If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Ingram School of Nursing, McGill University, Montreal, CanadaCentre for Nursing Research and Lady Davis Institute, Jewish General Hospital - CIUSSS West-Central-Montreal, Montreal, Canada
Ingram School of Nursing, McGill University, Montreal, CanadaCentre for Nursing Research and Lady Davis Institute, Jewish General Hospital - CIUSSS West-Central-Montreal, Montreal, Canada
Pain assessment in the intensive care unit (ICU) is challenging because many patients are unable to self-report or exhibit pain-related behaviors. In such situations, vital signs (VS) through continuous monitoring are alternative cues for pain assessment. This review aimed to describe the reliability and validity of VS for ICU pain assessment.
Design
Narrative review of the literature.
Data sources
Medline, Embase, CINAHL, Cochrane.
Review/Analysis methods
A narrative review was conducted with a comprehensive search in four databases. Search terms included VS, pain assessment, and ICU.
Results
Out of 1,359 results, 30 studies from 17 countries were included. Heart rate, blood pressure, and respiratory rate were most used for ICU pain assessment. Assessments were performed at rest before procedures, during nociceptive and non-nociceptive procedures, and after procedures. Increases in respiratory rate were clinically significant by more than 25% during nociceptive procedures (e.g., endotracheal suctioning, turning) compared with rest/pre-procedures in five studies. Correlations of VS with self-reported pain (reference standard measure) and behavioral pain scores (alternative measure) were absent or weak.
Conclusions
VS are not valid indicators for ICU pain assessment. Increases of respiratory rate may be a cue for the detection of pain. However, fluctuations in respiratory rate can be influenced by opioids or controlled ventilation mode. Our results dissuade the use of VS for pain assessment because of the lack of association with ICU pain reference standards. Other physiologic measures of pain in critically ill adults should be explored.
Most, if not all, critically ill adults experience pain during their stay in the intensive care unit (ICU) due to their reasons for admission and standard care procedures (i.e., endotracheal suctioning, catheter insertion/removal, turning/repositioning, etc.) (
). Suboptimal management of pain is associated with negative consequences such as prolonged duration of mechanical ventilation and ICU stay, higher mortality rate, and chronic pain development (
). Pain is described as an unpleasant sensory and emotional personal experience associated with, or resembling that associated with, actual or potential tissue damage (
International Association for the Study of Pain (IASP). (2020). Pain terms: A list with definitions and notes on usage. Retrieved April 16, 2022, from https://www.iasp-pain.org/resources/terminology/.
) which may lead to pain. The consequences of encoding noxious stimuli may be manifested through autonomic (e.g., fluctuations in vital signs) and behavioral responses (e.g., facial expressions) (
International Association for the Study of Pain (IASP). (2020). Pain terms: A list with definitions and notes on usage. Retrieved April 16, 2022, from https://www.iasp-pain.org/resources/terminology/.
). Thus, they may provide cues for pain assessment when communication ability is impaired.
Patient self-reporting is the reference standard measure of pain. However, many patients in the ICU are unable to self-report because of their critical condition, which may alter their level of consciousness (LOC) or require mechanical ventilation and sedation (
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
). In such situations, the “Pain, Agitation, Delirium, Immobility and Sleep disruption (PADIS)” practice guidelines recommend using behavioral scales as alternative pain measures (
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
). However, behaviors can be masked in heavily sedated patients, those receiving neuromuscular blocking agents, or those suffering from neurologic damages severely affecting their motor system. Therefore, other parameters must be explored for pain assessment. Although vital signs (VS) through continuous ICU monitoring are commonly used by nurses for pain assessment (
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
). However, no comprehensive review on the psychometric performance of VS for ICU pain assessment is available.
Aim and the Research Question
In this narrative review, we summarized and critically analyzed the literature on the psychometric performance of VS for pain assessment in critically ill adults able or not to self-report in the ICU. The research question was: “What are the reliability and validity of VS for ICU pain assessment in critically ill adults?”
