Accepted Manuscript

Diagnostic accuracy of delirium assessment methods in critical

care patients

Angkita Barman, Debasis Pradhan, Prithwis Bhattacharyya,

Samarjit Dey, Anirban Bhattacharjee, Sonali Shinde Tesia,
Jayanta Kumar Mitra

PII: S0883-9441(17)31007-9
DOI: doi:10.1016/j.jcrc.2017.10.013
Reference: YJCRC 52725
To appear in:

Please cite this article as: Angkita Barman, Debasis Pradhan, Prithwis Bhattacharyya,
Samarjit Dey, Anirban Bhattacharjee, Sonali Shinde Tesia, Jayanta Kumar Mitra ,
Diagnostic accuracy of delirium assessment methods in critical care patients. The address
for the corresponding author was captured as affiliation for all authors. Please check if
appropriate. Yjcrc(2017), doi:10.1016/j.jcrc.2017.10.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Our reference: YJCRC 52725

Article reference: JCRC_2017_909

Article title: Diagnostic accuracy of delirium assessment methods in

critical care patients

To be published in: Journal of Critical Care

PT
RI
TITLE: Diagnostic accuracy of delirium assessment methods in

SC
critical care patients
NU
Name of authors with affiliations:
MA

1. First author: Angkita Barman, North Eastern Indira Gandhi

ED

Institute of Health and Medical Sciences, Shillong, Meghalaya,

PT

India, dibemy@gmail.com
CE

2. Second author: Debasis Pradhan, North Eastern Indira Gandhi

Institute of Health and Medical Sciences, Shillong, Meghalaya,

AC

India, dmkcg85@gmail.com

3. Third and correspondence author: Prithwis Bhattacharyya,

North Eastern Indira Gandhi Institute of Health and Medical

Sciences, Shillong, Meghalaya, India, prithwisbhat123@gmail.com

 ACCEPTED MANUSCRIPT

4. Fourth author: Samarjit Dey, North Eastern Indira Gandhi

Institute of Health and Medical Sciences, Shillong, Meghalaya,

India, drsamar0002@gmail.com

5. Fifth author: Anirban Bhattacharjee, North Eastern Indira Gandhi

Institute of Health and Medical Sciences, Shillong, Meghalaya,

PT
India, onirbhatt@gmail.com

RI
6. Sixth author: Sonali Shinde Tesia, North Eastern Indira Gandhi

SC
Institute of Health and Medical Sciences, Shillong, Meghalaya,

India
NU
7. Seventh author: Jayanta Kumar Mitra, All India Institute of
MA

Medical Sciences, Bhubaneswar, Odisha, India

ED
PT
CE
AC
 ACCEPTED MANUSCRIPT
Abstract Page
Title of the article:

Diagnostic accuracy of delirium assement methods in critical care

patinets
Abstract:

Purpose: Delirium is a disorder of decreased ability to focus, sustain or

PT
shift attention, change in cognition and or perception. The main objective

was to evaluate the diagnostic accuracy of Confusion Assessment Method

RI
for the ICU (CAM-ICU) and Intensive Care Delirium Screening

SC
Checklist (ICDSC) among the nursing and medical staff in a
NU
multidisciplinary ICU.
MA

Methods and Material: Three hundred ten verbally communicating and

non-communicating patients (mean age in years 47.9, standard deviation

ED

[SD] 14.5, mean Acute Physiology and Chronic Health Evaluation II

score 13.8, SD 6.4 ) were assessed by a psychiatrist, nurse and intensivist

PT

for delirium. Inter-rater agreement was measured by Cohen’s kappa

CE

coefficient. Sensitivity, specificity, predictive values, likelihood ratios

AC

and diagnostic odds ratio (DOR) were calculated.

Results: CAM-ICU showed higher sensitivity and DOR (84%, 86.1)

compared to ICDSC (78%, 36.9). ICDSC had specificity and positive

predictive value (94.5%, 92%) equal to that of CAM-ICU. For both the

assessment methods (CAM-ICU and ICDSC), DORs for intensivists

(120.5, 53.0) were relatively higher than nurses (67.0, 27.0).

 ACCEPTED MANUSCRIPT

Conclusions: In our mixed ICU population, CAM-ICU remained more

sensitive than ICDSC. Though sensitivity and DOR were higher for

medical staff, other diagnostic parameters were similar for both medical

and nursing staff.

Key-words: Delirium, intensive care unit, assessment, sensitivity,

PT
specificity

RI
Introduction

SC
Delirium is a common manifestation of acute brain dysfunction in

critically ill patients which is associated with poor short-term outcomes

NU
and may result in adverse sequelae, even years after discharge from the
MA

intensive care unit (ICU). It is characterized by an acute disturbance of

consciousness and attention with cognitive or perceptual changes.

