VALIDATION OF THE CRITICAL-CARE PAIN

OBSERVATION TOOL IN ADULT PATIENTS

By Céline Gélinas, RN, PhD, Lise Fillion, RN, PhD, Kathleen A. Puntillo, RN, DNSc, Chantal Viens, RN, PhD, and
Martine Fortier, MPs. From School of Nursing, McGill University, Montreal, Quebec (CG), Faculty of Nursing,
Laval University, Quebec City, Quebec (LF, CV, MF), and Department of Physiological Nursing, University of
California, San Francisco, Calif (KP).

• BACKGROUND Little research has been conducted to validate pain assessment tools in critical care,
especially for patients who cannot communicate verbally.
• OBJECTIVE To validate the Critical-Care Pain Observation Tool.
• METHODS A total of 105 cardiac surgery patients in the intensive care unit, recruited in a cardiology
health center in Quebec, Canada, participated in the study. Following surgery, 33 of the 105 were evalu-
ated while unconscious and intubated and 99 while conscious and intubated; all 105 were evaluated
after extubation. For each of the 3 testing periods, patients were evaluated by using the Critical-Care
Pain Observation Tool at rest, during a nociceptive procedure (positioning), and 20 minutes after the
procedure, for a total of 9 assessments. Each patient’s self-report of pain was obtained while the patient
was conscious and intubated and after extubation.
• RESULTS The reliability and validity of the Critical-Care Pain Observation Tool were acceptable.
Interrater reliability was supported by moderate to high weighted κ coefficients. For criterion validity,
significant associations were found between the patients’ self-reports of pain and the scores on the Criti-
cal-Care Pain Observation Tool. Discriminant validity was supported by higher scores during position-
ing (a nociceptive procedure) versus at rest.
• CONCLUSIONS The Critical-Care Pain Observation Tool showed that no matter their level of con-
sciousness, critically ill adult patients react to a noxious stimulus by expressing different behaviors that
may be associated with pain. Therefore, the tool could be used to assess the effect of various measures
for the management of pain. (American Journal of Critical Care. 2006;15:420-427)

P ain is an important stressor for many patients in

critical care,1-3 and it is not unusual for the inten-
sity of the pain to be described as moderate to
severe.4-12 Pain assessment is the first step in proper pain
relief, an important goal in patients’ care. Although
sedative agents, mechanical ventilation, and changes in
the level of consciousness.13,14 Several pain scales have
been used to document self-reporting of pain in intubated
patients.7-9,12 In the absence of a patient’s self-report,
observable behavioral and physiological indicators
critical care clinicians strive to obtain each patient’s become important indices for the assessment of pain.13,15-17
self-report of pain, many factors compromise patients’ Preliminary research18-20 has been conducted to
ability to communicate verbally, including the use of validate instruments that include behavioral and/or
physiological indicators. Use of these instruments in
critical care practice is restricted because of the limi-
To purchase electronic or print reprints, contact The InnoVision Group, 101
Columbia, Aliso Viejo, CA 92656. Phone, (800) 809-2273 or (949) 362-2050 tations of the studies. Limitations include small sample
(ext 532); fax, (949) 362-2049; e-mail, [email protected]. sizes (<40 patients),18-20 lack of validation in intubated

420 AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 http://ajcc.aacnjournals.org
 Table 1 Description of the Critical-Care Pain Observation Tool

Indicator Description Score

Facial expression No muscular tension observed Relaxed, neutral 0
Presence of frowning, brow lowering, orbit tightening, Tense 1
and levator contraction
All of the above facial movements plus eyelid tightly Grimacing 2
closed
Body movements Does not move at all (does not necessarily mean Absence of movements 0
absence of pain)
Slow, cautious movements, touching or rubbing the Protection 1
pain site, seeking attention through movements
Pulling tube, attempting to sit up, moving limbs/ Restlessness 2
thrashing, not following commands, striking at staff,
trying to climb out of bed
Muscle tension No resistance to passive movements Relaxed 0
Evaluation by passive flexion and Resistance to passive movements Tense, rigid 1
extension of upper extremities Strong resistance to passive movements, inability to Very tense or rigid 2
complete them
Compliance with the ventilator Alarms not activated, easy ventilation Tolerating ventilator or 0
(intubated patients) movement
Alarms stop spontaneously Coughing but tolerating 1
Asynchrony: blocking ventilation, alarms frequently Fighting ventilator 2
OR activated
Vocalization (extubated patients) Talking in normal tone or no sound Talking in normal tone
or no sound 0
Sighing, moaning Sighing, moaning 1
Crying out, sobbing Crying out, sobbing 2
Total, range 0-8

