We conducted a multicentre open-label randomized clinical trial with participation of 5 primary care centres and the 2 regional hospitals located in the same health area of the Valencian Community, Spain. The study was carried out in adherence to the CONSORT guidelines.10 Patients were allocated in a 1:1 ratio to the intervention group (use of VR during blood draw) or the control group (standard care).

We calculated the sample size out of the total population aged 7–12 years who had undergone routine blood draws in our health area in 2019 required for a power of 90%. The calculated minimum was of 80 patients.

Eligible candidates were patients who visited participating centres to undergo blood tests during the recruitment period (March to October 2022).

We recruited children aged 7–12 years who assented and whose parents or legal guardians consented to participation in the study by signing the informed consent form. The minimum age was set at 7 years because visual development is complete, and the maximum at 12 years to avoid an overly broad age range, so that the same VR video could be used for the entire sample.

We excluded patients with psychomotor or neurodevelopmental delay, in whom assessment of anxiety would be complicated. Patients with visual or hearing impairments were also excluded, as we were unable to adapt the RV headset to their particular conditions.

To carry out the study, we randomised patients by means of the Research Randomizer software tool (https://www.randomizer.org/). The tool generated 7 sequences (one per participating centre) that randomly assigned the numbers 1 (intervention) and 2 (control) to 30 specific positions. After obtaining the 7 sequences, the principal investigator prepared 30 closed envelopes containing the words “intervention” or “control”, following the order established by the randomly generated sequences. The allocation envelopes were marked successively with numbers 1 through 30 so that the nurses, after recruiting patients, would open the corresponding envelope and know the group the patient had been allocated to. This way, nurses did not know which group the patient was going to be allocated to until they had been recruited.

The study protocol consisted of:

• 1

  Informing the patient.

• 2

  Obtaining signed informed consent.

• 3

  Opening of the envelop specifying the group allocation.

• 4

  Blood draw conducted by the standard procedure or with VR depending on group allocation.

At the time of enrollment, each patient was assigned a code for data pseudonymization.

In the intervention group, patients wore an Oculus Quest 2 VR headset (Meta, California, USA) that played a video from the moment the supplies started to be prepared to the moment pressure was applied to the puncture site. The video was a short film titled “Henry”. It is a cartoon aimed at young children in which the main character is a hedgehog that celebrates his birthday with animal-shaped balloons that come alive.

In the control group, venepuncture was carried out per the standard protocol (distraction method with nurses talking to the patient).

All the blood draws were performed by the same nurses. Before recruitment, all the nurses that would collaborate in the study attended a meeting. One or two nurses were included per participating centre; they were nurses usually tasked with performing routine blood draws in their respective facilities and the only ones who performed blood draws included in the study. In the meeting, they were trained on the use of the RV headset and informed of the study protocol. They also received all the material required for the study (headset and documentation).

Independent variables:

• -

  Age

• -

  Sex

• -

  History of recent venepuncture (<6 months)

• -

  Presence of family members

Outcome variables:

• none-

  Pain (visual analogue scale) (Fig. 1).

  
  

  Figure 1.

  Visual analogue scale.

  (0.12MB).
• none-

  Patient anxiety (Groningen Distress Scale) (Table 1).

  Table 1.

  Groningen distress scale.

  Rating	State of child	Muscle tension	Crying
  1	Calm, no crying	No	No
  2	Mild distress, no crying	Makes fists, pale knuckles, grinds teeth, closes eyes, contracted muscles, rigid body	No
  3	Distress and some crying	Same as 2	Intermittent
  4	Distress and continuous crying	Same as 2	Continuous
  5	Aggression/yelling/physical resistance	Agitation, violent movements of torso and extremities, considerable resistance to procedure	Continuous, screaming
• none-

  Number of attempts required for extraction.

• none-

  Time required for the technique (minutes), per attempt, from application of tourniquet to compression of puncture site.

• none-

  Degree of difficulty of the draw as perceived by the nurse (rated as easy, normal or difficult).

• none-

  Family satisfaction (numerical scale from 1 to 10, with 1 representing “not at all satisfied” and 10 “very satisfied”).

• none-

  Anxiety experienced by the nurse who carried out the procedure (numerical scale from 1 to 10, with 1 representing “totally calm” and 10 “very anxious or nervous”).

The study was approved by the competent ethics committee (CEIm 02/2022, project no 37/2021). All participants and their family members were informed and signed the consent form to participate in the study. The study was registered in ClinicalTrials.gov with NCT number NCT05902585.