Concept of Validity Testing
We focused on the reliability and validity of VS for the purpose of pain assessment. Reliability is a necessary condition of validity and refers to the consistency of VS values when the pain condition has not changed. Test-retest reliability is a strategy to evaluate whether VS parameters yield the same values over time in similar conditions (
). For example, pre- and post- procedure assessments may lead to similar values in patients with stable conditions. Validity refers to the ability of VS to measure pain (
). Relevant validation strategies of VS for ICU pain assessment include: (1) the ability of VS to discriminate between different conditions, e.g., nociceptive versus non-nociceptive procedures or rest conditions (i.e., discriminative validation); and (2) the association between VS and either patient self-reporting as the reference standard measure of pain, or an alternative behavioral measure of pain in the target population such as CPOT or BPS (i.e., criterion validation) (
). Thus, we examined the test-retest reliability, discriminative, and criterion validation of VS for pain assessment in critically ill adults.
Methods
A narrative review was the best approach as it is designed for answering research questions, which are relevant to clinical practice and are specific but in a broader sense than questions requiring systematic reviews (
). Our research question aimed for a comprehensive understanding of the current knowledge on the topic, contrasting opposite findings, and identifying gaps for future research (
). More specifically and in accordance with SANRA's recommendations for a high-quality narrative review, we specified our research question, included a comprehensive literature search, and provided evidence tables of relevant findings (
Studies were selected if they: (1) were conducted in adult ICU settings; (2) included participants aged 18 years or older; (3) reported VS as an indicator for pain assessment; and (4) used reliability or validation strategies. Studies were excluded if they were: (1) conference abstracts; (2) editorials; (3) case studies; (4) studies conducted in post-operative units; and (5) studies that reported VS for reasons other than pain assessment (e.g., hemodynamic stability). The studies from the previous PADIS practice guidelines were also included in this updated review. We excluded one study of PADIS (
) because the values of VS were reported only as percentage; thus, we were unable to include it in our analysis.
Search Strategy
The search strategy was developed by the authors and an experienced health research librarian. We initially searched Medline, Embase, CINAHL, and Cochrane health-related databases from their date of inception to April 6, 2022. We grouped the keywords under three concepts: (1) vital signs; (2) pain assessment; and (3) intensive care unit. Keywords were grouped as follows: (1) concept one: autonomic response, vital signs fluctuations, and each VS alone (i.e., heart rate, respiratory rate, blood pressure, SpO2[oxygen saturation]); (2) concept two: pain measurement/assessment, pain management, pain detection; and (3) concept three: intensive care unit, critical care, intensive care. Boolean operators “AND” and “OR” were used to combine the search terms. Additionally, the reference lists of the included studies were also searched for any further eligible studies. Results of the database searches were downloaded into Endnote X9. The search resulted in 1,359 citations. Duplicates were removed and the remaining (n = 976) citations were exported to the Rayyan software (
) for title and abstract screening against the eligibility criteria. The search process narrowed the results to a total of 30 articles, mostly in English but two were in Turkish (
). For languages not understood by the authors, Google translator was used, and multilingual colleagues were consulted to help with the translation when necessary.
Summary of Extracted Data
Data on sample characteristics, types of nociceptive and non-nociceptive procedures, reliability and validation strategies, and reference standards were collected from citations and described. The results of VS fluctuations during a nociceptive procedure compared with rest/pre-procedures were reported via calculation of mean change percentage (difference in VS value during nociceptive procedures and rest/pre-procedure X 100). The range of means was reported for each VS separately. The range of Pearson and Spearman's rho correlation coefficients of VS with reference standards (i.e., self-report, behavioral pain scores) were reported and identified as weak (≤.30), moderate (.40-.60), and strong (≥.70) (
). A total of 3,469 patients in the ICU were included in 30 studies. Table 1 summarizes characteristics of the study samples, VS that were examined, and types of procedures. Patients were admitted to the ICU with various diagnoses (75% medical-surgical, 12% cardiac, and 13% brain-injury), and had different LOC (34% conscious, 24% altered LOC, and 42% unconscious).
