The 6- and 12-Month Outcomes of Older Medical Inpatients Who
Recover from Subsyndromal Delirium
Martin G. Cole, MD, FRCP(C),?wJane McCusker, MD, DrPH,z§Antonio Ciampi, PhD,z§
and Eric Belzile, MScz
OBJECTIVES: To compare the 6- and 12-month outcomes
of patients who recovered from subsyndromal delirium
(SSD)by 8 weeks with the outcomes of patients who didnot
recover or did not have an index episode.
DESIGN: Secondary analysis of data collected for a cohort
study of the prognosis of delirium.
SETTING: University-affiliated primary acute care hospi-
PARTICIPANTS: Older medical inpatients with prevalent,
incident, or no SSD were classified into three mutually ex-
clusive groups at 8 weeks (SSD-recovered, SSD-not recov-
ered, no SSD) and followed up at 6 and 12 months.
MEASUREMENTS: The primary hierarchical composite
outcome was death, institutionalization, or cognitive or
functional decline at 6 and 12 months. In secondary ana-
lyses, components of the primary outcome were examined
RESULTS: Of the 129 patients assessed at 8 weeks, 51, 47,
and 31 met criteria for SSD-recovered, SSD-not recovered
and no SSD, respectively. At 6 and 12 months, the primary
and secondary outcomes of the SSD-recovered group were
better than the outcomes of the SSD-not recovered group
and, for the most part, intermediate between the outcomes
of the SSD-not recovered and no SSD groups.
CONCLUSION: Recovery from SSD appears to predict
better longer-term outcomes than no recovery. Efforts to
identify and treat SSD in older medical inpatients may im-
prove outcomes. J Am Geriatr Soc 56:2093–2099, 2008.
Key words: aged; subsyndromal delirium; recovery;
ness, inattention, disorientation, memory impairment,
disorganized thinking, and perceptual and motor distur-
bances.1However, it is common for patients to display one
or more of these symptoms without having the full syn-
drome of delirium. Such symptoms may precede or follow
an episode of full-blown delirium or never progress to full-
blown delirium. The latter condition is known as sub-
syndromal delirium (SSD), which occurs in 21% to 76% of
older medical inpatients.2–8
Thereare noofficially recognizeddiagnostic criteriafor
SSD. Two studies2,4have defined SSD as the presence of one
or more ‘‘core’’ symptoms of delirium, not meeting the cri-
teria for full delirium; the ‘‘core’’ symptoms included dis-
orientation, inattention, altered level of consciousness, and
perceptual disturbances. Three studies3,5,6have defined
SSD as the presence of one or more ‘‘core’’ symptoms of
delirium, not meeting the criteria for full delirium; the
‘‘core’’ symptoms included acute onset and fluctuating
course, inattention, disorganized thinking, and altered level
of consciousness. Finally, two studies7,8defined SSD as the
presence of one to three symptoms of delirium, not meeting
the criteria for full delirium; the symptoms included altered
level of consciousness, inattention, disorientation, halluci-
nations or delusions, agitation or retardation, inappropri-
ate speech or mood, sleep–wake cycle disturbance, and
symptom fluctuation. Irrespective of the diagnostic criteria
for SSD, SSD was consistently associated with adverse out-
comes such as greater cognitive impairment and functional
disability, longer acute care hospital stay, higher rates of
admission to long-term care institutions, and higher rates
It is unknown whether recovery from SSD affects these
adverse outcomes. Because knowledge of the outcomes of
patients who recover from SSD may inform decisions to
investigate and treat SSD, this analysis was designed to
compare the 6- and 12-month outcomes (cognition, basic
activities of daily living, institutionalization, mortality)
of patients who recovered from SSD by 8 weeks with the
elirium is a cognitive disorder characterized by acute
onset, fluctuating course, altered level of conscious-
Address correspondence to Martin G. Cole, MD, FRCP(C), Department of
Psychiatry, St. Mary’s Hospital Center, 3830 avenue Lacombe, Montreal,
Quebec, H3T 1M5, Canada. E-mail: firstname.lastname@example.org
From the?Department of Psychiatry;zClinical Epidemiology and Commu-
nity Studies, St. Mary’s Hospital Center;wDepartment of Psychiatry,
Montreal, Quebec, Canada; and§Epidemiology and Biostatistics, McGill
University, Montreal, Quebec, Canada
r 2008, Copyright the Authors
Journal compilation r 2008, The American Geriatrics Society
outcomes of patients whodid notrecoverordid nothavean
This study was a secondary analysis of data collected for a
cohort study of the prognosis of delirium9,10and used in a
previous study of the prognostic significance of SSD.4The
original study was conducted at St. Mary’s Hospital, a 400-
bed primary acute care university-affiliated hospital in
Montreal. A study nurse (SN) was responsible for patient
screening and enrollment. Patients aged 65 and older who
were admitted from the emergency department to medical
or geriatric services were included. Patients with a primary
diagnosis of stroke, those admitted to oncology (at St.
