Dynamic assessment of pain, fear, and physiological stress during pediatric intravenous cannulation: a randomized controlled trial of virtual reality distraction Authors: Shuai Xie, Feng Yin, Miyan Wang, Fan Wang, Yan Zhu, Yan Zhao, Yun Qiu, Xiaohong Chen, Yuwei Yang Published: 2026 ABSTRACT AbstractBackground and purposeMost studies confirm the effectiveness of virtual reality (VR) in alleviating pain and fear during pediatric venipuncture, with limited evidence regarding physiological stress markers and dynamic assessment. This study aims to investigate the effects of VR technology on pain, fear, and physiological responses in children, as well as family satisfaction during intravenous cannulation.MethodsThis single-blind randomized controlled trial was conducted in the Pediatrics Department of Mianyang Central Hospital, School of Medicine, University of Electronic Science and Technology of China. A total of 188 children requiring intravenous cannulation procedures and their families were allocated into: the control group (n = 96), which received the conventional verbal comfort; and the experimental group (n = 92), which wore VR headsets. The primary outcomes were dynamic fear/pain scores assessed by nurses and self-reported by patients, which were surveyed using the Children’s Emotional Manifestation Scale (CEMS) and the Children’s Fear Scale (CFS)/the Face, Legs, Activity, Crying, and Consolability scale (FLACC) and the Wong-Baker FACES Pain Rating Scale (Wong-Baker), respectively. The secondary outcomes were dynamic heart rate (HR) and peripheral oxygen saturation (SpO2) measurement, and family satisfaction levels. The pre-specified time points were pre-intervention, immediately post-intervention, at tourniquet application, at venipuncture, and one minute after venipuncture. Results were reported following the Consolidated Standards of Reporting Trials 2025 guidelines.ResultsVR significantly reduced children’s fear at two pre-venipuncture time points (nurse-assessed CEMS and child-reported CFS; partial η2 = 0.072 and 0.053, both p < 0.05). There were no significant group effects in nurse-assessed FLACC and child-reported Wong-Baker pain scores across three venipuncture time points. However, a significant time-by-group interaction was observed for nurse-assessed FLACC scores (partial η2 = 0.028, p = 0.006). Notably, nurse-assessed FLACC scores were lower than child-reported Wong-Baker scores in both control and experimental groups (Mean Diff.=-1.04 to -2.10, all p < 0.001). The dynamic changes in HR and SpO2 across five time points were lower within the experimental group (partial η2 = 0.106 and 0.045, both p < 0.01). Partial correlation analysis revealed that family satisfaction responded to children’s pain, fear, and physiological stress in a “process-sensitive” pattern in the control group but an “outcome-oriented” pattern in the experimental group.ConclusionsIn this trial, VR distraction effectively reduced procedural fear (but not pain) during venipuncture. However, the nurse assessment appears to suggest that the effects of VR distraction and conventional intervention on pain experience may vary over time. These findings underscore the importance of multidimensional pain/fear assessment in pediatrics.Implications for nursing managementOur study confirms that VR distraction can significantly reduce fear in children undergoing intravenous cannulation, though pain scores were not significantly lowered. Multidimensional assessment (including subjective fear measures) is therefore important. VR may improve the patient and family experience of painful procedures. We recommend further trials with larger samples to confirm these findings and refine VR interventions.Trial registrationChinese Clinical Trial Registry, ChiCTR2500113849 (Registration Date: Dec. 3, 2025; Retrospectively registered).Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12912-026-04906-3.Keywords: Pediatrics, Children, Intravenous indwelling needle, Virtual reality technology, Pain, Fear, Physiological stress Introduction Intravenous indwelling needle insertion is the most frequently performed invasive intravenous cannulation procedures in pediatric care and represents a primary source of pain and fear in children [1, 2]. Venipuncture constitutes the central procedural step and is followed by needle indwelling. Distress responses and negative emotions experienced during venipuncture may contribute to the development of needle phobia and negatively affect future medical compliance [3, 4]. To alleviate procedural pain, both pharmacological and non-pharmacological interventions are commonly used in clinical practice. Pharmacological approaches generally provide significant analgesic effects, including oral melatonin, topical anesthetics, opioids, alpha-2 agonists, N-methyl-D-aspartate receptor antagonists, and other analgesics [5, 6]; however, these agents often have a relatively long onset time and limited effectiveness in reducing fear [7, 8]. In contrast, non-pharmacological interventions, particularly distraction techniques, have become an essential component of pediatric pain management because of their rapid onset, noninvasive nature, and favorable safety profile [9]. In recent years, a range of non-pharmacological interventions has been widely used to alleviate procedural pain, including soap bubbles, coughing maneuvers, distraction cards or cartoons, the Buzzy device, and virtual reality (VR) techniques [10–13]. A randomized controlled trial revealed no significant differences in pain or fear reduction during pediatric venipuncture among soap bubbles, distraction cards or cartoons, and coughing [10]. The Buzzy device is an easy-to-use, reusable tool designed to alleviate injection pain in children. Shaped like a bee, it produces a “buzzing” sensation through vibration, and its wings deliver cold to the injection site, thereby helping to reduce injection pain [14–16]. However, whether its effectiveness in reducing pain and fear is superior to that of other methods remains controversial [11, 17]. As a highly immersive distraction modality, VR can effectively engage children’s visual, auditory, and cognitive channels, thereby competitively attenuating the brain’s perception and processing of pain signals [13, 18, 19]. This competitive perceptual inhibition results in a substantial reduction in self-reported pain and fear during pediatric medical procedures such as venipuncture, vaccination, wound dressing changes, and suturing [12, 20]. Overall, accumulating evidence from randomized controlled trials has demonstrated promising results for the use of VR to alleviate needle-related procedural pain in children [21, 22], although one study reported a nonsignificant effect [23]. Although the majority of existing studies have reported positive outcomes, they are often limited by several methodological constraints, including small sample sizes, reliance on nurse-administered or patient-reported pain assessments, limited incorporation of physiological outcome measures, lack of dynamic data collection at key procedural stages, and failure to tailor VR content