Figure 1Preferred Reporting Items for Systematic reviews and Meta-Analyses flow diagram.
Behavioral and physiologic indicators during a nociceptive procedure in conscious and unconscious mechanically ventilated adults: Similar or different?.
Pain assessments were performed at rest, pre-procedure, and during procedures in all studies, and 10 to 20 minutes post-procedure in 16 studies. The most common nociceptive procedures were endotracheal/tracheal suctioning (n = 17 studies) and turning/repositioning (n = 18 studies). Non-nociceptive procedures included non-invasive blood pressure using cuff inflation (NIBP) (n = 7 studies), eye care (n = 3 studies), and soft touch (n = 2 studies).
Vital Signs Used for Pain Assessment
Hemodynamic parameters such as heart rate [HR] (n = 30 studies) and blood pressure [BP] (i.e., systolic [SBP], diastolic [DBP], and mean [MBP]) (n = 29 studies) were the most frequent VS used for pain assessment in critically ill adults. Respiratory parameters including respiratory rate [RR] (n = 16 studies) and SpO2 (n = 13 studies) were also assessed in some studies for this purpose (Table 1). All VS were available via ICU monitoring devices.
Reliability and Validity of Vital Signs for ICU Pain Assessment
Reliability and validation strategies used for the evaluation of VS in each study and relevant findings are described in Table 2, 3, and 4. In the following sections, we synthesized the data from included studies to describe test-retest reliability, discriminative validation, and criterion validation results for each VS separately.
Test-retest reliability of HR pre-/post- nociceptive procedures was reported in 16 studies (Table 2). After nociceptive procedures, HR decreased to reach similar values found at pre-procedures; thus, supporting stable values pre- and post-procedures in most studies. However, HR values remained slightly higher (3-4 bpm) or lower (2 bpm) post-procedures in five studies.
Discriminative validation of HR was examined in 30 studies (Table 3). Statistically significant increases from 2% to 26% (range: 2-19 bpm) during nociceptive procedures compared with rest and/or pre-procedures were found in 23 studies. From these, only three studies reported significant increases of greater than 20% (range: 18-22 bpm) during endotracheal suctioning (
). Similar values of HR during nociceptive and non-nociceptive procedures were reported in four studies.
Criterion validation of HR was evaluated using patient self-reporting in nine studies and with the CPOT or BPS in eight studies (Table 4). Eight studies reported absence of or nonsignificant positive weak correlations (≤.30) between HR and self-reported pain intensity during nociceptive procedures.
found a significant positive weak correlation (Spearman's rho = .32) during wound care procedure in 62 neurosurgery ICU patients. Only one study reported a significant positive moderate correlation (Spearman's rho = .69) during turning in 99 cardiac surgery ICU patients (
). Eight studies reported absence of or nonsignificant weak correlations (≤.30) between HR and behavioral pain scores.
Blood pressure
BP consisted of three parameters: SBP (16 studies), DBP (15 studies), and MBP (17 studies). Test-retest reliability of SBP, DBP, and MBP at pre-/post- nociceptive procedures was examined in 15 studies (Table 2). These parameters remained stable at pre-/post- procedures only in 11 studies. Slight increases of up to 5% and decreases of 4% were also reported in BP values post-procedure compared with pre-procedure in four studies.
Discriminative validation of BP parameters was examined in 28 studies (Tables 3). Statistically significant slight increases up to 11% in SBP (range: 3-14 mm Hg), 12% in DBP (range: 5-10 mm Hg), and 13% in MBP (range: 2-10 mm Hg) during nociceptive procedures compared with rest and/or pre-procedures were reported in 18 studies. Significant increases in SBP by 19% and 17%, DBP by 27% and 15%, and MBP by 12% and 17% during nociceptive procedures were found in only two studies with samples of 24 and 12 ICU patients, respectively (
). Other studies reported either nonsignificant slight increases (<20%) or no changes in BP parameters during nociceptive procedures.
