Behavioral Pain Scale and Critical Care Pain Observation Tool for pain evaluation in orotracheally tubed critical patients. A systematic review of the literature

Objective: To describe the appropriateness of two behavioral scales, the Behavioral Pain Scale and the Critical Care Pain Observation Tool, for pain assessment in orotracheally intubated patients admitted to intensive care units. Method: Using the methodology recommended by the Cochrane Center, a systematic literature review was performed in the electronic database EBSCO Host (CINAHL Complete; MEDLINE® Complete; Nursing & Allied Health Collection: Comprehensive; Cochrane Central Register of Controlled Trials; Cochrane Database of Systematic Reviews; Cochrane Methodology Register; Library, Information Science & Technology Abstracts; MedicLatina). Two searches were conducted using the following English terms in the search field: “behavioral pain scale” AND “critical care pain observation tool” AND “behavioral pain scale” OR “critical care pain observation tool”. Two independent reviewers performed the critical evaluation and data extraction and synthesis. Results: Fifteen studies were included that showed that the Behavioral Pain Scale and the Critical Care Pain Observation Tool are valid and reliable scales for pain assessment in orotracheally intubated patients admitted to the intensive care unit. The scales showed similar psychometric properties and good reliability. Conclusion: Both scales are adequate for assessing pain in orotracheally intubated patients admitted to intensive care units; however, they exhibit limitations in specific populations, such as trauma, burn and neurosurgical patients. Further studies on the subject and in specific populations are suggested.

compromising patient recovery and well-being. (3) The accurate assessment of pain contributes to effective care management; improved adequacy of therapeutic measures, including analgesic and sedative use; and shorter IMV durations and lengths of stay in the ICU. (3,5) Pain control is a patient right and a duty of health care professionals, and the negation and devaluation of pain are ethical errors and failures of excellence in professional practice. (1,2) Thus, when patients cannot self-report pain, health care professionals must resort to the use of validated pain assessment scales, (3,5) such as the Behavioral Pain Scale (BPS), (6) which assesses indicators such as facial expression, upper limb movements and compliance with ventilation (Table 1), and the Critical Care Pain Observation Tool (CPOT), (5) which assesses indicators such as facial expression, body movements, muscle tension and compliance with ventilation in intubated patients or vocalization in extubated patients ( Table 2). These two observational and behavioral scales are indicated for the assessment of pain in critically ill patients who are sedated and/or unconscious and undergoing IMV and/or have difficulty with self-reporting pain. (7,8) This is the context of the present systematic literature review (SLR), which aims to identify the appropriateness of the BPS and the CPOT for assessing pain in noncommunicating patients admitted to the ICU.

METHODS
This SLR followed the methodology recommended by the Cochrane Center (9) and was guided by the following research question: How appropriate are two behavioral scales, BPS and CPOT, for assessing pain in orotracheally intubated patients admitted to the ICU? Selection criteria were defined and applied according to the PICo method: Participants (adult inpatients older than 18 years), Point of Interest (pain assessment scales: the BPS and the CPOT) and Context (the ICU).
The following exclusion criteria were established: studies conducted with children under 18 years, those including adult inpatients in non-ICU settings, those that used other scales/other strategies to assess pain, qualitative studies, and studies that were not original.
Thus, studies that directly compared the two scales or that mentioned the advantages of using each scale individually were included.

Search strategy
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist (10) was used in this SLR as a guide to meet the accepted standards for systematic reviews.
The search was conducted in the EBSCO Host database (Cumulative Index to Nursing and Allied Health  Two independent searches of the databases were performed. For both searches, the search terms "behavioral pain scale" and "critical care pain observation tool" were introduced. For the first search, the Boolean operator "AND" was used; for the second, the Boolean operator "OR" was used. The review was limited to studies published in Portuguese or English, and both searches were conducted without a time limit.
The titles were read first, followed by the abstracts and then the full text of the articles found in the search in order to select those that answered the research question.
The studies were read and their methodological quality (MQ) was evaluated by two independent researchers to ensure critical evaluation during article selection. In cases of disagreements between the researchers, a third evaluator was asked to review the article.
The MQ of the studies was assessed using instruments from the Joanna Briggs Institute (MAStARI). (11,12) Before the investigators conducted their evaluations, it was established that only studies with high MQ, i.e., those with a score of 8 to 10 on the MAStARI Checklist for Diagnostic Test Accuracy Studies, (12) would be included (Table 3).