to children’s developmental stages and age-specific needs. This study aimed to develop age-appropriate VR interventions and to dynamically assess pain, fear, and physiological stress during intravenous indwelling needle procedures in children. It also examined factors influencing family satisfaction by linking these dynamic assessments with procedural conditions. We hypothesized that (1) VR distraction would reduce procedural pain, fear, and physiological stress during pediatric intravenous cannulation, and (2) dynamic assessment would provide stronger empirical evidence to explain the pain experience. The findings are expected to provide empirical support for the use of VR technology to mitigate pediatric needle-related distress, enhance family satisfaction, promote the standardized implementation of VR in pediatric settings, and optimize the allocation of healthcare resources. Method Study design This single-center randomized controlled trial was pre-registered on the National Universal Health Security Information Platform of the National Health Commission (Registration No. MR-51-25-064041; Date: May 29, 2025), and retrospectively registered at the Chinese Clinical Trial Registry (Registration No.: ChiCTR2500113849; Date: Dec. 3, 2025). The study was reported in accordance with the Consolidated Standards of Reporting Trials statement checklist. Children and grouping From June to October 2025, children requiring intravenous cannulation were recruited from the Department of Pediatrics at Mianyang Central Hospital, School of Medicine, University of Electronic Science and Technology of China. The inclusion criteria were as follows: (1) age between 6 and 14 years; (2) ability to communicate effectively and understand instructions; (3) adequate peripheral vascular status and venous filling; and (4) verbal assent from the child and written informed consent from a parent or legal guardian, with voluntary participation and full cooperation. The exclusion criteria were as follows: (1) visual or hearing impairment; (2) a psychiatric diagnosis; (3) use of analgesics within the past week or the presence of acute or chronic pain; (4) inability to comply with study procedures, including refusal or failure to wear the VR headset or lack of cooperation during intravenous cannulation; and (5) withdrawal from the study at any time. The researcher generated six age- and sex-stratified randomization schedules (age groups: 6–8, 9–11, and 12–14 years; sex: male and female) and sealed them in opaque envelopes, which were provided to the intervention nurses responsible for participant recruitment and allocation. Based on these schedules, participants were randomly assigned to either the control group (Con group) or the experimental group (Exp group). All participants successfully completed the intravenous cannulation procedure within a few minutes, and no dropouts occurred. Therefore, a participant flow diagram was deemed unnecessary. The minimum sample size was calculated based on the primary outcome, namely the pain score during venipuncture. Using PASS software, we determined that at least 85 participants per group were required under the following assumptions: (1) an expected between-group difference of 0.5 under the null hypothesis versus 1.0 under the alternative hypothesis) (H₀: δ₀ = 0.5; H1: δ1 = 1.0); (2) a homogeneous within-group standard deviation not exceeding 1.0 (σ = 1.0); (3) a two-sided significance level of 0.05 (α = 0.05) and a statistical power of 0.90 (1 − β = 0.90); (4) equal allocation to both groups; and (5) no adjustment for dropout because no follow-up was conducted. VR-based intervention design From April to May 2025, three VR scenes were designed according to the developmental characteristics of different pediatric age groups. For children aged 6 to 8 years, the VR scenes incorporated bright colors, familiar cartoon imagery, and simple narratives to immerse them in interactive environments such as forest exploration, farm games, and “Paw Patrol: Mighty Pups Rescue.” Such content was intended to alleviate children’s fear of unfamiliar medical environments and clinical procedures. For children aged 9 to 11 years, the interactivity of the VR scenes was enhanced by providing more adventurous and exploratory experiences, such as virtual zoo tours or space adventures, along with simple interactive tasks (e.g., collecting stars or rescuing virtual animals) to improve focus and engagement. Additionally, clips from popular animated films, such as “The Monkey King: Hero Is Back,” “The Lion King,” and “Tom and Jerry,” were selected according to individual preferences, which helped children gradually relax and reduce fear. For adolescents aged 12 to 14 years, the VR scenes emphasized greater depth of interaction and environmental complexity, thereby encouraging immersion in challenging, task-based, or game-like adventures. Alternatively, immersive animated clips with rich storylines, such as “One Piece,” “Naruto,” and “My Neighbor Totoro,” were selected according to personal interests. By leveraging immersive adventures or highly engaging storylines to capture attention, these clips effectively diverted focus from the venipuncture procedure, thereby reducing perceived pain and fear. Intervention implementation and quality control A single-blind design was adopted, in which the outcome assessors were blinded, whereas the intervention nurses, children, and their families were not. The intervention measures were administered in two independent and enclosed rooms. In each room, one nurse performed standard intravenous indwelling needle insertion with family assistance, while another nurse monitored and recorded the child’s pain, fear, and physiological responses during the procedure. Before puncture, the control group received conventional verbal comfort supplemented by occasional distraction (e.g., watching cartoons or providing toys) when necessary, whereas the experimental group wore VR headsets displaying age- and interest-appropriate scenes tailored to their current emotional state. When children in the control group were calm or those in the experimental group were fully immersed in the virtual environment, the nurses performed intravenous indwelling needle insertion with assistance from family members. The puncture was performed as quickly and accurately as possible to minimize pain and discomfort. The puncturing nurse continuously monitored the child’s manifestations of pain and fear and sought assistance from the monitoring nurse, if necessary, to soothe the child. Finally, the blinded outcome assessors collected child-reported and nurse-assessed pain and fear scores based on the children’s verbal reports and the monitoring nurse’s records, along with data on family satisfaction. Data acquisition The primary outcomes were dynamic pain and fear scores, and the secondary outcomes were dynamic heart rate (HR), peripheral oxygen saturation (SpO2), and family satisfaction scores. All dynamic measurements were collected at five time points: pre-intervention (T0), immediately post-intervention (T1), at tourniquet application (T2), at