Many children visiting the emergency department (ED) experience pain. Pain in children has unintended but negative consequences. In the short term, acute pain heightens pain perception, anxiety, and fear, undermining the relationship of trust between child, family, and clinicians. In the long term, insufficient pain control can induce the avoidance of future medical procedures, such as vaccination and dental care. Insufficient pain control obstructs efficient and satisfying care in the ED. It costs more time and therefore obstructs workflow. It has negative effects on involved caregivers.1,2
Therefore, adequate pain relief is important for children visiting the ED. Although there are many reports about pain treatment of children, it is not known which interventions are the best for optimal pain treatment leading to complete pain relief. That was the reason to study research articles on clinical trials in which the treatment of pain in children in the ED was described. The primary interest was the outcome measurements used, whether complete pain relief was the aim, and which method had the best results. The ultimate goal was to find the best method to help children experiencing pain in the ED to get rid of the pain.
With the help of the hospital librarians, PubMed, the Cochrane Database of Systematic Reviews, and EMBASE, the author searched for articles, published from January 1, 2010, to September 30, 2020, of clinical trials and of randomized clinical trials on pain in children in the ED. Search terms included “pain,” “children (aged 1–18 years),” and “emergency department.” Only articles in English were used. Inclusion criteria consisted of predictable and identifiable pain, for example, after trauma or during procedures. Articles on possible recurrent pain, such as in sickle cell crises, were excluded. Studies on chronic pain and on pain with wide differential diagnoses, such as headache and abdominal pain, were also excluded. A formal protocol for this review was not developed; however, the checklist for PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) was used.
The continent and country where the study was performed were noted, as was the number of children included. The cause of pain was also noted as caused by a trauma or an intervention, such as blood sampling. The method of intervention (pharmacological or not), arms per study, comparator, time between assessments of pain, methods of pain measurement, and use of physiological parameters were collected. Furthermore, reporting statistically significant and clinically relevant differences between intervention and comparator were sought. Clinical relevance was defined as a difference of at least 2 points on the scale used, as suggested by some authors.3–5 The lowest outcome of pain scores was noted. If in doubt, the best outcomes, that is, the lowest pain scores, were used. For comparison, all outcomes were converted to a scale running from 0 to 100. Finally, it was ascertained whether the absence of pain was mentioned in the texts.
All data were stored in Excel (Microsoft Inc., Redmond, Wash). For the analysis of descriptive statistics, SPSS version 26 (IBM Corp, Armonk, NY) was used.
The literature search generated 1068 hits, and after removal of 448 double references, 620 titles and abstracts were read. Of these, 575 articles were excluded (Fig. 1). Forty-five articles were included (see Supplement S1, https://links.lww.com/PEC/A999 for references of included articles).
Table 1 shows the characteristics of these articles, for an overview (see Supplement S2, https://links.lww.com/PEC/A1000). Most studies were done in the United States and Canada (22; 49%); 6 (16%) came from Turkey. The number of participating children ranged from 28 to 501 (mean, 119; median, 85).
Characteristics of the 45 Clinical Trials on Pain in Children at the ED, Published in the Years 2015–2020
|Continent of origin|
|Country where the study was performed|
|United States||15 (33)|
|No. participating children, mean, median, range||119, 85, 28–501|
|Cause of pain|
|Venous access for blood sampling or cannulation||16 (36%)|
|Laceration repair||5 (11)|
|Intramuscular injection||2 (4)|
|Not specified||7 (16)%|
|2-Arm study||36 (80)|
|3-Arm study||9 (20)|
|Pharmacological/nonpharmacological intervention||20 (44)|
|Drugs studied||Fentanyl (nasal, parenteral); hydrocodone; ibuprofen, paracetamol; codeine, ketamine (systemic, nasal); lidocaine; morphine; midazolam, nitrous oxide|
|Nonpharmacological interventions studied|
|Information and instruction||5|
|Other (distraction, light, presence of dedicated person)||2|
|Tool used for pain measurement, performed by the children*|
|Physiological parameters studied†||6 (13): Heart rate (3), blood pressure (1), cortisol (2)|
|No. studies that measured pain during or directly after procedure||40 (89)|
|Pain measured in first 15 min after intervention||2 (4)|
|Pain measured 15–60 min after intervention||10 (22)|
|Pain measured 1–24 h after intervention||6 (13)|
*In some studies, more than 1 pain measurement tool was used.
†In some studies, more than 1 intervention was studied.
In 15 studies (33%), pain was caused by trauma. In 23 (51%), pain was caused by an intervention, mostly by gaining venous access for blood sampling or venous cannulation (16 [36%]). In 7 studies, the cause of pain was not specified.
Twenty studies (44%) reported a pharmacological intervention. Drugs studied included fentanyl, hydrocodone, ibuprofen, acetaminophen (paracetamol), codeine, ketamine, lidocaine, morphine, midazolam, and nitrous oxide. Nonpharmacological interventions included virtual reality, electronic devices, medical clowns, cartoons, information and instruction, music, cold, soap bubbles, and vibration.
To estimate the severity of pain, a visual analog scale (VAS) was used 25 times (56%); the FACES Pain Scale—Revised (FPS-R), 18 times (40%); and the Wong-Baker scale (WBS), 10 times (22%). Six studies (13%) used physiological measurements such as heart rate, blood pressure, and cortisol levels. In most studies (40 [89%]), pain was estimated during or directly after the procedure.
