Age and Ageing 2005; 34: 169–184Age and Ageing Vol. 34 No. 2 British Geriatrics Society 2005; all rights reserved
Predicting post-operative delirium in
elective orthopaedic patients: the Delirium
Elderly At-Risk (DEAR) instrument
SIR—Delirium is a common problem in hospitalised older
people, with prevalence estimates between 15% and 60%
[1–3]. Estimates of delirium in elective orthopaedic patients
tend to be in the 10–40% range [3–6]. As there is accumu-
lating evidence for delirium prevention [1, 7–10], it is timely
to identify elective orthopaedic patients who are at highest
risk. Prospective trials have identified a number of risk fac-
tors for post-operative delirium (POD) in elective surgery
patients, including older age, pre-existing cognitive impair-
ment, sensory impairment, psychoactive medications, type
of surgery, history of previous delirium and pre-operative
functional impairment [3, 4, 11–15].
Delirium is associated with poor outcomes, at hospital
discharge and several-month follow-up [2, 16, 17]. Delirium
has been associated with longer length of hospital stay,
nursing home placement, increased risk of death, and
decline in function and cognition [16, 18]. It may be of ben-
efit to have standardised pre-operative cognitive screening
to identify patients at higher risk of acute confusion, and to
facilitate post-operative diagnosis and management of cog-
nitive decline. Using data collected prior to elective surgery,
Marcantonio et al.  were able to stratify patients into risk
groups for developing delirium. Risk factors included age,
pre-operative cognitive and functional status, self-reported
alcohol abuse and laboratory abnormalities . This is sim-
ilar to predictors identified on an acute geriatric unit .
Although clearly of value in predicting POD, information
on baseline cognition and function is rarely systematically
collected in usual practice. We endeavoured to translate
existing information and test its feasibility when collected by
nurses in the course of usual pre-operative care.
The diagnosis of delirium is commonly under-recog-
nised by nurses and physicians [20–22]. The likelihood that
nurses will detect delirium during routine care is lower in
the presence of dementia and high baseline delirium risk
. Detection could be improved by documentation of
baseline cognitive status, incorporating cognitive screening
into routine clinical practice, education or clinical pathways
[20, 23]. Identifying high-risk patients in the course of rou-
tine pre-operative care may allow for interventions designed
to prevent or reduce the severity of POD.
Previous work supports collecting baseline information on
risk factors to target patients for intervention. However,
research protocols often employ personnel who are not avail-
able in the usual clinical setting . We incorporated known
predictors for POD  into routine pre-operative information
gathering using the Delirium Elderly At-Risk instrument
(DEAR). We assessed the feasibility of incorporating the
DEAR into routine nursing care of elective orthopaedic
patients and evaluated its usefulness in predicting POD.
Setting and patients
Baseline data were collected routinely on patients over the
age of 65, by nurses in the pre-operative arthroplasty clinic,
and systematically recorded on five delirium risk factors on
the DEAR (Appendix 1 available as supplementary data at the
journal website: www.ageing.oupjournals.org). These five
known risk factors were selected for their feasibility, as they
use information that is readily available during routine pre-
operative nursing assessment. Clinic nurses also performed
cognitive screening using the Mini-Mental State Exam
(MMSE) . Nurses were instructed in performing the
MMSE, by attending a 30-minute teaching session with a
geriatrician. Validity and reliability of the MMSE have been
shown in many studies . Although abnormal laboratory
values have been associated with the development of delir-
ium [18, 19, 26], this information is not routinely interpreted
by clinic nurses and may not be available at the time that the
DEAR is recorded. This study was approved by the
research ethics board.
Orthopaedic surgery was elective arthroplasty of the hip
or knee, performed between 1 day and 4 weeks after pre-
operative assessment. Prior to discharge from the ortho-
paedic service, delirium was diagnosed (by H.M. and S.F.)
based on criteria from the Confusion Assessment Method
(CAM), which is sensitive, specific, reliable and valid for
identification of delirium . The four CAM criteria are:
(1) acute onset of confusion and fluctuating course; (2) inat-
tention; (3) disorganised thinking; and (4) altered level of
consciousness. Information with respect to delirium was
compiled (by H.M.) from discussions with the charge nurse,
attendance at team rounds, nursing notes and the medical
record, and discussed with a geriatrician (S.F.) if necessary.
