Intensive and Critical Care Nursing (2016) 34, 111

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/iccn

REVIEW

Pain assessment in the critically ill adult:

Recent evidence and new trends
Cline Glinas

Ingram School of Nursing, McGill University, Montreal, QC H3A 2A7, Canada

Accepted 3 March 2016

KEYWORDS Summary Pain assessment in the critically ill adult remains a daily clinical challenge. Posi-
tion statements and practice guidelines exist to guide the ICU care team in the pain assessment
Pain;
process. The patients self-report of pain remains the gold standard measure for pain and
Pain assessment;
should be obtained as often as possible. When self-report is impossible to obtain, observational
Behaviours;
pain scales including the Behavioural Pain Scale (BPS) and the Critical-Care Pain Observation
Vital signs;
Tool (CPOT) have been recommended for clinical use in the critically ill adult. However, their
Pupillometry;
adaptation and validation in brain-injured and burn ICU patients is required. Family caregivers
Critically ill;
may help in the identication of pain-related behaviours and should be more involved in the
Intensive care
ICU pain assessment process. Fluctuations in vital signs should only be considered as cues for
further assessment of pain with appropriate tools, and may better represent adverse events of
severe pain. Other physiologic measures of pain should be explored in the ICU, and pupillometry
appears as a promising technique to further study. Implementation of systematic pain assess-
ment approaches using tools adapted to the patients ability to communicate and condition
has shown positive effects on ICU pain practices and patient outcomes, but randomised control
trials are needed to conrm these conclusions.
2016 Elsevier Ltd. All rights reserved.

Implications for Clinical Practice

Pain measures must be selected according to the patients ability to communicate.

Behavioural pain scales remain alternative measures in the ICU patient unable to self-report.
Fluctuations in vital signs could be considered as adverse events of severe pain rather than indicators for pain
assessment.
Pain management initiatives must involve the ICU inter-professional team for better practices and patient outcomes.

Tel.: +514 398 6157; fax: +514 398 8455.

E-mail address: [email protected]

http://dx.doi.org/10.1016/j.iccn.2016.03.001
0964-3397/ 2016 Elsevier Ltd. All rights reserved.
2 C. Glinas

Introduction family or caregiver about the patients pain behaviours; and

(4) attempt an analgesic trial when pain is suspected, and
Pain is an enduring concern and a common symptom in reassess for pain (Herr et al., 2011).
critically ill adults. Indeed, pain has been shown to be expe- The 010 Numeric Rating Scale (NRS) is commonly used
rienced at rest by more than 30% of patients (Chanques in clinical practice, and an enlarged visual format of the
et al., 2007), and this percentage exceeds 50% during NRS was found to be the most feasible and discriminative
common care procedures in the intensive care unit (ICU) self-report scale in comparison to other scales (i.e., visual
(Glinas, 2007a; Puntillo et al., 2001, 2014). Turning, endo- analogue scale, verbal descriptor scale) and formats (i.e.,
tracheal suctioning, tube or drain removal, wound care oral versus visual) for measuring pain intensity in critically
and arterial line insertion are procedures described as the ill adult patients (Chanques et al., 2010). The vertical ther-
most painful by ICU patients in large international studies mometer format has also been found to be easier to use by
(Puntillo et al., 2001, 2014). Nurses are facing numerous critically ill adults and older adults (Glinas, 2007b; Herr,
challenges with pain assessment in the ICU as many patients 2011). When a self-report of pain intensity is not possible,
are unable to self-report because of various factors such as a simple yes or no indicating the presence versus the
altered level of consciousness (LOC), sedation and mechan- absence of pain should be considered as a valid self-report
ical ventilation. Under-assessment of pain represents one (Herr et al., 2011). When self-report of pain is impossible
of the primary barriers to the adequate treatment of pain to obtain, validated behavioural pain scales should be used
in critically ill patients (Pasero et al., 2009). The evidence as alternative measures for pain assessment. The family can
demonstrating absent or incomplete assessment of pain is also provide the care team with relevant information on the
associated with negative patient outcomes such as longer patients pain behaviours to enhance the detection of pain.
duration of mechanical ventilation and ICU length of stay, However, not much is known about the family members
and increased mortality (Kastrup et al., 2009; Payen et al., perception of patients pain-related behaviours in the ICU
2009), is a particular concern for nurses. This paper might context. To our knowledge, only one qualitative study has
help nurses in reversing this situation through the review of been conducted with seven family members of ICU patients
steps for proper pain assessment, recent evidence and new with a traumatic brain injury (TBI) (Vanderbyl and Glinas,
trends in measures and indicators for the assessment of pain 2014). Interestingly, family members identied behaviours
in the critically ill adult. indicative of pain similar to those included in most scales
recommended for clinical use such as facial expressions,
body movements and muscle rigidity. More specically, they
Denition of pain and its gold standard described movements of the eyes and the mouth, tears or
measure eye weeping, touching or trying to touch the pain site and
visible muscle tension. They also provided meaningful expla-
As per the universal denition by the International Associa- nations of their perception in relation to their knowledge
tion for the Study on Pain (IASP, 1979), pain is described as of the patient. Another study examined the interrater reli-
an unpleasant sensory and emotional experience associated ability of proxy reporters of pain (Puntillo et al., 2012).
with actual or potential tissue damage or described in terms Family proxy reporters were found to be closer to ICU
of such damage. This denition has remained unchanged in patients self-reports when compared to nurses and physi-
the Kyoto protocol of IASP Basic Pain Terminology (Loeser cians. Specically in regards to pain, moderate intraclass
and Treede, 2008). In other words, pain is recognised as a correlation coefcient (ICC) of 0.43 and 0.45 were found
multidimensional subjective experience which implies that between patient-family member reports of pain intensity
pain has to be reported by the person who is experiencing and pain distress, respectively. Lower ICCs of 0.190.40
it. Such a denition of pain highlights that the patients self- were obtained between patient-nurse and patient-physician
report is the gold standard measure of pain and should be pain reports (Puntillo et al., 2012). Unfortunately, little evi-
obtained as often as possible. However, the IASP Task Force dence exists about proxy reporters reliability in the pain
on Taxonomy (1994) has acknowledged that The inabil- assessment process and further research is necessary.
ity to self-report does not negate the possibility that an The Communication Model of Pain (Hadjistavropoulos and
individual is experiencing pain and is in need of appropri- Craig, 2002; Hadjistavropoulos et al., 2011) also offers an
ate pain-relieving treatment. Therefore, pain assessment interesting conceptual foundation for the assessment of pain
methods must be adapted to the patients cognitive capacity (Fig. 1). This A B C model conceptualises pain as an inter-
and condition (Herr et al., 2011). nal state (A) that may be encoded in particular features of
expressive behaviours (B), allowing observers or raters to
draw inferences (C) about the nature of the patients expe-
A stepwise approach for pain assessment and rience. More specically, the processing of the pain stimulus
the Communication Model of Pain is modulated by intrapersonal and/or contextual factors (A)
inuencing the way pain is experienced in each individual.
Pain should be monitored routinely in all adult ICU patients Gender, age, ethnicity, health condition and the context in
(Barr et al., 2013). In their position statement, the American which this individual is evolving are examples of these fac-
Society for Pain Management Nursing (ASPMN) has proposed tors. The use of mechanical ventilation and the administra-
a 4-step approach for the assessment of pain which can be tion of analgesic and sedative agents are more specic to the
summarised as follows: (1) always attempt to obtain the critical care context and add challenges to pain assessment.
patients self-report of pain; (2) use a validated behavioural The pain stimulus is encoded in self-report (for patients able
pain scale or look for behavioural changes; (3) ask the to communicate) and behaviours (B) which are decoded by
Pain assessment in the critically ill adult 3

