Pain is a serious health problem worldwide; however, it is often underestimated. A study conducted in a Brazilian teaching hospital found that pain relief was delivered to 80% of surveyed inpatients but only a half of them showed signs of pain . A study found that nearly 80% of pediatric inpatients experiences at least one painful procedure per day; however, only 30% of those painful procedures were managed and documented . Children suffer worse pain than adults do, particularly in those admitted in Pediatric Intensive Care Units (PICU) where painful medical procedures are performed daily . Children admitted to PICUs were younger, stayed longer in PICUs and in hospitals, suffered more serious ills and mortality rate compared to those admitted to the general Intensive Care Units (ICU) .
Surgeries and burns are the leading causes of admission in children to ICUs and PICUs [4-6]. Pediatric surgical admission rate at PICU was reported at 25%, of which 52% was due to post-surgical admission [4, 5]. Pediatric surgical mortality was recorded at 34% in the ICU . A prospective cohort study found that one pediatric patient suffered 11 distressful and painful procedures per day in surgical PICU . Burn is the leading cause of hospital admission for injury in children with a report of 20% of admission and associated with painful and distressful experiences [6, 8, 9]. Medical advances have helped reduce mortality rate of burned children [9, 10]. Those who do not die still suffer painful burn care procedures such as wound dressing replacement, physiotherapy practice, skin grafts, and surgery to remove scar tissue [9, 11-13]. However, pain of burned children was improperly managed . Effective pain management, including continuously monitoring and reporting, regularly and accurately assessing, and successfully addressing, would help patients recover more quickly and keep complications at the minimum rate [9,10,14].
A valid and reliable tool is needed to manage pain effectively. There are two main types of pain assessment tools: self-reporting and behavioral scales . Self-reporting scales are considered the gold standard in assessing children’s pain; however, behavioral scales are more effective in assessing pain of children with verbal or visual impairments [16-18]. The Comfort Behavior Scale (Comfort-B) was adapted by Monique van Dijk et al in 2005  from the original version developed by Ambuel et al. in 1992 . The Comfort-B consists of six items including alertness, calmness/agitation, crying (spontaneously breathing children)/ respiratory response (mechanically ventilated children), physical movement, muscle tone, and facial tension. Each item is scored from 1 to 5, making the possible pain score of 6 (no pain) to 30 (severe pain). The Comfort-B is a valid and reliable tool and has been used globally for assessing pediatric pain in ICUs [19,21-29]. Users are not required to have any special skills but suggested to attend an online training module for sufficiently assessing pain .
Assessing pain in PICUs has not been a routine practice in many Vietnamese hospitals as few studies have been conducted to provide supporting evidences. Moreover, few valid and reliable tools were available in Vietnamese . Therefore, this study cross-culturally adapted the Comfort-B into Vietnamese clinical settings and assessed its psychometric properties. Findings of this study provide a useful reference for practitioners to manage pain in PICUs, and for students and researchers to do their studies.
This study was conducted at the Burn Intensive Care Unit (BICU) and Surgical Intensive Care Unit (SICU) of the Nhi Dong 1 Hospital, a provincial pediatric hospital, located in Ho Chi Minh City. The BICU and SICU have 61 beds and 57 nurses. There are 30 burned children and 120 critically ill children admitted in BICU and SICU monthly.
This study was approved by the Ethics Committee of the University of Medicine and Pharmacy at Ho Chi Minh City (506/DHYD-HDDD, October 17, 2019) and the Executive Board of the Nhi Dong 1 Hospital (2882/QD-BVND1, October 29, 2019).
After getting a permission for translation and use of the Comfort-B, the translation process followed the process of translation and adaptation of instruments from World Health Organization including forward translation to Vietnamese, Expert’s Committee Review, English-back translation, and pretesting . The study process is presented in Figure 1.
A registered nurse and an English professor independently translated the Comfort-B into Vietnamese. The two Vietnamese versions were discussed among researchers for a unified version which were then reviewed by an expert panel consisting of 23 doctors and nurses of the Nhi Dong 1 Hospital as suggested by Lynn et al. . Experts were asked to rate the relevance and clarity of 30 items of the Comfort-B Vietnamese version. Relevance was a 4-point Likert scale including 1 (cannot be used, not relevant), 2 (cannot be used, item needs some revisions), 3 (relevant, may be used with minor revisions) and 4 (very relevant). Clarity was a 4-point Likert scale including 1 (cannot be used, not clear), 2 (cannot be used, item needs some revisions), 3 (clear, may be used with minor revisions) and 4 (very clear).