venipuncture (T3), and one minute after venipuncture (T4). HR and SpO2 were recorded at all time points to enable comparison with baseline values. Fear was assessed only at T0 and T1 to evaluate the effects of the two interventions (verbal comfort versus VR headset use). Pain responses were evaluated from T2 to T4, as assessment at T0 and T1 had limited clinical relevance in children without pre-existing pain. Additional data were collected, including sociodemographic characteristics, venipuncture-related variables, and family satisfaction scores. Survey instruments Fear was assessed using the nurse-rated Children’s Emotional Manifestation Scale (CEMS) [24] and the child-reported Children’s Fear Scale (CFS) [25]. The original CEMS version was developed in Chinese [24], whereas the CFS started to be used after the simple translation of verbal labels corresponding to facial expressions [26]. The monitoring nurse used the CEMS to assess children’s fear across five dimensions: facial expression, vocalization, activity, interaction, and cooperation. Each dimension was scored from 1 to 5 using a 5-point Likert scale, with total scores ranging from 5 to 25. Higher scores indicated greater procedural fear. The reported interrater reliability, Cronbach’s alpha coefficient, and content validity index were 0.96, 0.92, and 0.96 among children in Hong Kong, China, respectively [24]. Child-reported fear was assessed after completion of intravenous cannulation. The CFS, adapted from the Faces Fear Scale [27], was displayed on a laminated card measuring 11 × 42 cm and depicted five facial expressions corresponding to different levels of fear, with scores ranging from 0 (no fear) to 4 (extreme fear). Children were asked to select the facial expression at T0 and T1 that best reflected their emotional state at those time points. Among Canadian children, the construct validity and interrater reliability of the CFS were supported by strong correlations with the Children’s Fear and Pain Scale (0.73 at venipuncture and 0.87 two weeks after venipuncture) and moderate correlations between parent- and child-reported CFS scores (0.51 at venipuncture and 0.60 two weeks after venipuncture), respectively [25]. Test–retest reliability was demonstrated by strong correlations in CFS scores between venipuncture and two weeks later (0.76 among children and 0.72 among parents) [25]. Pain was measured using the nurse-rated Face, Legs, Activity, Crying, Consolability (FLACC) scale [28] and the child-reported Wong–Baker FACES Pain Rating Scale (Wong–Baker) [29]. The FLACC was translated into a Chinese version by Liu etc [30]., and the Wong–Baker was widely used after the simple translation of verbal labels corresponding to facial expressions [31]. The monitoring nurse documented children’s immediate behavioral responses across the five domains of the FLACC scale. Each domain comprised three categories and was scored from 0 to 2. The total score ranged from 0 to 10, with higher scores indicating greater pain severity. The reported Cronbach’s alpha coefficient was 0.853, test-retest reliability ranged from 0.706 to 0.866, and inter-rater reliability ranged from 0.966 to 0.993 in China [32]. The Wong–Baker scale was presented on a laminated card measuring 9 × 48 cm and displayed six facial expressions (scored from 0 to 10 in 2-point increments) representing increasing pain intensity: “no pain,” “mild pain,” “moderate pain,” “severe pain,” “very severe pain,” and “extreme pain.” After completion of intravenous cannulation, children were asked to select the facial expression that best reflected their pain. The single-dimensional Wong–Baker scale demonstrated a high correlation of 0.922 with the FLACC scale and a test-retest reliability of 0.947 among Chinese children [31]. HR and SpO2 were measured using a CEM OX-97 finger pulse oximeter (China Everbest Machinery Industry Co., Ltd., Shenzhen, China). Sociodemographic characteristics of patients and their family members were extracted from the medical records, including the child’s age, sex, and clinical diagnosis, as well as the family member’s sex, age, educational level, and kinship with patient. Venipuncture-related data were recorded by the monitoring nurse, including puncture site, number of punctures, and puncture duration. Puncture duration was defined as the time from completion of the intervention to fixation of the needle. Family satisfaction was assessed across six domains: communication, pain management, facility completeness, service process, nursing staff attitude, and psychological support. Each domain was rated on a 5-point Likert scale, ranging from “very dissatisfied” (score 1) to “very satisfied” (score 5), yielding a total score between 6 and 30. Higher scores indicated greater satisfaction. Statistical analysis Data were analyzed using MedCalc software version 22.021 (MedCalc Ltd., Ostend, West Flanders, Belgium) and SPSS software version 22.0 (IBM Corp., Armonk, NY, USA). No missing data were identified. Continuous variables confirmed to be normally or approximately normally distributed with homogeneity of variance were presented as mean ± SD, and categorical variables were presented as frequency (percentage). The chi-square test was used to analyze differences in categorical data. Repeated-measures analysis of variance (ANOVA) was performed to assess time, group, and interaction effects in longitudinal multi-group data. When the assumption of sphericity was violated, Huynh–Feldt (Mauchly’s W > 0.75) or Greenhouse–Geisser (Mauchly’s W < 0.75) corrections were applied to adjust p-values, followed by Bonferroni-corrected pairwise t-tests. Effect sizes (partial η2) and confidence intervals (CIs) were calculated. The paired t-test was used to compare fear scores before and after the intervention and to examine paired differences between FLACC and Wong–Baker scores. Bivariate correlations between child-reported and nurse-assessed scores were visualized using scatter heat plots with Pearson regression lines and 95% CIs. Correlation strength was interpreted as follows: absolute correlation coefficient (|r|) < 0.200 (poor), 0.200–0.399 (mild), 0.400–0.599 (moderate), 0.600–0.799 (strong), and ≥ 0.800 (extremely strong). Partial correlations between family satisfaction and each indicator were estimated using backward linear regression, with variables removed if the p-value was > 0.1 or the variance inflation factor (VIF) was > 5. Statistical significance was set at p < 0.05. Results Sociodemographic characteristics of children and their companions and puncture outcomes A total of 188 children were included in this study, with 96 in the control group and 92 in the experimental group. There were no statistically significant differences between the two groups in age, sex, clinical diagnosis, or puncture-related outcomes. Similarly, no significant differences were observed in companions’ age, sex, educational level, or kinship with patient (all p > 0.05) (Table 1). Table 1Sociodemographic characteristics and puncture outcomes in the two groupsCon(n = 96)Exp(n = 92)χ2/t, pPatientAge (years)9.09 ± 2.339.18 ± 2.43-0.254, 