Introduction

Pain in newborns has been underestimated for decades, with a previous belief that neonates do not experience pain. Currently, it is well established that newborns can experience pain; evidence has shown that newborns experience an exaggerated pain response due to an immature nervous system and inhibitory pathways.1,2

Minor painful procedures (intramuscular injections, venipunctures, or heel pricks) are common in the neonatal population. The assessment of acute pain in neonates is challenging. It relies on scales developed based on behavioral, physiologic measures, or a combination of both to obtain a more objective evaluation. Standard pain scales for both preterm and term infants include the Neonatal Facial Coding system, the Neonatal Infant Pain Scale (NIPS), and the Premature Infant Pain Profile, among others.3 The NIPS scale is one of the well validated tools that relies on behavioral assessment for measuring procedural and post-procedural pain in both preterm and term neonates.

Studies have demonstrated that early pain experiences may alter pain responses later in life and lead to other long-term adverse outcomes.1,2 Hence, establishing an easy and reliable method to reduce pain in newborns is crucial. Pharmacological pain management is well described; however, there are concerns about its side effects and long-term neurodevelopmental outcomes. Therefore, non-pharmacological alternatives have been studied, such as oral sucrose,4 breastfeeding,5,6 parental skin-to-skin holding (Kangaroo mother care),7 and devices that mimic parental skin-to-skin holding.8 Besides the above interventions, alternatives such as music therapy through live music by music therapists, recorded music, and maternal singing have also been described to reduce pain during minor procedures.9,10 Music is reported to modulate pain response by causing sensorial saturation, leading to pain pathway blockage, distraction, and alteration of pain perception.11 Multiple other benefits of music in the NICU setting have also been demonstrated, such as positive changes in physiological parameters, behavioral distress, and parental anxiety.12,13,14

Small-scale prospective studies have shown that music causes a reduction in pain scores, primarily in preterm neonates who underwent minor procedures.10,15,16 However, a systematic review and meta-analysis have not found conclusive evidence of music’s analgesic efficacy, citing high heterogeneity and poor quality of studies.17,18 Furthermore, there continues to be a paucity of data from methodologically rigorous studies on music’s analgesic effect, specifically in term neonates who also experience several painful procedures, most commonly the routine newborn metabolic screening in the US.

Therefore, this study aims to assess the effectiveness of recorded music in a meticulously controlled setting as an adjuvant to the standard of care oral sucrose to relieve acute pain in term newborns undergoing heel pricks in a community hospital newborn nursery.

Material and methods

This randomized, controlled, blinded clinical trial was performed in the newborn nursery unit of Lincoln Medical & Mental Health Center, a community hospital in the Bronx, NY, from April 2019 to February 2020. Resident physicians in training performed the study under the guidance of an attending neonatologist. Inclusion criteria were any term neonate (born ≥37 weeks gestational age (GA)) undergoing painful minor procedures (heel pricks) in the nursery. Heel pricks were performed using an infant safety lancet (BD Quikheel™, Franklin Lakes, NJ) for either routine serum bilirubin check and/or newborn metabolic screening, in an otherwise well appearing newborn in the nursery. Exclusion criteria included GA <37 weeks, failed newborn hearing screening, major congenital malformation, neurologic disorders, or oxygen supplementation. This trial was approved by the local Institutional Review Board (Lincoln Medical & Mental Health Center) and was registered in the ClinicalTrials.gov registry (Identifier: NCT04313179). Written informed consent (approved by local IRB) was obtained from the infant’s legal guardian by an independent resident physician who did not assess the neonates during the intervention.

After informed consent was obtained, subjects were randomized via an online random sequence generator software by a person who was not going to assess pain, and the allocation was concealed until the study period was complete. The legal guardian was unaware of the group assignment until the intervention was completed. Subjects were randomized into two groups, either exposed to music (intervention group) or not exposed to music (control group).

Interventions

All infants in the study (i.e., in both groups) received a standard non-pharmacologic method of pain relief, 0.5 ml of 24% sucrose, 2 min before heel prick. In the music intervention group, neonates were exposed to recorded music starting 20 min before the heel prick to garner the full effect of music, and continued for 5 min after the procedure ended.19 The music chosen was an instrumental lullaby used in earlier studies—“Deep Sleep” track from “Bedtime Mozart: Classical Lullabies for Babies”.20 This music track was chosen due to its minor tones and soothing effect. Music was played via smartphone speakers (screen facing down) and placed near the head end of the bed near the baby’s vertex for binaural input. Sound levels were measured using a decibel meter (BAFX Products®, Muskego, WI) and maintained below the recommended AAP standard of 45dBA21 with the max transient volume of 60 dBA (1-s LMax). Music administration and regulation of sound levels were performed by an independent resident not performing pain assessment. After this setup, a separate investigator (pain assessor) entered the study room wearing an active noise-canceling Bluetooth headphones (also playing a random song to blind the intervention) and assessed the NIPS score. For the control group, a smartphone not playing music was placed in a similar position by the independent resident to improve the blinding of the pain assessor. A nursery nurse was present at the time of intervention but was not blinded to the intervention as they were not involved in the study.