Figure 1 Adapted from the Communication Model of Pain (Hadjistavropoulos and Craig, 2002).

observers (C). Self-report depends on higher mental process- called BPS-NI for non-intubated patients (Chanques et al.,
ing while behaviours are less subject to voluntary control 2009) includes an item related to vocalisation in place of
and are more automatic. Behaviours may be more difcult compliance with the ventilator for non-mechanically ven-
to decode by observers, and educational training is of great tilated patients. Both the BPS and CPOT were shown to
importance to support them in developing the competence discriminate between nociceptive procedures known to be
of adequately assessing pain behaviours using observational painful and non-nociceptive procedures (i.e. discriminant
measures. For example, standardised training for the use validation), to be related to the patients self-report of pain
of BPS and CPOT has led to appropriate tool utilisation into (i.e. criterion validation) and to lead to consistent scores
practice and consistent scoring among ICU nurses and physi- when used independently by different raters (i.e. interrater
cians (Chanques et al., 2006; Glinas et al., 2011a,b). reliability). In the Society of Critical Care Medicine (SCCM)
Such a conceptual framework may guide nurses in consid- practice guidelines (Barr et al., 2013), the BPS and the CPOT
ering different aspects of the patients situation and the are the two suggested scales for clinical use in ICU patients
context in which he/she is evolving in the pain assessment unable to self-report and whose motor function is intact and
process. It also provides a strong scientic basis for the behaviours are observable. Therefore, these scales are of
most appropriate methods to assess pain i.e., self-report and limited use in patients with a Glasgow Coma Scale (GCS)
behavioural observation to be interpreted by decoders (e.g., (Teasdale and Jennett, 1974) score of three or a Richmond
nurses and family caregivers). Furthermore, it ts nicely Agitation Sedation Scale (RASS) score of 5 (Sessler et al.,
with the current recommendations and guidelines for pain 2002) as such scores indicate that the patient is unrespon-
assessment (Barr et al., 2013; Herr et al., 2011). sive or unarousable. Some questions pertaining to their use
are discussed.
Challenges with the use of behavioural pain
scales How should we interpret behavioural pain scores?

As previously mentioned, when the patients self-report is Cut-off scores for the presence of pain have been estab-
impossible to obtain, behavioural pain scales should be used. lished for the BPS (>5) (Payen et al., 2007) and CPOT (>2)
To date, a total of eight behavioural pain scales have been (Glinas et al., 2009) which both represent more than two
developed and/or validated for use with nonverbal critically points from their respective minimal score. Behavioural pain
ill ICU adults (Glinas et al., 2013). From this recent criti- scores based on the nurses observation of the patient should
cal review and a previous one (Pudas-Thk et al., 2009), be interpreted differently from the patients self-report pain
both the 512 Behavioural Pain Scale (BPS: Payen et al., intensity scores. In fact, although they both represent pain
2001) and the 08 Critical-Care Pain Observation Tool (CPOT: scores, they are not measuring the same dimension of pain.
Glinas et al., 2006) were identied as the scales with the More specically, the patients self-report of pain intensity
most robust psychometric properties (i.e., validity and reli- relates to the sensory dimension of pain which refers to
ability) for detecting pain in medical, surgical and trauma the perception of pain by the person who is experiencing
ICU patients. Briey, the content of these two scales is it (Melzack and Casey, 1968). On the other hand, the nurses
similar to three items related to facial expression, body observational scores are associated with the behavioural
movements, and compliance with the ventilator. The CPOT component of pain i.e., behavioural responses exhibited by
has a fourth item assessing muscle rigidity, and compliance patients to express their pain (McGuire, 1992) and decoded
with the ventilator can be replaced with vocalisation in non- by an external observer (as illustrated in the C component of
mechanically ventilated patients. An adaptation of the BPS, Fig. 1). Self-reported pain intensity scores and behavioural
4 C. Glinas

Figure 2 Pain Intervention Algorithm Glinas C (2016). NRS, Numeric Rating Scale; CPOT, Critical-Care Pain Observation Tool.

pain scores move in the same direction (i.e., when one score The study of pain in ICU patients with a brain injury
increases, the other score increases as well) (Herr et al., from a traumatic and a non-traumatic source has received
2011; Glinas et al., 2013), but are not equal scores. More- some attention from research teams around the world.
over, while it is useful to classify the intensity of pain as mild, In a rst Canadian study by Glinas and Arbour (2009),
moderate and severe for the development of pain manage- unconscious ICU patients with a TBI (n = 43) exhibited dif-
ment protocols, it is important to know that behavioural ferent behaviours such as a relaxed face (70%), eye weeping
pain scales only allow the detection of the presence versus (14%), eye opening (16%) and relaxed muscles (72%) dur-
absence of pain. Indeed, in most studies, this criterion of ing turning. Not surprisingly, their CPOT score was lower
presence versus absence of pain was used to establish BPS (mean = 2.05) in comparison with surgical ICU patients
(Chen et al., 2014) and CPOT (Echegaray-Benites et al., (mean = 3.30) (p < 0.001). Similar ndings were found in
2014; Glinas and Johnston, 2007; Li et al., 2014) cut-off a study by a research team from Iran (Dehghani et al.,
scores. However, in a previous study with postoperative 2014) who found higher BPS scores during endotracheal
ICU patients, it was found that the CPOT cut-off score >2 suctioning (mean = 7.75) compared with a non-nociceptive
could better classify patients who self-reported moderate procedure (i.e., eye care) (mean = 3.28) in 50 critically ill
to severe pain (Glinas et al., 2009). Knowing this, opi- TBI patients (p < 0.001). Unfortunately, the authors did not
oid and/or non-opioid analgesic agents could be trialled provide details regarding item scores so it is difcult to
when cut-off scores of behavioural pain scales are reached. draw conclusions on their specic behavioural responses
An intervention algorithm with the CPOT was created to to the nociceptive procedure. Description of behavioural
help guide the ICU care team in their decision for pain responses in critically ill TBI patients (n = 45) was further
management (Fig. 2), and its implementation still needs to explored by the same Canadian research group (Arbour
be trialled in clinical practice. Nevertheless, pain manage- et al., 2014a). During turning, a high proportion of TBI
ment protocols must be adapted to take into account that patients showed neutral behaviours (i.e., relaxed face
behavioural pain scores cannot discriminate between mild, 48.9%, absence of body movements 44.4%, and relaxed mus-
moderate and severe levels of pain. cles 88.9%) and this was inuenced by their LOC. In fact,
87.5% of unconscious TBI patients (n = 8) were more likely
Can behavioural pain scales be used in any to exhibit a relaxed face and to remain immobile. How-
critically ill adult patient population? ever, signicant proportions of altered LOC and conscious
TBI patients exhibited pain-related behaviours or autonomic
Although behavioural pain scales have been developed for responses including moaning (>50%), brow lowering (>38%),
use in critically ill adults, some challenges remain in spe- face ushing (>33%), eye opening (>25%), eye weeping
cic patient groups. These include ICU patients with a brain (>25%) and limb exion (>23%). A positive high correlation
injury, burns, delirium and a cognitive decit. Evidence and (rs = 0.82; p < 0.001) was found between the number of pain-
challenges with the use of behavioural pain scales in these related behaviours/autonomic responses and the patients
groups are discussed. self-reported 010 pain intensity. Only four conscious TBI
Pain assessment in the critically ill adult 5