RESULTS
The first search with the Boolean operator AND resulted in 186 studies; 32 of them were excluded due to duplication, 109 after the title was read and 36 after the abstract was read. Of the 36 that were excluded after the abstract was read, 11 were excluded because the full text was not available, 3 because they were not in English or Portuguese, 2 because they were duplicates, 5 because they used other pain assessment scales, 12 because they were not original, 1 because it focused on pain control interventions, 1 because it focused on drug administration and 1 because it focused on a specific population that did not meet the inclusion criteria of the present review).
After reading the abstract, 9 studies were selected, and these studies remained in the sample after the full text was read ( Figure 1). The 9 studies showed high MQ according to the cited criteria ( Table 3).
The 9 selected studies were conducted in different countries: Sweden, (13) the Netherlands, (14) China, (15) Saudi Arabia, (16) the United States, (17,18) Canada, (19) and Italy. (20,21)  Although most of the indicators in both scales increased with a painful procedure, only those in the BPS increased in association with a nonpainful procedure Liu et al., (15) China Evaluate and compare the reliability and validity of the BPS and the CPOT for pain assessment in intubated and nonintubated critically ill patients Observational prospective study A total of 608 pain assessments were performed using the CPOT and BPS (BPS and BPS-NI) before and during painful and nonpainful procedures MQ = 9 Convenience sample of 117 critically ill adult patients admitted to the ICU The BPS and CPOT were found to be reliable and valid to assess pain in intubated and nonintubated patients The results showed good interrater reliability (ICC = 0.973 for the CPOT and ICC = 0.955 for the BPS); good internal consistency (α = 0.795 for the CPOT and 0.791 for the BPS) and reliability of 0.950 (CPOT) and 0.941 (BS) Al Darwish et al., (16) Saudi Arabia Determine the reliability and validity of nonverbal pain assessment tools in critically ill patients (BPS, NVPS and CPOT) Descriptive observational study with quantitative analysis Three pain assessment instruments -the BPS, CPOT, NVPS -were administered before, during and after painful and nonpainful procedures for a total of 240 evaluations MQ = 9 Convenience sample of 47 critically ill noncommunicating patients undergoing IMV admitted to the ICU The BPS was the most valid and appropriate instrument to assess pain in noncommunicating patients admitted to the ICU due to the characteristics of its subscales; however, the CPOT was considered an appropriate alternative. The results showed good interrater reliability (ICC = 0.80) and good internal consistency (α = 0.95 (CPOT); 0.95 (BPS); 0.86 (NVPS)) The NVPS was not sensitive to assess pain in these patients Continue...