ED

Standard assessment is performed by a psychiatrist using Diagnostic and

PT

Statistical Manual of Mental Disorders (DSM) criteria [1]. It usually

arises over a short span of time (hours to days) as a direct consequence of

CE

medical conditions, post-surgical state, substance intoxication or

AC

withdrawal, use of medications, toxin exposure, or a combination of these

factors [2]. Current critical care practice guidelines recommend routine

delirium screening [3]. It is highly disturbing for the nurses and family

members of the patient and also a frequent management problem in the

ICU associated with poor prognosis. Delirium in ICU is usually under

 ACCEPTED MANUSCRIPT

diagnosed and under recognised by nursing and medical staff with a

widely varying incidence rates ranging from 11% to 87% [4, 5].

Although the exact pathophysiological mechanisms involved in the

development and progression of delirium are unknown, it is thought to be

related to imbalances in the neurotransmitters that modulate the control of

PT
cognitive function, behaviour, and mood [6, 7]. Currently, central

RI
cholinergic deficiency is the leading hypothesized mechanism for

SC
delirium supported by several lines of evidence [8]. Delirium may occur

after administration of agents that impair cholinergic function or in

NU
conditions associated with cholinergic deficit [9]. Also, the
MA

administration of cholinesterase inhibitors may decrease the duration of

delirium [10].
ED

The development of internationally accepted diagnostic tools created the

opportunity to compare and verify the onset and progress of intensive
PT

care delirium without the need for consulting a psychiatrist. There are six
CE

delirium assessment instruments in the literature that have been validated

to evaluate delirium in an ICU setting and are based on the DSM IV
AC

criteria (Table 1). This prospective study was conducted to assesss

diagnostic accuracy of CAM-ICU and ICDSC in the diagnosis of
delirirum in a mixed ICU patient population and to find out the best scale
that can be used both by ICU nursing staff and physician.
Methods

Study design and population

 ACCEPTED MANUSCRIPT

This prospective study was conducted in a multidisciplinary ICU,

following approval from Institutional Ethics Committee and informed

consent from patients or their legal representative. Both ventilated and

non-ventilated medical and surgical adult patients were included. All

patients were assessed following 24 hours stay in ICU. Patients with

PT
Glasgow coma scale < 9, Richmond Agitation-Sedation Scale (RASS)

RI
score of ≤ -3 and patients with confounding co-morbid conditions (severe

SC
dementia, psychosis or progressive neurologic disease) were excluded.

Baseline information included age, sex, severity of illness (APACHE II-

NU
Acute Physiology and Chronic Health Evaluation II), diagnosis at
MA

admission, benzodiazepine usage, presence or absence of

dyselectrolytemia, sepsis, the need and duration of mechanical

ED

ventilation, blood urea nitrogen, serum creatinine and bilirubin levels.

PT

Data collection

Delirium was evaluated by using the Intensive Care Delirium

CE

Screening Checklist (ICDSC) and the Confusion Assessment Method for

AC

the ICU (CAM-ICU). Each patient was allowed to participate once and

readmissions were excluded. A trained ICU nurse and an intensivist

assessed the patients independently by using these two methods. Before

the commencement of the study, all the ICU nurses were educated with

regard to the appropriate use of the two delirium assessment methods. A

psychiatrist served as a reference rater and evaluated all the patients and
 ACCEPTED MANUSCRIPT

diagnosed delirium by using the established DSM IV criteria. The

assessments were done in a blinded fashion and were completed in a span

of four hours. The psychiatrist, nurse and intensivist did not have access

to patients’ records and other rater’s finding. All assessments were done

from 0800 hours to 1600 hours on weekdays.The rationale for allowing

PT
four hours for assessment was to balance the intent of having the

RI
evaluations as close together as possible to enhance the reliability and to

SC
avoid the difficulties introduced by patient care or procedures as well as

clinicians’ schedule.
NU
None of the raters had access to the other’s ratings of the assessment and
MA

the patient records. All patients were assessed by RASS during the first

twenty-four hours of delirium assessment. Based on the RASS value, all

ED

patients with delirium were classified into three motoric subtypes -

PT

hypoactive, hyperactive and mixed type. Hyperactive delirium was

defined when the RASS was persistently positive (+1 to +4 during all
CE

assessments). Hypoactive delirium was defined with a neutral or negative

AC

RASS (0 to -3 during all assessments). Delirium episode with both

positive and negative RASS values was defined as mixed type.