patients, 18 use of a subjective scale (eg, absence, they were 18 years or older, had been admitted for car-
slight, moderate, and extreme intensity of behaviors),18 diac surgery, understood French, were in the ICU after
confusion in the definition of behaviors (eg, body surgery, and were able to hear and to see. Patients
movements and muscle rigidity), and use of dependent were excluded if they had been admitted for a heart
observations (ie, statistical analysis of the observa- transplant or thoracic aortal aneurysm repair, received
tions rather than of the sample of patients).19 The aim medical treatment for chronic pain, had an ejection
of our study was to examine the reliability and validity fraction less than 0.25, had preexisting psychiatric or
of a newly developed instrument for pain assessment: neurological problems, had a dependence on alcohol
the Critical-Care Pain Observation Tool (CPOT). or drugs, received neuromuscular blockers following
surgery, or had complications after surgery (eg, hem-
orrhage, delirium, death).
This study was approved by the human research
Behavioral and physiological indicators committee of the health center. Recruitment was done
are important indices for assessment of the day before the surgery; the study was explained to
pain in patients unable to self-report. eligible patients, and informed consent was obtained.
At this time, patients were taught how to use the pain
intensity descriptive scale.

Method Instruments
Design, Sample, and Ethics Critical-Care Pain Observation Tool. The CPOT,
A repeated measures design was chosen for this developed in French, has 4 sections, each with differ-
quantitative study. A convenience sample of 105 cardiac ent behavioral categories: facial expression, body
surgery patients in the intensive care unit (ICU) at a car- movements, muscle tension, and compliance with the
diology health center in Quebec, Canada, was recruited ventilator for intubated patients or vocalization for
for the study. Patients were considered for inclusion if extubated patients (Table 1). Items in each section are

http://ajcc.aacnjournals.org AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 421
 Intensive care unit

First testing period Second testing period Third testing period

Unconscious Conscious intubated Conscious extubated
intubated patients patients patients
Consent
day before Operating
surgery room
T1 T2 T3 T4 T5 T6 T7 T8 T9
R P REC R P REC R P REC

Figure 1 Illustration of the time of consent and the 3 testing periods.

Abbreviations: P, positioning procedure; R, rest; REC, recovery (20 minutes after the positioning procedure).