A statistically significant difference between the intervention and the comparator (mostly standard care) was described in 24 (53%) of the studies. Pharmacological interventions lead to statistically significant pain reduction in 13 of the 20 studies (65%). Nonpharmacological methods resulted in a statistically significant reduction in pain in 6 studies (24%). Clinical relevance, defined as a difference of at least 2 points in the outcome measurement used, was reported in 21 (47%) of the articles. The mean of all lowest pain scores was 26 and the median was 30 (range 0–51). In 4 articles (9%), absence of pain was reached, sought, or mentioned in the text (Table 2).
Outcomes of the Intervention for the Treatment of Pain in 45 Clinical Trials (n = 45)
|Statistically significant difference between intervention studied and comparator (all studies)||24/45 (53)|
|Pharmacological interventions||13/20 (65)|
|Nonpharmacological interventions||6/25 (24)|
|Clinical relevance between intervention and comparator*||21 (47)|
|Absence of pain noted or mentioned||4 (9)|
|Lowest pain score of intervention, mean, median, minimum–maximum†||26, 30, 0–51|
*Clinical relevance was defined as an improvement of at least 2 points.
†For convenience and to facilitate comparison, all outcomes were converted to a scale running from 0 to 100.
In this study of 45 articles about the treatment of acute, predictable, and identifiable pain in children in ED, a reduction in pain scores that was considered statistically or clinically relevant was described in half of the studies. Even after successful interventions, the mean pain score reported was still elevated, indicating persisting pain. In only 4 articles, the authors mentioned the absence of pain as an important outcome.
Experiencing pain has negative effects in both the short and the long term. It is widely recognized that pain should be treated effectively and immediately. Many methods to treat pain have been studied, and not only pharmacological interventions have been evaluated. Some studies showed admirable creativity in introducing original nonpharmacological methods, such as medical clowns, cold, bubble blowing, and music. However, of these 25 nonpharmacological methods, only 6 (24%) resulted in a statistically significant reduction in pain.
Measuring the severity of pain in children is not easy; to date, there are no simple and reliable methods. In 1 study, one of the outcome measurements of successful pain management in venous cannulation was the reduction in the number of adults needed to hold the fighting child.6 This is not a sensitive parameter.
None of the 6 studies that included physiological measurements found statistically significant differences. This is easy to understand: even in healthy children, the parameters used (eg, blood pressure, heart rate, and salivary cortisol levels) tend to vary widely.
Most of the included studies used semiquantitative methods, such as VAS, FPS-R, and WBS. These methods were developed to help children describe their pain experience more precisely.7,8 However, these instruments were never developed for research purposes.9 Moreover, these instruments differ from each other, making direct comparison between the outcomes of studies almost impossible. Recently, the literature about tools for pain measurements in children was reviewed. The authors recommend using the FPS-R for acute pain. Evidence for sufficient reliability of other instruments was low.10
Using the VAS, FPS-R, and WBS, 24 studies (53%) found a statistically significant reduction in pain. Twenty-one articles (47%) described a clinically relevant reduction. Only 1 article expressed disappointment in not succeeding in making children completely free of pain.5 Another article stated explicitly that the aim was to find a dose to make children pain-free.11 These authors were successful in reaching that goal. It seems that, in most other studies, the most important outcome was finding statistical significance.
In recent years, a few Cochrane reviews on pain in children in the ED were published. Harrison et al12 reviewed the studies on the effect of sweet-tasting drinks in needle-related pain in children. This study summarized differences between the means of semiquantitative studies. Murphy et al13 studied the trials on the effect of intranasal fentanyl. Again, the reduction in pain scores was the outcome of interest. Also, in the systematic review by Lambert et al,14 the effect of using virtual reality was measured as a reduction in pain scores. According to these authors, none of these systemic reviews offer enough evidence for certain conclusions. Several studies and guidelines on pain in children in the ED have been published.1,2,14–17 Most of them recommend assessing pain and optimizing treatment. However, in none of them a firm statement advising complete pain relief was found.
Before a conclusion is reached, possible limitations of this work should be discussed. The first limitation is that relevant articles could be missed. However, by using 3 databases, it is probable that the most read, important, and relevant studies are included. Another limitation is that only 1 person retrieved information, which can introduce bias. Moreover, information could have been misinterpreted. This possibility cannot be excluded, although effort was put to use clear numbers and figures from the articles. All data used can be checked because the list of included articles is available. A further limitation is that articles dealing with pain such as headache and abdominal pain were excluded. However, it is unlikely that these articles may include methods for immediate and definite pain relief not used in other circumstances.
Taking the limitations into account, questions arise. Are statistical significance and a clinical relevance difference of 2 points really important for children? Are children happy when the mean score of an intervention reaches statistical significance compared with the mean score of a comparator? What do these outcomes mean if the instruments used are not designed for that purpose? Is a clinically relevant reduction of 2 points on a 6- or 10-point scale what children want?
Children want to get rid of the pain; they do not seek statistical significance. Although it might be too idealistic aiming for complete pain relief in all individual situations, reaching complete pain relief should be the aim in clinical trials about pain in children visiting the ED. The lowest score on any scale should be the primary outcome. The FPS-R should be used to measure acute pain. Individual tailored pain management should contain effective pharmacological and nonpharmacological interventions. Modern research on pain treatment in children should look at combinations of interventions with the only relevant outcome measure: no pain at all.
In most studies about pain in children visiting the ED, the aim was to find a statistically significant reduction; instruments are used that were not developed for research purposes. Too few articles mention the absence of pain. Children deserve better. Complete pain relief for children should be the aim of research in this field.
The author thanks the hospital librarians for their help in performing the literature search and retrieving articles. Scribbr Editing Service helped improve the English language. Brita de Jong gave helpful comments on an earlier version of the manuscript.