Statistical analysis

Descriptive statistics were used for evaluation of baseline

characteristics of demographic and clinical variables. Normality of data

distribution was tested by D’Agostino-Peasron test. Inter-rater aggrement

 ACCEPTED MANUSCRIPT

was measured by Cohen’s kappa coefficient. The two observers were

compared to each other and the results (non-parametric data) were

computed using the Mann Whitney U test. The diagnostic accuracy of the

tools were described as sensitivity, specificity, positive predictive value

(PPV), negative predictive value (NPV), likelihood ratios and diagnostic

PT
odds ratio (DOR) using simple 2 × 2 tables. The statistical analysis was

RI
done using MedCalc for Windows, version 12.5 (MedCalc Software,

SC
Ostend, Belgium) and Microsoft Excel 2007 for Windows (Redmond,

Washington, USA).
NU
Results
MA

A total of 515 patients were evaluated over a period of twelve

months for possible inclusion in this study. From these, we excluded 152
ED

patients based on the criteria mentioned in the methods (73-deeply

PT

sedated and 79- comatose patients). We could not obtain informed

consent in 17 patients and 8 patients were deaf, so they were excluded.

CE

Therefore, 338 patients were assessed for the purpose of the study. In 28
AC

patients, all the three assessors could not complete the evaluations. After

all exclusion, the final study population constituted 310 patients whose

observations were available for final analysis (Figure 1). The study

population was comparable with respect to their demographics and

diagnosis at admission (P > 0.05) (Table 2). As per psychiatrist’s

evaluation based on DSM IV criteria, prevalence of delirium was 45%

 ACCEPTED MANUSCRIPT

(140 out of 310). Delirious patients were significantly more ill with

average (± 2 X standard deviation) APACHE II score of 18.0 ± 5.2

compared to non delirious patients (9.8 ± 6.1, P < 0.0001). Though,

female patients were less prone to be delirious, yet it was not statistically

significant (P = 0.078). Delirum was significantly less prevalent in

PT
postoperative patients compared to other groups. Out of 310 patients 162

RI
(52.2 %) were nonventilated and verbally communicating.

SC
Based on RASS assessment, 22 (15.7%), 91 (65%) and 27 (19.3%)

patients were found to have hyperactive, hypoactive and mixed type

NU
respectively. The inter rater agreement (kappa coefficient) between nurse
MA

and intensivist was 0.86 (95% CI 0.8-0.9) and 0.89 (95% CI 0.83-0.96)

for CAM-ICU and ICDSC respectively. Overall sensitivity and

ED

specificity (Table 3) were 84.4 % (95% CI: 70.5%-93.5%) and 94.6 %

PT

(95% CI: 84.9%- 98.9%) for CAM-ICU, 77.8% (95% CI: 62.9%-88.8%)

and 94.6 % (95% CI: 84.9%-98.9%) for ICDSC. Overall DORs was
CE

higher for CAM-ICU (86.1) than ICDSC (36.9) and intensivist

AC

assessment (120.5, 53.0) had higher DORs compared to nurse assessment

(67.0, 27.0) for CAM-ICU and ICDSC respectively (Table 3). Subgroup

analysis was performed and observer specific diagnostic accuracies for

verbally communicating and non-communicating patients were calculated

(Table 4). In verbally communicating patients, CAM-ICU was more

specific as reported by both nurse and intensivist (96.2% and 95.3%

 ACCEPTED MANUSCRIPT

respectively). In verbally non-communicating group, CAM-ICU had

good sensitivity (84.5 %, 95% CI: 75%-91%) when used by nurse and

most sensitive (94.1%, 95% CI: 87%-98%) when used by intensivist.

Patients with coexisting factors like a diagnosis of sepsis, benzodizepine

use, need for mechanical ventilation, and APACHE II >16 were found to

PT
have significantly high relative risk (RR) (Table 5) of being delirious.

RI
Postoperative patients were found to have low relative risk of delirium

SC
(RR - 0.57, P - 0.03).

Discussion
NU
Compared to general hospital population, delirium in critically ill
MA

patients is more severe and of longer duration with long standing bad

impact on the post discharge quality of life. This is because of the lower
ED

threshold for the development of delirium (due to a higher number of risk

PT

factors) in critically ill patients and the exposure of more intense

modifiable precipitating factors (medications, pain, altered sleep-wake

CE

cycle, metabolic disturbances). Accurate identification and prompt

AC

modification of the risk factors may prevent many adverse outcomes

associated with this phenomenon. In 2002, Joint Task Force of the

American College of Critical Care Medicine (ACCM) of the Society of

Critical Care Medicine (SCCM), American Society of Health System

Pharmacists (ASHP), American College of Chest Physicians proposed

clinical practice guidelines for the sustained and appropriate use of

 ACCEPTED MANUSCRIPT

sedatives and analgesics in the critically ill adults and had recommended

routine assessment of delirium to be a part of routine patient care in ICU

[3]. Despite of various guidelines, routine use of delirium assessment tool

had remained low as reported by various studies, such as 7% of all Dutch

ICUs (Van Eijk et al., 2008) [11] and 9% of ICUs in Australia and New

PT
Zealand (Shehabi et al., 2008) [12] used a validated screening tool, in

RI
England 25% of intensivists routinely screen for delirium and 14% used a

SC
validated tool in mechanically ventilated patients (Mac Sweeney et al.,

2010) [13]. Two surveys from North America had reported routine
NU
screening of delirium by 40–59% of the intensivists and only 16-33% of
MA