0.800Male/Female (n/n) *74/2266/260.705, 0.401Diagnosis [n(%)] *0.603, 0.740 Respiratory system diseases59(61.5)64(69.6) Digestive system diseases24(25.0)21(22.8) Urinary system diseases11(11.5)9(9.8)Family memberAge (years)42.17 ± 9.3941.65 ± 9.100.381, 0.703Male/Female (n/n) *32/6432/600.044, 0.834Education [n(%)] *2.322, 0.313 Primary school degree24(25.0)21(22.8) High school degree24(25.0)23(25.0) College degree48(50.0)48(52.2)Kinship with patient [n(%)] *0.855, 0.836 Mother44(45.8)43(46.7) Father32(33.3)32(34.8) Grandma20(22.8)17(18.5)Puncture duration (seconds)50.11 ± 20.8652.08 ± 29.08-0.459, 0.646Puncture site [n(%)] *0.864, 0.649 Left hand dorsum37(38.5)34(37.0) Right hand dorsum46(47.9)49(53.3) Others13(13.5)9(9.8)Experienced secondary punctures [n(%)] **8(8.3)2(2.2)2.422, 0.120Note: Con, control group; Exp, experimental group. * Using the chi-square test; ** Using the continuous correction chi-square test; the else used the t-test Sociodemographic characteristics and puncture outcomes in the two groups Note: Con, control group; Exp, experimental group. * Using the chi-square test; ** Using the continuous correction chi-square test; the else used the t-test Fear manifestations before and after intervention implementation Both at T0 and T1, nurse-assessed CEMS scores and child-reported CFS scores were strongly correlated [r (95% CI) = 0.693 (0.610, 0.761) and 0.689 (0.605, 0.757), both p < 0.001]. Similar strong correlations were observed before and after the verbal comfort intervention in the control group [r (95% CI) = 0.601 (0.453, 0.718) and 0.755 (0.654, 0.830), both p < 0.001] (Fig. 1A), as well as before and after the VR intervention in the experimental group [r (95% CI) = 0.759 (0.656, 0.834) and 0.609 (0.459, 0.726), both p < 0.001] (Fig. 1B). Bivariate pearson correlation between child-reported and nurse-assessed fear at (A) pre-intervention and (B) post-intervention. Note: Pearson r (95% CI) and p-values indicated. Con, the control group; Exp, the experimental group; CEMS, the Children’s Emotional Manifestation Scale; CFS, the Children’s Fear Scale; T0, pre-intervention; T1, immediately post-intervention. The black solid line and purple dashed lines illustrated the correlation coefficient and its 95% confidence intervals. The heat plot visually represented the density distribution of data, with increasing density indicated by a color gradient ranging from white, through blue, green, and orange, to red. There is a strong correlation between the nurse-assessed CEMS and child-reported CFS scores strong correlation at two pre-venipuncture time points Repeated-measures ANOVA indicated that nurse-assessed CEMS and child-reported CFS scores satisfied the sphericity assumption (both Mauchly’s W = 1.000, p = 1.000), with significant time effects (partial η2 = 0.205 and 0.034, df = 1, F = 48.069 and 6.591, both p < 0.001), group effects (partial η2 = 0.072 and 0.053, df = 1, F = 10.606 and 7.434, both p < 0.001), and time-by-group interaction effects (partial η2 = 0.183 and 0.171, df = 1, F = 41.704 and 38.343, both p < 0.001). At baseline (T0), nurse-assessed CEMS scores did not differ significantly between the two groups (partial η2 = 0.010, df = 186, t = 1.399, padj = 0.164), whereas child-reported CFS scores were significantly higher in the experimental group (partial η2 = 0.053, df = 186, t = 3.234, padj = 0.001) (Table 2), indicating greater baseline fear. After the intervention (T1), nurse-assessed CEMS scores decreased significantly in the experimental group (partial η2 = 0.061, df = 186, t = − 4.808, padj = 0.006), whereas no significant between-group difference was observed in child-reported CFS scores (partial η2 = 0.002, df = 186, t = − 0.598, padj = 0.550). Although both nurse-assessed CEMS and child-reported CFS scores at T1 were lower than those at T0 in multiple comparisons from repeated-measures ANOVA (Supplementary Table 1), paired t-tests showed significant decreases in the experimental group (Mean Diff. = −1.76 and − 0.48, df = 183, t = − 8.236 and − 7.011, both padj < 0.001), whereas increases were observed in the control group (Mean Diff. = 1.05 and 0.20, df = 191, t = 2.482 and 2.340, padj = 0.014 and 0.021). Table 2Nurse-assessed and child-reported fear before and after the interventionItemConExptdf p partial η2CEMS Pre-intervention (T0)6.81 ± 2.727.47 ± 3.651.3991860.1640.010 Post-intervention (T1)7.85 ± 2.755.71 ± 2.32-4.8081860.0060.061 t2.629-8.236 df9591 P0.009< 0.001CFS Pre-intervention (T0)1.25 ± 1.091.83 ± 1.343.2341860.0010.053 Post-intervention (T1)1.45 ± 1.311.35 ± 0.94-0.5981860.5500.002 t2.340-7.011 df9591 P0.021< 0.001Note: Con, control group; Exp, experimental group; CEMS, Children’s Emotional Manifestation Scale; CFS, Children’s Fear Scale; T0, pre-intervention; T1, immediately post-intervention. Paired t-test compared pre- and post-intervention outcomes within each group; independent t-test compared the two groups at each time point. These results indicated the exacerbation of fear in the control group and the alleviation of fear in the experimental group, leading to a significant difference of fear scores between the two groups after the intervention Nurse-assessed and child-reported fear before and after the intervention Note: Con, control group; Exp, experimental group; CEMS, Children’s Emotional Manifestation Scale; CFS, Children’s Fear Scale; T0, pre-intervention; T1, immediately post-intervention. Paired t-test compared pre- and post-intervention outcomes within each group; independent t-test compared the two groups at each time point. These results indicated the exacerbation of fear in the control group and the alleviation of fear in the experimental group, leading to a significant difference of fear scores between the two groups after the intervention Pain response during indwelling needle procedures Nurse-assessed FLACC scores and child-reported Wong–Baker scores demonstrated mild-to-moderate correlations at three time points [r (95% CI) = 0.308 (0.172, 0.732) at T2, 0.568 (0.462, 0.657) at T3, and 0.568 (0.463, 0.658) at T4, all p < 0.001]. Subgroup analyses showed that in the control group, no significant correlation was observed at T2 [r (95% CI) = 0.147 (− 0.060, 0.337), p = 0.153] (Fig. 2A), a mild correlation at T3 [r (95% CI) = 0.392 (0.207, 0.549), p < 0.001] (Fig. 2B), and a strong correlation at T4 [r (95% CI) = 0.644 (0.509, 0.748), p < 0.001] (Fig. 2C). In the experimental group, moderate correlations were observed at T2 [r (95% CI) = 0.460 (0.281, 0.607), p < 0.001] and T4 [r (95% CI) = 0.480 (0.305, 0.623), p < 0.001] (Fig. 2A and C), with a strong correlation at T3 [r (95% CI) = 0.732 (0.620, 0.815), p < 0.001] (Fig. 2B). Notably, nurse-assessed FLACC scores were consistently lower than child-reported Wong–Baker scores in both groups at all time points (Con and Exp groups at T2: Mean Diff. = −1.94 and − 1.57, t = − 6.367 and − 5.845, df = 95 and 91, both p < 0.001; at T3: Mean Diff. = −2.10 and − 1.72, t = − 7.072 and − 7.215, df = 95 and 91, both p < 0.001; at T4: Mean Diff. = −1.04 and − 1.49, t = − 4.950 and − 5.994, df = 95 and 91, both p < 0.001). Bivariate