scored from 0 to 2, with a possible total score ranging Procedure

from 0 to 8. The CPOT was developed as follows. Three testing periods, each including 3 assess-
Some items and their operational definitions were ments for a total of 9 pain assessments (T1-T9) with
derived from previously described instruments for the CPOT, were completed during each patient’s early
pain assessment.18-21 In addition, pain indicators were postoperative course (Figure 1). For each patient, the
described by using findings from a chart review of the f irst 3 assessments (T1-T3) were done while the
medical files of 52 critically ill patients22 and from 9 patient was intubated and still unconscious (ie, with a
focus groups with 48 critical care nurses and inter- sedation score of 5 or 6 on the Ramsay Scale). T1 was
views of 12 physicians.23 done with the patient at rest, approximately 2 hours
Content validity of the CPOT was established with after the end of surgery. T2 was completed a few min-
4 physicians and 13 critical care nurses. The physicians utes after T1 during positioning of the patient. Posi-
and nurses completed a questionnaire on the relevance tioning represented a previously confirmed nociceptive
of the inclusion of these indicators in the CPOT by procedure.9 On the basis of the patient’s needs, endo-
using a Likert scale (1 = not at all, 2 = a little, 3 = moder- tracheal suctioning often was performed at the same
ately, and 4 = very much). Content validity indices, time as positioning. Finally, T3 was done at recovery,
which are the proportion of participants who answered 20 minutes after the positioning procedure.
3 or 4 on the Likert scale, were calculated. All indica- The second testing period (assessments T4-T6)
tors had indices of 0.88 to 1.00. Content validity indices was 3 hours after the first testing period. During this
greater than 0.80 were sufficiently satisfactory24 to con- time, the patient was still intubated but conscious.
sider including all these indicators in the CPOT. Patients were considered conscious if they had a score
Pain Intensity Descriptive Scale. A previously vali- of 2, 3, or 4 on the Ramsay Scale.
dated pain intensity descriptive scale (0 = none, 1 = mild, Finally, the third testing period (assessments T7-
2 = moderate, 3 = severe, 4 = unbearable) was used. T9) was after the patient was extubated, approximately
This scale has been used in previous studies18,25 in 5 hours after the second testing period. The position-
acute and critical care. ing procedure at T8 sometimes occurred with ambula-
Confusion Assessment Method for the Intensive tion and/or respiratory exercises, which were part of
Care Unit. The Confusion Assessment Method for the the postoperative care protocol.
Intensive Care Unit (CAM-ICU) was used to assess For each of the 3 testing periods, patients were
delirium. The instrument has good sensitivity and speci- evaluated with the CPOT for 1 minute at rest both
ficity for assessing delirium in critically ill patients.26,27 before and after positioning and for the duration of the
Two modifications were made in the CAM-ICU to positioning procedure. This standardization of proce-
adapt it to the sedation scale used in our study and to dures was based on the work of Puntillo et al.9 One of
facilitate assessment of patients’ inattention. First, the us (C.G.) and a critical care nurse (G.N.) evaluated the
Ramsay Scale28 was used to assess the level of sedation. patients. Upon completion of the CPOT during the
Second, patients’ inattention was verified by assessing second testing period (ie, assessments T4-T6), intu-
their capacity to concentrate on the pain intensity bated patients communicated the presence or absence
descriptive scale used in our study. of pain by nodding their heads (yes or no) to the

422 AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 http://ajcc.aacnjournals.org
 Table 2 Description of reliability and validity methods examined in this study

Psychometric
property Description Coefficient or analysis Level of acceptability*
Interrater Interrater reliability is the consistency κ coefficient (proportion <0 Poor
reliability with which 2 raters agree on their of responses in which the 0-0.20 Slight
measurement/observation (ie, the CPOT) 2 raters agreed) 0.21-0.40 Fair
of a phenomenon (ie, pain) 0.41-0.60 Moderate
Two raters assessed the patients in this 0.61-0.80 Substantial
study: the principal investigator and 1 0.81-1.00 Almost perfect
critical care nurse
Criterion Criterion validity refers to the relationship Intubated patients (T4-T6): P ≤ .01
validity between the instrument (ie, the CPOT) analysis of variance
and the gold standard measure of pain
(ie, the patient’s self-report) Extubated patients (T7-T9):
In this study, yes/no and pain intensity Spearman correlation
were used as the gold standard self-report
measures
Discriminant Discriminant validity refers to evidence Rest time compared with P ≤ .01
validity that instruments measuring 2 different positioning for all three
constructs should not correlate testing periods: paired t test
In this study, we examined whether the
CPOT could be used to discriminate
between pain during positioning and
lack of pain at rest

Abbreviation: CPOT, Critical-Care Pain Observation Tool.

* Levels of acceptability for interrater reliability scores from Landis and Koch.30