patients (n = 16; 25%) exhibited a full grimace and tried (median = 0; range 06) (p < 0.001). Patients who reported
to reach their pain site during turning. Interestingly, in pain during turning had higher CPOT scores than those
those able to self-report their pain (n = 13/16), nine of them who reported no pain (p < 0.001), and a moderate positive
reported pain during turning but only one exhibited a grim- correlation was obtained with their 010 NRS self-report
ace and muscle rigidity; such ndings are opposite to the (rs = 0.57; p < 0.001). A CPOT cut-off score 2 showed sensi-
Thunder Project in which these two behaviours were the tivity and specicity ndings of 76.9% and 73.3% respectively
most frequent in ICU patients who reported pain (Puntillo and the ability to discriminate (Area Under the Curve or AUC)
et al., 2004). between those with or without pain was high at 86.4%.
To our knowledge, only three studies addressed pain- Only one American study was conducted in a burn pop-
related behaviours in ICU patients with a non-traumatic ulation (Wibbenmeyer et al., 2011) and evaluated two pain
brain injury. In a rst study by a research team from scales including the CPOT (Glinas et al., 2006) and the Non-
Korea (Lee et al., 2013), pain was assessed at rest and Verbal Pain Scale (NVPS: Odhner et al., 2003). The NVPS
during endotracheal suctioning using the CPOT on ve spe- includes three behavioural items (i.e., facial expression,
cic days (i.e., 1, 3, 6, 9 and 14) after ICU admission activity and muscle guarding) and two physiologic ones in
in 31 brain-injured patients most of them with a cere- relation to changes in vital signs (i.e., blood pressure and
bral haemorrhage and two with a brain tumour. None had heart rate) and others (i.e., skin temperature, diaphoresis,
a GCS of 1315 and patients were unable to self-report. dilated pupils and ushing) for a possible total score ranging
Mean CPOT scores decreased over time, and were higher from 0 to 10 (Odhner et al., 2003). A total of 225 paired
in ICU patients who did not undergo brain surgery (n = 9) assessments were performed in 38 burn patients at rest,
after including analgesic use as a covariate which was before and after daily activities (either wound cleansing or
common in the surgery group (n = 22). Consistent with pre- physical/occupational therapy). Flame or ash burn was the
vious validation studies, mean CPOT scores were higher most common aetiology (76.3%) and the average burn size
ranging from 2.58 to 3.26 during endotracheal suctioning. was 10.4% (SD = 12.1%) of body surface area. CPOT and NVPS
Therefore, the CPOT cut-off score of >2 for the presence scores were shown to discriminate between pain at rest and
of pain was reached during a painful procedure in this during activity but remained low with mean scores <1. In this
patient group. In another study by a research team from sample, patients were able to self-report and mild positive
Switzerland (Roulin and Ramelet, 2014), pain behaviours correlations were obtained between behavioural scores and
were described in 116 ICU patients with a non-traumatic the 010 NRS scores with Pearson coefcients of 0.36 (for
brain injury (i.e., 66% with intracranial haemorrhage, 30.2% CPOT) and 0.38 (for NVPS) (p < 0.01). It is worth mentioning
with anoxic or ischaemic origin and others). Patients were that means of 4.27 and 5.41 of self-reported 010 NRS scores
clustered into three groups according to their LOC: (a) were described at rest and during activity respectively show-
stereotyped responses i.e., limb exion/extension or no ing moderate perceived pain in this sample. Again, the item
response to a nociceptive stimulus (n = 37); (b) localised scores were not presented making it impossible to iden-
responses to a nociceptive stimulus (n = 33); and (c) those tify the behavioural responses exhibited by these patients.
able to self-report and to follow commands (n = 46). Some Moreover, the body area was not specied which would have
differences in pain behaviours were observed during turning been useful in better understanding the pain scores. For
across the three groups. For instance, brow lowering, eye example, we could expect that patients with burns located
closure and touching the pain site were observed more often on the face would have altered facial expressions in relation
in the self-report group (63%, 73% and 29% respectively) than with pain.
in the localised group (29%, 65% and 7%) and the stereo- Finally, behavioural pain scales were validated in deliri-
typed group (51%, 35% and 4%) (p < 0.05). Interestingly, face ous ICU patients in only two studies. A research team
ushing was more frequently identied in the stereotyped from France validated the use of the BPS-NI in 30 med-
group (43%) than in the localised (19%) and the self-report ical or surgical ICU patients with 84% being positive for
groups (21%) (p < 0.05). Opposite to what was found in TBI delirium (Chanques et al., 2009) as per the Confusion Assess-
patients (Arbour et al., 2014a; Glinas and Arbour, 2009), ment Method (CAM)-ICU (Ely et al., 2001). Patients were
muscle rigidity was observed in similar proportions (3248%) assessed for pain during a non-nociceptive procedure (i.e.,
in all three groups. Overall, median number of behaviours catheter dressing change) and a nociceptive procedure (i.e.,
was higher during turning than at rest (p < 0.01). In addition, turning). Higher BPS-NI scores were found during turning
a moderate positive correlation was found between number (median = 6) compared with rest (median = 3) and the non-
of behaviours and the patients self-report of pain inten- nociceptive procedure (median = 3) (p < 0.001). The effect
sity using the 010 NRS (rs = 0.53) in the self-report group size for responsiveness was large for the three items of the
(n = 30/46) during turning. However, the grimace which is a BPS-NI with 3.64 for vocalisation, 2.82 for facial expression,
key facial expression of pain was not described. 1.47 for upper limb movements and 3.46 for the total BPS-
The last study conducted in non-traumatic brain injury NI scores. Recently, the use of the CPOT was validated in 40
ICU patients was led by a Canadian team (Echegaray-Benites delirious ICU patients (i.e., as per screening with CAM-ICU)
et al., 2014) with elective brain surgery ICU patients (n = 43) by a Canadian research team (Kanji et al., 2016). Simi-
who underwent either a craniotomy (79%) or a craniec- larly, higher CPOT scores during painful procedures (i.e.,
tomy (21%). Pain was assessed with the CPOT before, during turning, endotracheal suctioning or dressing change) com-
and after two procedures: (a) non-nociceptive: non-invasive pared with baseline and a non-painful procedure (mean
blood pressure (NIBP) and (b) nociceptive: turning. CPOT difference = 3.13; p < 0.001, and effect size Cohen D = 2.0).
scores were higher during turning (median = 2; range 06) These study ndings suggest that the BPS-NI and the CPOT
in comparison with rest (median = 0; range 03) and NIBP are valid tools for the assessment of pain in ICU patients with
6 C. Glinas