Pearson correlation between child-reported and nurse-assessed pain at (A) tourniquet application, (B) venipuncture, and (C) one-minute post-venipuncture. Note: Pearson r (95% CI) and p-values indicated. Con, the control group; Exp, the experimental group; FLACC, the Face, Legs, Activity, Crying, Consolability scale; Wong-Baker, the Wong-Baker FACES Pain Rating Scale; T2, at tourniquet application; T3, at venipuncture; T4, one minute after venipuncture. The black solid line and purple dashed lines illustrated the correlation coefficient and its 95% confidence intervals. The heat plot visually represented the density distribution of data, with increasing density indicated by a color gradient ranging from white, through blue, green, and orange, to red. There was an overall moderate correlation between nurse-assessed FLACC and child-reported Wong-Baker scores, except for the situation at tourniquet application in the control group (showing no correlation) and after venipuncture in the two groups (exhibiting a strong correlation) Repeated-measures ANOVA indicated that neither nurse-assessed FLACC scores nor child-reported Wong–Baker scores satisfied the sphericity assumption (Mauchly’s W = 0.770 and 0.937, p < 0.001 and p = 0.003, respectively). Consequently, the Huynh–Feldt correction was applied in subsequent analyses. From T2 to T4, both FLACC and Wong–Baker scores demonstrated significant time effects, with an inverted V-shaped trend (partial η2 = 0.393 and 0.443, df = 1.911 and 1.648, F = 120.66 and 147.77, both p < 0.001; Fig. 3). The highest scores were observed at T3 (Supplementary Table 2), which differed significantly from those at T2 (Mean Diff. = 0.62 and 0.77, both df = 185, t = 6.084 and 6.811, both padj < 0.001) and T4 (Mean Diff. = 1.55 and 2.20, both df = 185, t = 14.071 and 13.931, both padj < 0.001). The lowest scores occurred at T4, which differed significantly from those at T2 (Mean Diff. = −0.93 and − 1.42, both df = 185, t = − 10.563 and − 12.655, both padj < 0.001) and T3 (Mean Diff. = −1.55 and − 2.20, both df = 185, t = − 14.071 and − 13.931, both padj < 0.001). However, no significant group effects were found for either nurse-assessed FLACC or child-reported Wong–Baker scores (partial η2 = 0.005 and 0.008, both df = 1, F = 0.882 and 1.416, p = 0.349 and 0.236, respectively). Furthermore, a significant time-by-group interaction effect was observed only for nurse-assessed FLACC scores (partial η2 = 0.028, df = 1.911, F = 5.403, p = 0.006), but not for child-reported Wong–Baker scores (partial η2 = 0.008, df = 1.648, F = 1.450, p = 0.237). The dynamic FLACC and Wong-Baker scores. Note: Con, the control group; Exp, the experimental group; T2, at tourniquet application; T3, at venipuncture; T4, one minute after venipuncture; FLACC, the Face, Legs, Activity, Crying, Consolability scale; Wong-Baker, the Wong-Baker FACES Pain Rating Scale. The dot represents the mean value, while the I-shaped symbol denotes its 95% confidence intervals. The nurse-assessed FLACC scores were consistently lower than the child-reported Wong-Baker scores in both the control and experimental groups across all time points Dynamic changes in HR and SpO2 throughout the procedure Dynamic monitoring data for HR and SpO2 violated the sphericity assumption (Mauchly’s W = 0.448 and 0.698, both p < 0.001); therefore, the Greenhouse–Geisser correction was applied in subsequent repeated-measures ANOVA. SpO2 showed a significant group effect (partial η2 = 0.045, df = 1, F = 8.759, p = 0.003), with lower levels in the experimental group (Mean Diff. = −0.63, df = 186, t = − 2.958, padj = 0.003), but no significant time effect (partial η2 = 0.003, df = 2.741, F = 0.537, p = 0.641) or time-by-group interaction effect (partial η2 = 0.003, df = 2.741, F = 0.482, p = 0.678). HR also demonstrated a significant group effect (partial η2 = 0.106, df = 1, F = 21.976, p < 0.001), with lower levels in the experimental group from T1 to T4 (T1: partial η2 = 0.099, t = − 4.513, padj < 0.001; T2: partial η2 = 0.130, t = − 5.264, padj < 0.001; T3: partial η2 = 0.137, t = − 5.423, padj < 0.001; T4: partial η2 = 0.075, t = − 3.870, padj < 0.001), as well as significant time effects (partial η2 = 0.207, df = 3.314, F = 48.589, p < 0.001) and a significant time-by-group interaction effect (partial η2 = 0.066, df = 3.314, F = 13.235, p < 0.001). HR increased from T0 to T3, peaking at T3 (vs. T0–T2 and T4: Mean Diff. = 9.6, 12.1, 5.5, and 14.5; df = 183; t = 6.590, 9.904, 5.031, and 13.590; all padj < 0.001), and then decreased sharply at T4 to the lowest level (vs. T0–T3: Mean Diff. = −4.9, − 2.4, − 8.9, and − 14.5; df = 183; t = − 3.703, − 2.285, − 8.711, and − 13.590; padj = 0.003, 0.234, < 0.001, and < 0.001, respectively) (Supplementary Table 3). Notably, from T0 to T2, the control group demonstrated a continuous increase in HR, whereas the experimental group displayed a V-shaped HR pattern (Fig. 4). The dynamic HR measurements. Note: Con, control group; Exp, experimental group; HR, heart rate; T0, pre-intervention; T1, immediately post-intervention; T2, at tourniquet application; T3, at venipuncture; T4, one minute after venipuncture. The dot represents the mean value, while the I-shaped symbol denotes its 95% confidence intervals. Before venipuncture, the control group exhibited a progressive increase in heart rate, whereas the experimental group demonstrated a V-shaped decline Partial correlations between family satisfaction and dynamic process indicators Backward elimination was applied in linear regression, with p < 0.05 as the criterion for variable retention and p > 0.10 for variable removal. Given the potential correlations between nurse- and child-reported pain and fear scores, variables with a VIF > 5 were excluded to address multicollinearity. The results (Table 3) indicated that, in the control group, family satisfaction was significantly associated with several whole-process evaluation variables: nurse-assessed CEMS score at T1 (Std.Beta = − 0.374, rpartial = − 0.228, p < 0.001), SpO2 at T1 (Std.Beta = 0.368, rpartial = 0.376, p < 0.001), HR at T2 (Std.Beta = − 0.336, rpartial = − 0.327, p = 0.001), SpO2 at T4 (Std.Beta = 0.197, rpartial = 0.225, p = 0.035), child-reported Wong–Baker scores at T2 (Std.Beta = − 0.578, rpartial = − 0.517, p < 0.001), T3 (Std.Beta = − 0.211, rpartial = − 0.254, p = 0.017), and T4 (Std.Beta = − 0.354, rpartial = − 0.330, p = 0.002), as well as puncture duration (Std.Beta = − 0.401, rpartial = − 0.450, p < 0.001). Notably, a positive association was observed between satisfaction and child-reported CFS score at T0 (Std.Beta = 0.492, rpartial = 0.416, p < 0.001). In contrast, in the experimental group, family satisfaction was significantly associated only with selected outcome variables: HR at T4 (Std.Beta = − 0.280, rpartial = − 0.280, p = 0.008), nurse-assessed FLACC score (Std.Beta = − 0.225, rpartial = − 0.255, p = 0.016), child-reported Wong–Baker score at T4 (Std.Beta = − 0.246, rpartial = − 0.218, p = 0.040), and puncture duration (Std.Beta = − 0.522, rpartial = − 0.495, p < 0.001). Table 3Partial correlation between