them used specific tools for delirium assessment (Ely et al., 2004 [14],

Patel et al., 2009 [15]). Increased recognition of delirium is the first step
ED

in decreasing the incidence of delirium and maximizing patient’s comfort

PT

in the ICU.

The CAM-ICU was developed by Ely et al. [16] to detect delirium in

CE

mechanically ventilated or restrained ICU patients. It uses nonverbal

AC

tasks such as picture recognition, vigilance task, simple yes/no logic

questions and commands to rate the features of the CAM-ICU algorithm.

The following four key delirium criteria and an algorithm are used to

determine the presence of delirium: (a) acute mental status change; (b)

inattention; (c) disorganized thinking; and (d) altered level of

consciousness. Delirium is considered to be present when criteria (a) and

 ACCEPTED MANUSCRIPT

(b) along with either criteria (c) or (d) are present. In 2001, Bergeron et

al. [17] created the ICDSC, which includes eight items based on the DSM

criteria and features of delirium (including: inattention; disorientation;

hallucination-delusion psychosis; psychomotor agitation or retardation;

inappropriate speech or mood; sleep/wake cycle disturbances; and

PT
symptom fluctuation), which has a total score ranging from 0 to 8 points.

RI
A patient with more than 4 points is diagnosed as delirium-positive.

SC
In the face of low rates of utilization of delirium assessment tools in

routine ICU care, a variable rate of delirium prevalence has been found
NU
among intensive care patients. In the present study, prevalence of
MA

delirium was found to be 45.2% as diagnosed by a psychiatrist as a

reference rater using DSM IV criteria. Differential reporting of delirium

ED

from various ICUs can be attributed to various factors like heterogeneity

PT

of patient population, methods used for assessment, methods incorporated

into regular ICU practice, the day and time of assessment, presence or
CE

absence of any undiagnosed confounding illness and variable usage of

AC

sedatives and analgesics. A systematic review incluidng six original

papers had identified twenty five risk factors for delirium development

and four of them (respiratory disease, advanced age, history of alcohol

abuse, and dementia) were considered as predisposing factors. The

remaining twenty one were reported as precipitating risk factors and

included dyselectrolytemia, fever, need for a vasopressor, higher dose of

 ACCEPTED MANUSCRIPT

opioid usage and metabolic acidosis [18]. In the present study, age did

not affect the prevalence of delirium. This is similar to the findings of

Ouimet et al [19] and Inouye et al [20], which can be attributed to relative

younger study population (33-62 years). But earlier studies including a

systematic review had reported advanced age as an important established

PT
risk factor for development of delirium [18, 21].

RI
Increased severity of illness as indicated by higher APACHE II score has

SC
a positive assosiation with delirium. In a study involving hospitalized

elderly medical patients, a composite score defined by a nurse rating of

NU
“severe” or an APACHE II score of more than 16 was considered to
MA

represent “severe illness”. Patients with severe illness had a relative risk

(RR) of 3.5 (95% confidence interval [CI]: 1.5-8.2) in developing

ED

delirium [20]. In the present study, 136 (44%) patients with delirium and
PT

31 (10%) patients without delirium had severe illness (APACHE II >16)

and the relative risk was 2.3 (95% CI: 1.4-3.6). Similar to earlier studies,
CE

patients receiving benzodiazepines [19, 22], mechanical ventilation [23],

AC

those diagnosed with sepsis [24], dyselectrolytemia and BUN/creatinine

>18 [17] had a higher risk of development of delirium (Table 5).

We found a high kappa coefficient (confirming a good interobserver

agreement) when the two methods were used by nurses and intensivists

for delirium assessment. This is similar to the findings of earlier studies

by Ely et al. [16] with k = 0.84 (95% CI 0.63-0.99) between the

 ACCEPTED MANUSCRIPT

intensivist and the nurse and 0.95 (95% CI 0.84 - 1.00) between the two

nurses. Plaschke et al [25] had reported a good agreement between CAM-

ICU and ICDSC (Kappa coefficient 0.80, CI 95%: 0.78-0.84;P <0.001 ).

Ease of administration and acceptance by the nursing and medical staff

are critical to implementation of delirium screening in an ICU setting.