Experts were also asked to comment on the language to improve the clarity of the scale. Any comment was noted then discussed among researchers for a final Vietnamese version. This version was then translated back to an English version separately by another registered nurse and English professor. This English-back translated version was assessed against the original version by an expert panel consisting of five English-native healthcare professionals. Differences were reviewed and discussed between researchers and the expert panel until a consensus was reached.
Registered nurses, who are working at the BICU and the SICU of the Nhi Dong 1 Hospital and responsible for caring for children, were invited for the pilot study that assessed whether the Comfort-B was 1) clear, 2) understandable, 3) easy to use, 4) convenient for use, 5) not time-consuming, 6) helpful for nurses to decide medical care, 7) feasible and applicable to clinical settings, and 8) able to classify pain level. Each item, a 5-point Likert scale, was rated from 1 (strongly disagree) to 5 (strongly agree). The mean and standard deviation of eight items were calculated. These items had been successfully used in previous studies [34-36].
Eligible nurses were excluded if they were probationary or experienced less than one working year at the hospital. All nurses of the BICU and SICU were approached and screened for their eligibility by a researcher. Eligible nurses who agreed to attend the study were asked to sign an informed consent after receiving an information sheet explaining the study. They were trained to use the Vietnamese Comfort-B following the instruction of Monique van Dijk et al. (https://www.comfortassessment.nl/web/index.php)  before using it in the pilot.
Five children hospitalized in the BICU and prescribed wound dressing replacement were consecutively selected for video-recording. They were not recruited if they were mechanically ventilated, had unstable vital signs, had hearing impairment, or were unconscious. A researcher approached and explained the purpose of the study to parents of eligible children and asked their permissions for video-recording their children at three periods before, during and after wound dressing replacement. If parents permitted their children to join the study, they were asked to sign an informed consent before researchers recorded videos.
Nurses of the BICU and SICU were invited to watch 15 recorded videos and used the Vietnamese Comfort-B to assess pain at two time-points which were two weeks apart from each other. The two-week interval was considered appropriate to assess intra-rater reliability as it had been selected by previous studies [34,37]. The inclusion criteria were nurses working at the BICU or SICU of the Nhi Dong 1 Hospital and responsible for caring for children. Participants were excluded if they were probationary or experienced less than one working year. Thirty-four nurses agreed to participate in the first assessment but only 15 nurses agreed to do the second assessment. A laptop was used to show videos and record duration to complete an assessment. All informed consents were signed before any data was collected.
All content validity measures including the item content validity index (I-CVI), the scale content validity index universal agreement (S-CVI/UA), the scale content validity index average (S-CVI/Ave), and Kappa (K) were calculated.
where A is the number of experts rating 3 or 4 for relevancy/clarity, and N is the total number of experts [38,39]. I-CVI was appropriate if it is higher than 0.8 as suggested by previous studies [33, 38, 39]. S-CVI/UA was “the proportion of items on an instrument that achieved a rating of 3 or 4 by all the content experts” [33,38,40]. S-CVI/Ave was the average of all I-CVIs [33,38,40].
All nurses were asked to use the Vietnamese Comfort-B for assessing pain of 15 videos of five children recorded before, during and after wound dressing replacement. Two weeks after the first assessment, they were asked to re-assess the videos. The Cronbach’s alpha coefficients were calculated at the first and second assessment separately. The alpha coefficients were acceptable if they were above 0.7 .
Intra-rater reliability was calculated to assess a difference in the Comfort-B mean score of 15 raters between two assessments (first and second assessment) at three periods (before, during and after wound dressing replacement) using group-average intra-class correlation coefficient (ICC) with two-way mixed-effects absolute agreement model . Intrarater reliability of each of 15 raters between two assessments at three periods was also calculated using individual ICC with two-way mixed-effects absolute agreement model .
Inter-rater reliability was assessed using individual ICC with two-way random-effect absolute agreement model [31,43-46]. ICC was classified as poor (ICC <0.5), moderate (0.5 ≤ ICC ≤ 0.75), good (0.75 < ICC ≤ 0.9), and excellent (ICC > 0.9) reliability .