family satisfaction and child’s responses during intravenous cannulationBeta (95%CI)SEStd. Betarpartial (95%CI)t p VIF Con group ( n = 96) (Constant)-14.460(-46.957, 18.037)16.580-0.8720.386Child-reported CFS score at T01.359(0.714, 2.004)0.3290.4920.416(0.234, 0.595)4.131< 0.0012.577Nurse-assessed CEMS score at T1-0.486(-0.772, -0.200)0.146-0.374-0.228(-0.403, -0.055)-3.331< 0.0013.167SpO2 at T10.844(0.421, 1.267)0.2160.3680.376(0.177, 0.555)3.911< 0.0011.844HR at T2-0.047(-0.074, -0.020)0.014-0.336-0.327(-0.497, -0.159)-3.4210.0011.721SpO2 at T40.311(0.027, 0.595)0.1450.1970.225(0.432, 0.006)2.1400.0351.638Child-reported Wong-Baker score at T2-1.084(-1.451, -0.717)0.187-0.578-0.517(-0.660, -0.342)-5.800< 0.0012.037Child-reported Wong-Baker score at T3-0.360(-0.650, -0.070)0.148-0.211-0.254(-0.452, -0.068)2.4360.0172.006Child-reported Wong-Baker score at T4-0.664(-1.066, -0.262)0.205-0.354-0.330(-0.528, -0.115)3.2450.0023.520Puncture duration-0.042(-0.060, -0.024)0.009-0.401-0.450(-0.606, -0.263)-4.547< 0.0011.465 Exp group ( n = 92) (Constant)33.073(30.580, 35.566)1.27225.998< 0.001HR at T4-0.036(-0.061, -0.011)0.013-0.283-0.280(-0.471, -0.078)-2.7150.0081.284Nurse-assessed FLACC score at T4-0.591(-1.061, -0.121)0.240-0.268-0.255(-0.421, -0.079)-2.4600.0161.405Child-reported Wong-Baker score at T4-0.244(-0.473, -0.015)0.117-0.246-0.218(-0.452, -0.045)-2.0850.0401.646Puncture duration-0.036(-0.050, -0.022)0.007-0.522-0.495(-0.636, -0.329)-5.316< 0.0011.139The final model fitting: r = 0.647, R2 = 0.419, Adjusted R2 = 0.396, SE of the estimate value = 2.193Note: Beta, regression coefficient; CI, confidence interval; SE, standard error; Std. Beta, standard regression coefficient; VIF, variance inflation factor; CFS, the Children’s Fear Scale; CEMS, the Children’s Emotional Manifestation Scale; SpO2, peripheral oxygen saturation; HR, heart rate; Wong-Baker, the Wong-Baker FACES Pain Rating Scale; FLACC, the Face, Legs, Activity, Crying, Consolability scale. Using stepwise method with adjusting by children and their companions’ sociodemographic characteristics. In the verbal comfort group, family satisfaction exhibited a typical “process-sensitive” characteristic, closely correlating to multiple response factors during the medical procedure. In contrast, family satisfaction in the VR group follows an “outcome-oriented” pattern, closely correlating to the several outcomes one minute after venipuncture Partial correlation between family satisfaction and child’s responses during intravenous cannulation The final model fitting: r = 0.647, R2 = 0.419, Adjusted R2 = 0.396, SE of the estimate value = 2.193 Note: Beta, regression coefficient; CI, confidence interval; SE, standard error; Std. Beta, standard regression coefficient; VIF, variance inflation factor; CFS, the Children’s Fear Scale; CEMS, the Children’s Emotional Manifestation Scale; SpO2, peripheral oxygen saturation; HR, heart rate; Wong-Baker, the Wong-Baker FACES Pain Rating Scale; FLACC, the Face, Legs, Activity, Crying, Consolability scale. Using stepwise method with adjusting by children and their companions’ sociodemographic characteristics. In the verbal comfort group, family satisfaction exhibited a typical “process-sensitive” characteristic, closely correlating to multiple response factors during the medical procedure. In contrast, family satisfaction in the VR group follows an “outcome-oriented” pattern, closely correlating to the several outcomes one minute after venipuncture Discussion Although most domestic and international studies have shown that VR technology can effectively alleviate pain and reduce fear during pediatric venipuncture [11–13, 20–23, 33, 34], a limited number of studies have reported ineffective findings [35–37]. Moreover, the underlying physiological and psychological mechanisms require further empirical clarification. In addition, previous studies have primarily focused on nurses’ and/or patients’ subjective evaluations of pain and fear, with limited consideration of physiological stress and dynamic assessment. This dynamic study, incorporating both subjective reports and physiological measures, yielded four main findings. First, VR alleviated children’s fear as reported by both nurses and children and indirectly supported by objective physiological data, particularly the V-shaped decline in HR before and after wearing VR headsets and the smaller dynamic HR fluctuations observed in the experimental group. These objective findings support the role of VR in diverting attention, suppressing negative perceptions, and consequently reducing tension. Second, a subjective discrepancy was observed between nurse-assessed and child-reported pain. Whether this reflects underestimation by nurses or overreporting by children remains unclear and warrants further investigation. Third, the absence of significant group effects in both nurse-assessed and child-reported dynamic data suggests that VR distraction does not exert a direct physiological analgesic action. Fourth, family members’ concerns during conventional venipuncture exhibited a pattern of process sensitivity, whereas after children’s immersion in VR, their concerns shifted toward an outcome-oriented perspective. VR intervention alleviates children’s subjective fear responses The strong correlations between CEMS and CFS scores, both pre- and post-intervention, indicate high consistency between nurses’ observations of children’s fear behaviors and children’s self-reported feelings. This finding is consistent with the conclusions of a systematic review and supports the feasibility and effectiveness of multidimensional assessment in pediatric fear management [38]. The pre- and post-intervention comparisons revealed a continued increase in fear following conventional verbal comfort, but a significant decrease following VR immersion. Previous studies have reported that conventional distraction techniques often fail to alleviate fear in children with limited cognitive comprehension [32]. Consequently, in high-stress medical situations, verbal comfort alone may be insufficient and may even exacerbate fear by directing children’s attention toward the procedure itself. In contrast, VR technology has a greater capacity to capture children’s attention through immersive engagement with rich visual content and compelling narratives, thereby effectively reducing subjective fear. This finding has been supported by multiple studies [38–40]. Notably, although no baseline difference was observed in nurse-assessed CEMS scores between the two groups, children in the experimental group reported higher baseline fear, as indicated by CFS scores. This discrepancy may reflect sampling variation. Alternatively, limited familiarity with and insufficient hands-on experience using VR equipment may have heightened fear by introducing cognitive uncertainty. Because baseline self-reported fear was significantly higher in the experimental group, these fear-related findings should be interpreted with caution. VR intervention effectively alleviates physiological stress responses in pediatric patients Compared with scale-based