The study was done in a quiet, dimly lighted room, ambient room temperature, in a bassinet, without a pacifier to reduce other sensory inputs and assess the pure effect of music. Neonates were not cuddled by their parents during the study period, as interventions such as kangaroo care are known to help calm the infant and potentially confound the effects of music.7

Measurement of pain

Pain was measured using the NIPS scale that assesses facial expression, crying, breathing patterns, limb movements, and arousal. Prior to the start of the study, individual assessors were trained by an attending neonatologist and nursery charge nurse to assess pain in neonates using NIPS scoring system. Since NIPS is a standardized pain scoring system with reported good inter-rater reliability, we did not test the inter-rater reliability among the assessors of this study. A NIPS score greater than 3 indicates the need for pain control.2 In this study, NIPS scores were assessed 5 min before heel prick, at the time of heel prick, and at 1-min intervals for 5 min post-heel prick.

The primary outcome of this study was to assess the effectiveness of music as reflected by a reduction in NIPS scores by 5 min post-procedure. A reduction of NIPS score by 3 points was considered clinically significant.

Sample size

Our initial sample size calculation resulted in 16 infants per group, based on a two-sided alpha of 0.05 and a power of 0.80 to detect a 3 units difference between groups, assuming a standard deviation (SD) of 3.22 We opined that this estimated sample size was small and hence reviewed other studies such as the RCT by Zhu et al.23 Based on the literature search and experts’ inputs, we concluded that a sample size of 100 infants per group (total of 200) would be ideal to detecting small differences in the primary outcome.

Statistical analysis

For primary outcome analysis, we determined that Generalized Linear Model24 was the most appropriate test for repeated measures to compare groups across time. In this model, NIPS score was treated as an ordinal measure, and the group result is summarized as an odds ratio of intervention to control group with a 95% confidence interval.

As supporting analyses, NIPS scores were also compared between groups at each timepoint using nonparametric Mann–Whitney U tests [medians (interquartile ranges) of NIPS scores are reported]. NIPS scores were also categorized into binary ʻno painʼ versus ʻpainʼ (<3 vs. 4–7), and longitudinal measurements of subjects’ binary measurements were analyzed using Generalized Estimating Equations,25 and odds ratio of intervention to control along with 95% confidence interval is reported.

Baseline mother-infant characteristics are reported as frequencies or means (with SD) and were analyzed using either the χ2 test for the categorical variables or the t-test for continuous variables. Statistical significance for all analyses was considered achieved if p < 0.05. Statistical analysis was performed using SPSS software version 25.0 (IBM SPSS Statistics for Windows, Armonk, NY).

Results

The study was conducted between April 2019 and February 2020. Due to the COVID-19 pandemic, further enrollment was not feasible, so the study was closed prior to the recruitment of the planned sample size. One hundred twelve eligible infants’ mothers were initially approached for the study. A total of 100 infants were finally enrolled, and all were randomized, resulting in 46 subjects in the control group and 54 in the music intervention group (shown in Fig. 1). Infants received the interventions per randomization and all 100 were analyzed using the intention to treat principle. The pain assessor accidentally noted the assignment for one neonate due to headphone failure, but this neonate was included in the final analysis in the assigned group.

Fig. 1: Flow diagram of patient allocation as per CONSORT guidelines.
figure 1

CONSORT flow diagram.

Full size image

Infant characteristics

There was no statistical difference in baseline characteristics between the study participants (Table 1). Among those neonates, 61% were of Hispanic ethnicity, and 53% were males. The overall mean GA was 39.2 weeks, with the procedure performed primarily at the postnatal age of 2 days. The procedure duration among both groups did not differ significantly and lasted for an average of 113 seconds. None of the infants underwent any painful procedure before the study, and no male infants underwent circumcision before the intervention. There were no adverse events or side effects in the intervention group.

Table 1 Demographics and observations.
Full size table

Pain scores

When NIPS measurements were analyzed as repeated measures, the music group had significantly lower scores [OR = 0.42 (0.31,0.56), p < 0.001]. Mean NIPS scores across all times are shown in Fig. 2. Baseline NIPS score at 5 min before heel prick was also noted to be similar across groups. When the duration of heel prick was incorporated into the model, statistical significance between groups was preserved (p < 0.001). There was no difference in NIPS response based on infant sex or time (no interaction effect).