question, Do you have pain? This procedure was extubated patients’ self-reports of pain intensity (ordi-
selected because many intubated patients during this nal descriptive scale) and the CPOT scores (assess-
phase of their recovery were unable to use the pain ments T7-T9). Finally, discriminant validity 31 was
intensity descriptive scale. Before the third testing examined by performing paired t tests between assess-
period, at T7, patients were evaluated by using the ments with the CPOT taken at rest and during posi-
CAM-ICU to determine the presence of delirium. tioning (T1 with T2, T4 with T5, and T7 with T8).
Three patients were excluded because of delirium.
During the third testing period (ie, assessments T7- Results
T9) after completion of the CPOT, the extubated Characteristics of the Sample
patients used the pain intensity descriptive scale to A total of 131 patients were approached for consent
grade their pain. the day before surgery, and 117 (89%) agreed to partici-
pate in the study. Reasons for refusal were as follows:
Data Analysis anxious about the surgery (n = 9 patients), not interested
Statistical analyses were completed by using ver- (n = 3), undecided (n = 1), and bad experience with
sion 11.5 of SPSS for Windows (SPSS Inc, Chicago, research (n = 1). During the course of the study, 8
Ill). Descriptive statistics were computed for all vari- patients were excluded because of postoperative compli-
ables. Interrater reliability was examined. Weighted κ cations (hemorrhage, delirium, death), 3 because their
coefficients were calculated for all assessments (T1- surgery was canceled, and 1 because of extubation right
T9).29 To test validity of the CPOT, we determined cri- after surgery. The final sample size was 105 patients
terion and discriminant validity (Table 2). Criterion enrolled during a 3-month period. Table 3 gives the
validity was examined by measuring the relationship demographic characteristics of the patients.
between the CPOT scores and the patients’ self- Anesthesia was similar for all patients, and all
reports, the gold standard measure of pain. Analysis were receiving continuous infusions of propofol after
of variance was used to examine the differences between surgery (mean dose 85.4 mg/h, SD 39.7 mg/h). For
the intubated patients’ self-reports of pain (yes or no) each patient, this medication was tapered off and
and the CPOT scores (assessments T4-T6). Also, stopped 1 to 3 hours after the patient’s arrival in the
Spearman correlations31 were calculated between the ICU. Thus, all patients were receiving propofol during

http://ajcc.aacnjournals.org AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 423
Table 3 Description of the study sample (n = 105)* Conscious
4.0 intubated
Variable Value T5 Conscious
3.5
Age, mean (SD) 60 (8) extubated
3.0 T8
Sex, No. (%) of patients Unconscious
Male 83 (79) 2.5 intubated

Mean score
Female 22 (21) T2
2.0
Type of cardiac surgery, No. of patients (%)
Coronary artery bypass graft 83 (79)
1.5
Valvular repair or replacement 11 (10)
T4 T6
Coronary artery bypass graft and valvular 9 (9) 1.0
surgery T9
T7
Interauricular/interventricular 2 (2) 0.5 T3
T1
communication repair
0.0
*All patients had sternal incisions. Rest Procedure Recovery

Time

the first testing period (ie, assessments T1-T3). All Testing period Assessment Mean SD
patients also were receiving continuous infusions of 1 Unconscious T1 Rest 0.55 1.03
fentanyl when they were admitted to the ICU from intubated patients T2 Procedure 2.70 1.36
surgery. The mean dosage of fentanyl decreased from (n = 33) T3 Recovery 0.67 0.89
73.7 µg/h (SD 21.8) at the first testing period to 50.7 2 Conscious T4 Rest 1.21 1.23
µg/h (SD 31.1) at the third testing period. Rarely, intubated patients T5 Procedure 3.38 1.38
patients (n = 4) received an intravenous bolus of fen- (n = 99) T6 Recovery 1.35 1.42
tanyl before positioning. 3 Conscious T7 Rest 0.69 0.87
extubated patients T8 Procedure 2.79 1.31
Sample at the 3 Testing Periods (n = 105) T9 Recovery 0.87 1.04
First Testing Period. For assessments T1 to T3, data
were collected on 33 of the 105 intubated patients who
were unconscious, a criterion for testing during this Figure 2 Mean scores and standard deviations of the Critical-
period. CPOT scores were higher during the positioning Care Pain Observation Tool for the 3 testing periods (N = 105
patients). The scores can range from 0 to 8.
procedure (T2) than during rest (T1) or recovery (T3;
Figure 2).
Second Testing Period. For assessments T4 to T6,
data were collected on 99 of the awake 105 intubated Table 4 Weighted κ coefficients for each assessment from
patients. The remaining 6 patients were extubated before T1 to T9*
the completion of this testing period. Again, CPOT No. of
scores were higher during the positioning procedure Assessment patients Weighted κ coefficient
(T5) than during rest (T4) or recovery (T6). Moreover, T1 12 0.83
in this testing period, patients had the highest scores T2 12 0.63
on the CPOT (Figure 2).
T3 14 0.85
Third Testing Period. Finally, for assessments T7 to
T9, all 105 patients were assessed after they were extu- T4 29 0.52
bated. The CPOT scores were similar to those of the 2 T5 33 0.85
previous testing periods (Figure 2). T6 33 0.88
T7 34 0.62
Interrater Reliability
Together, the principal investigator and the critical T8 33 0.77
care nurse (C.G. and G.N.) completed the CPOT at all T9 34 0.71
9 assessments and were blinded to each other’s scores. *Assessments made by using the Critical-Care Pain Observation
The sample sizes for interrater reliability differed for Tool were independently completed by the principal investigator
and the critical care nurse when both were present.
each time, reflecting the times when both were present.

