Digital Technology Distraction for Acute Pain in Children: A Meta-analysis

We followed a predetermined protocol (International Prospective Register of Systematic Reviews CRD42017077622). Given the large number of included studies, we report only on the primary outcomes of pain and distress.

A research librarian implemented a peer-reviewed¹⁶  search strategy in October 2017 in ePub Ahead of Print, In-Process and Other Nonindexed Citations, Ovid Daily Medline Daily, and Ovid Medline; Ovid Embase; Wiley Cochrane Library; Cumulative Index to Nursing and Allied Health Literature (CINAHL) Plus with full text via Elton B. Stephens Co. host (EBSCOhost); Ovid PsycINFO; Institute of Electrical and Electronic Engineers Xplore; Ei Compendex via Engineering Village; and Web of Science Core Collection via Clarivate Analytics (Supplemental Information). We searched for gray literature via health technology assessment agency Web sites (Canadian Agency for Drugs and Technologies in Health [CADTH] and Turning Research Into Practice [TRIP]), clinical trial registries (clinicaltrials.gov) and Web search engines (Google). In November 2018, we updated the search in Medline, Embase, CINAHL, Ei Compendex, Web of Science, the CADTH Web site, the TRIP database, and Google. We also scanned reference lists (included studies and relevant reviews) and contacted 6 external content experts and the authors of ongoing research. We used EndNote (version X5; Clarivate Analytics, Philadelphia, PA) to manage records and remove duplicates.

Table 1 shows the eligibility criteria for inclusion of studies in the review. We included all quantitative primary studies reporting on digital technologies used as a distraction for children with acutely painful medical conditions or having painful procedures in any clinical setting. Birnie et al¹⁷  defined distraction as “a shifting of attention away from pain … to stimuli that are more engaging or enjoyable.” We defined acute pain as any pain related to an injury, harm, or repair that was of short duration (ie, <30 days), typically occurring as part of a single treatable incident.¹⁸  We included any medical procedure that could cause pain.²  Comparators were (1) usual care (could have included spontaneous caregiver or health care provider interaction without study-provided guidance) or (2) another form of distraction (digital or nondigital; eg, blowing bubbles or nonmedical conversation); we also included studies with no comparison group. We allowed for the use of similar cointerventions (eg, topical anesthetics) among groups being compared. The primary outcomes were child pain and distress (including anxiety, stress, and fear).

Using a piloted form, 2 reviewers independently screened all records by title and abstract then by full text in Microsoft Office Excel (version 2016; Microsoft Corporation, Redmond, WA). The final inclusion of studies was determined by consensus. We sought the opinion of a third reviewer with methodologic or content expertise when needed.

Using a piloted form in Microsoft Excel 2016, 1 reviewer independently extracted study characteristics, population characteristics, setting, intervention and comparator(s), and outcome data. This reviewer also categorized the studies on the basis of predetermined subgroups of interest (Supplemental Table 8). When needed, we estimated data points from graphs using Plot Digitizer software (http://plotdigitizer.sourceforge.net/). A second reviewer verified the extraction and subgroup categorizations, and a statistician independently verified all data entered into meta-analyses. Reviewers resolved disagreements by discussion. We contacted study authors twice via e-mail when data were missing or unclear.

We assessed the risk of bias of the included studies in duplicate using the following tools: Cochrane’s risk-of-bias tool¹⁹  for randomized controlled trials (RCTs) and nonrandomized controlled trials, an adapted tool developed by the Cochrane Effective Practice and Organization of Care Group²⁰  for before-after and time series studies, and a Modified Newcastle-Ottawa Quality Assessment Scale²¹  for cross-sectional studies. The final rating was determined by consensus.

We synthesized data from randomized trials separately from the data obtained from other study designs. We first pooled data from RCTs and quasi RCTs via pairwise meta-analysis using the DerSimonian and Laird method random-effects model^22,23  in Review Manager (version 5.3; Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen, Denmark). The randomized trials were the main focus of our analyses because these offer higher internal validity than the other study designs included.

