blackwell, 2003. 20. Cullum S, Tucker S, Todd C, Brayne C. Screening for depression in older medical inpatients. Int J Geriatr Psychiatry 2006; 21: 46976. 21. Koenig H, Kuchibhatla M. Use of health services by hospitalized medically ill depressed elderly patients. Am J Psychiatry 1998; 155: 85960. 22. Di-Iorio A, Longo A, Mitidieri CA et al. Factors related to the length of in-hospital stay of geriatric patients. Aging (Milano) 1999; 11: 1504. 23. Bertozzi B, Barbisoni P, Franzoni S, Rozzini R, Frisoni GB, Trabucchi M. Factors related to length of stay in a geriatric evaluation and rehabilitation unit. Aging (Milano) 1996; 8: 1705. 24. Ingold BB, Yersin B, Wietlisbach V, Burckhardt P, Bumand B, Bula CJ. Characteristics associated with inappropriate hospital use in elderly patients admitted to a general internal medicine service. Aging (Milano) 2000; 12: 4308. 25. von Ammon CS, Furlanetto LM, Creech SD, Powell LH. Medical illness, past depression, and present depression: a predictive triad for in-hospital mortality. Am J Psychiatry 2001; 158: 438. 26. Lally F, Crome P. Understanding frailty. Postgrad Med J 2007; 83: 1620. 27. Unutzer J, Katon W, Callahan CM et al. Collaborative care management of late-life depression in the primary care setting: a randomized controlled trial. JAMA 2002; 288: 283645. 28. Bruce ML, Ten Have TR, Reynolds CF III et al. Reducing suicidal ideation and depressive symptoms in depressed older primary care patients: a randomized controlled trial. JAMA 2004; 291: 108191. 29. Lin EH, Katon W, Von Korff M et al. Effect of improving depression care on pain and functional outcomes among older adults with arthritis: a randomized controlled trial. JAMA 2003; 290: 24289. 30. Tan RSH, Barlow RJ, Abel C, Reddy S et al. The effect of low dose lofepramine in depressed elderly patients in general medical wards. Br J Clin Pharmacol 1994; 37: 3214. 31. Evans M, Hammond M, Wilson K, Lye M, Copeland J. Placebocontrolled treatment trial of depression in elderly physically ill patients. Int J Geriatr Psychiatry 1997; 12: 81724. 32. Shah A, Odutoye K, De T. Depression in acutely medically ill elderly inpatients: a pilot study of early identication and intervention by formal psychogeriatric consultation. J Affect Disord 2001; 62: 23340. 33. Cole MG, McCusker J, Elie M, Dendukuri N, Latimer E, Belzile E. Systematic detection and multidisciplinary care of depression in older medical inpatients: a randomized trial. CMAJ 2006; 174: 3844. 34. Baldwin R, Pratt H, Goring H, Marriott A, Roberts C. Does a nurse-led mental health liaison service for older people reduce psychiatric morbidity in acute general medical wards? A randomised controlled trial. Age Ageing 2004; 33: 4728.
Received 11 August 5 September 2008 2007; accepted in revised form

5.

6.

7.

8.

9.

Downloaded from http://ageing.oxfordjournals.org/ by guest on February 4, 2012

10.

11.

12.

13.

14. 15.

16.

17.

18.

19.

695