delirium; however, further testing by other research teams support tools on the evaluation of pain were provided to
is required to support such conclusions. ICU nurses, physicians and residents; pocket cards were
Assessment tools can only be shown to be valid for a spe- distributed to nurses, and posters were placed in every
cic purpose, in a determined group of respondents and in a patients room. After patients were assessed for pain with
given context (Streiner et al., 2014). As specied in the rec- the NRS or the BPS by ICU nurses, physicians were noti-
ommendation related to BPS and CPOT in the PAD guidelines ed of a pain event (NRS > 3 or BPS > 5) or an agitation
(Barr et al., 2013), their use should be considered with cau- event (RASS > 2) and could intervene accordingly. Decreases
tion in brain-injured ICU patients. Indeed, recent evidence in mechanical ventilation duration and nosocomial infection
has shown that ICU patients with a brain injury affecting rate were demonstrated in the intervention group (p < 0.05)
their LOC appear to have different behavioural reactions to but no differences in ICU length of stay and mortality
pain, such as the absence of grimace and muscle rigidity in were found. Another research team from Australia imple-
most of them. Therefore, existing behavioural pain scales mented the BPS along with RASS in a general ICU (Williams
may not be applicable to those patients and further adapta- et al., 2008) using a pre/post-test study design. Educa-
tion of these scales may be necessary. Insufcient evidence tion on the tool use was provided to all staff. Outcomes
is available in burn patients (Wibbenmeyer et al., 2011). were documented in 369 and 400 mechanically ventilated
However, we may anticipate that burns on the face and limbs patients before and after the tool implementation, respec-
may affect facial expressions and limb movements which tively. The proportion of ICU patients who received sedatives
would necessitate adaptation of existing pain scales. The with or without analgesics was greater after tool implemen-
BPS-NI and CPOT appear to be valid for use in ICU patients tation (88%) than before (57%) (p < 0.001). No difference
with delirium but this has to be replicated by other research in duration of mechanical ventilation was found between
teams to support external validity. Yet no evidence on the the two groups (Williams et al., 2008). In a pre/post
validity of the use of behavioural pain scales in ICU patients prospective study from a German research team (Radtke
with cognitive decit is available and research is denitely et al., 2012), the NRS and BPS were also implemented
needed. along with a sedation (RASS: Sessler et al., 2002) and
delirium (Delirium Detection Score: Otter et al., 2005)
scales in three ICUs. A total of 619 ICU patients were
Impact of the implementation of pain
included i.e., 241 in the pre-training period, 228 in the
assessment tools on ICU practices and patient post-training period and 150 in the follow-up period. This
outcomes latter period representing the extended training (with three
consecutive educational sessions, and an ICU support team
A recent systematic review described the impact of pain available during the implementation phase) was compared
assessment on critically ill patients outcomes (Georgiou with standard training (one educational session of 45-minute
et al., 2015). A total of 10 eligible studies were identied. duration) in the post-training period and was found to
Overall, there was evidence of positive effects on the detec- lead to more signicant changes in ICU practices (i.e.,
tion and management of pain, and on patient outcomes more frequent documentations of pain assessments) and
including duration of mechanical ventilation, ICU length of to their sustainability at 1-year follow-up (p < 0.01). Tool
stay, adverse events and mortality. However, an observa- implementation had no impact on mechanical ventilation
tional pre-experimental design was used in all studies, and a duration and ICU length of stay. Pain monitoring was asso-
higher level of evidence research (i.e., experimental design) ciated with a decrease in mortality (OR = 0.35 with 95% CI:
is necessary to draw rm conclusions. Some key studies for 0.140.86).
which methodological quality was evaluated as moderate More recently, in a quality improvement project led by
to strong in the systematic review were selected for fur- Chanques team, systematic pain assessments with NRS or
ther discussion of implementation strategies and outcomes. BPS were completed by ICU nurses during the rst turning of
These studies targeted the implementation of systematic the day (De Jong et al., 2013), and appropriate pain man-
pain assessment approaches using the BPS, CPOT and NRS. agement interventions were provided in collaboration with
Although its methodology quality was considered weak by physicians. Similar educational strategies (i.e., training in
Georgiou et al. (2015), the implementation study of the small group sessions, posters) from the rst study (Chanques
NVPS by Topolovec-Vranic et al. (2010) is also discussed as et al., 2006) were used, and the clinical information system
it described patient satisfaction which was rarely addressed software was updated. A total of 630 care procedures were
previously. analysed in 193 ICU patients i.e., 53 at baseline (n = 184 pro-
The rst study to evaluate the impact of a system- cedures), 90 during implementation (n = 299 procedures) and
atic pain and sedation assessment approach was conducted 50 post-implementation (n = 149 procedures). The incidence
by a research team from France led by Chanques et al. of severe pain as dened by NRS > 6 or BPS > 5 decreased
(2006) who implemented the NRS and BPS (Payen et al., signicantly from 16% at baseline to 6% during the imple-
2001) along with RASS (Sessler et al., 2002) in a medico- mentation phase, and 2% post-implementation (p < 0.05).
surgical ICU. Using a pre/post prospective controlled Incidence of at least one severe adverse event (i.e., cardiac
design (n = 100 in control group and n = 130 in interven- arrest, arrhythmias, tachycardia, bradycardia, hyperten-
tion group), they documented decreased incidence of pain sion, hypotension, desaturation, bradypnoea or ventilator
(63% versus 42%, p = 0.002) and agitation events (29% versus distress) also signicantly decreased from 37% at baseline to
12%, p = 0.002), and increased therapeutic changes (esca- 17% during implementation, and 21% post-implementation
lation and de-escalation) of analgesics and psychoactive (p < 0.05). Patients who had severe pain were more likely to
drugs post-implementation. Verbal information and written experience a severe adverse event (OR = 2.74 (1.54; 4.89),
Pain assessment in the critically ill adult 7

