The incidence of pain in adult critical care patients is higher than 50% and it can be experienced at rest or during routine clinical care procedures1–5. Inadequate management of pain is associated with a stress response that includes changes in physiological parameters, neuro-endocrine secretion and immunomodulation6,7. It is also associated with agitation8 and sleep pattern alterations9, and it can be a contributing risk factor for delirium10 and post-traumatic stress disorder in critical care patients11.

The systematic evaluation of pain integrated in the pain-agitation-delirium (PAD) management protocols has been associated with reductions in the incidence of pain, use of analgesics, duration of mechanical ventilation (MV) and length of stay (LOS) in the Intensive Care Unit (ICU) 1,12–15. Adequate pain treatment necessarily requires the use of reliable tools that aid in its detection and measurement16,17.

Recent international guidelines recommend the use of scales based on behavioral indicators of pain for patients who are unable to self-report, provided that their motor function is preserved and behaviors are observable18,19. For these patients, the behavioral pain scale (BPS)20 and the critical care pain observation tool (CPOT)21,22 are recommended due to their psychometric properties17,23.

In 2011, Latorre et al.24 validated a new tool, the behavioral indicators of pain scale or ESCID from its acronym in Spanish, as a useful tool for assessing pain in mechanically ventilated and unable to self-report critically ill adult patients. The ESCID scale was developed after the original Campbell scale, proposed by The Analgesia and Sedation Working Group of the Spanish Society of Intensive Care Medicine and Coronary Units, taking in to account that the Campbell scale could be an appropriate instrument to assess pain and quantify its intensity in non-communicative, critically ill patients25. However, the Campbell scale has two limitations: it is not been subjected to validation studies in critically ill patients and it is not applicable to patients with mechanical ventilation because it does not include it as an indicator. In the validated version of the ESCID scale, the indicator ‘vocalization’ from the original Campbell scale was replaced by ‘compliance with mechanical ventilation’. Additionally, each indicator of the scale was more precisely defined and quantified to minimize subjectivity of the observer. The ESCID scale provides a greater number of behavioral indicators, which can help decrease the artifacts produced by causes unrelated to pain, considering the positive correlation between the number of behavioral indicators exhibited by the patient and pain shown26.

The ESCID scale has shown good psychometric properties: the internal consistency measured with the Cronbach's Alpha Coefficient was 0.70–0.80; a high correlation between ESCID and BPS scales; and good interrater and intrarater reliability. However, the ESCID scale validation study has some limitations that affect its external validity, as is the fact that it was conducted in a single ICU with a sample of 42 critical care patients with exclusively medical pathologies24.

The interest of the current study is to determine the psychometric properties of the ESCID scale when applied to a wide range of medical and surgical critical care patients in various ICUs of different hospitals27. Our primary hypothesis is that the ESCID scale is a valid and reliable tool for measuring pain in mechanically ventilated and unable to self-report critical care patients.

This study pretends to validate the ESCID scale by addressing the following research objectives: (a) to assess interrater agreement when using pain scales; (b) to assess intrarater agreement when using pain scales in repeated measures for the same patient; (c) to examine the correlation between the ESCID scale and the BPS while using them simultaneously in the same patient and at the same time, when patients are exposed to a non-nociceptive procedure and two nociceptive procedures; (d) to examine the relationship between applying a painful routine care procedure and the changes in physiological indicators of pain; (e) to evaluate the possible differences in the application of the scales for measuring pain in patients with medical and surgical pathologies; and (f) to evaluate the possible differences in the application of scales to measure pain in patients with different levels of sedation.

Multicenter, prospective and observational study to validate a scale-measuring instrument was conducted in 20 Intensive Care Units of 14 Spanish National Health System hospitals with different types of patients admitted for medical and surgical pathologies. The study was initiated in January 2012 and finalized in January 201427. The protocol was approved by the Clinical Investigation Ethics Committee of Puerta de Hierro Majadahonda University Hospital (Spain), the coordinating center (University Hospital Ethics Committee, record n° 273, January 23, 2012), and all corresponding Ethics Committees in each hospital participating in the study.

Written informed consent was obtained from the next of kin or the patient's legal guardian, and was ratified afterwards by the patient itself. To maintain anonymity and guarantee confidentiality, all participants were identified by a code number that was maintained throughout the study.

Two hundred eighty-six patients from 20 ICUs were included during the recruitment period. During the study period, 4386 observations were recorded by 150 observers. A consort flow chart of patient enrolment is shown in Fig. 1. The patients’ characteristics are shown in Table 2.

Figure 1.

Study flow chart.

(0.61MB).

Finally, 190 patients completed the pain assessments in all of three procedures proposed in the study, resulting in 3420 observations being recorded and included in the analysis (Fig. 1). These patients had similar characteristics to those defined for the study population (Table 2).