Mary’s Hospital, these patients are in a terminal phase of
their illness), those admitted to the intensive care unit or
cardiac monitoring unit (an ethics requirement) unless they
were transferred to a medical ward within 48 hours of ad-
mission, and those who did not speak English or French
The SN screened all patients using the Short Portable
Mental Status Questionnaire (SPMSQ)11and examined
their nursing notes daily during the first week of hospital-
ization. The SN administered the Confusion Assessment
Method (CAM)12to patients whose SPMSQ score was 3 or
more or whose nursing notes indicated symptoms of delir-
ium. The SN used various data sources to complete the
CAM (medical chart, family, nursing staff, cognitive as-
sessment using theMini-Mental
(MMSE)) and assessed the patient at several points in time,
if necessary. Delirium was diagnosed if the patient met Di-
agnostic and Statistical Manual of Mental Disorders, Third
Edition, Revised (DSM-III-R) criteria for delirium.1A ran-
dom (systematic) sample of subjects without delirium was
selected from patients screened for delirium but free of this
condition. To balance the distributions of age and prior
cognitive impairment between patients with and without
of 3 to 10 and a one in five sample of patients aged 70 and
older with a SPMSQ score of 0 to 2 were selected. Patients
with delirium or dementia were asked for assent to partic-
ipate in the study, and a family member was asked for in-
formed consent; cognitively intact patients were asked for
informed consent. The hospital research ethics committee
approved the study.
A research assistant (RA) assessed patients at enrollment
and 2, 6, and 12 months later and also interviewed a family
member. The 2-, 6-, and 12-month follow-ups were nor-
mally completed in the community.
At enrollment, the SN collected demographic and clin-
ical data (age, sex, marital status, education, living ar-
rangements before admission, presence of visual and
hearing impairment) and completed a Clinical Severity of
Illness measure. The RA completed the Informant Ques-
tionnaire on Cognitive Decline in the Elderly (IQCODE),13
Delirium Index (DI),14MMSE,15and Barthel Index (BI).16
At 8 weeks and 6 and 12 months, the RA completed the DI,
MMSE, and BI. Institutional status and mortality were de-
termined at 6 and 12 months. The Charlson Comorbidity
Index17and Acute Physiology Score (APS)18were deter-
mined according to chart review.
Definition of SSD-Recovered, SSD-Not Recovered, and
No SSD Cohorts
In a previous study,4patients without delirium according to
the DSM-III-R-criteria during the first week of hospitaliza-
tion were classified into three mutually exclusive groups:
prevalent SSD, incident SSD, or no SSD. Prevalent SSD was
defined as the presence of two or more of four core symp-
toms (clouding of consciousness, inattention, disorienta-
tion, and perceptual disturbances as measured using
the CAM or DI) at admission. Incident SSD was defined
as the presence of one or more new (not present at admis-
sion) core symptoms of delirium (as measured using the DI)
during the first week of hospitalization in those who did not
meet criteria for prevalent SSD. No SSD was defined as the
absence of prevalent and incident SSD.