Criterion validation was evaluated using patient self-reporting of pain in nine studies and the CPOT or BPS in a total of eight studies (Table 4). Eight studies reported absence of or nonsignificant positive weak correlations (≤.30) between BP parameters and self-reported pain intensity during different nociceptive procedures. Only
Vital signs fluctuations and their relationship with pain in the brain-injured adult critically ill - a repeated-measures descriptive-correlational study.
reported a moderate positive correlation between SBP (Spearman's rho=.62) and self-reported pain intensity during turning in 24 brain-injured ICU patients. Eight studies reported absence of or nonsignificant positive weak correlations (≤.30) between BP parameters and behavioral pain scores. Only
Comparison between critical-care pain observation tool and physiologic indicators for pain assessment in the critically ill, mechanically ventilated adult patients.
reported a significant weak correlation (r = .15) between DBP and the CPOT scores during positioning in 60 sedated medical-surgical ICU patients.
Respiratory rate
Test-retest reliability of RR at pre-/post- nociceptive procedures was examined in 12 studies (Table 2). RR remained stable at pre-/post- nociceptive procedures in most studies. RR was higher (8% increase) or lower (9% decrease) at post-procedures compared with pre-procedures values in four studies (
Validation of the critical-care pain observation tool and vital signs in relation to the sensory and affective components of pain during mediastinal tube removal in postoperative cardiac surgery intensive care unit adults.
Journal of Cardiovascular Nursing.2016; 31: 425-432
Vital signs fluctuations and their relationship with pain in the brain-injured adult critically ill - a repeated-measures descriptive-correlational study.
Discriminative validation of RR was examined in 16 studies (Table 3). Statistically significant increases from 6% to 30% (range: 1-5 brpm) during nociceptive procedures compared with rest and/or pre-procedures were found in 14 studies.
reported significant RR increases of 75% during turning in 43 brain-injured ICU patients. Only one study reported no changes in RR during suctioning in 36 neurosurgery ICU patients (
). In two studies where statistically nonsignificant findings were found, RR increased by 25% (range: 2-4) during suctioning and turning in 34 and 190 mechanically ventilated ICU patients unable to self-report, respectively (
Psychometric validation of the behavioral indicators of pain scale for the assessment of pain in mechanically ventilated and unable to self-report critical care patients.
Criterion validation of RR was evaluated using patient self-reporting of pain in eight studies and the CPOT in one study (Table 4). Absence of or nonsignificant positive weak correlations (≤.30) were found between RR and self-reported pain intensity during nociceptive procedures in eight studies. Spearman's rho correlations of .32 (turning), .36 (wound care), and .53 (mobilization) were found between RR and self-reported pain intensity scores in 84 neurosurgical ICU patients (
Validation of the critical-care pain observation tool and vital signs in relation to the sensory and affective components of pain during mediastinal tube removal in postoperative cardiac surgery intensive care unit adults.
Journal of Cardiovascular Nursing.2016; 31: 425-432
Test-retest reliability of SpO2 at pre-/post- nociceptive procedures was examined in 11 studies and showed stable values (Table 2).
Discriminative validation of SpO2 was examined in 13 studies (Table 3). Statistically significant decreases of 2% (range: 1-2% in oxygen saturation value) during nociceptive procedures compared with rest and/or pre-procedures were reported in only three studies (
). No changes or statistically nonsignificant findings were found in 12 studies. Contrarily, one study found that SpO2 decreased more than 2% during NIBP procedure, but it did not change during turning procedure in 43 brain-injured patients (
Criterion validation of SPO2 was evaluated using patient self-reporting of pain in eight studies and the CPOT in one study (Table 4). Seven studies reported absence of or nonsignificant positive weak correlations (≤.30) between RR and self-reported pain intensity during nociceptive procedures. A negative weak correlation of RR with self-reported pain intensity (r = -.20) during turning in 43 brain-injured ICU patients was found in one study (
Validation of the critical-care pain observation tool and vital signs in relation to the sensory and affective components of pain during mediastinal tube removal in postoperative cardiac surgery intensive care unit adults.
