Background Neonatal pain assessment generally requires access to facial expression. Improved neonatology practices, such as greater protection against bright lights and non-invasive mask ventilation, have made facial observation more difficult.
Objective To validate a ‘faceless’ acute neonatal pain scale (FANS), which does not depend on facial expression.
Methods In a prospective, multicentre study, 24–40-week-old neonates were videotaped during a painful procedure (heel prick). Three investigators then scored the pain using FANS and a previously validated scale: DAN (Douleur aiguë du Nouveau-né). FANS is based on assessment of limb movement, cry and autonomic reaction. Reliability was assessed by inter-rater agreement and internal consistency (Cronbach's α). Validity was evaluated by agreement between scales (intraclass correlation coefficient (ICC)). The Wilcoxon test evaluated the FANS score differences between conditions. Results are expressed as medians (25th and 75th percentiles). Ranges are presented for outcome parameters.
Results From April 2006 to September 2007, 53 preterms of 32 (30–35) gestational weeks and 1500 (1000–2200) g were observed. Cronbach's α was 0.72. The ICC was 0.92 (0.9–0.98) for inter-rater agreement and 0.88 (0.76–0.93) for agreement between scales.
Conclusion FANS, which is reliable and valid, is the first scale to score pain in preterm newborns when facial expression is not accessible.
Statistics from Altmetric.com
The pathways for pain conduction are present as of 23 weeks of gestation,1 but pain management is often delayed in neonatal intensive care units (NICUs) because the detection of pain and the assessment of its intensity can be difficult. Different scales are used, all based on the evaluation of several factors but having in common the study of facial expression, which is considered to be the most discriminating element in the evaluation of neonatal pain.2 3 However, facial expression has become less accessible over the past 15 years because of changes in treatment practices, particularly non-invasive mask ventilation and limited light exposure during medical procedures.4 A recent investigation in our unit revealed that facial expression was inaccessible in 39% of the newborns during acute care, either because they were being protected from bright lights (2/3 cases) or because they were wearing non-invasive ventilation masks (1/3) (personal communication). Many pain scales have been developed, and the detection and prevention of pain have been greatly improved.5 The validation of a scale adapted to the changes in NICU care thus seemed necessary.
What is already known on this topic
▶. Premature neonates experience pain. Treatment is based on reliable evaluation and is adapted to the clinical situation of each infant.
▶. All scales for evaluating neonatal pain incorporate items on facial expression.
▶. Neonatal facial expression is often inaccessible to the care giver because current neonatology practices include protection against bright lights and non-invasive facial masks for ventilation.
What this study adds
▶. This study proposes the first scale for evaluating acute pain in newborns when the face is not accessible.
▶. This simple and easy to use scale can be used at the bedside.
▶. This scale was validated by comparison with a validated scale (Douleur aiguë du Nouveau-né (DAN)), with an intraclass correlation coefficient of 0.88 (0.76–0.93).
The objective of this study was therefore to validate a new scale to evaluate pain in neonates whose faces were not visible to the care giver and to determine whether this scale was able to differentiate a painful procedure from non-nociceptive stimulation.
Design and setting
We carried out this multicentre prospective study in the level III NICUs of Montpellier, Nimes and Perpignan, France. All neonates born at 24–40 weeks of gestation were eligible if their faces were visible. Thus, newborns who were intubated or who were on non-invasive ventilation were not eligible.
One videotape per newborn was made. Only the tapes of sufficient quality to clearly observe facial expression and limb movement and to hear vocal expressions were retained. The newborns received analgesia according to our unit's pain treatment policy (sucrose sucking: 24% of 0.1 ml before 34 weeks and 1 ml from 34 to 40 weeks, 3 min before the acute care).6 The newborns were videotaped for 3 min before the procedure and throughout the procedure. In all conditions, heart rate and oxygen saturation were monitored continuously (Viridia M3M3046A; Philips, Suresnes, France). The videotapes were anonymised and then viewed by three independent observers who filled out the Douleur aiguë du Nouveau-né (DAN) scale. A month later, the videotapes were randomly viewed a second time, and the pain was assessed using the faceless acute neonatal pain scale (FANS). The observers were blind from the DAN score previously realised. The procedures carried out in this study are current practices in our neonatology unit and thus did not require specific approval from an ethics committee, according to French law. The parents were informed of the study, however, and gave signed consent to have their children videotaped.
FANS does not depend on the observation of facial expression. It had been constructed during a preliminary study of 20 newborns hospitalised in the Montpellier University Hospital neonatology unit. This instrument, which was inspired by other validated scales (table 1),7,–,14 evaluates behavioural items (body movements, vocal expression) and physio logical items (variation in heart rate, bradycardia or desaturation) (table 2) on a scale from 0 (no pain) to 10 (maximal pain). The increase in heart rate, a reflection of sympathetic nervous system activation, is a very sensitive item but its specificity is less so, and thus vagal tone was also taken into consideration by inclusion of a ‘malaise’ item that comprised bradycardia and desaturation.15 16 To increase the specificity of this scale, behavioural elements were given particular weighting,17 18 as they are strongly associated with acute pain in the preterm baby.19
The reference scale for comparing FANS was DAN,20 which has been validated in the premature newborn, discriminates pain well, and presents good inter-rater correlation and good internal consistency (table 1).21
In our preliminary study of 20 newborns, the intraclass correlation coefficient (ICC) between FANS and DAN was 0.7 (0.38–0.87). From this value, we calculated that it would be necessary to include at least 40 patients in order to reduce the CI to (0.5 to 0.8). The inter-rater agreement was calculated by randomly analysing the scale filled out by one of the observers and then comparing it to the data reported by the two other observers.