The collected data were entered in a database created for the purpose with the software SPSS version 26 (IBM Corp, Armonk, NY, USA), safeguarding the confidentiality of the patients.

We carried out a descriptive analysis of the data. Quantitative variables were expressed as mean and standard deviation if the data followed a normal distribution, or otherwise as median and range. Qualitative variables were expressed as absolute frequency and percentage distributions.

To assess the association between two variables, we compared frequencies by means of the Fisher exact test in the case of comparing 2 dichotomous variables and the χ2 test in the case of comparing variables with more than 2 categories. Since a nonparametric test was required, we used the Mann–Whitney U test to assess if there were differences in individual variables between groups, and we analysed the association between two variables with the Spearman correlation coefficient.

To estimate the number needed to treat (NNT), we calculated the reduction in absolute risk for mild pain and anxiety between the intervention and control groups. The level of statistical significance was established at 0.05 for all statistical tests.

The sample included 83 patients with a mean age of 9.7 years (SD, 1.7), of who 44.6% were female.

After random allocation, 40 patients were included in the control group (standard care) and 43 in the intervention group (VR). There were no losses to follow-up.

Both groups were comparable, as there were no statistically significant differences in their characteristics (Table 2); 21.7% of participants had undergone venepuncture in the past 6 months. Ninety-five percent of the children were accompanied by a relative during the procedure.

The perceived pain, assessed with the visual analogue scale, was milder in the VR group (Fig. 2), a difference that was statistically significant (P = .012) (Table 3). In the VR group, 83.7% of children reported absent or mild pain compared to 57.5% of controls. This corresponds to a 26% reduction in the frequency of pain in children, and therefore, the NNT for one child to report mild pain is of 4 children.

Figure 2.

Perceived pain reported with the visual analogue scale in the intervention and control groups.

(0.16MB).

The assessment of anxiety by means of the Groningen Distress Scale found calm (rating of 1) or mild anxiety (rating of 2) in 81% of the children, without statistically significant differences between groups (P = .08) (Table 3). However, in the control group there were 6 patients with scores greater than 3 on the Groningen Distress Scale, while there were no patients with this score in the intervention group (Fig. 3). Of the children who wore the headsets, 93% reported no anxiety or mild anxiety, compared to 72% of controls. Therefore, the NNT to reduce anxiety in one child with VR is of 6 children.

Figure 3.

Level of anxiety reported with the Groningen Distress Scale in the intervention and control groups.

(0.15MB).

In 93% of cases, the blood draw was successful on the first attempt, only five children required a second attempt and one more than 2 attempts. The mean time elapsed from the application of the tourniquet to compression after completion of the blood draw was of 1 min and 40 s, with no differences between groups (Table 2).

Based on the nurse reports, 58% of the blood draw procedures were easy, 30% had the average level of difficulty and 12% were difficult.

Family members reported a satisfaction greater than or equal to 9 in 93% of cases in the VR group compared to 72.5% in the control group (P = .026). None of the patients experienced adverse effects from the use of VR, although some participants reported that the headset felt too heavy.

Nurses rated their level of anxiety below 5 points (on a scale from 1 to 10) for more than 90% of blood draws in both groups (P = .13) (Table 3). However, nurses reported difficulty performing the procedure less frequently with the use of VR headsets (P = .01).

In respect of the care setting, 77% of blood draws were performed in primary care and 23% in hospital. We did not find differences in the pain ratings based on the setting where the procedure was performed (P = .13). Children also did not experience greater anxiety at the hospital (P = .37). There were also no differences in the anxiety or the difficulty reported by the nurse performing the procedure (P = .15 and P = .13, respectively). Family satisfaction was high in both care settings (P = .10) in both the intervention and control groups.

In the overall sample, we found a significant correlation between the pain and the anxiety experienced by the child (rs = 0.4; P = .001 [2-tailed]), and an inverse correlation between the pain reported by the child and the satisfaction of the family (rs = −0.39; P = .001 [2-tailed]) (Table 4). Anxiety in the nurse correlated to anxiety in the child (rs = 0.5; P = .001 2-tailed) and both were inversely correlated to family satisfaction, although the association was only significant with nurse anxiety (rs = −0.5; P = .001 [2-tailed]). These associations persisted in the analysis by group, both in children who used VR headsets and in children who received standard care (Table 5), with a weaker correlation in the intervention group.

The history of a previous blood draw in the past 6 months was not associated with the degree of anxiety (P = .39) or pain (P = .48) in the children, nor with the difficulty of the procedure (P = .33).