PT
In the present study CAM-ICU had a higher overall sensitivity while

RI
specificity, predictive value and likelihood ratio were similar for both the

SC
scales (Table 3). Similar higher sensitivity of CAM-ICU in diagnosing

delirium had been reported by various earlier studies [5, 26]. A

NU
systematic review involving nine studies (969 patients) had reported
MA

higher pooled sensitivity and specifcity of 80% (95% CI: 77%-0.83%)

and 96% (95% CI: 95%-97%) for CAM-ICU when compared to ICDSC
ED

with pooled sensitivity 74% (95% CI: 65%-81%) and specificity 82%
PT

(95% CI: 76%-86%) [27]. In the present study, we found overall equal

specificities between the two methods and ICDSC had higher specificity
CE

compared to the findings of earlier meta-analysis. Use of previous

AC

twenty-four hours data in the face of delirium being an acute fluctuating

situation may explain this variable finding. When used by nurses, CAM-

ICU had a higher specificity of 96.2% (95% CI: 91%-99%) in verbally

communicating patients and higher sensitivity of 84.52% (95% CI: 75%-

91%) in verbally non-communicating patients. Similar to earlier studies,

 ACCEPTED MANUSCRIPT

CAM-ICU had a higher sensitivity in the non-communicating subgroup

[28].

As hypothesized in a previous meta-analysis [27], sedation may affect the

assessment of the first component i.e., acute onset of mental status

changes while increasing the sensitivity of CAM-ICU in verbally non-

PT
communicating patients. Very few studies have compared the diagnostic

RI
accuracy of sedation scales between different users. Similar to the

SC
findings of Chuang et al [29], higher sensitivity of 82.1 % (95% CI: 79%-

91%) in verbally communicating group and 94.1% (95% CI: 87%-98%)

NU
in verbally non-communicating group was reported when CAM-ICU was
MA

used by intensivist. While assessing inter-observer comparison, we found

the highest DOR of 120.5 (95% CI- 53.5-271.5) for CAM-ICU when it
ED

was used by intensivist. Overall, the DORs were higher for CAM-ICU
PT

and for the physician observers similar to the findings of earlier meta-

analysis which reported DORs for CAM-ICU and ICDSC as 103.2 (95%
CE

CI: 39.6 to 268.8) and 21.5 (95% CI: 8.51 to 54.4) respectively [27].
AC

When comparing various assessment methods, there are some advantages

and disadvantages related to each method. In general, the CAM-ICU has

a fast application (2-5 minutes) and does not depend exclusively on the

verbal response, thus being relevant for patients on mechanical

ventilation. In contrast, the ICDSC is a checklist implemented over 24

hours, with easy administration and high sensitivity [19]. The patients are
 ACCEPTED MANUSCRIPT

involved actively in CAM-ICU assessment, but indirectly involved in the

ICDSC checklist. An ideal delirium screening tool for clinical use must

be performed rapidly at the bedside and should not have complicated

scales to identify delirium. Sedation assessment is a key component of

delirium assessment, regardless of the delirium scale chosen, and thus it

PT
is important to ensure that nursing staff are routinely evaluating all

RI
patients with a sedation scale (e.g., RASS) before a delirium assessment

SC
is completed.

However, this study has some limitations. First, delirium being an acute
NU
fluctuating condition the time of entry into the study affects the outcome.
MA

Second, we did not assess the patient during the rest of ICU stay for the

study purpose. Therefore, the overall delirium prevalence can not be

ED

made based on a single assessment. Third, we have tried to minimize the

PT

time between nurse, intensivist and psychiatrist evaluations; however,

this was not always possible due to logistic difficulties. But this has been
CE

a very tricky area, because as we minimize the time gap we tend to

AC

intorduce patient memory based bias and on the other hand when we try

to keep them apart, we may pick up a new onset delirium by one of the

observer. Fourth, we could not perform the outcome assessment of the

education session. So, the interobserver bias could not be ruled out

completely. Fifth, we routinely use CAM-ICU assessment as a part of

nursing care documentation. This may have an impact on the outcome.

 ACCEPTED MANUSCRIPT

Sixth, this result should not be generalised to all types of health care

settings and ICUs. Lastly, We could not determine the impact of our

findings on the long-term outcomes as patients were not followed beyond

ICU.

Conclusions

PT
CAM-ICU was found to be a better delirium assessment tool than

RI
the ICDSC. Frequency of delirium was found to be higher in patients

SC
receiving sedatives, mechanical ventilation, and those who were more

severely ill and lower in the postoperative patients.

NU
The results of the assessment tools whether performed by the intensivists
MA

or the nurses were comparable and similar to those of the reference rater.

The hospital settings and necessary training of nursing and medical staff
ED

should be considered while choosing an appropriate method for routine

PT

delirium assessment. Incorporation of delirium assessment into clinical

practice in the intensive care unit using a validated tool may improve
CE

patient care.
AC
 ACCEPTED MANUSCRIPT

Conflicts of interest

None.