Interrater and intrarater reliability

There was a high interrater and intrarater conformity when the ESCID scale was applied at the three moments of each procedure (Fig. 2). There were no statistically significant intrarater differences (p=0.241), however differences were observed between procedures (p<0.001) given that the values were similar for painful procedure but differed for non-painful ones. The values show evaluation of increased pain during the application of painful procedures which normalized in the measurement taken 15min after. However, there were no changes in the pain levels at the three moments during non-painful procedures.

Figure 2.

Pain measured with ESCID scale in mechanically ventilated, unable to self-report critically ill patients, by two independent observers before, during and after three different procedures. This figure shows the median scores of the ESCID scale evaluated by two observers according to different situations: before, during and after turning/repositioning, tracheal suctioning and non-painful procedure (rub with a gauze cloth over a portion of healthy skin tissue of the patient). The numbers show a significant increasing of the level of pain during the two painful procedures, that normalized in the assessment after 15min finishing this procedure. However, there were no modifications in the pain level in the three assessment times during the non-painful procedure. There is a good inter-observer concordance. The points are outliers or values outside these limits. Note: Proced: procedure; Obs: observer.

(0.24MB).

The median punctuation of the ESCID and BPS scales (Table 3) demonstrate a difference of 0.0 or 0.1 points between the two observers at the three measuring time points. These results demonstrate a high intrarater and interrater conformity.

Table 3.

Analysis of variance for repeated measures (ANOVA) for BPS and ESCID.

	Turning/repositioning	Tracheal suctioning	Non-painful procedure
	n=190	n=190	n=190
	Mean (SD)	Mean (SD)	Mean (SD)
BPS scalea
Before
Observer 1	3.1 (0.4)	3.2 (0.5)	3.2 (0.5)
Observer 2	3.2 (0.5)	3.2 (0.5)	3.2 (0.6)
During procedure
Observer 1	4.7 (1.6)	5.4 (1.8)	3.2 (0.6)
Observer 2	4.8 (1.6)	5.3 (1.8)	3.3 (0.7)
After
Observer 1	3.2 (0.7)	3.2 (0.5)	3.1 (0.4)
Observer 2	3.3 (0.7)	3.2 (0.5)	3.2 (0.5)
ESCID scalea
Before
Observer 1	0.2 (0.7)	0.3 (0.8)	0.2 (0.7)
Observer 2	0.3 (0.8)	0.3 (0.8)	0.3 (0.9)
During procedure
Observer 1	2.6 (2.5)	3.4 (2.5)	0.3 (0.9)
Observer 2	2.7 (2.4)	3.3 (2.4)	0.4 (1.0)
After
Observer 1	0.3 (1.0)	0.3 (0.8)	0.2 (0.6)
Observer 2	0.4 (1.1)	0.3 (0.8)	0.3 (0.8)

a

The response to pain in BPS and ESCID changes over time, increasing during the procedure (p<0.001). There are differences between painful and non-painful procedures (p<0.001). There are not inter-observer differences. There is no significative effect of the interaction neither among time and observer (p=0.720) nor procedure and observer (p=0.083). SD: standard deviation.

The main finding of this study is that the ESCID scale has good psychometric properties. Considering the results obtained in relation to the internal consistency, interrater and intrarater reliability as well as the discriminant validation of the ESCID scale, it can be concluded that this tool is a valid and reliable tool for measuring pain in unable to self-report mechanically ventilated critically ill patients. These results are consistent with those reported in the first validation of the ESCID scale performed with a small sample of patients in a single medical ICU24.

The psychometric properties of the ESCID scale are similar to those shown by the BPS20,23,28,29 and the CPOT21–23 scales, which is why the use of such behavioral scales has been recommended in recent guidelines18,19. The internal consistency of the ESCID scale had a Cronbach-α coefficient of 0.85, similar to other studies such as that of Chanques et al.23, which reported a Cronbach-α coefficient of 0.80 for the BPS and 0.81 for the CPOT. On the other hand, the ESCID scale demonstrated a high internal consistency in its five domains with Cronbach-α coefficients of 0.70–0.87.

In both painful and painless procedures, ESCID shows a high degree of correlation with the BPS, the tool of reference which has been used in different studies1,20,23,28,29, in addition to demonstrating a high degree of interrater and intrarater conformity in the measurements performed.

The ESCID scale has two additional domains with respect to the BPS and is also innovative with respect to other scales, such as the BPS and the CPOT16,17,23, with the introduction of the “consolability” domain which reflects the patient's reaction to interaction with the observer through verbal and/or tactile stimuli. The “Consolability” domain has proven useful and reliable in other pain behavioral scales, such as FLACC in pediatric patients31 and PAINAD in patients with advanced dementia32, but has not previously been tested in scales to measure pain in critically ill patients.

The fact that the “consolability” domain had a high Cronbach-α coefficient reinforces the internal consistency of the ESCID scale, especially knowing that 150 different observers participated in the study with high interrater and intrarater concordance.