Journal of Cardiovascular Nursing.2016; 31: 425-432
This is the first narrative review to comprehensively report the validity of VS for pain assessment in adult ICU patients during 20 years of research. Overall, 30 studies from 17 countries across six continents were included. The most common VS used for ICU pain assessment were HR and BP. Most VS showed stable values at pre- and post- procedures, demonstrating test-retest reliability. Discriminative validation of all VS was supported in most studies but did not reach clinical significance for most. Moreover, criterion validation of VS using reference standards was not supported for ICU pain assessment.
Discriminative changes of VS in response to nociceptive procedures may be triggered by different mechanisms. Many ICU standard care procedures are noxious stimuli and may trigger not only the nociception process but also the biologic stress response. This leads to the release of catecholamines including norepinephrine from the terminals of sympathetic nerves and epinephrine from the adrenal cortex (
). The immediate effects of catecholamines are manifested through increases in VS including HR (chronotropic effect of epinephrine on cardiac activation), BP (peripheral vasoconstriction caused by norepinephrine), and RR (caused by binding of epinephrine to beta receptors) (
); therefore, hypoxemia (i.e., decrease in blood oxygen saturation) may be detected by decreases in SpO2. Therefore, VS values may change in response to nociceptive procedures as was demonstrated in many included studies of this review. If enough time post-procedure (15-20 minutes) is given to allow elimination of catecholamines (i.e., epinephrine and norepinephrine) (
), then the VS values should be similar to their pre-procedure values, which was demonstrated as test-retest reliability in some studies.
Despite statistically significant changes of VS during some nociceptive ICU procedures, their clinical significance is negligible. Clinically significant changes are of paramount importance to guide clinical practice (
). One approach for establishing clinical significance refers to the interpretation of change scores into meaningful clinical data perceived as effective and beneficial (
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
). Only a small number of studies have shown clinically significant changes in some VS values during nociceptive procedures. For example, HR increased more than 25% only in three of 23 studies, but it also remained unchanged in seven studies. No clinically significant changes were reported for BP parameter in 28 studies. RR has shown clinically significant increases of more than 25% during nociceptive procedures in five of 14 studies. Yet, attention should be given to conditions that hinder respiratory function (i.e., opioid-induced respiratory depression or controlled ventilation assistance alter the RR). A systematic review highlighted that pain may influence respiration by increasing RR and volume; nonetheless, hyperventilation may also reflect the aversive feelings related to pain and fear (
). Similar to other VS, increases in RR are not specific to pain. Indeed, pain, fear, and anxiety may coexist, which emphasizes the importance of using assessment tools that are more sensitive and specific to pain (
). Although the clinical significance for SpO2 has not been specified, it was the weakest parameter in responding to nociceptive procedures as its changes were negligible or absent. In summary, our results did not support discriminative validation of VS for ICU pain assessment due to inconsistency in clinically significant changes and frangibility of these parameters to other clinical conditions (e.g., shock, hypovolemia, arrythmias), treatments, and medications (e.g., beta-blockers, vasopressors) (
Lack of correlation between VS and pain reference standards hinders criterion validation of their use for ICU pain assessment. Similar findings were demonstrated in other contexts of care such as emergency and pre-hospital care, where weak correlations between VS and self-reported pain were not considered to be clinically significant for pain assessment (
). The lack of correlation of VS with the reference standard measures of pain (i.e., self-reporting, behavioral scales) represents a major limitation. Therefore, absence of criterion validation of VS to pain compromises their use for ICU pain assessment.
Pain assessment remains challenging in patients in the ICU unable to self-report. To date, no reliable and valid physiologic parameters have been identified for ICU pain assessment. Future studies are required to explore new technologies that are developed for the purpose of nociception and pain monitoring. Some studies have shown that the simultaneous combination of multiple physiologic parameters through AI algorithms may be superior to individual parameters (e.g., HR and BP) for the monitoring of nociception and related pain in anesthetized adults (
). The Nociception Level Index (NOLTM) is the only multi-parameter physiological technology that incorporates several VS in its algorithm to estimate the level of nociception and related pain (
Exploration of a multi-parameter technology for pain assessment in postoperative patients after cardiac surgery in the intensive care unit: The nociception level index (NOL)™.