Funding

This research did not receive any specific grant from funding

agencies in the public, commercial, or not-for-profit sectors.

PT
Acknowledgements

RI
This work would not have been possible without the collaboration

SC
of the nurses of the Intensive care unit. We deeply thank Dr Elantamilan

Durairajan for his help in data analysis and statistics.

NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

References

1. Tucker GJ: The diagnosis of delirium and DSM-IV. Dement

Geriatr Cogn Disord 1999;10:359-363.

2. Ely EW, Pun BT. The Confusion Assessment Method for the ICU

(CAM-ICU) training manual. ICU Delirium and Cognitive

PT
Impairment Study Group. Vanderbilt University Medical Center,

RI
Nashville. 2002.

SC
3. Jacobi J, Fraser GL, Coursin DB, Riker RR, Fontaine D, Wittbrod

Tet ,et al, Clinical practice guidelines for the sustained use of
NU
sedatives and analgesics in the critically ill adult. Crit Care Med
MA

2002; 30:119-141.

4. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP,
ED

Horwitz RI. Clarifying confusion: The confusion assessment

PT

method. A new method for detection of delirium. Ann Intern Med

1990;113(12):941-8.
CE

5. Ely EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L,

AC

Truman B, Speroff T, Gautam S, Margolin R, Hart RP, Dittus R:

Delirium in mechanically ventilated patients: validity and

reliability of the confusion assessment method for the intensive

care unit (CAM-ICU). JAMA 2001;286(21):2703-10.

6. Justic M. Does “ICU psychosis” really exist? Crit Care Nurse

2000;20:28-37.
 ACCEPTED MANUSCRIPT

7. Meagher DJ, Trzepacz PT. Motoric subtypes of delirium. Semin

Clin Neuropsychiatry 2000;5:75-85.

8. Hshieh TT, Fong TG, Marcantonio ER, Inouye SK. Cholinergic

deficiency hypothesis in delirium: a synthesis of current evidence.

J Gerontol A Biol Sci Med Sci 2008;63:764-72.

PT
9. Han L, McCusker J, Cole M, Abrahamowicz M, Primeau F, Elie

RI
M. Use of medications with anti cholinergic effect predicts clinical

SC
severity of delirium symptoms in older medical inpatients. Arch

Intern Med 2001; 161:1099-105.

NU
10. Liptzin B, Laki A, Garb JL, Fingeroth R, Krushell R. Donepezil in
MA

the prevention and treatment of post-surgical delirium. Am J

Geriatr Psychiatry 2005;13:1100-6.

ED

11. Van Eijk MMJ, Kesecioglu J, Slooter AJ. Intensive care delirium
PT

monitoring and standardised treatment: a complete survey of Dutch

intensive care units. Intensive Crit Care Nurs 2008;24:218–21.

CE

12. Shehabi Y, Botha JA, Boyle MS, Ernest D, Freebairn RC, Jenkins
AC

IR, Roberts BL, Seppelt IM. Sedation and delirium in the intensive

care unit: an Australian and New Zealand perspective. Anaesth

Intensive Care 2008;36(4):570–8.

13.Mac Sweeney R, Barber V, Page V, Ely EW, Perkins GD, Young

JD, McAuley DF. A national survey of the management of

delirium in UK intensive care units. QJM: An International Journal

 ACCEPTED MANUSCRIPT

of Medicine 2010;103(4):243–51.

14.Ely EW, Stephens RK, Jackson JC, Thomason JW, Truman B,

Gordon S, Dittus RS, Bernard GR. Current opinions regarding the

importance, diagnosis, and management of delirium in the

intensive care unit: a survey of 912 healthcare professionals. Crit

PT
Care Med 2004;32:106–12.

RI
15.Patel RP, Gambrell M, Speroff T, Scott TA, Pun BT, Okahashi J,

SC
Strength C, Pandharipande P, Girard TD, Burgess H, Dittus RS,

Bernard GR, Ely EW. Delirium and sedation in the intensive care
NU
unit: survey of behaviors and attitudes of 1384 healthcare
MA

professionals. Crit Care Med 2009;37:825–32.

16.Ely EW, Margolin R, Francis J, May L, Truman B, Dittus R,

ED

Speroff T, Gautam S, Bernard GR, Inouye SK. Evaluation of

PT

delirium in critically ill patients: validation of the Confusion

Assessment Method for the Intensive Care Unit (CAM-ICU). Crit

CE

Care Med 2001;29:1370-9.

AC

17.Bergeron N, Dubois MJ, Dumont M, Dial S, Skrobik Y. Intensive

Care Delirium Screening Checklist: evaluation of a new screening

tool. Intensive Care Med 2001;27:859-64.