p < 0.001). The administration of at least one analgesic drug before and 32 (82%) patients after tool implementation had
was higher at post-implementation compared with baseline recollection of their ICU stay and completed the patient
(33% versus 53%, p < 0.01). satisfaction survey. Patients reported decreased retrospec-
Regarding the implementation of the CPOT (Glinas tive pain ratings (8.5 pre versus 7.2 post, p = 0.04) and
et al., 2006), two before-after prospective studies led by were satised with pain control before and after the pain
two Canadian research teams were published (Glinas et al., tool implementation. Fifty-three surveys (89% of all dis-
2011a; Rose et al., 2013). In the rst study in a medi- tributed) and 32 (50%) were completed by ICU nurses pre
cal/surgical/trauma ICU, a 90-minute standardised training and post-implementation of NVPS, respectively. Most nurses
for the use of the CPOT which also included practice with (78%) ranked the tool as easy to use. Implementation of
patient videos, and the creation of a support ICU champion the tool increased nurses condence in assessing pain
team to provide feedback and answer ICU nurses ques- in nonverbal, sedated patients (57% pre versus 81% post-
tions were used (Glinas et al., 2011a). The CPOT was implementation, p = 0.02). However, no differences were
also incorporated into the ICU nursing ow sheet prior to found in nurses condence in managing patients pain
conducting the study. Findings showed an increase in the fre- before and after the tool implementation. The main barriers
quency of documented pain assessments per patient per day described by nurses included the physicians pain man-
between pre and post-implementation groups (n = 30/group) agement practices, personal beliefs and attitudes towards
(median of 3 at pre, and 10.5 at 3-month post) which was pain, and characteristics that limit the patients ability to
maintained at 12-month post (median of 12). Pain reassess- self-report.
ments following the administration of an analgesic and/or Although these study results are promising, further
sedative agents were also higher post-implementation of research is clearly needed to evaluate the uptake of an
the CPOT (10% at pre versus 43% and 59% at 3 and 12- ICU pain management initiative by the inter-professional
month post respectively), and a decrease in the use of team, and to measure its impact on sustainability of practice
sedative agents was also observed (p < 0.05). At 12 month changes and patients outcomes. Indeed, earlier studies pri-
post-implementation, 92100% of ICU nurses (n = 38) agreed marily focused on the effects of clinicians training on the
that the CPOT was quick and easy to use, and 87% men- use of pain assessment tools. Whether this really translates
tioned that it helped them to adequately evaluate pain. into the day-to-day use of such tools by the ICU inter-
However, less than 50% of ICU nurses mentioned that the professional team, changes in clinical practices that are
tool helped them to effectively communicate pain assess- sustained over time, improved inter-professional collabora-
ment results with physicians, and this was identied as tion and better patient outcomes needs to be investigated
an area for improvement (Glinas et al., 2014). Regard- more closely. In addition, very little is known about the
ing clinical outcomes, only a signicant decrease in the impact of such initiatives on the patients experience of pain
number of complications was found in a small subsam- in the ICU and many initiatives were led by nursing teams
ple of trauma ICU patients (n = 15) at post-implementation only (Glinas et al., 2011a; Rose et al., 2013; Topolovec-
(p < 0.05) (Arbour et al., 2011). Using the same standard- Vranic et al., 2010). Involving the ICU inter-professional
ised training (Glinas et al., 2011a) in addition to other team in the implementation process is necessary to ensure
strategies (i.e., incorporating the CPOT in protocols, bed- their collaboration and support in the pain management
side and web portal tools), another study was conducted process. Indeed, inter-professional collaboration is a key
by Rose and colleagues in two ICUs (i.e. cardiovascular component in health care delivery and can lead to better
and medical/surgical/trauma) with 189 and 184 patients patient outcomes (Rose, 2011). Finally, pain and sedation
before and after CPOT implementation, respectively. Sim- assessments should be performed concomitantly as it has
ilar ndings were found in both settings with frequencies been found to guide clinicians in their decision making pro-
of documented pain assessments being four times higher at cess and to lead to better use of analgesic and sedative
post-implementation than at pre (p < 0.001). Interestingly, agents (Chanques et al., 2006).
decreases in the use of analgesic (5 mg to 4 mg; p = 0.02) and
sedative (12 mg to 2 mg; p < 0.001) medication were found
in the cardiovascular ICU. However, mixed ndings were Caution with vital signs
obtained in the two ICUs according to length of stay and
mechanical ventilation duration. In comparison to pain related-behaviours, vital signs have
In another prospective pre/post study, the use of the received less attention in research in critically ill adults,
Non-Verbal Pain Scale (NVPS: Odhner et al., 2003) was imple- and inconsistent ndings were reported. In most studies,
mented in a trauma/neurosurgery ICU (Topolovec-Vranic heart rate (HR) or blood pressure (BP) indicators were found
et al., 2010). In-service small group training sessions of to increase when ICU patients were exposed to painful
1520 minutes were offered by the study investigator who procedures such as endotracheal suctioning, mediastinal
reached 90% of the ICU nurses. Pocket cards were dis- tube removal, turning and wound care (Aissaoui et al.,
tributed to nurses, and a poster was placed in the ICU 2005; Arbour and Glinas, 2010; Arroyo-Novoa et al., 2008;
for staff reference. Nurses documented pain assessments Boitor et al., 2015; Chanques et al., 2009; Chen and Chen,
on a NVPS documentation tool created for this study. Sim- 2015; Glinas and Arbour, 2009; Glinas and Johnston, 2007;
ilar to other studies, the number of documented pain Kapoustina et al., 2014; Li et al., 2009; Payen et al., 2001;
assessments increased per day in the ICU after the imple- Stotts et al., 2004). However, these increases (<20%) were
mentation of the tool (2.2 pre versus 3.4 post, p = 0.02). not considered to be clinically signicant by the authors.
A total of 64 patients (25 pre and 39 post-implementation Conversely, these vital signs could either be found to uctu-
of NVPS) were approached and of these, 20 (80%) patients ate both during painful and non-painful procedures (Arbour
8 C. Glinas