In the present study, the DI was used to determine the
presence of SSD at 8 weeks. In patients who had prevalent
SSD during the first week of hospitalization, SSD at 8 weeks
was defined as the presence of two or more core symptoms.
In patients who had incident SSD, SSD at 8 weeks was
defined as the presence of one or more new (not present at
admission) core symptoms. In both groups, patients who
the first week of hospitalization but who progressed to full-
blown delirium were excluded from the SSD cohort,
because their initial delirium symptoms represented a prod-
rome to full-blown delirium rather than SSD. After deter-
mination of SSD status at 8 weeks, patients were classified
into one of three cohorts. The no SSD cohort was defined
according to the absence of prevalent or incident SSD dur-
ing the first week of hospitalization and at 8 weeks. The
SSD-recovered cohort was defined according to the pres-
ence of prevalent or incident SSD during the first week of
hospitalization but no SSD at 8 weeks. The SSD-not recov-
ered cohort was defined according to the presence of prev-
alent orincident SSD during
hospitalization and SSD at 8 weeks. The 8-week follow-
up point was chosen because the maximum rate of recovery
occurred by 8 weeks, andthere was little change thereafter.4
the SN diagnosis of delirium had a sensitivity of 89% and a
specificityof 100%comparedwith a consensus diagnosis.19
The DI,14derived from the CAM, measures the severity of
seven symptoms of delirium (inattention, disorganized
thinking, altered level of consciousness, disorientation,
memory impairment, perceptual disturbances, psychomo-
tor activity). The SPMSQ11is a widely used, observer-rated,
10-item questionnaire that evaluates orientation, memory,
and concentration; scores range from 0 (no impairment) to
10 (severe impairment). The IQCODE13was used to assess
the presence of dementia before admission. Family mem-
bers were asked to report cognitive changes during the pre-
vious 10 years. A cutoff score of 3.5 points was used.
Comorbidity at admission was assessed according to chart
COLE ET AL.
NOVEMBER 2008–VOL. 56, NO. 11JAGS
review using the Charlson Comorbidity Index.17The SN
assessed Clinical Severity of Illness;20scores range from 1
(minimal) to 9 (most severe). Physiological severity of ill-
ness was assessed using an APS derived from the Acute
Physiology and Chronic Health Evaluation II scale.18The
MMSE15measures cognitive function on a scale of 0 (poor)
to 30 (excellent). The BI16measures independence in per-
sonal care activities; a modified scoring system21that
ranged from 0 (dependent) to 100 (independent) was used.
Information related to institutionalization and mortality
was obtained from the hospital database and from follow-
up interviews with patients and informants.
Baseline characteristics of the three groups (SSD-recovered,
SSD-not recovered, and no SSD) were compared. For dis-
crete variables, the Pearson chi-square test was used, except
when expected cell sizes were smaller than five, in which
case the Fisher exact test was used. For continuous vari-
ables, the F-test one-way analysis of variance and the ap-
propriate contrast were used to compare the three groups.
A primary hierarchical composite outcome, repre-
sented using a binary variable (in this order: death, insti-
tutionalization, decline of ?3 points from enrollment on
the MMSE or ?10 points on the BI), was defined to include
all possible negative outcomes at 6 and 12 months. Specifi-
cally, patients were said to experience the composite neg-
ative outcome if they died, did not die but were
institutionalized, or did not die and were not institution-
alized but experienced a decline of 3 or more points from
enrollment on the MMSE or 10 or more points on the BI.
Crude proportions with their standard deviations were
calculated, and two logistic regression models were fitted
(one at 6 months and one at 12 months) to predict outcome
odds ratios (ORs) of SSD-recovered versus SSD-not recov-
ered groups and SSD-recovered and SSD-not recovered ver-
sus no SSD groups without adjustment and adjusting for
age, sex, marital status, education, dementia status, clinical
severity of illness, acute physiology, and Charlson scores. In
secondary analyses, components of the composite were ex-
amined outcome separately. For binary outcome variables,
the logistic model (SAS/PROC LOGISTIC, SAS Institute,
Inc., Cary, NC) was used.22For continuous outcome vari-
ables, ordinary linear regression (SAS/PROC REG) was
used.23ORs that fell within the range of generally accepted
cutpoints for clinical significance (?0.67 or ?1.5)24were
considered to be potentially clinically significant; mean
differences between groups of 3 or more points on the
MMSE and 10 or more points on the BI were considered to
be minimally clinically significant.