18.Van Rompaey B, Schuurmans MJ, Shortridge-Baggett LM, Truijen

S, Bossaert L. Risk factors for intensive care delirium: a systematic

review. Intensive Crit Care Nurs 2008;24(2):98–107.

 ACCEPTED MANUSCRIPT

19.Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y. Incidence, risk

factors and consequences of ICU delirium. Intensive Care Med

2007;33(1):66–73.

20. Inouye SK, Viscoli CM, Horwitz RI, Hurst LD, Tinetti ME. A

predictive model for delirium in hospitalized elderly medical

PT
patients based on admission characteristics. Ann Int Med

RI
1993;119(6):474-81.

SC
21.Lee HB, Mears SC, Rosenberg PB, Leoutsakos JM, Gottschalk A,

Sieber FE: Predisposing factors for postoperative delirium after hip

NU
fracture repair in individuals with and without dementia. J Am
MA

Geriatr Soc 2011, 59(12):2306–2313.

22.Pandharipande P, Shintani A, Peterson J, Pun BT, Wilkinson GR,

ED

Dittus RS, Bernard GR, Ely EW. Lorazepam is an independent risk

PT

factor for transitioning to delirium in intensive care unit patients.

Anesthesiology 2006;104(1):21–6.
CE

23. Lin SM, Huang CD, Liu CY, Lin HC, Wang CH, Huang PY, Fang
AC

YF, Shieh MH, Kuo HP. Risk factors for the development of early-

onset delirium and the subsequent clinical outcome in

mechanically ventilated patients. J Crit Care 2008;23(3):372–9.

24.Angles EM, Robinson TN, Biffl WL, Johnson J, Moss M, Tran

ZV, Moore EE. Risk factors for delirium after major trauma. Am J

Surg 2008;196(6):864-70.
 ACCEPTED MANUSCRIPT

25.Plaschke K, Von Haken R, Scholz M, Engelhardt R, Brobeil A,

Martin E, Weigand MA. Comparison of the confusion assessment

method for the intensive care unit (CAM-ICU) with the Intensive

Care Delirium Screening Checklist (ICDSC) for delirium in critical

care patients gives high agreement rate (s). Intensive Care Med

PT
2008;34(3):431-6.

RI
26.van Eijk MM, van den Boogaard M, van Marum RJ, Benner P,

SC
Eikelenboom P, Honing ML, van der Hoven B, Horn J, Izaks GJ,

Kalf A, Karakus A. Routine use of the confusion assessment

NU
method for the intensive care unit: a multicenter study. Am J
MA

Respir Crit Care Med 2011;184(3):340-4.

27.Gusmao-Flores D, Salluh JI, Chalhub RÁ, Quarantini LC. The

ED

confusion assessment method for the intensive care unit (CAM-

PT

ICU) and intensive care delirium screening checklist (ICDSC) for

the diagnosis of delirium: a systematic review and meta-analysis of

CE

clinical studies. Crit Care 2012;16(4):R115.

AC

28.Toro AC, Escobar LM, Franco JG, Diaz-Gomez JL, Munoz JF,

Molina F, Bejarano J, Yepes D, Navarro E, Garcia A, Ely WE,

Esteban A. Spanish version of the CAM-ICU (Confusion

Assessment Method for the Intensive Care Unit). Pilot study of

validation. Med Intensiva 2010;34:14-21.

 ACCEPTED MANUSCRIPT

29.Chuang WL, Lin CH, Hsu WC, Ting YJ, Lin KC, Ma SC.

Evaluation of the reliability and validity of the Chinese version of

the confusion assessment method for the intensive care unit. Hu Li

Za Zhi 2007;54:45-52.

PT
RI
Cognitive Test For Delirium7
Cognitive Test For Delirium – abbreviated version8
Confusion Assessment for Intensive Care Unit(CAM-ICU)9,10

SC
Intensive Care Delirium Screening Checklist(ICDSC)11
The Neelon And Champagne Confusion Scale(NEECHAM) 12,13
Delirium Detection Score14
NU
Table 1: Delirium assessment methods commonly used in ICU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

TOTAL DELIRIUM NO DELIRIUM

P
(n = 310) (n = 140) (n = 170)
Age (mean ± SD) 47.87 ± 14.55 48.06 ±14.47 47.70 ± 14.74 P = 0.902
Male, n (%) 189 96 (68.8 %) 93 (54.54 %) 0.89
Female, n (% ) 121 43 (31.11 %) 78 (45.45 %) 0.078
APACHE II 13.8 ±6.39 9.8 ± 6.1 18.0 ± 5.2 P<0.0001
Diagnosis at admission, n (%)
Postoperative 121 37 (26.6 %) 84 (49.09 %) 0.0163
Renal 74 25 (17.7 %) 49 (29.09 %) 0.1024
Gastrointestinal 43 28 (20 %) 15 (9.09 %) 0.2852
Respiratory 28 16 (11.1 %) 12 (7.27 %) 0.7391