Patients were seen daily on weekdays.
Descriptive statistics are reported. Factors related to incidence
of delirium were compared using the chi-square test or t-test.
The sensitivity and specificity of the DEAR in predicting
delirium were calculated and receiver operating characteristic
(ROC) analyses were performed. Logistic regression was
performed with delirium as the dependent variable. Odds
ratios and 95% confidence intervals (CI) are reported.
Baseline characteristics and information on POD were
collected on 132 elective arthroplasty patients (Table 1).
The mean age was 76.8 years (±8). Mean pre-operative
by guest on May 14, 2011
MMSE score was 27.7 (±3.2). The MMSE was not com-
pleted pre-operatively on two patients, but information on
other delirium risk factors was recorded. No patient refused
POD occurred in 18 patients (incidence 13.6%). Among
elective arthroplasty patients, having two or more risk fac-
tors was associated with an eight-fold increase in the inci-
dence of delirium (chi-square=6.33, P=0.01) and with an
increased length of stay (9.3 days versus 6.7 days; t=2.18,
P=0.031). ROC analysis gave an area under the curve
(AUC) of 0.77 (95% CI=0.64–0.0.87) for the DEAR in pre-
dicting delirium. The pre-selected cut-off of two or more
risk factors had a sensitivity of 0.61 and a specificity of 0.76.
The corresponding positive likelihood ratio was 2.58 and
the negative likelihood ratio was 0.51.
Logistic regression was used to explore the association
between pre-operative factors and POD. In the univariate
analysis, factors associated with the development of POD
included substance use (OR=4.71, 95% CI=1.63–13.57) and
cognitive impairment (OR=6.96, 95% CI=2.41–20.07).
Both substance use (OR=9.98, 95% CI=1.96–24.66) and
cognitive impairment (OR=8.26, 95% CI=2.44–27.99)
remained predictive of delirium with adjustment for other
baseline factors (see Table 2).
We studied 132 elective arthroplasty patients to identify
risks for delirium. The delirium rate in this study, 13.6%,
was within the expected range [3–6]. Nurses were able to
incorporate systematic risk factor recording and cognitive
screening into pre-operative care throughout the study
period, and have continued this practice as routine there-
after. The DEAR, which relies on baseline information on
known delirium risk factors, can be used to identify individ-
uals who are at greatest risk of POD. A score of two or
more on the assessment instrument was associated with a
greatly increased risk of delirium.
Our data must be interpreted with caution. Given the
relatively small number of events, logistic regression analy-
sis may be relatively unreliable. However, the factors in the
final model are consistent with well-recognised important
predictors. We did not attempt to measure delirium severity.
Identifying patients who fall anywhere on the spectrum of
delirium severity may be valuable for managing post-operative
care. Our primary intent was to apply existing information
to routine clinical care.
Cognitive impairment before surgery appears to be an
important predictor of POD, and should be routinely
screened. Pre-operative cognitive impairment, as meas-
ured by the MMSE [3, 4, 11, 18] or the Clock-Drawing
Test , has been found to be an important predictor for
Given the poor outcomes associated with delirium [16,
17, 28], it has been suggested that assessment of risk for
delirium should be incorporated into pre-operative evalua-
tion [3–6, 7, 29]. Nurses have a key role in early recognition
of delirium, yet without standardised risk factor screening
and cognitive assessment have demonstrated a low sensitiv-
ity for the detection of delirium . Despite this, chart
documentation of symptoms of delirium by nurses has been
found to be fairly sensitive , suggesting that nurses take
note of changes in the condition of delirious patients but
may not always make the correct interpretation.
Part of the special challenge in delirium is not only
that we do not know what best to do, but also that, quite
commonly, we do not do those things which we know we
ought to do, and we do too much of what we know we
ought not to do. In consequence, the translation of exist-
ing information into better routine care is an important
practical challenge, and one that largely has not been met.