assessments, HR and SpO2 provide more objective indicators of children’s tension. During the pre-venipuncture phase, HR dynamic trends differed markedly between groups: the conventional verbal comfort group showed a steady increase, whereas the VR group exhibited a V-shaped pattern. This disparity suggests that VR effectively diverts attention and mitigates tension, thereby attenuating sympathetic nervous system activation and alleviating fear [41]. In contrast, verbal comfort alone may intensify children’s tension because of limited cognitive development and reduced comprehension capacity [42, 43]. Overall, HR dynamics across the entire procedure demonstrated significant effects of time, group, and their interaction. The consistently lower HR levels in the VR group indicate an inhibitory effect of VR on stress perception. These findings provide objective physiological evidence supporting the effectiveness of VR in alleviating medical stress in children. In contrast to HR, SpO2 dynamic changes showed only a group effect, with slightly lower levels in the experimental group. This finding may be related to alterations in breathing rhythm associated with different levels of anxiety. Compared with verbal comfort, immersive VR experiences may more effectively relieve children’s tension, potentially reducing respiratory rate and subtly modifying breathing patterns [44, 45]. Nevertheless, all SpO2 values in both groups remained within the normal clinical range (all > 94%), which may explain the absence of significant time effects or time-by-group interactions. These findings suggest that minor alterations in breathing patterns during venipuncture do not produce clinically meaningful fluctuations in SpO2. Therefore, when implementing VR during venipuncture, SpO2 may have limited clinical value for assessing tension. In contrast, routine HR monitoring may provide greater clinical relevance. Subjective disparities in pain scores and the impact of VR intervention Because none of the enrolled children had baseline pain, pain was assessed only during the intravenous cannulation procedure. In both the control and experimental groups, nurse-assessed FLACC scores were consistently lower than child-reported Wong–Baker scores at all time points, with an overall moderate correlation that was lower than previously reported values [37]. Further subgroup analyses indicated that correlations at tourniquet application in the control group (showing no significant correlation) and after venipuncture in both intervention groups (showing a strong correlation) deviated from the overall moderate pattern. These discrepancies may be explained by two factors. First, younger children may overreport pain because of limited psychological maturity and heightened fear of medical procedures [46–48]. Second, older children may attempt to appear courageous, thereby deliberately concealing pain behaviors [49, 50]. Such behaviors may result in substantial discrepancies between the two scoring methods and consequently influence correlation estimates. Fortunately, nurse-assessed FLACC and child-reported FACES scores demonstrated highly consistent dynamic patterns over time. Specifically, both measures indicated that pain began at tourniquet application (T2), peaked at venipuncture (T3), and rapidly diminished after venipuncture (T4). This pattern was consistent with the intensity changes of nociceptive stimuli during the venipuncture procedure, suggesting adequate temporal sensitivity of both measures. Apart from the significant time effect, neither measure showed a significant group effect in dynamic pain changes, indicating that neither verbal comfort nor VR intervention fundamentally reduced pain perception from a physiological perspective. Nevertheless, the significant time-by-group interaction observed in nurse-assessed FLACC scores suggests that the effects of VR distraction and conventional intervention on pain experience varied over time. VR intervention may more effectively divert attention through immersive distraction, thereby reducing tension and facilitating physical relaxation. This immersive distraction contributes to children’s pain tolerance, resulting in faster temporal changes in pain behaviors observed by nurses. Its mechanism differs from that of analgesics, which directly attenuate nociceptive processing and pain perception. In contrast to nurse-assessed FLACC scores, no significant time-by-group interaction was detected in child-reported Wong–Baker scores. Overall, FLACC scores are relatively more objective because they focus on observable pain-related behaviors. In comparison, Wong–Baker scores are more subjective, as they may capture both physiological sensations and psychological experiences, including perceived pain intensity and coping tolerance. These findings further emphasize that, in pediatric pain assessment, nurse-observed behavioral measures and child self-reports have complementary clinical value and should not be used interchangeably. VR intervention alters the focus of family satisfaction VR intervention improved children’s experiences during venipuncture and significantly increased family satisfaction. Notably, the determinants of family satisfaction differed between the VR and conventional verbal comfort groups. In the verbal comfort group, family satisfaction exhibited a typical process-sensitive” characteristic. It is closely correlated to multiple response factors during the medical procedure: pre-intervention fear, physiological responses, child-reported pain, and puncture duration. These factors reflected children’s real-time experiences at key time points. Interestingly, higher pre-intervention fear reported by children was positively correlated with family satisfaction in this group. This seemingly paradoxical finding may be related to families’ perceptions that children overreport pre-intervention fear [51]. When interpreted in the context of the child’s experiences and performance throughout the procedure, pre-intervention fear showed an unexpected positive association with family satisfaction. In contrast, family satisfaction in the VR group followed an outcome-oriented pattern. Satisfaction was correlated only with post-procedure outcomes at T4, including heart rate, pain manifestations, child-reported pain, and puncture duration. Puncture duration influenced family satisfaction in both groups, highlighting procedural efficiency as a central concern. Unlike the conventional verbal comfort group, in which families could continuously observe the child’s immediate responses, VR intervention limited direct observation of children’s reactions during venipuncture. It shifted family attention from focusing on what the child was experiencing during the procedure to emphasizing the child’s ultimate post-procedural status. From a clinical perspective, VR technology not only distracted children but also reshaped the family experience. This shift may have alleviated family anxiety and altered the way satisfaction was evaluated. Therefore, when implementing VR interventions, in addition to monitoring children’s responses, nurses should also