A randomized controlled trial. Based on the findings of a previous study examining differences in subjective pain perception between male and female children [24] , sex-based stratified randomization was performed. Using a table of random sampling numbers, the researchers randomized subjects into Music and Control groups in the order of reservation.

The study was approved by the Ethics Committee of Hiroshima University (approval number: C-103) and registered with the UMIN Clinical Trials Registry (UMIN000023582). It conforms to the CONSORT guidelines for the reporting of randomized controlled trials.

The subjects’ assent was obtained from their parents as their legal representatives. When the parents of candidate children made reservations for an influenza immunization by telephone, staff of the collaborative study facility explained the study objective to them, and confirmed their children’s intention to participate. To the parents of those who assented to participate in the study, explanatory documents outlining the study for parents and children, informed assent and assent cancellation forms, and a questionnaire regarding immunizations for children (the numbers of past NICU admissions and surgeries) were mailed or directly handed over by staff of the collaborative study facility. On the day of immunization, the researchers provided sufficient written and oral explanations of the study to obtain written assent from the parents. The children were also provided with explanation using a picture book to obtain their oral assent.

The subjects were recruited at A clinic in Hiroshima from 9:00 to 18:00 each day during the period between October and December 2016. The inclusion and exclusion criteria were as follows: inclusion: (1) preschoolers aged 5 or 6 and scheduled only for an influenza immunization, and (2) those approved for the immunization by the physician in charge; and exclusion: (1) those with a history of treatment using analgesics, (2) those with auditory or visual impairment, (3) those with intellectual or developmental impairment, and (4) those regarded as inappropriate for the study by the physician in charge.

When determining the sample size, a study conducted by Ogawa et al., involving 32 children [25] , was referred to. This previous study compared the children’s dental treatment-related stress based on their sAA levels, and revealed significant decreases in such levels after treatment. Considering that a comparison of sAA levels similar to this would be feasible using controls, the size of each group was set at 50 in the present study.

The sex and age were used as demographic variables, and the following items were used as confounders: the numbers of past NICU admissions, hospitalizations, surgeries, immunizations, injections, influenza vaccinations during this season, and elder and younger siblings who attended; minimum and maximum sound levels in the treatment and waiting rooms; room temperature and humidity; and the time to initiate the study, period from the initiation to immunization, and duration.

The primary outcome variable was the Face Scale score. The Face Scale developed by Wong-Baker shows a series of faces expressing 6 different levels of pain from (0) to (5) [26] . The scale has been shown to have sufficient reliability and validity [27] . Its Japanese version has also been proven to have sufficient concurrent validity and reliability based on re-test results [28] . The guidelines for the assessment of acute pain in children [29] recommend the use of the scale from the age of 3. It is particularly frequently used in pediatrics [26] , such as pain assessment for children during blood sampling [15] [30] .

The secondary outcome variables were: the heart rate (HR) and oxygen saturation (SpO₂) and salivary alpha amylase (sAA) levels. The HR and SpO₂ level are used as indices to assess pain in children undergoing painful medical procedures, as they were shown to markedly increase after immunization, blood sampling, or acupuncture in previous studies [13] [16] [21] [31] . The sAA level that increases with stress [32] is used for stress assessment for children undergoing blood sampling [30] or receiving dental treatment [25] .

The demographic variables, confounders, Face Scale score, HR, and sAA level were compared between the Music and Control groups. The categorical variables were examined using Fisher’s exact test, while the continuous variables were analyzed using the Wilcoxon rank sum test or unpaired t-test based on the results of the Shapiro-Wilk test.

Among the results obtained through 30 repetitions of measurement at intervals of 2 seconds for 1 minute immediately after pulse oximeter attachment, the maximum HR and minimum SpO₂ level were adopted as baseline values. In addition, the maximum HR and minimum SpO₂ level obtained through 15 repetitions of measurement at intervals of 2 seconds for 30 seconds during the period from 1 minute after pulse oximeter attachment to 10 minutes after immunization were adopted as subsequent values [36] .

To examine the changes of the sAA level or Face Scale score between baseline and each of time point, the paired t-test or signed rank sum test was conducted based on the Shapiro-Wilk test results, and the Bonferroni correction was performed. Furthermore, the HR and SpO₂ level were compared using one-way analysis of variance (Dunnett test). For analysis, the statistical software SAS9.4 was used.