Our study shows that this can be done. POD is an impor-
tant problem, and readily available baseline risk factors
can be used to identify patients at increased risk of acute
confusion and longer lengths of stay on an orthopaedic
unit. It is possible to incorporate the systematic docu-
mentation of POD risk factors, using the DEAR, into
routine pre-operative orthopaedic nursing care. Pre-
operative cognitive screening by MMSE is particularly
valuable for targeting patients at increased risk of post-
operative confusion, and provides a useful baseline when
evaluating POD. The nursing staff is in the best position
to perform routine cognitive screening on admission to
orthopaedics, and this task can be incorporated into usual
Further research is required to evaluate the DEAR in
other orthopaedic populations (e.g. hip fractures), and to
design and implement delirium prevention interventions
targeted at orthopaedic patients who are at increased risk of
Table 1. Baseline risk factors and length of stay for POD
Delirium (n=18) No delirium (n=114)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gender (% female) 65%
Pre-op MMSE (max 30)a28 (2.8)
Risk factors (% patients
with risk factor)
Age ≥80 years 44%
Sensory impairment 70.2%
Dependence in ≥1 ADL 13.2%
Cognitive impairmenta 8.4%
Substance use 29.8%
Length of stay (days) 6.7
aAnalysis excluded two patients owing to missing data.
61.1% χ2=15.4, P<0.0001
66.7% χ2=9.3, P=0.0023
9.3 t=2.18, P=0.031
Table 2. Multiple logistic regression analysis of DEAR risk
factors for POD
aThe −2 log likelihood improved from 92 to 80 with the additional variables.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.96 (2.41, 20.07)
Substance use 4.71 (1.63, 13.57)
Age ≥80 years 1.28 (0.47, 3.46)
Dependence in ≥1 ADL 1.89 (0.55, 6.50)
Sensory impairment 2.13 (0.58, 7.82)
Crude OR (CI)Adjusted OR (CI)
8.26 (2.44, 27.99)
6.98 (1.98, 24.66)
2.67 (0.72, 9.91)
1.83 (0.40, 8.31)
0.86 (0.18, 4.02)
by guest on May 14, 2011
• Delirium is common in post-operative orthopaedic
• Patients at high risk of developing POD can be identified
using risk factor data available pre-operatively.
• Systematic collection of baseline delirium risk factors can
be accomplished by using the DEAR instrument, and
can be incorporated into routine pre-operative nursing
SUSAN H. FRETER1*, MICHAEL J. DUNBAR2, HEATHER MACLEOD3,
MICHELLE MORRISON4, CHRIS MACKNIGHT1,
1Dalhousie University, Department of Medicine, Division of
Geriatric Medicine, Halifax, Nova Scotia, Canada
Fax: (+1) 902 473 7133
2Dalhousie University, Department of Surgery, Division of
Orthopaedics, Halifax, Nova Scotia, Canada
3Capital District Health Authority, Department of
Occupational Therapy and
4Nursing, Halifax, Nova Scotia, Canada
*To whom correspondence should be addressed
1. Cole MG, Primeau F, McCusker J. Effectiveness of interven-
tions to prevent delirium in hospitalized patients: a systematic
review. Can Med Assoc J 1996; 155: 1263–8.
2. Edlund A, Lundström M, Lundström G, Hedqvist B,
Gustafson Y. Clinical profile of delirium in patients treated for
femoral neck fractures. Dement Geriatr Cogn Disord 1999;
3. Galanakis P, Bickel H, Gradinger R, Von Gumppenberg S,
Förstl H. Acute confusional state in the elderly following hip
surgery: incidence risk factors and complications. Int J Geriatr
Psychiatry 2001; 16: 349–55.
4. Dupplis GS, Wikblad K. Acute confusional states in patients
undergoing hip surgery. Gerontology 2000; 46: 36–43.
5. Litaker D, Locala J, Franco K, Bronson DL, Tannous Z. Pre-
operative risk factors for postoperative delirium. Gen Hosp
Psychiatr 2001; 23: 84–9.
6. Williams-Russo P, Urquhart BL, Sharrock NE, Charlson ME.
Post operative delirium: predictors and prognosis in elderly
orthopedic patients. J Am Geriatr Soc 1992: 40; 759–67.
7. Inouye SK, Bogardus ST, Charpentier PA et al. A multicompo-
nent intervention to prevent delirium in hospitalized older
patients. N Engl J Med 1999; 340: 669–76.