To validate FANS, we assessed the reliability of the scale (that is, the inter-rater agreement and internal consistency) by calculating Cronbach's α. The coefficient varies from 0 to 1 and determines the contribution of each item to the global power of the scale, with each item being compared with the totality of the items.22 Internal consistency is considered to be high when the coefficient is between 0.77 and 0.99.23
We assessed the validity of FANS (that is, its capacity to describe a painful event) by comparing the FANS results with those obtained using the reference scale (DAN). To do so, we calculated the ICC between the two scales. The Bland–Altman analysis assessed the agreement between the two scales.24 The results are expressed as medians (25th–75th percentiles). For the outcome parameters, ranges are presented. The statistical analyses were performed with Statistical Analysis Systems software (SAS Institute, Cary, North Carolina, USA).
Between April 2006 and September 2007, 67 newborns were videotaped during acute care. We rejected 14 tapes for technical reasons (insufficient light, face not visible) of the 53 analysed recordings.
The infants were born at 32 (30–35) weeks of gestation; birth weight was 1500 (1000–2200) g. Approximately one out of four infants (26%) had a birth weight < −2 SD25 and 31 (58%) were boys. The videotapes were made at 6 (2–18) days postnatally. The median scores during heel pricking were, respectively, 2 (1–4) (range: 0–7) and 1 (0.5–4) (range: 0–9) for FANS and DAN. Inter-rater agreement was 0.92 (0.9–0.98); the internal consistency of the items evaluated by Cronbach's α was 0.72. The ICC coefficient between the FANS and DAN scores was 0.88 (0.76–0.93) (figure 1). The difference between FANS and DAN remained stable whatever the gestational age (figure 2).
The Bland–Altman analysis showed that the disagreement between the two scales, estimated by the 95% CI, was 2 (95% CI 1.5 to 2.7). The pain scores provided by FANS tended to be higher than those provided using the DAN scale for moderate pain (average score below 3) (figure 3).
Our results indicate that the new pain scale is reliable, well correlated with the reference scale, and able to discriminate reliably between painful procedures and non-painful stimulation.
The construction of the scale from four items (heart rate variation, malaise, limb movement and cries) seemed consistent because Cronbach's α between the different items was high. The number of participants evaluated was similar to that in other scale validation studies.7 10 The inter-rater correlation was very high and comparable to that of other scales.
In line with other studies, we compared FANS to a reference scale (DAN), as there are no objective markers of pain, and in particular no validated biological marker to evaluate transitory nociceptive stimulation in the newborn. Grunau et al observed a diminution in cortisol secretion in response to stress in preterms of less than 28 gestational weeks, but this population was subjected to a high number of painful procedures.26 Moreover, physical measures such as transfontanel pressure or cerebral blood flow, skin conductance, electroencephalogram (EEG), functional MRI and near infrared spectroscopy are invasive techniques that are not routinely performed and that require clinical validation.27,–,30 A limitation of FANS is that it is appropriate only in infants who are not intubated, so that vocal expressions can be scored.
The expression of pain varies with gestational age, and pain assessment tools must thus be pertinent for 24-week-old preterms presenting poor motor control, as well as for near-term infants. The odds that preterm babies would demonstrate changes in their behavioural and physiological pain indicators were observed to be 0.94 times greater per week of postconceptional age at birth.31 32 In a recent study, however, Stevens et al noted that these behavioural and physiological differences between groups of newborns with different gestational ages were neither clinically nor statistically significant.33 In our population, the correlation between the scores given by FANS and DAN remained stable for all gestational ages between 30 and 35 weeks.
Most of the FANS scores were at the lower end of the scale, between 1 and 4. This was explained in part by the type of intervention (capillary or venous puncture), which caused moderate pain, and also by the use of an effective analgesic, in line with the current recommendations.34 These moderately painful procedures are the most frequently practised in neonatology.35,–,38 FANS seems to be quite suitable to screen for pain as it slightly overestimates moderate pain (by 1–2 points) when compared to the DAN scale. This is related to the particular sensitivity of the variation in heart rate, which is first to change in moderate pain.38
To our knowledge, FANS is the first scale that permits the acute pain of non-intubated, premature newborns to be scored when the face is not visible, a situation that is increasingly encountered. This simple and easy to use scale can be used at the bedside. It shows good inter-rater agreement and is well correlated with DAN, especially for moderate pain. In conclusion, it is well adapted to the improved care of preterm neonates seen in NICUs over the last several years.
The authors thank C Stott-Carmeni for translating the manuscript and for editorial assistance. The authors are grateful to the nurses who participated to this study Catherine Bouchbacher; Patricia Founier and Leticia Laurent.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed
Patient consent Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.