To avoid bias, as in several previous studies28,29, a single measurement was performed per patient and procedure. In addition, all observations were made by 2 observers who were blinded to each other's evaluation.

The reliability of the ESCID scale according to the type of ICU admission was high in both cases, although Cronbach-α coefficient was better in patients with medical compared to surgical pathologies.

The internal consistency of the ESCID scale displayed a significant decrease in very profoundly sedated patients (RASS score -5) with a Cronbach-α coefficient of 0.63, due to bias or abolition of behavioral indicators, therefore its use may be limited in these patients. This finding was consistent with the literature where other studies have established the existence of some limitations in the use of the BPS and the CPOT, which make them inapplicable in cases of deep sedation, neuromuscular block, tetraplegia or polyneuropathy. This concurs with the guidelines published by Barr et al. 18, which established that behavioral scales should be used in patients whose motor functions are intact and in whom behavioral indicators are observable.

The ESCID scale is the only scale which has been subjected to a psychometric validation process in the Spanish language. This is relevant because it increases the possibility of integrating it into PAD management protocols when implementing these in Spanish-speaking countries. The use of a tool which is validated and adapted to the language is crucial18 because this can produce improvements in the outcomes of critical care patients1,3,12–15.

The nurses responsible for data collection were previously trained in the use of the ESCID and BPS scales. While some reliability studies have shown lower interrater conformity values in the use of the BPS, in which the subjective interpretation of the included items has been established as its main limitation17,33, prior knowledge of behavioral scale use could promote correct application23, as demonstrated in the good interrater conformity obtained using the ESCID and BPS scales in the present study. The ESCID scale also features a user guide that furthers the knowledge and application of this scale aiming to minimize observer bias27.

As in previous studies, the presence of procedural pain is evident. Major studies, such as the Thunder Project II26 and the Europain5, demonstrated the presence of pain during turning/repositioning and tracheal suctioning procedures. In this study, the patients presented increased pain scores when the procedures were carried out: 1.6 (BPS) and 2.4 (ESCID) during turning/repositioning, and 2.2 (BPS) and 3.1 (ESCID) during tracheal suctioning.

This result could justify the conclusion that the analgesia used prior to performing such procedures was insufficient. In fact, in our study, only 9% of patients received a bolus of analgesia prior to painful procedures. It is important to establish pre-emptive analgesia34–36 for improved pain control and, consequently, decrease adverse events related to pain and increase the adequate administration of analgesics and sedatives37.

An effort is required to implement validated tools for the detection of pain, which are currently underused38–40, to avoid pain going undetected and untreated as has been demonstrated in some studies3,41,42.

In our sample, no significant differences were observed between the level of pain and the type of admission (medical or surgical). Similarly, no significant differences were found between the level of sedation and the type of admission. This confirms the findings of studies into procedural pain in both medical and surgical patients, regardless of the reason for admittance2,3,5,26.

This study has several limitations, such as the exclusion of patients with suspicion or diagnosis of delirium. Moreover, the diagnosis of delirium is not possible in deeply sedated patients (RASS score −4 or −5), in whom the CAM-ICU tool is not applicable, so it is impossible to determine whether these patients had delirium during the measurements of pain. Although the presence of delirium was considerer an exclusion criteria, delirium screening was not routinely done in all of the patients so the possibility exists that delirious patients might have been included. Consequently, subsequent studies should be performed to determine the influence of delirium on the ESCID scale domains.

The ESCID scale has been validated in the Spanish language in non-communicative medical and surgical patients. Using the ESCID scale in other ICU patient populations and translating the same into foreign languages will require further validation testing.

The ESCID scale offers good psychometric characteristics for pain assessment in non-communicative, mechanically ventilated critical care patients. As shown in our study, which included a large sample of patients with different types of admission, the ESCID scale is a valid, reliable and reproducible tool, and its integration into PAD management protocols could help to prevent significant adverse events resulting from poor pain management.

Pain monitoring should be performed routinely, especially in potentially painful procedures that are performed on critical care patients in routine clinical practice.

This study was financed by the Spanish Health Research Fund (PI11/00766, Health Ministry. Spanish National Research, Development and Innovation Plan 2008–2001. Spanish title: “Estudio multicéntrico de validez y fiabilidad de la Escala de Conductas Indicadoras de Dolor ESCID para medir el dolor en pacientes críticos, no comunicativos y sometidos a ventilación mecánica”) after a peer-reviewed funding process. Also co-financed by FEDER Fund and Economy and Competitiveness Ministry.

I.L., M.S., M.A., L.H., study concept and design. I.L., M.S., data analysis and interpretation. I.L., M.S., M.A., L.H., manuscript preparation and drafting. I.L., M.S., statistical methods, statistical data analysis. C.L., M.M.S., M.W., J.d.P., G.R., R.d.B., M.J.F., C.G., J.d.F., C.C., acquisition of the data and manuscript critique and review. All authors approved the manuscript submitted.

The authors declare that they have no conflict of interest.