Self-report is the preferred measure of assessment for children who are able to describe their own pain, but it is not free from limitations.²⁴  For this reason, studies often provide self-report alongside other measures (eg, observer reports and behavioral assessments; ie, by using a tool designed to assess pain- or distress-related behaviors). We therefore performed separate pain and distress meta-analyses for each type of assessment (ie, self-reported, observed, or behavioral) using mean differences (MDs) or standardized mean differences (SMDs) (results of studies that measure outcomes using different tools are standardized to a uniform scale) and 95% confidence intervals (CIs). We described non-RCTs and studies with inadequate data for meta-analysis in tables and compared findings to those of the RCTs.

We preferentially used data recorded during the first (if there were multiple) painful procedure or the closest time point thereafter. We preferentially used first-period data from crossover trials. When data were provided for multiple time points or observers, we produced a composite score.²⁵  We computed SDs assuming a correlation of 0.5 between measurements unless data were available to compute correlations manually. To facilitate the interpretation of the clinical significance of the findings, we reexpressed SMDs using the mean control group SD for tools commonly used by studies within the meta-analyses and presented the absolute and relative differences in results tables.

We used the I² statistic,²⁶  the degree of overlap in CIs, and the direction of effect estimates across studies to quantify heterogeneity. We explored heterogeneity using the following between-study subgroup analyses: age category, type of patient population (added post hoc on the recommendation of a peer reviewer), setting, condition or procedure, type of digital technology distraction, active (requiring child engagement) versus passive (engagement not required) distraction,²⁷  presence of cointerventions, and measurement scale. We conducted sensitivity analyses on the basis of whether the authors received industry funding as well as study design. We tested for small-study bias when analyses included at least 9 studies using funnel plots and Egger’s regression test.²⁸ 

Two reviewers independently judged the certainty of the body of evidence for each meta-analytic comparison by following the Grading of Recommendations Assessment, Development and Evaluation approach.²⁹  Reviewers resolved disagreements by discussion.

We identified 3247 unique records, assessed the full text of 423, and included a total of 106 studies (reported in 109 publications)^30–138 ; 70 of these studies (reported in 73 publications) contributed to the quantitative synthesis (Fig 1; excluded studies in Supplemental Information).^*

Table 2 provides a summary of the characteristics of the included studies (study-level details are available in Supplemental Table 9). The included studies reported on a total of 7820 (median 56; range 5–400) participants, and more than two-thirds were RCTs (n = 73 of 106; 68.9%). Most studies took place in an outpatient setting (n = 81 of 106; 76.4%). The most common age group studied was children aged 2 to 11 years (n = 57 of 106; 53.8%), followed by a mix of children and teenagers (aged 2–21 years; n = 40 of 106; 37.7%). The participants were most often previously healthy children (n = 68 of 106; 64.2%). The studies reported on children with a variety of painful procedures (most commonly venipuncture [n = 32 of 106; 30.2%] or dental procedures [n = 28 of 106; 26.4%]); we located no studies reporting on acutely painful conditions. The most common digital technology distractions were traditional audiovisual aids (ie, watching a movie, cartoon, or program; n = 45 of 106; 42.5%), followed by virtual reality (n = 22 of 106; 20.8%) and audiovisual eyeglasses (ie, goggles through which children can watch television or movies; n = 22 of 106; 20.8%). In approximately half of the studies (n = 51 of 106; 48.1%), additional pharmacologic treatments for pain were provided to all study participants.

Supplemental Tables 10 through 13 show the detailed risk-of-bias judgments for the included studies. We judged the overall risk of bias to be high among all studies. Common sources of potential bias included a lack of blinding, inadequate randomization and allocation concealment, the use of crossover designs, block randomization in unblinded trials, baseline imbalances, and a lack of control groups in the observational studies.