During the study enrollment period, 1,552 patients were
screened for delirium (Figure 1). One hundred sixty-one
patients met DSM-III-R criteria for prevalent delirium on
admission. Of the remaining 1,391 patients who did not
meet criteria for delirium, 200 were selected for inclusion in
the study. Of these 200 patients, 129 had prevalent SSD and
71hadnoSSD.Bytheendof thefirst week,23patientswith
noSSDinitiallyhaddevelopedincidentSSD.Bythe endof 8
weeks, 25 patients diagnosed with prevalent SSD initially
had developed incident delirium, 18 patients in the prev-
alent and incident SSD cohorts had died, 12 had with-
drawn, six were lost to follow-up, and 10 patients with no
SSD initially had developed SSD. The 25 patients who were
diagnosed with prevalent SSD initially but who developed
incident delirium were excluded from the SSD cohort be-
cause their initial delirium symptoms represented a prod-
rome to full-blown delirium rather than SSD. Of the 129
patients assessed at 8 weeks, 51, 47, and 31 met criteria for
SSD-recovered, SSD-not recovered, and no SSD, respec-
tively. There were significant differences in the clinical pro-
files of patients in the three groups (Table 1).
The outcomes are presented in Table 2. Of the 129
patients,datawereavailablefor 123at 6monthsand125at
12 months. At 6 and 12 months, the crude rates of the
primary hierarchical composite outcome of patients in the
SSD-recovered group were lower than the rates of patients
in the SSD-not recovered group and intermediate between
the rates of patients in the SSD-not recovered and the no
SSD groups. At 6 months, the adjusted OR between SSD-
recovered and the no SSD groups and between the SSD-not
recovered and no SSD groups were clinically, but not sta-
tistically, significant; at 12 months, the adjusted ORs be-
tween the SSD-recovered and the SSD-not recovered
groups, between the SSD-recovered and the no SSD groups,
and between the SSD-not recovered and the no SSD groups
were all clinically, but not statistically, significant.
In secondary analyses, components of the primary
composite outcome were examined separately. At 6 and 12
months, the crude rates of death or institutionalization of
the SSD-recovered group were lower than the rates of the
SSD-not recovered group and intermediate between the
rates of patients in the SSD-not recovered and no SSD
groups. At 6 months, the adjusted ORs between the SSD-
recovered and the SSD-not recovered groups and between
the SSD-recovered and no SSD groups were clinically, but
not statistically, significant; the OR between the SSD-not
recovered and the no SSD groups was clinically and
statistically significant. At 12 months, the adjusted ORs
for all comparisons were clinically, but not statistically,
At 6 and 12 months, the mean MMSE scores of the
SSD-recovered group were higher than the mean scores of
the SSD-not recovered group and similar to the mean scores
of the no SSD group. At 6 and 12 months, the ad-
justed differences in mean MMSE scores between the SSD-
recovered and the SSD-not recovered groups were clinically
and statistically significant. At 6 and 12 months, the ad-
justed differences in mean MMSE scores between the SSD-
not recovered and the no SSD groups were clinically and
statistically significant, whereas the adjusted differences in
mean MMSE scores between the SSD-recovered and the no
SSD groups were small and not clinically or statistically
At 6 and 12 months, the mean BI scores of the SSD-
recovered group were higher than the mean scores of the
SSD-not recovered group but lower than the mean scores of
the no SSD group. The adjusted differences in mean BI
scores between the SSD-recovered and the SSD-not recov-
ered groups were not clinically or statistically significant.
The adjusted differences in mean BI scores between the
SSD-not recovered and the no SSD groups were clinically,
OUTCOMES OF SUBSYNDROMAL DELIRIUM
2095JAGSNOVEMBER 2008–VOL. 56, NO. 11
Figure 1. Flow chart of study subjects. DSM3R5Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised;
Table 1. Characteristics of the Three Groups at Enrollment
(n547) No SSD (n531)P-Value?