et al., 2014b; Young et al., 2006) or to remain stable during index can only be utilised in sedated patients and it would
painful ones (Glinas et al., 2011b; Sifeet et al., 2007). In not be applicable to patients who are awake, conscious and
other studies, respiratory rate (RR) (Chanques et al., 2009; unable to self-report.
Kapoustina et al., 2014) and end-tidal CO2 were found to Regarding pupillary reexes, pupil size signicantly
increase during a painful procedure (Arbour and Glinas, increased by 16% during a noxious procedure (i.e., endotra-
2010; Glinas and Arbour, 2009), while SpO2 (oxygen sat- cheal suctioning or repositioning) and returned to baseline
uration) was found to decrease (Boitor et al., 2015; Glinas ve minutes after the procedure (p < 0.001) in 48 sedated
and Johnston, 2007). Except for associations found between and mechanically ventilated cardiac surgery ICU patients (Li
vital signs (i.e., HR, RR and SpO2 ) and the self-report of et al., 2009). Interestingly, in a sample of 100 postopera-
pain in cardiac surgery ICU patients (Arbour and Glinas, tive patients who awoke from general anaesthesia (Aissou
2010) and critically ill TBI patients (Arbour et al., 2014b); et al., 2012), a high positive correlation of 0.88 (p < 0.001)
vital signs failed to be related to the patients self-report of was found between pain scores (5-item verbal rating scale
pain (Arbour and Glinas, 2010; Boitor et al., 2015; Chen and or VRS) and pupil dilatation reex (PDR). In patients (n = 39)
Chen, 2015; Glinas and Arbour, 2009; Glinas and Johnston, who reported a VRS > 1, PDR before and after morphine titra-
2007; Kapoustina et al., 2014). Such ndings are not surpris- tion was 35% and 12% respectively (p < 0.001). Moreover, the
ing as many other sources of distress besides pain can make PDR threshold value of 23% was associated with a sensitiv-
them uctuate (Herr et al., 2011). ity of 91% and a specicity of 94% (Aissou et al., 2012). In a
However, as described in a wide Canadian survey, most recent study by Lukaszewicz et al. (2015), a percentage of
ICU nurses (n = 733/796, 92%) consider vital signs moderately pupil size variation >19% was found to predict the presence
to extremely important for the assessment of pain (Rose of pain by a BPS score >3 with a sensitivity of 100% and a
et al., 2012). Nurses need to be more aware of the evidence specicity of 77% before surgical dressing changes in 37 ICU
showing the lack of validity related to the use of vital signs patients with cellulitis. Similar ndings were obtained in a
in the ICU pain assessment process. Indeed, the recommen- previous study by Paulus et al. (2014) in which a threshold
dation in both ASPMN (Herr et al., 2011) and SCCM (Barr value of PDR >5% during a 20 mA tetanic stimulation could
et al., 2013) calls for caution with their use, and that they predict with an AUC of 0.78 or 78% (95%CI: 0.610.91) insuf-
should only be used as cues for further assessment of pain cient analgesia (dened by an increase 1 point score on
with appropriate validated tools i.e., self-report measures the BPS) before endotracheal suctioning in 34 sedated ICU
or behavioural pain scales. patients. Overall, these results suggest that pupillometry
may potentially guide clinicians to adjust analgesia before
nociceptive procedures in critically ill patients.
New trends in pain-related indicators
Future avenues in research
Besides vital signs, other indicators have been studied in
relation to the detection of pain in the ICU. From those,
Behavioural pain scales have undergone numerous validation
the BISpectral Index (BIS), and pupillary reexes have been
studies, but adaptation in their content is required in brain-
examined. These indicators were mainly studied in the con-
injured or burn ICU patient populations. Also, not much
text of anaesthesia or sedation, and their interest of use
is known on the validity of the use of these scales in ICU
in the ICU pain assessment process is relatively new. The
patients with delirium or cognitive decit. Although an
BIS index is a single number computed from complex algo-
effort was made to include items relevant to the patients
rithmic equation based on the electroencephalogram (EEG)
condition (i.e., those who are mechanically ventilated or
data obtained from healthy subjects undergoing general
those able to vocalise), those items are not representative
anaesthesia (Johansen and Sebel, 2000). The BIS index value
of patients with a tracheostomy who are not on a venti-
can vary from 0 (complete EEG suppression) to 100 (fully
lator and unable to vocalise (Bambi and Solaro, 2012). An
awake) and its primary purpose is for the titration of anaes-
alternative item could be considered for this patient sub-
thetic agents during surgery. In three studies with small
group. There is also need for valid physiologic measures of
samples of medical, surgical and trauma ICU patients (over-
pain especially in ICU patients too heavily sedated or paral-
all n of 82 in 3 studies), the BIS index value was found to
ysed as behavioural responses cannot be observed in such
signicantly increase (>5%) during nociceptive procedures
situations. Research on ICU proxy reporters of pain is scarce
such endotracheal suctioning and turning (Arbour et al.,
and experimental design is needed to evaluate the effects
2015; Glinas et al., 2011b; Li et al., 2009). Interestingly
of the implementation of pain assessment and management
in Arbour and colleagues study (2015), the bilateral BIS
approaches on patient outcomes, ICU practices and their
monitor was utilised in 25 critically ill TBI patients, and the
sustainability over time.
increases in the BIS index values were contralateral to the
brain lesion. More specically, increases in BIS-Right were
more pronounced in left-sided TBI patients than those with Conclusion
right-sided injury. Also, BIS-Right uctuations in left-sided
TBI patients were highly positively correlated (rs = 0.99, Despite major discoveries in relation to pain assessment in
p < 0.001) with the number of pain behaviours observed dur- the critically ill adult in the last three decades, pain remains
ing turning. Although these ndings appear promising, the a problem. Pain practice guidelines exist (Barr et al.,
primary purpose of the BIS index is not to measure pain, and 2013; Herr et al., 2011), and their implementation into ICU
further research in larger samples is needed to conrm its practice is urgently needed. Inter-professional ICU teams
utility in the ICU pain assessment process. Moreover, the BIS must adjust their pain management protocols to available
Pain assessment in the critically ill adult 9