There were no significant differences between the groups in the Face Scale score, HR, or sAA level at the baseline (Figure 3, Figure 4 and Figure 5). As the Control group showed a significantly higher baseline SpO₂ level (P = 0.0297), analysis of covariance was used to adjust the value for comparison between them (Figure 6).

At the 12 points from 1 minute after the initiation of measurement to immediately before immunization, there were no significant differences between the Music and Control groups in the HR or SpO₂ level. Similarly, there were no significant differences between them in the Face Scale score or sAA level at 5 minutes after the initiation of musical intervention (6 minutes after the initiation of measurement in the Control group).

None of the 3 items, the Face Scale score, HR, and SpO₂ level, markedly varied

^aChi square test P < 0.05, ^bUnpaired t-test P < 0.05, ^cFisher’s exact test P < 0.05, ^dWilcoxon rank sum test P < 0.05.

between the Music and Control groups immediately after immunization. Only the SpO₂ level was significantly higher at 30 seconds after immunization in the Control group than in the Music groups (P = 0.0489). The HR did not markedly vary between the groups at any of the 19 points from 30 seconds to10 minutes after immunization. There were no significant differences in the Face Scale score or sAA level between them at 10 minutes after immunization.

The results were the same when including and excluding 4 preschoolers with a history of NICU admission.

In both the Music and Control groups, the Face Scale score and HR markedly increased immediately after immunization, supporting the findings of the previous studies (reporting pain in children receiving immunization, blood sampling, and acupuncture, respectively) [16] [21] [30] . Thus, influenza immunization was painful for the children.

There were no significant differences between the Music and Control groups in the Face Scale score, HR, or SpO₂ level at any point, excluding the third item measured 30 seconds after immunization, indicating that music was not effective to reduce immunization-related pain.

This result may have been due to a low level of pain, in addition to the small sample size. For example, in a previous study, music from speakers effectively reduced pain in children, including preschoolers, undergoing blood sampling [13] . As their HRs and Face Scale scores immediately after blood sampling were higher than the values in the present study [13] , it is likely that the procedure was more painful than immunization. The sample size of another study, in which music from headphones reduced immunization-related pain in preschoolers was 200 [19] , while those of 2 studies that did not achieve such effects were 99 and 62, respectively [21] [22] . In contrast, the sample size of the present study was 48, which is even smaller than the latter. This suggests that sufficient pain-relieving effects were not achieved using music, as the level of immunization-related pain was lower than that due to blood sampling, and the sample size was insufficient.

The sAA level did not markedly vary between the Music and Control groups at any point. Furthermore, in both groups, it did not show a significant difference from the baseline value before/after immunization, revealing that music did not reduce immunization-related stress. As a factor associated with this, the high number of past immunizations among the subjects (28.83 ± 6.04), which had possibly made this procedure familiar to them, should be noted. In a previous study involving 11 children aged 4 - 15, their sAA levels during blood sampling were markedly higher than those after it and the standard values, indicating that the procedure was stressful for them [30] . In contrast, as the Control group’s sAA level did not markedly change after the baseline in the present study, immunization may have been less stressful, making it difficult to achieve clear pain-relieving effects. In short, a stress level lower than that when undergoing blood sampling, as well as an insufficient sample size, may have prevented music from effectively reducing immunization-related stress. It is also undeniable that such stress was not accurately evaluated, as sAA measurement was not performed immediately after immunization.

As study limitations, both the sample size and measurement variables were insufficient. Among previous studies using scales to measure distress, 2, in which music reduced immunization-related pain in preschoolers, used the Child-Adult Medical Procedure Interaction Scale-Revised (CAMPIS-R) and Observational Scale of Behavioral Distress (OSBD), respectively [21] [34] . These scales are multi-item objective indices with sufficient reliability and validity. Therefore, it may be necessary to further examine the pain-relieving effect of music using them, with immunization-related distress incorporated.

Furthermore, as topical local anesthetics (EMLA CREAM) have also been reported to reduce immunization-related pain [37] , intervention methods combining multiple approaches, such as these drugs and music that may show pain-relieving effects, should also be considered to reduce immunization-related pain and distress in preschoolers.