Categorical variables, n (%)
Education secondary or less
Continuous variables, mean ? standard deviation
Severity of illness
Charlson comorbidity index
Acute physiology score
82.3 ? 6.6
4.0 ? 1.4
1.9 ? 2.0
3.0 ? 2.6
84.5 ? 7.1
4.3 ? 1.4
2.4 ? 1.5
4.2 ? 2.8
81.2 ? 5.6
3.8 ? 1.2
2.0 ? 2.2
1.7 ? 2.2
?Chi-square or Fisher exact test for categorical variables and F test from one-way analysis of variance for continuous variables.
COLE ET AL.
NOVEMBER 2008–VOL. 56, NO. 11JAGS
Table 2. Outcomes at 6 and 12 Months?
SSD-Recovered Versus No SSD
SSD-Not Recovered Versus No SSD
Odds Ratio or bz(95% Confidence Interval)
At 6 months
outcome, n (%)§
Death or institutionalization,
MMSE, mean ? SD
22.6 ? 5.2
18.3 ? 5.5
24.6 ? 3.6
BI, mean ? SD
74.6 ? 28.4
71.9 ? 23.6
86.1 ? 18.7
At 12 months
outcome, n (%)§
Death or institutionalization,
MMSE, mean ? SD
23.1 ? 4.8
18.9 ? 5.7
24.3 ? 4.6
BI, mean ? SD
74.9 ? 28.9
63.2 ? 28.9
82.6 ? 23.4
Note: Boldface indicates statistically significant at Po.05.
?Univariate and multivariate logistic regression models for 6- and 12-month outcomes.
wAdjusted for age, sex, marital status, education, dementia status, Clinical Severity of Illness score, Acute Physiology Score, and Charlson comorbidity score.
zOdds ratio (OR) for hierarchical composite outcome and death or institutionalization and b for Mini-Mental State Examination (MMSE) and Barthel Index (BI).
§Hierarchical composite outcome is defined in this order: death, institutionalized, decline of 3 or more points on MMSE, decline of 10 or more points on BI.
SSD5subsyndromal delirium; SD5standard deviation.
OUTCOMES OF SUBSYNDROMAL DELIRIUM
2097JAGSNOVEMBER 2008–VOL. 56, NO. 11
but not statistically, significant at 6 months and neither
clinically nor statistically significant at 12 months, whereas
the adjusted differences in mean BI scores between the SSD-
recovered and the no SSD groups were not clinically or
statistically significant at 6 or 12 months.
SSD is a condition in which one or more symptoms of de-
lirium do not progress to full-blown delirium. This study
proposed to compare the 6- and 12-month outcomes of
older medical inpatients who recovered from SSD by 8
weeks with the outcomes of patients who did not recover or
did not have an index episode of SSD. The primary and
secondary outcomes of patients in the SSD-recovered group
were better than the outcomes of patients in the SSD-not
recovered group and, for the most part, intermediate be-
tween the outcomes of patients in the SSD-not recovered
and no SSD groups. These findings are similar to those of
studies that report that patients who recovered from full-
blown delirium had better outcomes than patients who did
not recover10,25andthatpatients whorecovered by8 weeks
had outcomes similar to those of patients who did not have
an index episode.26
This study had three strengths. First, patients were as-
sessed at several time points using reliable and valid mea-
sures. Second, the follow-up of the SSD-recovered, SSD-not
recovered, and no SSD cohorts was relatively complete.
Third, the study included different outcome measures and
adjusted for potential confounders of the outcomes.