consider families’ needs and perceptions and promptly communicate procedural outcomes and the child’s status. Such measures may further enhance overall healthcare satisfaction. Further speculation on VR distraction and pain experience Overall, our findings may be interpreted in light of theories such as attention modulation, cognitive load, and pain neuromatrix theories. The observed reductions in fear and physiological stress, together with dynamic fluctuations in nurse-assessed pain experience during VR-based distraction, are consistent with the principles of attention modulation, cognitive load theory, and pain neuromatrix theory. According to pain neuromatrix theory, pain experience is a protective output generated by the brain through the integration of pain perception, emotional response, memory formation, and pain empathy [52]. Compared with the control group, VR distraction was associated with greater improvements in objective physiological responses and subjective fear evaluations, suggesting that immersive VR experiences effectively capture children’s attention through visual and auditory stimulation and alleviate negative emotions [53]. These effects may attenuate activation in neural circuits involved in pain perception and emotional response. Moreover, cognitive load theory proposes that limited attentional and cognitive resources must be dynamically allocated among concurrent tasks through competitive mechanisms based on resource trade-offs [54]. The redirection of children’s attention and cognition toward VR scenarios may weaken attention and cognition toward pain, thereby reducing activation in neural circuits involved in memory formation and pain-related affective processing. Therefore, the interplay among attention regulation, cognitive load theory, and pain neuromatrix theory may help explain the beneficial effects of VR distraction. Cognitive load theory clarifies the constraints and allocation principles of attentional and cognitive resources, whereas pain neuromatrix theory provides a framework for understanding the psychological and neural reconstruction of pain experience. Further empirical validation of these psychological mechanisms is warranted. In conclusion, the psychophysiological regulation associated with VR distraction has important clinical implications. First, VR distraction effectively redirected children’s attention away from procedural stimuli, thereby reducing fear rather than pain. Second, VR distraction facilitated a cooperative state, thereby improving procedural efficiency and success rates and ultimately alleviating children’s psychological distress. Third, VR distraction was associated with a result-oriented mode of family satisfaction evaluation, which may contribute to establishing a more harmonious nurse–patient–family relationship. Limitations (1) These findings were derived from a single-center study with a relatively small sample size and require validation in multicenter studies with larger samples. (2) Unlike continuous physiological measurements, pain and fear were assessed at discrete time points because of the short procedural duration; future studies should incorporate fully dynamic, whole-process assessments to provide additional insights. (3) The inability to blind nurses and pediatric participants may have introduced bias into the assessment of subjective outcomes, and baseline imbalances in fear may also have influenced the observed reductions; therefore, these conclusions require further rigorous validation. (4) Family perspectives other than satisfaction were not assessed and should be incorporated in future research to evaluate applicability across stakeholders. (5) Only short-term procedural responses were evaluated; long-term outcomes, including medical fear, pain memory, and treatment compliance, should be examined in longitudinal follow-up studies. (6) The absence of significant group differences in pain outcomes may indicate that the study was underpowered to detect small but clinically meaningful effects on pain; 7) Given limited validation evidence for CFS in Chinese children, the complexity and cross-cultural variability of emotional facial expressions, may compromise the interpretation of CFS findings. Conclusion This study suggests that VR distraction may enhance children’s experiences during venipuncture. VR significantly reduced children’s fear (but not pain) and stabilized heart rate fluctuations during the procedure. In addition, VR shifted the focus of family satisfaction from children’s immediate procedural responses to post-procedural outcomes. Overall, integrating VR technology into pediatric invasive procedures not only improved children’s comfort but also enhanced the experience of accompanying family members, reflecting a humanistic approach to care. Future clinical practice should strengthen pre-intervention adaptive guidance, employ multidimensional assessments, and incorporate family feedback to optimize care quality. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We express our sincere appreciation to Editage for their professional English editing and proofreading. Abbreviations Repeated-measures analysis of variance The Children’s Emotional Manifestation Scale the Children’s Fear Scale The Face, Legs, Activity, Crying, and Consolability scale Heart rate Peripheral oxygen saturation Variance inflation factor Virtual reality The Wong-Baker FACES Pain Rating Scale Author contributions S.X., Y.Q., and Y.Y. proposed the conceptualization and obtained funding support. X.C., M.W., and F.Y. wrote the main text. Y.Z., F.W., and Y.Z. performed the investigation and conducted the data curation. Y.Y. graphed the figures. All the authors reviewed the manuscript. S.X. and F.Y. contributed equally to this work. Funding This study was supported by the Mianyang Municipal Health Commission (2024038), the Health Humanities Research Center of Zigong Key Research Base of Philosophy and Social Sciences (JKRWY25-22), and the Patient Safety Research Center of Zigong Key Research Base of Philosophy and Social Sciences (HZAQ-2025-06). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Please contact Yuwei Yang (E-mail: yyw318@sc-mch.cn) for data access. Declarations Ethics approval and consent to participate This study was in accordance with the principles of the Declaration of Helsinki, and approved by the Ethics Committee of Mianyang Central Hospital (Approval No.: S20250237-01). Verbal assent was obtained from the children, and written informed consent was obtained from the parents or legal guardians. Consent for publication Not applicable. This manuscript does not contain any individual privacy in any form (including individual details, images, or videos). Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Shuai Xie, Feng Yin and Miyan Wang contributed equally to this work. Contributor Information Yun Qiu, Email: 1606585803@qq.com. Xiaohong Chen, Email: hongxiaoch@aliyun.com. Yuwei Yang, Email: yyw318@sc-mch.cn. References Associated Data Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Please contact Yuwei Yang (E-mail: yyw318@sc-mch.cn) for data access.