The Comfort-B was successfully translated into Vietnamese. One out of 23 Vietnamese members of the experts’ committee suggested that “ngủ không sâu” (item A2) and “tỉnh táo và hoạt bát” (item A4) should be replaced by “ngủ gà” and “tỉnh và chơi”. However, all members agreed to keep the original wording after the discussion. The experts’ committee suggested that “vận động” in items D1 to D5 should be replace by “cử động” and “phản ứng” in item A1 to A5 replaced by “đáp ứng”, “phản ứng quá mức” in item A5 should be replaced by “tăng đáp ứng”, “trẻ” replaced by “trẻ tỏ vẻ” in item B1, B3 and B4, “đau đớn” replaced by “lo lắng” in item B5, “thở êm” replaced by “thở đều” in item C1, “giãn hoàn toàn” replaced by “hoàn toàn thư giãn” in item E1, “sức kháng giảm” replaced by “kháng lực giảm” in item E2, “co gấp” replaced by “gấp chặt” in item E4, “gồng cứng” replaced by “co cứng” in item E5, “căng cơ ở vài vị trí cơ mặt” replaced by “căng cơ ở vài nhóm cơ mặt” in item F3, and “các cơ khắp mặt” replaced by “toàn bộ nhóm cơ mặt” in item F4.
All five English members of the experts’ committee agreed that the English-back translated version has similar meanings to the original version and can be used for research and clinical practice. An explanation “clearly noticeable during the 2 minutes observation” should be added to F3 and F4.
The I-CVIs of the Vietnamese Comfort-B for the relevance and clarity were above 0.95 for all items. S-CVIs/Ave were 1 for the relevance and 0.99 for the clarity. S-CVIs/UA was 0.93 (28/30) for the relevance and 0.87 (26/30) for the clarity. Modified Kappas of all items were above 0.95.
Table 2 shows the characteristics of the Vietnamese Comfort-B. All characteristics were rated 4 and above by 34 nurses of the hospital. Nurses take 1.6 (SD=0.6) minutes on average to complete the Comfort-B. They need more time to complete the Comfort-B during wound dressing replacement (2 minutes (SD=0.5)) compared to before (1.2 minutes (SD=0.4)) and after (1.5 minutes (SD=0.5)) wound dressing replacement.
The Cronbach’s alpha coefficients were 0.97 at the first and second assessment. The inter-rater reliability between 34 nurses was measured at the first assessment, ICCs indicated good agreement before (ICC=0.82; 95%CI: 0.61 to 0.97) and during wound dressing replacement (ICC=0.79; 95%CI: 0.56 to 0.97) and excellent agreement after wound dressing replacement (ICC=0.91; 95%CI: 0.78 to 0.99). The inter-rater reliability between 15 nurses was measured at the second assessment, ICCs indicated good agreement before (ICC=0.86; 95%CI: 0.66 to 0.98) and during (ICC=0.90; 95%CI: 0.74 to 0.99) and excellent agreement after wound dressing replacement (ICC=0.98; 95%CI: 0.94 to 1.00).
The intra-rater reliability of 15 nurses between the first and second assessment was measured using group-average ICCs indicating good agreement during wound dressing replacement (ICC=0.90; 95%CI: 0.78 to 1.01), but excellent agreement before (ICC=0.94; 95%CI: 0.78 to 1.10) and after (ICC=0.97; 95%CI: 0.92 to 1.02) wound dressing replacement (Table 3).
* Two-way random-effects absolute agreement model was used to assess inter-rater reliability between 34 nurses at the first assessment and between 15 nurses at the second assessment with before, during and after wound dressing replacement.
** 5 videos were assessed by 15 nurses assessed at two times; the second assessment was two weeks away from the first assessment. Average-group intraclass correlation coefficient was reported for the group intra-rater using two-way mixed effects absolute agreement model.
The multi-level linear regression generated no statistical difference in the Vietnamese Comfort-B mean score between the first and second assessment (b=-0.05; 95%CI: -0.61 to 0.51; p=0.86), but higher scores during (b=14.45; 95%CI: 13.77 to 15.13; p<0.01) and after (b=4.75; 95%CI: 4.07 to 5.43; p<0.01) compared to before wound dressing replacement (table 4).
|Second vs. First assessment||-0.05||0.86||-0.61; 0.51|
|During vs. Before wound dressing replacement||14.45||<0.01||13.77; 15.13|
|After vs. Before wound dressing replacement||4.75||<0.01||4.07; 5.43|
Coef.: regression coefficient; 95%CI: 95% Confident Interval Multi-level linear regression random effect.