424 AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 http://ajcc.aacnjournals.org
 Table 5 Differences in scores on the Critical-Care Pain Observation Tool according to patients’ self-reports of pain in the
second testing period (T4, T5, and T6)

Patients’ self-reports of pain: Scores ANOVA

Assessment pain present or absent* Mean SD F df
Yes, present (n = 53) 1.62 1.38 11.83† (1, 92)
T4
No, absent (n = 41) 0.78 0.85
Yes, present (n = 79) 3.65 1.31 22.18† (1, 95)
T5
No, absent (n = 18) 2.11 0.90
Yes, present (n = 54) 2.07 1.40 38.89† (1, 91)
T6
No, absent (n = 39) 0.49 0.88
Abbreviation: ANOVA, analysis of variance.
*Because of intermittent drowsiness postoperatively, 5 patients at T4, 2 at T5, and 6 at T6 were unable to give their self-reports of pain by
nodding their heads.
†P ≤ .001. Alpha is adjusted to 0.01 because 3 comparisons were made on the same subjects.

Weighted κ coefficients were moderate to high at all

assessments (Table 4). CPOT scores were associated with
patients’ self-reports of pain.
Criterion Validity
Mean CPOT scores according to patients’ self-
reports of the presence or absence of pain during the When patients were intubated during the second
second testing period (ie, assessments T4-T6) and the testing period, CPOT scores differed significantly
analysis of variance are presented in Table 5. At each between those who reported pain and those who did
assessment in this testing period, CPOT scores were not. Moreover, when patients were extubated during
significantly higher for intubated patients reporting the third testing period, the higher a patient’s self-
pain than for those who had no pain. report of pain was, the higher was the patient’s score
During the third testing period (ie, assessments on the CPOT. These results are consistent with those
T7-T9), mean pain intensity scores were significantly of previous studies18,32 in which self-reports of pain of
higher during the positioning procedure at T8 (2.01) patients in a postanesthesia care unit were moderately
than during rest at T7 (1.71) and recovery at T9 related to pain behaviors. Our results support the crite-
(1.40). Spearman correlations of 0.49, 0.59, and 0.40 rion validity of the CPOT because the indicators were
(P ≤ .001) at T7 to T9 showed that the patients’ self- tested against the most valid measurement of pain;
reported pain intensity scores were moderately corre- that is, the patients’ self-reports.
lated with the CPOT scores.

Discriminant Validity
At the 3 testing periods, CPOT scores were signif- CPOT scores were higher during painful
icantly higher during positioning than during the rest procedures, lending support to its validity.
periods. Table 6 gives the results of the paired t tests.

Discussion Discriminant validity was supported by the find-

Our f indings validated the CPOT, which was ing that CPOT scores were higher during positioning
developed specifically to assess pain in ICU patients. than at rest in the 3 testing periods. Payen et al19 also
Interrater reliability was high for most assessments found higher behavioral scores during positioning
and moderate at T4. Payen et al19 obtained a weighted than at rest in unconscious critically ill patients. Such
κ coefficient of 0.74 when they compared behavioral results emphasize that pain behaviors are observable
pain scores between pairs of evaluators. A total of 46 even if a patient cannot report pain.
nurses and nurse’s aides, 1 physical therapist, and 1 Our study, however, is the first to document differ-
physician participated in that study.19 In our study, ences in pain behavior scores according to levels of
only 2 evaluators used the instrument, which is a limi- activity during different states of consciousness and
tation to the examination of interrater reliability, and intubation: unconscious and intubated, conscious and
results cannot be generalized to other ICU nurses. intubated, and then awake and extubated. These results