Table 3 presents the summary of findings for each meta-analysis by outcome (see the Supplemental Information for all forest plots); within this table, we have also reexpressed SMDs on common scales for ease of interpretation. Table 4 provides a summary of findings for studies not included in the meta-analyses because of design (not an RCT or quasi RCT) or a lack of suitable data. In most cases, the findings of these studies did not differ substantially from the results of the meta-analyses. Table 5 shows the findings from additional comparisons that involved only 1 study each as well as studies without a comparator.

Compared with usual care, the use of digital technology distraction may result in a modest but clinically important reduction in pain for children undergoing painful procedures. We observed a small reduction in self-reported pain (SMD −0.48; 95% CI −0.66 to −0.29; I² = 72%; 46 RCTs; 2609 children; very low certainty),^† moderate reduction in observer-reported pain (SMD −0.68; 95% CI −0.91 to −0.45; I² = 83%; 17 RCTs; 1199 children; low certainty),^‡ and small reduction in behavioral measures of pain (SMD −0.57; 95% CI −0.94 to −0.19; I² = 89%; 19 RCTs; 1173 children; very low certainty)^§ with digital technologies.

Compared with usual care, the use of digital technology distraction may result in a modest but clinically important reduction in distress, as reported by children undergoing painful procedures and their observers (eg, caregivers and health care providers). We observed a moderate reduction in self-reported distress (SMD −0.49; 95% CI −0.70 to −0.27; I² = 78%; 19 RCTs; 1818 children; very low certainty)^|| and a small reduction observer-reported distress (SMD −0.47; 95% CI −0.77 to −0.17; I² = 73%; 10 RCTs; 826 children, low certainty)^¶ with digital technology. The reduction in behavioral distress observed was small and potentially not clinically important (SMD −0.35; 95% CI −0.59 to −0.12; I² = 72%; 17 RCTs; 1264 children; very low certainty).^#

Within-study analyses by age (n = 31)^** and/or sex (n = 25)^†† generally showed little to no difference in pain or distress scores; when a difference was noted, in most cases, younger children and girls experienced more pain and/or distress than did older children and boys.

When assessed by using behavioral measures of pain (eg, Face, Legs, Activity, Cry, Consolability scale [FLACC]), digital technology with active distraction (ie, requiring child engagement) resulted in reduced pain compared with usual care (SMD −0.92; 95% CI −1.52 to −0.31; I² = 95%; 6 RCTs; 387 children), but there was little to no effect for passive distraction (SMD −0.06; 95% CI −0.27 to 0.14; I² = 22%; 8 RCTs; 492 children). Sensitivity analyses showed no significant change in effects with the removal of quasi RCTs and crossover trials or industry-funded studies except for behavioral measures of pain, for which the effect for industry-funded studies appeared to be slightly greater (SMD −0.13; 95% CI −0.58 to 0.31; I² = 59%; 4 RCTs; 207 children) than for nonindustry-funded studies (SMD −0.05; 95% CI −0.30 to 0.20; I² = 30%; 6 RCTs; 384 children). We did not detect small-study bias.

Compared with nondigital distractors, the use of digital technology may result in a modest but clinically important reduction in pain for children undergoing painful procedures. We observed a small reduction in self-reported pain (SMD −0.27; 95% CI −0.56 to 0.02; I² = 71%; 8 RCTs; 771 children; low certainty)^‡‡ and moderate reduction in observer-reported pain (SMD −0.66; 95% CI −1.08 to 0.23; I² = 45%; 2 RCTs; 181 children; very low certainty)^41,77  with digital technology. However, the evidence for these outcomes was uncertain, and CIs included the potential for little to no difference. There were no data available for meta-analysis for the behavioral pain outcome.