PT
Others 44 35 (24.4 %) 9 (5.45 %) 0.0325

RI
Table 2: Demographic parameters and diagnosis at admission

SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

CAM-ICU ICDSC
Sensitivity 84.4% 70.5%-93.5% 77.8% 62.9%-88.8%
Specificity 94.6% 84.9%- 98.9% 94.6% 84.9%-98.9%
Positive Likelihood Ratio 15.48 5.12 - 46.86 14.26 4.7-43.3
Negative Likelihood Ratio 0.16 0.08 - 0.33 0.24 0.14-0.41
Disease Prevalence 45.0% 35.0%- 55.3% 45.0% 35.0%-55.3%
Positive Predictive Value 92.7% 80.1%- 98.5% 92.1% 78.6%-98.3%
Negative Predictive Value 88.1% 77.1%- 95.1% 83.9% 72.3%- 92.0%
Diagnostic odds ratio
Overall 86.1 49.7-149.2 36.9 22.9-59.3

PT
Intensivist 120.5 53.5-271.5 53.0 25.2-110.2
Nurse 67.0 31.2-143.8 27.0 14.4-50.7

RI
Table 3: Diagnostic value of CAM-ICU and ICDS with 95% CI

SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Sensitivity, Specificity, PPV, NPV, PLR NLR

%(95% CI) %(95% %(95% %(95% (95% CI) (95%CI)
CI) CI) CI)
CAM-ICU (verbally communicating)
Nurse 75.0(62-86) 96.2(91-99) 91.3(79-98) 87.9(80-93) 19.9(7.5-52.6) 0.26(0.2-0.4)
Intensivist 82.1(79-91) 95.3(89-98) 90.2(79-97) 91.0(84-96) 17.4(7.3-41.3) 0.2(0.1-0.3)
ICDSC (verbally communicating)
Nurse 73.2(60-84) 89.6(82-95) 78.8(65-89) 86.4(78-92) 7.1(3.9-12.6 0.3(0.2-0.5)
Intensivist 76.8(63-87) 92.4(86-97) 84.3(71-93) 88.3(81-94) 10.2(5.1-20.1) 0.25(0.2-0.4)
CAM-ICU (verbally non-communicating)

PT
Nurse 84.52(75-91) 90.6(81-96) 92.2(83-97) 81.7(71-90) 9(4.2-19.4) 0.17(0.1-0.28)
Intensivist 94.1(87-98) 90.6(81-96) 92.9(85-97) 92.1(82-97) 10.03(4.7- 0.07(0.03-
21.5) 0.15)

RI
ICDSC (verbally non- communicating)
Nurse 76.19(65.7- 90.6(81-96) 91.43(82- 74.36(63- 8.13(3.8-17.6) 0.26(0.2-0.4)
84.8) 97) 83)

SC
Intensivist 79.8(70-88) 95.3(87-99) 95.7(88-99) 78.2(67-87) 17(5.6-51.6) 0.21(0.1-0.3)

Table 4: Observer specific diagnostic parameters of assessment methods,

NU
CI- Confidence interval, PPV- Positive predictive value, NPV- Negative
predictive value, PLR - Positive likelihood ratio, NLR- Negative
likelihood ratio
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Relative 95 % CI Significance
Risk Level
Sepsis 1.8333 1.2142 to 2.7681 P = 0.0039
Benzodiazepine use 1.7143 1.1177 to 2.6293 P = 0.0135
Mechanical ventilation 1.6064 1.0326 to 2.4991 P = 0.0355
Dyselectrolytemia 1.7051 1.1331 to 2.5659 P = 0.0105
BUN/ S.creatinine > 18 1.6731 1.0772 to 2.5987 P = 0.0220
APACHE II >16 2.2745 1.4455 to 3.5789 P = 0.0004
Postoperative cases 0.5688 0.3366 to 0.961 P = 0.0350

PT
Table 5: Relative risk of various factors along with its significance

RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
ED
PT
CE

Figure 1. Patient flow diagram.

AC
 ACCEPTED MANUSCRIPT

HIGHLIGHTS

1. CAM-ICU was found to be a better delirium assessment tool than

the ICDSC.

2. The results of the assessment tools whether performed by the

PT
intensivists or the nurses were comparable.

RI
3. Sepsis, benzodiazepine use, mechanical ventilation, and APACHE

SC
II >16 were found to have significantly high relative risk of being

delirious.
NU
MA
ED
PT
CE
AC