http://ajcc.aacnjournals.org AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 425
 Table 6 Differences in scores on the Critical-Care Pain tioned. However, when conscious intubated patients
Observation Tool measured at rest before the procedure were positioned during assessment T5, 18 of them
(T1, T4, and T7) and during the procedure (T2, T5, and T8) who did not report pain had high CPOT scores (mean
No. of 2.11, SD 0.90). Perhaps for these patients the position-
Assessments patients t df ing was a distressful or an uncomfortable experience1,9,37
T1-T2 33 -9.01* 32 rather than a painful one. Also, the endotracheal tube
T4-T5 -12.07* 98
may have caused coughing during the positioning pro-
99
cedure, leading to higher CPOT scores in the absence
T7-T8 105 -15.96* 104
of reported pain. This finding suggests that behaviors
*P < .001. Alpha is adjusted to 0.01 because 3 comparisons were observed by using the CPOT may be an indicator of
made on the same subjects.
more than pain. Further research is warranted to deter-
mine the sensitivity and specificity of the CPOT as a
suggest that patients, whatever their levels of con- measure of pain.
sciousness, may demonstrate pain behaviors in response
to a nociceptive procedure. Whether a behavioral
response to a noxious procedure is accompanied by Behaviors observed by using the CPOT
perception of pain in an unconscious patient is
may indicate more than pain.
unknown. Until there is evidence to the contrary,
experts recommend that healthcare providers assume
that unconscious patients may have pain, especially if
behavioral responses to a known noxious stimulus We also found that CPOT scores were similarly low
occur. The experts33,34 recommend that these patients for both unconscious and conscious extubated patients.
be treated the same way as conscious patients when the This result may have occurred because the patients were
patients are exposed to sources of pain. highly sedated while unconscious and may have been
experiencing the residual effects of anesthesia.19 Once
extubated, they could have experienced less severe pain
Experts recommend we assume than they did when they were intubated.
that unconscious patients have pain, Data collection in this study was completed in the
especially if behavioral responses to 8 hours after surgery, a period when intermittent drowsi-
noxious stimuli occur. ness can be expected. Previous studies7-9,12,20 in which
intubated patients provided self-reports of pain were
conducted in periods varying from 12 to 72 hours after
Indeed, in a study by Lawrence,35 formerly uncon- the end of surgery. Those patients might have had more
scious patients revealed that they could hear, under- time to recuperate from the residual effects of anesthe-
stand, and respond emotionally to what was being said sia than our patients did. In the study by Ferguson et
while they were unconscious. In light of this finding, al,5 patients’ self-reports of pain after coronary artery
perhaps the CPOT can be used to assess pain in other bypass graft surgery were collected more than 8 hours
populations of critical care patients. This hypothesis after the end of surgery, and 9 (21%) of 43 patients
requires confirmation in future studies. were unable to communicate their self-report of pain
Interestingly, patients seemed to have higher CPOT because of drowsiness. The occurrence of the same
scores when they were conscious and intubated than inability to communicate in the patients in our study is
they did when they were unconscious or extubated. not surprising.
The presence of the endotracheal tube as a potential
source of pain6,36 for cardiac surgery patients, in addi- Limitations
tion to the sternal incision,5,8 may help to explain these This study was not without limitations. First, data
findings. Moreover, mechanical ventilation modes were collected by only 2 persons. More raters should
provided a positive inspiratory pressure to allow dis- be used in tests of interrater reliability in subsequent
tribution of oxygen throughout the lungs.17 This pres- evaluations of the CPOT. Second, data could be col-
sure may cause stretching of the sternal incision, lected for only 33 of the 105 patients while the patients
which does not occur in nonsurgical patients, and were unconscious. Third, postoperative drowsiness led
which can be painful to some patients. to missing data for some patients. Finally, cardiac
In our study, 65 conscious intubated patients had surgery patients are a relatively healthy ICU group
endotracheal suctioning while they were being posi- and may not represent most ICU patients, who are