When compared with nondigital distraction, the impact of digital technology on distress among children undergoing painful procedures is unclear. We observed a moderate reduction in self-reported distress (SMD −0.92; 95% CI −1.43 to −0.41; I² = 69%; 3 RCTs; 235 children; very low certainty)^77,109,111  with digital technology. Observer-reported distress was slightly higher for children distracted by digital technology compared with nondigital distractors (SMD −0.32; 95% CI −0.98 to 0.34; I² = 78%; 3 RCTs; 209 children; very low certainty),^77,94,99  but this effect was very uncertain, and the CI included the potential for no important difference. Behavioral measures of distress were reduced in children distracted with nondigital distractors compared with digital ones, but the difference was small enough that it may not be clinically important (SMD 0.21; 95% CI −0.02 to 0.45; I² = 0%; 4 RCTs; 288 children; very low certainty).^48,91,94,99 

One study commented on within-study differences in pain by age and sex, finding no difference between groups.⁴¹  The small number of included studies for each comparison hindered our ability to identify any credible between-study subgroup differences. One RCT (n = 89) excluded from the meta-analysis showed no difference when comparing digital technologies to active nondigital distraction (ie, singing or blowing bubbles) but favored digital technology when compared with passive nondigital distraction (ie, being read a fairy tale),⁴³  whereas another nonrandomized study showed little to no difference between groups.⁸⁹  Sensitivity analyses showed no significant change in effects with the removal of quasi RCTs and crossover trials or industry-funded studies. Small-study bias was not detected.

The findings of 2 studies showed no important difference in self-reported pain for distraction via audiovisual eyeglasses compared with a projected movie during minor dental procedures (MD 0.01; 95% CI −2.36 to 2.39; I² = 99%; 2 RCTs; 150 children; very low certainty).^63,64  No other outcomes of interest were available in these studies.

Three RCTs (n = 117) examined distraction by using video games compared with virtual reality (2 for needle-related procedures and 1 for burn dressing changes).^{68,101,102,137}  The CIs for these analyses were wide; thus, it was impossible to determine with any certainty whether any clinically important differences in self-reported, observer-reported, or behavioral pain might exist between the interventions.^{68,101,102,137}  No distress-related outcomes were reported.

The small number of included studies for each comparison hindered our ability to identify between-study subgroup differences, perform sensitivity analyses, or test for small-study bias.

There is a large and rapidly accumulating body of evidence exploring the use of digital technology distractors for children’s pain and distress during painful medical procedures (especially needle-related procedures, minor dental procedures, and burn treatments) when compared with usual care. We identified no studies reporting on acutely painful conditions. On the basis of the included studies, distraction appears to have a positive effect on both pain and distress, but digital technology does not seem to confer an obvious advantage over nondigital distractors. Certainty in the evidence ranged from low to very low, driven largely by risk of bias and inconsistency in effect estimates. More head-to-head trials are needed to clarify which type of distraction might work best, for which children and procedure, and in what setting.

There is little consensus in the literature on how to define a clinically significant difference in pain.¹³⁹  On the basis of available literature,^101,140,141  the reductions in pain associated with use of digital technology distractors may be considered clinically important (estimated between 10% and 20% on commonly used scales), although CIs were sometimes wide (including small effects). Large, high-quality trials are needed to improve the precision of current effect estimates. We did not specifically investigate the occurrence of adverse events, but these seemed to be limited to mild “simulator sickness” from virtual reality devices.^120,123  Knowing that inadequate pain management can have both short- and long-term consequences for children,^2,7  the potential for small to moderate reductions in pain and distress in the presence of what appears to be minimal risk supports the use of digital technology distraction for children and teenagers undergoing acutely painful procedures. Given proposed mechanisms,^17,142  it seems plausible that digital technology distraction could be beneficial for children with acutely painful conditions, but this cannot be confirmed by our review. Understanding the potential value of digital technology distraction for acutely painful conditions should be a future research focus given the ubiquitous nature of this presentation.