Author/country Objectives Methods/MQ (12) Participants Results
Rahu et al., (17) United States Identify the most appropriate scale for pain assessment in intubated patients Determine the validity and sensitivity of six pain scales Descriptive study Observations were made by two independent investigators in communicating and noncommunicating intubated patients before and during nonpainful and painful procedures using six scales (NVPS; BPS; Comfort; Faces; Face, Legs, Activity, Cry, Consolability; NRP) MQ = 9 Convenience sample of 50 communicating patients and 100 patients who were unable to communicate verbally who were intubated and undergoing IMV All pain scales had moderate to high correlation with the self-report of patients during endotracheal suction (painful procedure) All scales were sensitive in obtaining the patient's pain response in all phases (p < 0.001) Both the patients and the investigators assessed the highest pain on the Faces scale, which reveals that some caution in its use is necessary Chanques et al., (18) United States Compare the psychometric properties of three pain assessment scales (the BPS/ BPS-NI, CPOT and NVPS) in intubated and nonintubated patients unable to self-report pain A total of 258 assessments of pain, sedation (RASS) and delirium (CAM-ICU) were performed by at least one investigator and one nurse in 30 patients before, during and 10 minutes after routine procedures MQ = 10 Convenience sample of 30 adult patients; RASS > -4 in patients who were unable to self-report their pain intensity The three scales showed good psychometric properties in the assessment of pain in intubated and nonintubated patients unable to self-report their pain intensity The BPS and CPOT showed better reliability (κ = 0.81 for both) and internal consistency (α = 0.80 -BPS; α = 0.81 -CPOT). The BPS was classified as the most feasible scale, with the highest score on the category of "easiest to remember" Bourbonnais et al., (19)  Convenience sample of 20 critically ill patients admitted to the ICU (10 intubated and 10 nonintubated patients) The Swedish version of the BPS is adequate for pain assessment in patients unable to self-report pain The discriminant validity for the assessments before, during and after the procedure obtained a percentage agreement of 28%, with 95% CI (relative position of -0.08 to +0.02; relative concentration of -0.06 to +0.08; relative variance of classification of 0.000 -0.002) and a reliability of 85%.
Frandsen et al., (23)  The CPOT has a greater validity than the NVPS-R for pain assessment in critically ill noncommunicating patients, particularly those with neurological and trauma injuries The interrater reliability was higher for the CPOT (0.60 -0.97) than for the NVPS-R They were published in 2011, (13) 2014, (18) 2015, (14,15,17) 2016 (16,19,21) and 2017. (20) The second search with the Boolean operator "OR" resulted in 853 studies; 208 of them were excluded due to duplication, 601 after the title was read, 22 after the abstract was read and 12 after they were read in full. In the various elimination stages, studies were excluded because they were not in English or Portuguese, because they focused on a specific population other than the one considered in this review (children or nonintubated patients, for example), because they were duplicates, because they were not original studies, or because they were studies on other topics (validations of other scales, for example).
After the full text was read, ten studies were selected, four of which were eliminated because they had already been included in the first search with the Boolean operator AND; thus, six studies were included in the sample ( Figure  2), and all of them had high MQ scores (12) (Table 3).
The six selected studies were conducted in different countries: Brazil, (7) Sweden, (22) Denmark, (23) Iceland (24) and Canada. (5,25) They were published in 2006, (5) 2013, (24,25) 2014 (7) and 2016. (22,23) After the results of the two searches were compared, 15 studies were included in this SLR. All of the studies were quantitative, and their sample sizes ranged from 20 (22) to 150 (17) patients. The BPS was developed and tested in 2001 by Payen et al. (26) on a sample of 30 patients with medical and surgical diagnoses undergoing IMV (269 observations), and it showed good validity and reliability in the study population.
There were differences in the correlations between assessments performed before, during and after a painful procedure, (15,22,26) and the discriminant validity for assessments performed before, during and after the procedure obtained a percentage agreement of 28%. However, in one of the studies, the discriminant validity was less supported because it increased during exposure to a nonpainful stimulus. (14) Whereas in the original validation study, there were negative correlations between the pain score and the administered midazolam and fentanyl doses, (26) in the validation study for Saudi Arabia, routine procedures, such as secretion aspiration, caused pain in all patients regardless of the administration of analgesia. (16) The CPOT was developed in 2006 in French, in Quebéc, Canada, by Gélinas et al. (5) and was validated on a convenience sample of 105 intubated cardiac surgery patients (33 unconscious and 99 conscious) before and after extubation. The CPOT showed good interrater reliability (ICC: 0.52 -0.88), adequate content validity (0.88 -1.0) and criterion validity (patients who reported pain (1.62 -3.65)), and discriminant validity was evidenced by higher CPOT scores during painful procedures than at rest (t = -9.01 --15.96; p < 0.001). Spearman correlations of 0.40 -0.59 (p < 0.001) showed that the patients' selfreported pain intensity scores were moderately correlated with the CPOT scores. (5) The review found that the CPOT had been validated/ revalidated with communicating and noncommunicating patients admitted to the ICU. It was validated for the Swedish population on a sample of 40 intubated patients, (13) for Iceland on a sample of 30 patients intubated after cardiac surgery, (24) for Canada on a sample of 23 intubated patients, (19) for the United States on a sample of 30 intubated patients, (18) in Toronto on 66 trauma and neurosurgical patients (34 communicating and 32 noncommunicating), (25) for the Netherlands on 68 patients undergoing mechanical ventilation, (14) for Italy on a sample of 101 patients (41 conscious and 60 unconscious), (21) for China on a sample of 117 critically ill ventilated patients, (15) for Denmark on 70 patients undergoing IMV without sedation, (23) and for Saudi Arabia on 47 noncommunicating critically ill patients. (16) Validation of the CPOT revealed good reliability (ICC of 0.75 (14) -0.95 (15) ) and internal consistency (Cronbach' alpha of 0.70 (23) -0.973 (15) ) and good criterion validity. There was a significant correlation between the pain intensity scores reported by communicating patients and the CPOT scores; (5) the scores increased significantly when patients were exposed to painful procedures rather than nonpainful procedures, indicating criterion validity. (13)(14)(15)(16)20,(23)(24)(25) Significant correlations were also found between CPOT scores and mean arterial pressure (p = 0.32 -0.45). (25) and general vital signs before, during and after a painful procedure; (20) however, further studies should be conducted to explore the role of vital signs in pain. (25) Table 3 shows the objectives, methods and assessment of MQ, (12) the participants and the results of the selected studies.