This study translated and adapted the Comfort-B into Vietnamese using the standard process guided by the World Health Organisation . The standard process used forward and back-translation with Vietnamese-English bilingual and English-native expert committees that has been highly recommended . This process has been used sucessfully in a previous study validating Neonatal Infant Pain Scale (NIPS) conducted in Tien Giang General hospital, 70 km away from the south of Ho Chi Minh City .
The Vietnamese Comfort-B scale was strictly reviewed and approved by experienced nurses and doctors of the Nhi Dong 1 Hospital. Any disagreement on the Vietnamese Comfort-B between raters were carefully discussed to reach agreement. The English-back translation of the Comfort-B was strictly reviewed and approved by the English-native experienced healthcare professionals, including authors of the original Comfort-B, to ensure clear and correct meanings. The I-CVIs were all above 0.95 indicating relevance and clarity of translated items [33,38,39]. S-CIV/Aves and S-CIV/UAs were above 0.8 indicating that the scale was appropriately validated [33,38,39]. All modified Kappas were above 0.95 indicating an excellent agreement between raters [38,41]. All content validity and agreement measures of the Vietnamese Comfort-B indicating satisfactory translation.
All 57 nurses, working at BICU and SICU of the Nhi Dong 1 hospital, had at least one working year of experience and satisfied the study’s eligibility, of which 23 refused to participate. Nearly 60% (34/57) of nurses did the first assessment and 26% (15/57) did both first and second assessment that met the sample size requirement for a reliability study [38,39,43]. The sample nurses (34) agreed that the Vietnamese Comfort-B is clear, understandable, easy to use, convenient for nurses, and not time-consuming. Moreover, it helps nurses classify pain level and make decisions on medical care. Those nurses agreed that the Vietnamese Comfort-B is applicable and feasible in the hospital context. The Comfort-B is applicable to assess children’s pain in ICUs as suggested by previous studies [22,23,48]. A previous study showed that nurses might spend 30 seconds to complete the Comfort-B due to heavy workloads that lead to underscore the Comfort-B . It was suggested that nurses should take two minutes to complete the Comfort-B [19,49]. This study found that nurses took two minutes to complete the Comfort-B during wound dressing replacement but shorter time before and after wound dressing replacement.
The Vietnamese Comfort-B scale provides Cronbach’s alpha coefficients of 0.97 for the first and second assessment indicating that the scale has appropriate internal consistency. Tavakol and Dennick recommended that the scale should be shortened if its alpha coefficient is higher than 0.9 . However, Bland and Altman noted that alpha coefficient should be 0.95 as a minimum value for clinical application .
The average ICCs of 34 nurses at the first assessment and 15 nurses at the second assessment indicated good inter-rater reliability before, during and after wound dressing replacement (ICCs ≥0.79). The average-group ICCs of 15 nurses in a 2-week interval indicated good intra-rater reliability before, during and after wound dressing replacement (ICCs ≥0.9). The Vietnamese Comfort-B scale had 14.45 points (95CI%: 13.77 to 15.13) higher during wound dressing replacement compared to before wound dressing replacement indicating that the scale can classify the pain level at the cut-off point of 15. Its narrow confidence interval implies that the result is reliable . Previous studies suggested the cut-off point of 17 for the Comfort-B to classify pain and no pain [19,23,53], however, the Chinese Comfort-B suggested a cut-off point of 13 .
The original Comfort-B was developed to assess children’s pain and distress in PICUs, including those with mechanical ventilation; however, this study did not validate the ‘respiratory response’ category. There are various care procedures done in ICUs, this study focused only on wound dressing replacement in children, a burn care procedure. Another limitation of this study was that the Comfort-B can be used to assess pain in critically ill children, however, this study excluded children who had hearing impairment, unstable vital signs, or were unconscious. Further studies should be conducted to address these limitations. This study was conducted in one of three pediatric hospitals in Ho Chi Minh City, including Nhi Dong 1, Nhi Dong 2 and City Children’s Hospitals. The Nhi Dong 1 Hospital is one of the four best pediatric hospitals in Vietnam, the generalizability of the findings to other hospitals in the city and other provinces should be cautiously considered.
In conclusion, this study showed that the Vietnamese Comfort-B can detect pain at a cut-off point of 15 in burned children undergoing wound dressing replacement. As few Vietnamese pain assessment tools have been validated, this study provides preliminary evidences to support the Vietnamese Comfort-B for use in clinical practices and research at the local context.