8. Milisen K, Foreman MD, Abraham IL et al. A nurse-led inter-
disciplinary intervention program for delirium in elderly hip-
fracture patients. J Am Geriatr Soc 2001; 49: 523–32.
9. Marcantonio ER, Flacker JM, Wright J, Resnick NM. Reduc-
ing delirium after hip fracture: a randomized trial. J Am
Geriatr Soc 2001; 49: 516–22.
10. Gustafson Y, Brännström B, Berggren D et al. A geriatric-
anesthesiologic program to reduce acute confusional states in
elderly patients treated for femoral neck fractures. J Am
Geriatr Soc 1991; 39: 655–62.
11. Edlund A, Lundström R, Brännström B, Bucht G, Gustafson Y.
Delirium before and after operation for femoral neck fracture.
J Am Geriatr Soc 2001; 49: 1335–40.
12. MacDonald JB, Dutton MJ, Stott DJ, Hamblen DL. Evalua-
tion of pre-admission screening of elderly patients accepted
for major joint replacement. Health Bull 1992; 50: 54–60.
13. Fisher BW, Flowerdew G. A simple model for predicting
postoperative delirium in older patients undergoing elective
orthopedic surgery. J Am Geriatr Soc 1995; 43: 175–8.
14. Brauer C, Morrison S, Silberzweig SB, Siu AL. The cause of
delirium in patients with hip fracture. Arch Intern Med 2000;
15. Dai YT, Lou MF, Yip PK, Huang GS. Risk factors and inci-
dence of postoperative delirium in elderly Chinese patients.
Gerontology 2000; 46: 28–35.
16. Inouye SK, Rushing JT, Foreman MD, Palmer RM, Pompei P.
Does delirium contribute to poor hospital outcomes? J Gen
Intern Med 1998; 13: 234–242.
17. Rockwood K. The occurrence and duration of symptoms in
elderly patients with delirium. J Gerontol Med Sci 1993; 48:
18. Marcantonio ER, Goldman L, Mangione CM et al. A clinical
prediction rule for delirium after elective noncardiac surgery.
JAMA 1994; 271: 134–9.
19. O’Keeffe ST, Lavan JN. Predicting delirium in elderly patients:
development and validation of a risk-stratification model. Age
Ageing 1996; 25: 317–21.
20. Rockwood K, Cosway S, Stolee P et al. Increasing the recogni-
tion of delirium in elderly patients. J Am Geriatr Soc 1994; 42:
21. GustafsonY, Brännström B, Norberg A, Bucht G, Winblad B.
Underdiagnosis and poor documentation of acute confusional
states in elderly hip fracture patients. J Am Geriatr Soc 1991;
22. Inouye SK, Foreman MD, Mion LC, Katz KH, Cooney LM.
Nurses’ recognition of delirium and its symptoms. Arch Intern
Med 2001; 161: 2467–73.
23. Inouye SK, Schlesinger MJ, Lydon TJ. Delirium: a symptom
of how hospital care is failing older persons and a window to
improve quality of hospital care. Am J Med 1999; 106: 565–73.
24. Folstein MF, Folstein SE. Mini-mental state: a practical guide
for grading the cognitive state of patients for the clinician.
J Psychiatr Res 1975; 12: 189–98.
25. Tombaugh TN, McIntyre NJ. The mini-mental state examina-
tion: a comprehensive review. J Am Geriatr Soc 1992; 40:
26. Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A
predictive model for delirium in hospitalized elderly medical
patients based on admission characteristics. Ann Intern Med
1993; 119: 474–81.
27. Inouye SK, van Dyck CH, Alessi CA et al. Clarifying confu-
sion: the Confusion Assessment Method. Ann Intern Med
1990; 113: 941–8.
28. O’Keeffe S, Lavan J. The prognostic significance of delirium
in older hospitalized patients. J Am Geriatr Soc 1997; 45:
29. Andersson EM, Gustafson L, Hallberg IR. Acute confusional
state in elderly orthopaedic patients: factors of importance
for detection in nursing care. Int J Geriatr Psychiatry 2001;
30. Rolfson DB, McElhaney JE, Jhangri GS, Rockwood K.
Validity of the Confusion Assessment Method in detecting
postoperative delirium in the elderly. Int Psychogeriatr 1999;
by guest on May 14, 2011