DISCUSSION
The detection, quantification and treatment of pain in critically patients have long been a concern of health care professionals. Nevertheless, pain is common in critically ill patients, regardless of their clinical condition, and its accurate assessment using appropriate instruments allows for the use of more beneficial therapeutic measures.
The BPS and CPOT have good psychometric properties and good reliability in intubated and nonintubated patients in the ICU who are unable to self-report their pain, (8,(14)(15)(16)18,20,21,23,24) and both scales should be used for pain assessment in this patient population. (18,21) The BPS is considered applicable to critically ill patients who are sedated, unconscious or have difficulty self-reporting pain, especially those undergoing IMV, given that one of its three domains pertains specifically to compliance with ventilation. In turn, the CPOT, in addition to the domain intended for patients undergoing mechanical ventilation, has a vocalization domain; thus, it can also be used with extubated patients, unlike the BPS, which is only intended for use with patients undergoing IMV. (8) Some authors have stated that the BPS is the more viable, (18) specific, (21) reliable, valid (16) and sensitive tool for assessing patient pain; however, the CPOT is a good alternative. (16,21) Other authors consider the CPOT the scale of choice because its discriminant validity is more supported because the score does not increase with a nonpainful stimulus, contrary to what has been observed with the BPS. (14) In addition, there was an increase in the pain assessment scores of both the BPS and the CPOT when a painful procedure was evaluated. (13)(14)(15)(16)(17)20,(23)(24)(25) Both instruments are sensitive to painful procedures, with an observed increase in the various indicators that constitute both scales. (14)(15)(16)(17) The main parameters with major changes are facial expression in the BPS (18,21) and muscle tension/ stiffness, facial tension and ventilator tolerance/cough in the CPOT. (19,21) During painful procedures, there is a significant correlation between pain scores and vital signs, specifically blood pressure;, i.e., the higher the pain score, the higher the blood pressure will be. (13,20) Although this correlation has been observed, some authors suggest that further studies be conducted to explore the relationship between vital signs and pain. (25) Although some consider the CPOT to have good validity for assessing pain in noncommunicating critically ill patients, particularly those with neurological and traumatic injuries, and in neurosurgical patients, (25) others argue that its application to patients with brain injuries, (27,28) cognitive deficits or burns (28,29) is a limitation.
There is still no consensus regarding levels of consciousness, sedation and analgesia because some authors found no significant differences in the application of the two scales to patients with different levels of sedation or analgesia; (20) however, other authors found significant differences during and after nursing care in conscious and unconscious patients. (21,28) In their practical application, both the BPS and CPOT are considered by health care professionals to be useful tools in the ICU setting because they are easy to use and remember. (7,(18)(19)(20)25) Their application contributes to an increased frequency of assessments and, consequently, to the reduced administration of analgesics and sedatives. (3,30) Researchers have suggested using the two scales simultaneously, as doing so may result in more accurate pain detection and assessment. (21) Although the use of the BPS and CPOT scales has been shown to have positive effects on pain management in patients admitted to the ICU, experimental studies are recommended. (28) CONCLUSION This systematic literature review found several studies validating the Behavioral Pain Scale and the Critical-Care Pain Observation Tool for use with orotracheally intubated critically ill patients from various cultures, and both instruments were found to be valid and reliable for assessing pain in intubated patients admitted to intensive care units.
Both instruments were sensitive when applied during painful procedures, showing increases in various indicators: facial expression on the Behavioral Pain Scale and muscle tension/stiffness, facial tension and ventilator tolerance/cough on the Critical-Care Pain Observation Tool, and blood pressure on both scales.
There is, however, no agreement regarding the administration of the scales in patients with different levels of consciousness, sedation and analgesia. However, it appears that the use of at least one of the scales helps to increase the frequency of assessments and, consequently, reduces the administration of analgesics and sedatives. In this regard, it is essential that health care professionals use at least one of the two analyzed scales for pain assessment in intubated patients, with the goal of improving the care provided.
Further studies with an experimental design covering different critically ill patients populations admitted to intensive care units, namely, trauma, burn and neurosurgical patients, are suggested. . Foram realizadas duas pesquisas com os seguintes termos em inglês no campo de pesquisa: "behavioral pain scale" AND "critical care pain observation tool" AND "behavioral pain scale" OR "critical care pain observation tool". Dois revisores in-dependentes realizaram a avaliação crítica, a extração e a síntese dos dados.