Exploration of the nociception Level (NOL) index for pain assessment during endotracheal suctioning in mechanically ventilated patients in the intensive care unit: An observational and feasibility study.
) but future studies are required to further validate its use in the ICU context.
Limitations
Narrative reviews are often subjected to selection bias. To mitigate this issue, we uniformly applied our eligibility criteria. Although we used a comprehensive list of keywords and a systematic search strategy in four databases, it is likely that some relevant studies might have been missed. Nevertheless, we illustrated our search strategy, structured, and analyzed studies with similar findings (
). Also, we used the SANRA as a guiding framework in this narrative review.
Some limitations must be highlighted. First, the non-invasive blood pressure cuff inflation has been considered as a non-nociceptive procedure in primary studies included in this review. However, attention should be paid to chronic pain etiologies such as allodynia and fibromyalgia. With the presence of these conditions, the cuff inflation could have caused significant pain. Yet, this information was not available in included studies. Pain resulting from chronic pain conditions should be explored in future research. A second limitation relates to the heterogeneity of samples across studies e.g., different LOC and diagnostic categories. Nonetheless, our analysis of reliability and validity findings were consistent across studies. A third and last limitation, we did not extract data on the administration of analgesics before procedures which could have influenced VS. It is worth mentioning that this information was reported in only a few studies included in this review. Moreover, an optimal strategy to examine the influence of analgesia on VS would be to compare their fluctuations pre- and post-analgesic administration. However, such a strategy was not commonly examined in included studies. This objective could be explored in future research.
Conclusion
Consistent with clinical recommendations and practice guidelines (
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
), findings of our narrative review suggest that VS available through continuous ICU monitoring devices lack reliability and validity for ICU pain assessment. Even though VS values fluctuate during nociceptive ICU procedures, their changes are often not clinically significant. Moreover, criterion validation of individual VS was not supported for ICU pain assessment due to lack of correlations with reference standards (i.e., the patient's self-report of pain intensity, behavioral pain measures: CPOT or BPS). To conclude, ICU nurses should only use VS as cues to begin further pain assessment using valid methods such as the patient's self-report of pain whenever possible, or behavioral pain measures. For a comprehensive list of valid pain assessment tools, readers can refer to
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
. Alternative physiologic measures are urgently needed to guide pain detection in critically ill adults in whom none of the reference standard pain methods can be used.
References
Abdelhakeem A.K.
Amin A.
Hasanin A.
Mukhtar A.
Eladawy A.
Kassem S.
Validity of pulse oximetry-derived peripheral perfusion index in pain assessment in critically ill intubated patients.
Validation of the critical-care pain observation tool and vital signs in relation to the sensory and affective components of pain during mediastinal tube removal in postoperative cardiac surgery intensive care unit adults.
Journal of Cardiovascular Nursing.2016; 31: 425-432
Vital signs fluctuations and their relationship with pain in the brain-injured adult critically ill - a repeated-measures descriptive-correlational study.
Clinical practice guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU.
Behavioral and physiologic indicators during a nociceptive procedure in conscious and unconscious mechanically ventilated adults: Similar or different?.
Exploration of a multi-parameter technology for pain assessment in postoperative patients after cardiac surgery in the intensive care unit: The nociception level index (NOL)™.
Comparison between critical-care pain observation tool and physiologic indicators for pain assessment in the critically ill, mechanically ventilated adult patients.
Psychometric validation of the behavioral indicators of pain scale for the assessment of pain in mechanically ventilated and unable to self-report critical care patients.
Exploration of the nociception Level (NOL) index for pain assessment during endotracheal suctioning in mechanically ventilated patients in the intensive care unit: An observational and feasibility study.
International Association for the Study of Pain (IASP). (2020). Pain terms: A list with definitions and notes on usage. Retrieved April 16, 2022, from https://www.iasp-pain.org/resources/terminology/.
00025-5&id=gr1.jpg){kind=link}