This study had seven potential limitations. First, the
original cohort (200 patients) was a selected sample of
medical admissions. This selection may have increased the
frequency of SSD in the sample, but it is unlikely that it
affected the outcomes. Second, 18 patients died, 18 with-
drew (or were lost to follow-up), and six developed incident
delirium between1 and8weeks, creating potential biasesin
the follow-up cohorts. Third, there may be overlap of
symptoms between prevalent SSD, defined as the presence
of two or more of four core symptoms of delirium (clouding
of consciousness, inattention, disorientation, perceptual
disturbances) and dementia; consequently, some patients
with dementia may have been misdiagnosed with prevalent
SSD, although in a previous study,4the two most common
symptoms of SSD were disorientation and inattention, and
in the present study, rates of recovery were similar in pa-
tients with prevalent or incident SSD (50%). Fourth, the
absence of SSD was based on one assessment at 8 weeks.
Because SSD is probably a fluctuating disorder, some pa-
tients may have been misclassified as recovered, although
such misclassification would have decreased differences in
outcomes between the SSD-recovered and the SSD-not
recovered groups and increased differences between the
SSD-recovered and the no SSD groups. Fifth, some of the
patients with SSD in this study may have met DSM-IV cri-
teria for delirium, but when such patients were excluded
from the analysis in the previous study of the prognostic
significance of SSD, the results were unchanged.4Sixth, the
IQCODE was used to determine dementia status. The sen-
sitivity and specificity of this instrument are high, but it has
not been validated in patients presenting with delirium.
Finally, many of the clinically significant differences in out-
comes between groups were not statistically significant,
perhaps because of the small numbers of subjects and lack
of statistical power. The power at the .05 level of signifi-
cance for detecting ORs of 2.5, 2.0, and 1.5 for the primary
outcome was 52%, 33%, and 15%, respectively; conse-
quently, it is probable that statistically significant differ-
ences between groups were missed by chance.
If replicated, these findings have two important impli-
cations. First, recovery from SSD appears to predict better
longer-term outcomes than no recovery and should be the
primary goal of SSD management. Second, systematic in-
terventions to improve the longer-term outcomes of SSD
could involve identification and treatment of patients with
SSD, especially those who have not recovered by 8 weeks.
Although the results of trials that have involved systematic
identification and treatment of full-blown delirium have
been modest27–29or negative,30efforts to identify and treat
SSD in older medical inpatients may improve outcomes.
Conflict of Interest: The editor in chief has reviewed the
conflictof interestchecklistprovided bythe authorsandhas
determined that the authors have no financial or any other
kind of personal conflicts with this manuscript. Funded by
Canadian Institutes for Health Research Grants MOP82494
Author Contributions: M. Cole and J. McCusker were
involved in allaspectsof manuscript preparation.A. Ciampi
and E.Belzilewereinvolvedinanalysisand interpretation of
data and preparation of the manuscript.
Sponsor’s Role: None.
1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental
Disorders, 3rd Ed. Washington, DC: American Psychiatric Association, 1987.
2. Levkoff SE, Liptsin B, Cleary PD et al. Subsyndromal delirium. Am J Geriatr
3. Marcantonio ER, Kiely DK, Simon SE et al. Outcomes of older people admitted
to postacute facilities with delirium. J Am Geriatr Soc 2005;53:963–969.
4. Cole MG, McCusker J, Dendukuri N et al. The prognostic significance of
subsyndromal delirium in elderly medical inpatients. J Am Geriatr Soc 2003;
5. Kiely DK, Bergmann MA, Murphy KM et al. Delirium among newly admitted
postacute facility patients: Prevalence, symptoms, and severity. J Gerontol A
Biol Sci Med Sci 2003;58A:441–445.
6. Bourdel-Marchasson I, Vincent S, Germain C et al. Delirium symptoms and
low dietary intake in older inpatients are independent predictors of institu-
tionalization: A 1-year prospective population-based study. J Gerontol A Biol
Sci Med Sci 2004;59A:350–354.
7. Ouimet S, Riker R, Bergeon N et al. Subsyndromal delirium in the ICU: Ev-
idence for a disease spectrum. Intensive Care Med 2007;33:1007–1013.
8. Marquis F, Ouimet S, Riker R et al. Individual delirium symptoms: Do they
matter? Crit Care Med 2007;35:2533–2537.