426 AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 http://ajcc.aacnjournals.org
much sicker. Future research on the effectiveness of tralian intensive care unit. J Adv Nurs. 1997;26:1065-1072.
6. Puntillo KA. Pain experiences of intensive care unit patients. Heart Lung.
the CPOT as a nonverbal measure of pain in other 1990;19:526-533.
sicker ICU patients is warranted. 7. Puntillo KA. Dimensions of procedural pain and its analgesic management
in critically ill surgical patients. Am J Crit Care. 1994;3:116-122.
Despite these limitations, this study was innovative 8. Puntillo KA, Weiss JS. Pain: its mediators and associated morbidity in crit-
in several aspects. First, development of the CPOT was ically ill cardiovascular surgical patients. Nurs Res. 1994;43:31-36.
9. Puntillo KA, White C, Morris AB, et al. Patients’ perceptions and responses
based on previous research of others as well as on to procedural pain: results from Thunder Project II. Am J Crit Care.
descriptive data from 2 preliminary studies that led to 2001;10:238-251.
10. Stanik-Hutt J, Soeken KL, Fontaine DK, Gift AG. Pain experiences of
selection of the behavioral indicators. Second, the traumatically injured patients in a critical care setting. Am J Crit Care.
relationship between intubated patients’ self-reports of 2001;10:252-259.
11. Valdix SW, Puntillo KA. Pain, pain relief and accuracy of their recall after
pain and behavioral indicators was explored for the cardiac surgery. Prog Cardiovasc Nurs. Summer 1995;10:3-11.
second time. Finally, data were obtained from patients 12. Whipple JK, Lewis KS, Quebbeman EJ, et al. Analysis of pain manage-
ment in critically ill patients. Pharmacotherapy. 1995;15:592-599.
at different levels of consciousness. 13. Hamill-Ruth RJ, Marohn L. Evaluation of pain in the critically ill patient.
Future research should be conducted to determine Crit Care Clin. 1999;15:35-53.
14. Shannon K, Bucknall T. Pain assessment in critical care: what have we
if CPOT scores can be used to differentiate pain from learnt from research? Intensive Crit Care Nurs. 2003;19:154-162.
other conditions. Also, receiver operating characteristic 15. Christoph SB. The problem of pain in the critically ill patient. Crit Care
Nurs Clin North Am. 1991;3:11-16.
curve analysis could be performed to examine the speci- 16. Kwekkeboom KL, Herr KA. Assessment of pain in the critically ill. Crit
ficity and the sensitivity of the CPOT as a measure of Care Nurs Clin North Am. 2001;13:181-194.
17. Kaiser KS. Assessment and management of pain in the critically ill trauma
pain. This further testing could substantiate the CPOT as patient. Crit Care Nurs Q. 1992;15:14-34.
a valid, reliable, and useful tool for measuring pain in 18. Mateo OM, Krenzischek DA. A pilot study to assess the relationship
between behavioral manifestations and self-report of pain in postanaesthe-
critically ill patients who are unable to self-report. sia care unit patients. J Post Anesth Nurs. 1992;7:15-21.
19. Payen JF, Bru O, Bosson JL, et al. Assessing pain in critically ill sedated
patients by using a behavioral pain scale. Crit Care Med. 2001;29:2258-2263.
Conclusions 20. Puntillo KA, Miaskowski C, Kehrle K, Stannard D, Gleeson S, Nye P.
The CPOT had acceptable reliability and validity Relationship between behavioral and physiological indicators of pain, criti-
cal care patients’ self-reports of pain, and opioid administration. Crit Care
in this sample of cardiac surgery ICU patients. How- Med. 1997;25:1159-1166.
ever, the tool needs to be further validated in different 21. Ambuel B, Hamlett KW, Marx CM, Blumer JL. Assessing distress in pedi-
atric intensive care environments: the COMFORT Scale. J Pediatr Psy-
populations of critically ill patients. Appropriate pain chol. 1992;17:95-109.
assessment is an important part of quality care for 22. Gélinas C, Fortier M, Viens C, Fillion L, Puntillo KA. Pain assessment and
management in critically ill intubated patients: a retrospective study. Am J
critically ill patients, and use of validated measures of Crit Care. 2004;13:126-135.