Evidence from our meta-analyses showed little to no difference in the effect of various digital technologies on pain compared with nondigital distractors; there was a possible greater effect on distress reduction with digital technologies, but this was supported by only 3 trials. Similarly, a 2018 Cochrane Review on the effect of both digital (eg, watching cartoons or movies and using handheld video games) and nondigital (eg, music, playing with a toy, and presence of a medical clown) distraction on pain during needle procedures showed that a variety of distractors are likely to be equally effective.¹³  Although we focused on digital distraction, our findings extend those of Birnie et al’s¹³  recent systematic review, which showed that distraction works, but it is unclear which type of distraction is best suited for children on the basis of age, stage of development, or clinical setting. In the absence of concrete empirical evidence, health care providers should focus on implementing the type of distraction that is most feasible in their particular setting while taking into account the needs of their patient population (eg, child preference, type of procedure, and available resources and time).

Feasibility and resources should be taken into account when deciding whether to implement digital technology distractors. The purchase of digital devices can be costly, as is the replacement of damaged devices or upgrades for novel technologies that may rapidly become out of date. However, digital technologies (eg, virtual reality, tablets, and humanoid robots) may offer some unique advantages compared with nondigital distractions (eg, bubbles, reading a story, and puppets). Many digital technology distractions require little if any training to implement. Many families commonly have access to their own portable handheld devices (eg, smartphones and tablets), which can be used for various forms of distraction (eg, game applications [apps] and videos),¹⁴³  and their use avoids important infection control issues that are of critical importance in health care settings. Distraction provided by virtual reality and humanoid robots might offer the most novel and immersive experience,^144,145  at least for children using them infrequently. A few individual studies from our review suggest that active forms of distraction (eg, video games and virtual reality) might be more effective than passive distraction (eg, cartoons and movies),^76,77,80  but this is not yet firmly established.^31,35 

Although evidence on the use of digital technologies is rapidly accumulating, many of the included studies were likely underpowered to detect statistically significant differences. An essentially insurmountable weakness is that it is impossible to blind children and personnel to the intervention. This should not overshadow the other prevalent methodologic and reporting deficiencies that trialists could address (eg, inadequate description of allocation concealment, baseline imbalances, crossover effects, and heterogeneity in outcome assessment tools). To improve the quality of future research, we suggest that trialists consult established guidance (eg, Consolidated Standards of Reporting Trials¹⁴⁶  and the Standards for Research in Child Health¹⁴⁷ ) during trial planning and conduct. In addition, consideration of the recommendations of the Pediatric Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials¹⁴⁸  regarding standardization of outcome measures will facilitate the synthesis of findings across studies. The inclusion of head-to-head comparisons and/or within-study subgroup analyses (eg, by age, sex, type of procedure, type of distractor, setting) would provide valuable information on the applicability of the findings to particular patient groups.

The review was limited by the quality of the included studies. Generalization of the findings to young children (eg, toddlers) should be made cautiously because we found few studies for these populations. It is not possible to generalize the findings to children with acutely painful conditions at this time.

The findings of this review have important clinical and research implications. Health care providers should strongly consider digital technology and other forms of distraction as a pain- and distress-reduction strategy for children and teenagers during common painful procedures (eg, venipuncture, minor dental procedures, and burn treatments). In the absence of sufficient data to make recommendations for specific distractors or populations, health care providers should choose distractors that are feasible to implement in their setting and desirable to the individual. Because of a lack of evidence, it is not possible at this time to know whether digital technology distraction would be a useful strategy to reduce pain and distress among children with acutely painful conditions. Research efforts should be focused on the planning of large, adequately powered trials that will contribute meaningfully to clinical recommendations.

We thank MacKinna Hauff for article retrieval.

App

application

CADTH

Canadian Agency for Drugs and Technologies in Health

CI

confidence interval

CINAHL

Cumulative Index to Nursing and Allied Health Literature

EBSCOhost

Elton B. Stephens Co. host

ED

emergency department

FLACC

Face, Legs, Activity, Cry, Consolability scale

MD

mean difference

RCT

randomized controlled trial

SMD

standardized mean difference

TRIP

Turning Research Into Practice