9. McCusker J, Cole MG, Dendukuri N et al. Delirium in older medical inpa-
tients andsubsequentcognitive and functional status:A prospectivestudy.Can
Med Assoc J 2001;165:575–583.
10. McCusker J, Cole MG, Dendukuri N et al. The course of delirium in older
medical inpatients: A prospective study. J Gen Intern Med 2003;18:696–704.
11. Pfeiffer E. A short portable mental status questionnaire for the assessment of
organic brain deficit in elderly patients. J Am Geriatr Soc 1975;23:433–441.
12. Inouye SK,VanDyck CH, Alessi CAet al.Clarifying confusion: The Confusion
Assessment Method, a new method for detection of delirium. Ann Intern Med
13. Jorm AF. A short form of the Informant Questionnaire on Cognitive Decline in
theElderly(IQCODE):Developmentand cross-validation.Psychol Med1994;
COLE ET AL.
NOVEMBER 2008–VOL. 56, NO. 11 JAGS
14. McCusker J,ColeM,DendukuriNetal.TheDeliriumIndex:A measureofthe Download full-text
severity of delirium: New findings on reliability, validity and responsiveness.
J Am Geriatr Soc 2004;52:1744–1749.
15. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-Mental State’’: A practical
method for grading the cognitive state of patients for the clinician. J Psychiatr
16. Mahoney FI, Barthel DW. Functional evaluation: The Barthel Index. MD State
Med J 1965;14:61–65.
17. Charlson ME, Pompei P, Ales KL et al. A new method of classifying prognostic
comorbidity in longitudinal studies: Development and validation. J Chronic
18. Knaus WA, Draper EA, Wagner DP et al. APACHE II: A severity of disease
classification system. Crit Care Med 1985;13:818–829.
19. Zou Y, Cole MG, Primeau FJ et al. Detection and diagnosis of delirium in the
elderly: Psychiatrist diagnosis, confusion assessment method or consensus di-
agnosis? Int Psychogeriatr 1998;10:303–308.
20. Charlson ME, Sax FL, MacKenzie R et al. Assessing illness severity: Does
clinical judgment work? J Chronic Dis 1986;39:439–452.
21. Shah S, Vanclay F, Cooper B. Improving the sensitivity of the Barthel Index for
stroke rehabilitation. J Clin Epidemiol 1989;42:703–709.
22. Hosmer DWJ, Lemeshow S. Applied Logistic Regression. New York: Wiley,
23. Neter J, Kutner MH, Nachtsheim CJ et al. Applied Linear Statistical Models.
Homewood, IL: Irwin, 1996.
24. Sackett DL, Haynes RB, Guyatt GH et al. Clinical Epdemiology: A Basic
Science for ClinicalMedicine.Boston, MA:Little,Brown andCompany,1991.
25. McAvay GJ, VanNess PH, Bogardus STet al. Older adults discharged from the
hospital with delirium: 1-year outcomes. J Am Geriatr Soc 2006;54:1245–
26. Cole MG, You Y, McCusker J et al. The 6 and 12 month outcomes of older
medical inpatients who recover from delirium. Int J Geriatr Psychiatry
27. Cole MG, Primeau FJ, Bailey RF et al. Systematic intervention for elderly
inpatients with delirium: A randomized trial. Can Med Assoc J 1994;151:965–
28. Lundstrom M, Edlund A, Karlsson S et al. A multifactorial intervention pro-
gram reduces the duration of delirium, length of hospitalization, and mortality
in delirious patients. J Am Geriatr Soc 2005;53:622–628.
29. Pitkala KH, Laurila JV, Strandberg TE et al. Multicomponent geriatric inter-
vention for elderly inpatients with delirium: A randomized, controlled trial.
J Gerontol A Biol Sci Med Sci 2006;61A:176–181.
30. Cole MG, McCusker J, Bellavance F et al. Systematic detection and multidis-
ciplinary care of delirium in older medical patients: A randomized trial. Can
Med Assoc J 2002;167:753–759.
OUTCOMES OF SUBSYNDROMAL DELIRIUM
2099JAGSNOVEMBER 2008–VOL. 56, NO. 11