pain could aid in the evaluation of multidisciplinary 23. Gélinas C, Viens C, Fortier M, Fillion L. Les indicateurs de douleur en
soins critiques [Pain indicators in critical care]. Perspect Infirm. 2005;2:12-
pain management techniques for nonverbal critically 14, 16-20, 22.
ill patients. 24. Waltz CF, Strickland OL, Lenz ER. Measurement in Nursing Research.
2nd ed. Philadelphia, Pa: FA Davis Co; 1991.
25. Choinière M, Melzack R, Girard N, Rondeau J, Paquin M-J. Comparisons
ACKNOWLEDGMENTS between patients’ and nurses’ assessment of pain and medication efficacy
Special thanks to Gaëlle Napert, the critical care nurse who contributed to data col- in severe burn injuries. Pain. 1990;40:143-152.
lection, and to François Harel, the statistician who reviewed the analyses. Thanks to 26. Ely EW, Inouye SK, Bernard GR, et al. Delirium in mechanically venti-
lated patients: validity and reliability of the Confusion Assessment Method
all intensive care unit nurses and physicians for their support and collaboration in the for the Intensive Care Unit (CAM-ICU). JAMA. 2001;286:2703-2710.
performance of this study. Special thanks to Drs Jean Bussières and Mathieu Simon 27. Ely EW, Margolin R, Francis J, et al. Evaluation of delirium in critically ill
for their valuable collaboration in the conception and the conduction of this study. patients: validation of the Confusion Assessment Method for the Intensive
Also, thanks to Celeste Johnston, RN, DEd, and Denise Li, RN, for their helpful comments Care Unit (CAM-ICU). Crit Care Med. 2001;29:1370-1379.
on this article. Finally, thanks to all the patients who participated in this study. 28. Ramsay M, Savege T, Simpson B, Goodwin R. Controlled sedation with
The research was performed at the Faculty of Nursing, Laval University, Quebec alphaxalone-alphadolone. Br Med J. 1974;2:656-659.
City, Quebec, Canada. This study was supported by a nursing fellowship research 29. Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide
grant from the Heart and Stroke Foundation of Canada and a doctorate research to Their Development and Use. 3rd ed. New York, NY: Oxford University
Press Inc; 2003.
grant from the Fondation de Recherche en Sciences Infirmières du Québec. 30. Landis RJ, Koch GG. The measurement of observer agreement for categor-
ical data. Biometrics. 1977;33:159-174.
REFERENCES 31. Polit DF, Hungler BP. Nursing Research: Principles and Methods. 6th ed.
1. Hallenberg B, Bergbom-Engberg I, Haljamäe H. Patients’ experiences of Philadelphia, Pa: Lippincott; 1999.
postoperative respirator treatment: influence of anaesthetic and pain treat- 32. Webb MR, Kennedy MG. Behavior responses and self-reported pain in
ment regimens. Acta Anaesthesiol Scand. 1990;34:557-562. postoperative patients. J Post Anesth Nurs. 1994;9:91-95.
2. Turner JS, Briggs SJ, Springhorn HE, Potgieter PD. Patients’ recollection 33. Bushnell MC. Perception of pain in the persistent vegetative state [letter]?
of intensive care unit experience. Crit Care Med. 1990;18:966-968. Eur J Pain. 1997;1:167.
3. Wilson VS. Identification of stressors related to patients’ psychologic 34. Klein M. Perception of pain in the persistent vegetative state [letter]? Eur J
responses to the surgical intensive care unit. Heart Lung. 1987;16:267-273. Pain. 1997;1:165-167.
4. Desbiens NA, Wu AW, Broste SK, et al. Pain and satisfaction with pain 35. Lawrence M. The unconscious experience. Am J Crit Care. 1995;4:227-232.
control in seriously ill hospitalized adults: findings from the SUPPORT 36. Gries ML, Fernsler J. Patient perceptions of the mechanical ventilation
research investigations. Crit Care Med. 1996;24:1953-1961. experience. Focus Crit Care. 1988;15:52-59.
5. Ferguson J, Gilroy D, Puntillo KA. Dimensions of pain and analgesic 37. Bergbom-Engberg I, Haljamäe H. Assessment of patients’ experience of
administration associated with coronary artery bypass grafting in an Aus- discomforts during respirator therapy. Crit Care Med. 1989;17:1068-1072.

http://ajcc.aacnjournals.org AMERICAN JOURNAL OF CRITICAL CARE, July 2006, Volume 15, No. 4 427