Fig. 2: Mean NIPS scores measured across time and compared between groups.
figure 2

Mean NIPS scores.

Full size image

Mann–Whitney test at each time also resulted in statistically significant lower scores in the music group, except at +3 min post-heel prick [Table 2]. A review of NIPS +3 min data did not show any significant outliers to explain the non-significance. When NIPS measurements were categorized as pain vs. no pain, there continued to be statistically significant lower NIPS scores in the music group compared to control [OR = 0.39 (0.24, 0.64), p < 0.001].

Table 2 NIPS scores across groups.
Full size table

Discussion

To the best of our knowledge, this is the first randomized, controlled, blinded trial evaluating the effectiveness of music intervention as an adjuvant to the standard of care oral sucrose in reducing acute pain in term newborns undergoing minor painful procedures in a community hospital newborn nursery. Our study showed a statistically and clinically significant reduction in pain scores. The significance persisted during the investigated period and supports music intervention as an adjuvant pain-relieving therapy in stable, healthy term neonates.

Shah et al. compared the effectiveness of music, oral sucrose, and combination therapy in neonates >32 weeks of postmenstrual age undergoing a heel prick procedure in the NICU, enrolling 35 neonates with an average postmenstrual age of 35 weeks.20 There was no difference in pain scores between the music and sucrose groups, but there was better pain relief during heel prick when combined. The authors mentioned that prior painful exposures or analgesia during the NICU stay might have modulated pain perception and subsequent response. Our study is unique in studying only the term neonates’ pain response to music, and none of the neonates had any prior painful exposure.

In our study, consistent with the literature, NIPS scores were significantly lower in the music intervention group than in the control group across time. But in the supporting analysis using Mann–Whitney test, there was no statistical significance at 3 min post-procedure. However, this analysis does not consider time as a factor; hence, the main effect of music across time, as reflected by the primary analysis, is more relevant. We also noted a lower NIPS score at baseline in the intervention group and it is likely secondary to earlier exposure to music prior to intervention.

Studies have examined variations concerning cochlear sensitivity based on the infant’s sex, observing significant differences.14,16 Female infants were more responsive than male infants to music therapy, a diversity replicated in small studies.16 However, in our larger RCT setting, the difference in music response based on the infant’s sex was not replicated or statistically significant. Future trials should continue to explore this difference.

The authors acknowledge that this study is not without limitations. The use of oral sucrose for pain relief can potentially confound music’s analgesic effect. However, sucrose is our institution’s standard of care for a minor painful procedure in term neonates, and it would have been unethical not to provide any baseline analgesia to study the music’s unbiased effect. In addition, it has been shown that multisensorial stimulation achieves better pain relief than a single intervention.26 We also acknowledge that dissuading mothers from cuddling their neonates is not part of family-centered neonatal care and not our institution’s practice, but this was necessary to find the music’s discrete effect on pain. Future studies will need to consider the role of families as partners and healing touch in addition to music intervention.

Another limitation was the inability to completely ensure that the participants were not fed for at least 1 h before the procedure. Parents were aware of this need, but the feeding patterns could not be entirely coordinated due to infants rooming in with their mothers before the intervention. However, this made our study pragmatic. Other limitations include not evaluating physiologic responses to pain such as heart rate changes; the reason is to extrapolate our findings to community units with limited resources, using the NIPS scale that accounts only for behavioral responses, thus leading to easy widespread usage. Only one investigator scored the infant at a time in our study, with no video recording of the procedure, thus limiting additional evaluation ability. Video recording and a second review by an additional investigator such as an expert assessor or neonatal nurse could have avoided any potential observer bias. We also did not collect other important variables that could have affected the pain scores, such as breast or bottle feeding, and if skin to skin was performed prior to intervention of interest.

Despite the limitations stated, our trial showed a statistically and clinically significant reduction in the NIPS pain scores in term neonates undergoing heel sticks when exposed to recorded music. We suggest that future studies should also strongly consider exploring the effects of similar interventions, such as recorded parental voice instead of the lullaby music used in our study. Involving parents as partners in neonatal care is an underutilized approach and incorporating parental voice during painful procedures may be an effective alternative as well.

Conclusions

Music intervention is an easy, reproducible, and inexpensive tool for pain relief from minor procedures in healthy, term newborns. The study results can be applied to term newborn nurseries with limited resources in the US and potentially worldwide.