pain measures (i.e., self-report scores versus behavioural self-report: an adaptation of the Behavioral Pain Scale. Intensive
scores) so they adequately represent score interpretation Care Med 2009;35(12):20607.
of pain levels or presence versus absence of pain. Fam- Chanques G, Sebbane M, Barbotte E, Viel E, Eledjam JJ, Jaber
ily caregivers should be systematically consulted to better S. A prospective study of pain at rest: incidence and charac-
understand the patients behavioural responses to pain (Herr teristics of an unrecognized symptom in surgical and trauma
versus medical intensive care unit patients. Anesthesiology
et al., 2011). A change in nursing ICU practice in relation to
2007;107(5):85860.
the use of vital signs in the ICU pain assessment process Chanques G, Viel E, Constantin JM, Jung B, de Lattre S, Carr J, et al.
must happen. Fluctuations in vital signs should be consid- The measurement of pain in intensive care unit: comparison of
ered as adverse events to severe pain (De Jong et al., 2013) 5 self-report intensity scales. Pain 2010;151(3):71121.
rather than indicators for pain assessment as they are not Chen HJ, Chen YM. Pain assessment: validation of the physio-
valid for this purpose (Barr et al., 2013; Herr et al., 2011). logic indicators in the ventilated adult patient. Pain Manag Nurs
Finally, pain assessment and management is a collaborative 2015;16(2):10511.
effort and pain management initiatives must involve the ICU Chen YY, Lai YH, Shun SC, Chi NH, Tsai PS, Liao YM. The Chinese
inter-professional team for better outcomes. Behavior Pain Scale for critically ill patients: translation and
psychometric testing. Int J Nurs Stud 2014;48:43848.
Dehghani H, Tavangar H, Ghandehari A. Validity and reliability
of behavioral pain scale in patients with low level of con-
References sciousness due to head trauma hospitalized in intensive care
unit. Arch Trauma Res 2014;3(1):e18608, http://dx.doi.org/
Aissaoui Y, Zeggwagh AA, Zekraoui A, Abidi K, Abouqal R. Validation 10.5812/atr.18608, 4 p. Open Access.
of a behavioral pain scale in critically ill, sedated, and mechan- De Jong A, Molinari N, De Lattre S, Gniadek C, Carr J, Conseil M,
ically ventilated patients. Anesth Analg 2005;101(5):14706. et al. Decreasing severe pain and serious adverse events while
Aissou M, Snauwaert A, Dupuis C, Atchabahian A, Aubrun F, Beaussier moving intensive care unit patients: a prospective interventional
M. Objective assessment of the immediate postoperative anal- study (the NURSE-DO project). Crit Care 2013;17(2):R74, Open
gesia using pupillary reex measurement: a prospective and Access http://ccforum.com/content/17/2/R74.
observational study. Anesthesiology 2012;116(5):100612. Echegaray-Benites C, Kapoustina O, Glinas C. Validation of the
Arbour C, Choinire M, Topolovec-Vranic J, Loiselle C, Puntillo K, Critical-Care Pain Observation Tool in brain surgery patients dur-
Glinas C. Detecting pain in traumatic brain injured patients ing common procedures in the Intensive Care Unit. Intensive Crit
exposed to common procedures in the ICU: typical or atypical Care Nurs 2014;30(5):25765.
behaviors? Clin J Pain 2014a;30(11):9609. Ely EW, Margolin R, Francis J, May L, Truman B, Dittus R, et al.
Arbour C, Choinire M, Topolovec-Vranic J, Loiselle C, Glinas C. Can Evaluation of delirium in critically ill patients: validation of the
vital signs uctuations be used for pain assessment in nonverbal Confusion Assessment Method for the Intensive Care Unit (CAM-
ICU patients with a traumatic brain injury? Pain Res Treat- ICU). Crit Care Med 2001;29(7):13709.
ment 2014b., http://dx.doi.org/10.1155/2014/175794, Article Glinas C. Management of pain in cardiac surgery ICU patients:
ID 175794, 11 p. Open Access. have we improved over time? Intensive Crit Care Nurs
Arbour C, Glinas C. Are vital signs valid indicators for the assess- 2007a;23:298303.
ment of pain in postoperative cardiac surgery ICU adults? Glinas C. Le thermomtre dintensit de douleur: un nouvel outil
Intensive Crit Care Nurs 2010;26(2):8390. pour les patients adultes en soins critiques [The Faces Pain Ther-
Arbour C, Glinas C, Loiselle C, Bourgault P. An exploratory study mometer: a new tool for critically ill adults]. Perspect Inrmire
of the bilateral bispectral index for pain detection in trau- 2007b;4(4):1220.
matic brain-injured-patients with altered level of consciousness. Glinas C, Arbour C. Behavioral and physiologic indicators dur-
J Neurosci Nurs 2015;47(3):16677. ing a nociceptive procedure in conscious and unconscious
Arbour C, Glinas C, Michaud C. Impact of the implementation mechanically ventilated adults: similar or different? J Crit Care
of the CPOT on pain management and clinical outcomes in 2009;24(4):628e717.
ventilated trauma ICU patients: a pilot study. J Trauma Nurs Glinas C, Arbour C, Michaud C, Vaillant F, Desjardins S. The imple-
2011;18(1):5260. mentation of the critical-care pain observation tool on pain
Arroyo-Novoa CM, Figueroa-Ramos MI, Puntillo KA, Stanik-Hutt J, assessment/management nursing practices in an intensive care
Thompson CL, White C, et al. Pain related to tracheal suctioning unit with nonverbal critically ill adults: a before and after study.
in awake acutely and critically ill adults: a descriptive study. Int J Nurs Stud 2011a;48:1495504.
Intensive Crit Care Nurs 2008;24(1):207. Glinas C, Tousignant-Laamme Y, Robitaille A, Bourgault P. Explor-
Bambi S, Solaro M. CPOT: is there a missing link? (Letter to the ing the validity of the Bispectral Index, the Critical-Care Pain
editor). Pain Manag Nurs 2012;13(1):67. Observation Tool and vital signs for the detection of pain in
Barr J, Fraser GL, Puntillo KA, Ely EW, Glinas C, Dasta JF, sedated and mechanically ventilated critically ill adults: a pilot
et al. Clinical practice guidelines for the management of pain, study. Intensive Crit Care Nurs 2011b;27(1):4652.
agitation, and delirium in adult ICU patients. Crit Care Med Glinas C, Johnston C. Pain assessment in the critically ill
2013;41(1):263306. ventilated adult: validation of the Critical-Care Pain Observa-
Boitor M, Lachance JF, Glinas C. Validation of the Critical- tion Tool and physiologic indicators. Clin J Pain 2007;23(6):
Care Pain Observation Tool and vital signs in relation to the 497505.
sensory and affective components of pain during mediasti- Glinas C, Fillion L, Puntillo K, Viens C, Fortier M. Validation of the
nal tube removal in postoperative cardiac surgery intensive Critical-Care Pain Observation Tool in adult patients. Am J Crit
care unit adults. J Cardiovasc Nurs 2015., http://dx.doi.org/ Care 2006;15(4):4207.
10.1097/JCN.0000000000000250 [in press]. Glinas C, Harel F, Fillion L, Puntillo KA, Johnston C. Sensitivity
Chanques G, Jaber S, Barbotte E, Violet S, Sebbane M, Perrigault JF, and specicity of the Critical-Care Pain Observation Tool for the
et al. Impact of systematic evaluation of pain and agitationin an detection of pain in intubated adults after cardiac surgery. J Pain
intensive care unit. Crit Care Med 2006;34:16919. Symptom Manag 2009;37(1):5867.
Chanques G, Payen J-F, Mercier G, de Lattre S, Viel E, Jung B, et al. Glinas C, Puntillo KA, Joffe A, Barr JA. A validated approach to
Assessing pain in non-intubated critically ill patients unable to evaluating psychometric properties of pain assessment tools for
10 C. Glinas

use in non-verbal critically ill adults. Sem Respir Crit Care Med Odhner M, Wegman D, Freeland N, Steinmetz A, Ingersoll GL.
2013;34(2):15368. Assessing pain control in nonverbal critically ill adults. DCCN
Glinas C, Ross M, Boitor M, Desjardins S, Vaillant F, Michaud Dimens Crit Care Nurs 2003;22(6):2607.
C. Nurses evaluations of the CPOT use at 12-month post- Otter H, Martin J, Basell K, van Heymann C, Hein OV, Bollert P, et al.
implementation in the intensive care unit. Nurs Crit Care Validity and reliability of the DDS for severity of delirium in the
2014;19(6):27280. ICU. Neurocrit Care 2005;2:1508.
Georgiou E, Hadjibalassi M, Lambrinou E, Andreou P, Papathanas- Pasero C, Puntillo K, Li D, Mularski RA, Grap MJ, Erstad BL, et al.
soglou EDE. The impact of pain assessment on critically ill Structured approaches to pain management in the ICU. Chest
patients outcomes: a systematic review. Biomed Res Int 2015., 2009;135(6):166572.
http://dx.doi.org/10.1155/2015/503830, Article ID 503830, 18 Paulus J, Ruquilly A, Beloeil H, Thraud J, Asehnoune K, Lejus
p. Open Access. C. Pupillary reex measurement predicts insufcient analgesia
Hadjistavropoulos T, Craig KD. A theoretical framework for under- before endotracheal suctioning in critically ill patients. Crit Care
standing self-report and observational measures of pain: a com- 2014;17(4):R161, Open Access http://ccforum.com/content/
munication model. Behaviour Res Therapy 2002;40(5):55170. 17/4/R161.
Hadjistavropoulos T, Craig KD, Duck S, Cano A, Goubert L, Jackson Payen JF, Chanques G, Mantz J, Hercule C, Auriant I, Leguillou JL,
PL, et al. A biopsychosocial formulation of pain communication. et al. Current practices in sedation and analgesia for mechani-
Psychol Bull 2011;137(6):91039. cally ventilated critically ill patients: a prospective multicenter
Herr K. Pain assessment strategies in older patients. J Pain patient-based study. Anesthesiology 2007;106(4):68795.
2011;12(3):S313. Payen JF, Bosson JL, Chanques G, Mantz J, Labarere J. Pain
Herr K, Coyne PJ, McCaffery M, Manworren R, Merkel S. Pain assessment is associated with decreased duration of mechani-
assessment in the patient unable to self-report: position state- cal ventilation in the intensive care unit: a post Hoc analysis of
ment with clinical practice recommendations. Pain Manag Nurs the DOLOREA study. Anesthesiology 2009;111(6):130816.
2011;12(4):23050. Payen JF, Bru O, Bosson JL, Lagrasta A, Novel E, Deschaux I, et al.
International Association for the Study of Pain (IASP), Subcommittee Assessing pain in critically ill sedated patients by using a behav-
on Taxonomy. Pain terms: a list with denitions and notes on ioral pain scale. Crit Care Med 2001;29(12):225863.
usage. Pain 1979;6:24952. Pudas-Thk SM, Axelin A, Aantaa R, Lund V, Salanter S.
IASP Task Force on Taxonomy Part III: Pain Terms. A current list with Pain assessment tools for unconscious or sedated intensive
denitions and notes on usage. In: Merskey H, Bogduk N, editors. care patients: a systematic review. J Adv Nurs 2009;65(5):
Classication of chronic pain. 2nd ed. Seattle: IASP Press; 1994. 94656.
p. 20914. Puntillo KA, Max A, Timsit JF, Vignoud L, Chanques G, Robleda G,
Johansen J, Sebel P. Development and clinical application of et al. Determinants of procedural pain intensity in the inten-
electroencephalographic bispectrum monitoring. Anesthesiol- sive care unit. The Europain study. Am J Respir Crit Care Med
ogy 2000;93:133644. 2014;189(1):3947.
Kanji S, MacPhee H, Singh A, Johanson C, Fairbairn J, Lloyd T, et al. Puntillo KA, Morris AB, Thompson CL, Stanik-Hutt J, White CA,
Validation of the Critical-Care Pain Observation Tool in critically Wild LR. Pain behaviors observed during six common procedures:
ill patients with delirium: a prospective cohort study. Crit Care results from Thunder Project II. Crit Care Med 2004;32(2):4217.
Med 2016., http://dx.doi.org/10.1097/CCM.0000000000001522 Puntillo KA, Neuhaus J, Arai S, Paul SM, Gropper MA, Neal H,
[in press]. et al. Challenge of assessing symptoms in seriously ill inten-
Kapoustina O, Echegaray-Benites C, Glinas C. Fluctuations in vital sive care unit patients: can proxy reporters help? Crit Care Med
signs and behavioral responses of brain surgery patients in the 2012;40(10):27607.
intensive care unit: are they valid indicators of pain? J Adv Nurs Puntillo KA, White CA, Morris AB, Perdue ST, Stanik-Hutt J,
2014;70(11):256276. Thompson CL, et al. Patients perceptions and responses to pro-
Kastrup M, von Dossow V, Seeling M, Ahlborn R, Tamarkin A, cedural pain: results from Thunder Project II. Am J Crit Care
Conroy P, et al. Key performance indicators in intensive care 2001;10(4):23851.
medicine. A retrospective matched cohort study. J Int Med Res Radtke FM, Heymann A, Franck M, Maechler F, Drews T, Luetz A,
2009;37(5):126784. et al. How to implement monitoring tools for sedation, pain and
Lee K, Oh H, Suh Y, Seo W. Patterns and clinical correlates delirium in the intensive care unit: an experimental cohort study.
of pain among brain injury patients in critical care assessed Intensive Care Med 2012;38(12):197481.
with the critical-care pain observation tool. Pain Manag Nurs Rose L. Interprofessional collaboration in the ICU: how to dene?
2013;14(4):25967. Nurs Crit Care 2011;16(1):510.
Li D, Miaskowski C, Burkhardt, Puntillo K. Evaluations of physiologic Rose L, Haslam L, Dale C, Knechtel L, McGillion M. Behavioral pain
reactivity and reexive behaviors during noxious procedures in assessment tool for critically ill adults unable to self-report pain.
sedated critically ill patients. J Crit Care 2009;24:472.e913. Am J Crit Care 2013;22(3):24654.
Li Q, Wan X, Gu C, Yu Y, Huang W, Li S, et al. Pain assessment using Rose L, Smith O, Glinas C, Haslam L, Dale C, Knechtel L, et al.
the critical-care pain observation tool in Chinese critically ill Canadian critical care nurses pain assessment and management
ventilated adults. J Pain Symptom Manag 2014;48(5):97582. practices: a national survey. Am J Crit Care 2012;21(4):24755.
Loeser JD, Treede R-D. The Kyoto protocol of IASP basic pain termi- Roulin MJ, Ramelet AS. Behavioral changes in brain-injured critical
nology. Pain 2008;137(3):4737. care adults with different levels of consciousness during noci-
Lukaszewicz AC, Dereu D, Gayat E, Payen D. The relevance of pupil- ceptive stimulation: an observational study. Intensive Care Med
lometry for evaluation of analgesia before noxious procedures in 2014;40:111523.
the intensive care unit. Anest Anal 2015;120(6):1297300. Sessler CN, Gosnell MS, Grap MJ, Brophy GM, ONeal PV, Keane KA,
McGuire DB. Comprehensive and multidimensional assessment and et al. The Richmond Agitation-Sedation Scale: validity and reli-
measurement of pain. J Pain Symptom Manag 1992;7(5):3129. ability in adult intensive care unit patients. Am J Respir Crit Care
Melzack R, Casey KL. Sensory, motivational, and central control Med 2002;166(10):133844.
determinants of pain: a new conceptual model. In: Kenshalo D, Sifeet J, Young J, Nikoletti S, Shaw T. Patients self-report
editor. The skin senses. Springeld: Chas. C. Thomas; 1968. p. of procedural pain in the intensive care unit. J Clin Nurs
42343. 2007;16(11):21428.
Pain assessment in the critically ill adult 11

Stotts NA, Puntillo K, Bonham MA, Stanik-Hutt J, Thompson CL, report. Montreal, QC, Canada: Ingram School of Nursing, McGill
White C, et al. Wound care pain in hospitalized adult patients. University; 2014.
Heart Lung 2004;33(5):32132. Wibbenmeyer L, Sevier A, Liao J, Williams I, Latenser B, Lewis
Streiner D, Norman GR, Cairney J. Health measurement scales: a R, et al. Evaluation of the usefulness of two established
practical guide to their development and use. 5th ed. New York: pain assessment tools in a burn population. J Burn Care Res
Oxford University Press; 2014. 2011;32(1):5260.
Teasdale G, Jennett B. Assessment of coma and impaired conscious- Williams TA, Martin S, Leslie G, Thomas L, Leen T, Tamaliunas S,
ness. A practical scale. Lancet 1974;2(7872):814. et al. Duration of mechanical ventilation in an adult intensive
Topolovec-Vranic J, Canzian S, Innis J, Pollmann-Mudryj MA, McFar- care unit after introduction of sedation and pain scales. Am J
lan AW, Baker AJ. Patient satisfaction and documentation of Crit Care 2008;17(4):34956.
pain assessments and management after implementing the adult Young J, Sifeet J, Nikoletti S, Shaw T. Use of a Behav-
nonverbal pain scale. Am J Crit Care 2010;19(4):34554. ioral Pain Scale to assess pain in ventilated, unconscious
Vanderbyl B, Glinas C. Family perspectives of traumatically brain and/or sedated patients. Intensive Crit Care Nurs 2006;22(1):
injured patient pain behaviors in the intensive care unit. Internal 329.