OBJECTIVE: Hospitalized children experience significant pain despite improvement efforts. This study was undertaken to better understand the epidemiology of acute pain in hospitalized children and the extent to which existing measures reveal targets for improving pain management.
METHODS: A cross-sectional survey was used to audit pain assessment, intensity, prevalence, source, and treatment hospital-wide on a single day in 2011. Chart audits were performed on patients aged 0 to 21 years. All patients had the option to participate in a structured interview about their pain experience.
RESULTS: The audit included 112 children, 47 of whom were interviewed. Pain prevalence obtained by child/parent interview (72%) was more than twice that documented by nurses (30%). Infants, but not cognitively impaired children, had significantly lower rates of pain detection and analgesic ordering than older age groups. Procedural pain was the most frequently cited source of pain among interviewed patients and was poorly addressed in the medical record. Fifty percent of children with documented moderate-to-severe pain received scheduled pain medications. More than one-third of interviewed patients would have wanted more pain medication if it could have been safely given.
CONCLUSIONS: Specific gaps remain in the quality of pain management provided to hospitalized children. Focus on infant pain detection, assessment and management of procedural pain, and scheduled analgesic ordering are sensible targets for future process improvement efforts.
Pain is a nearly universal experience among hospitalized children.1 Efforts to prevent and alleviate a child’s pain can shape families’ overall experience in the hospital setting.2–4 Inadequate pain control causes unnecessary suffering and is associated with a number of harms including postoperative complications, chronic pain development, and needle phobia.2,5,6 Pediatricians and others caring for children have an ethical obligation to treat and, if possible, prevent children’s pain.7–10
Fortunately, much is now known about the prevention and safe treatment of pediatric pain.2,11–16 Opioids, nonopioids, and dermal analgesics are accessible for children in the United States, and effective multimodal techniques extend pain management beyond pharmacotherapy. High-quality pain management is becoming increasingly recognized and rewarded by institutions such as the Joint Commission, U.S. News and World Report, and the ChildKind initiative, which recognizes pediatric institutions making a significant commitment to pediatric pain relief.
As a result, many hospitals have adopted processes to better identify and treat children’s pain. Yet despite these efforts, studies show that hospitalized children are still experiencing moderate to severe pain in high numbers.1,17–22 Additionally, review of our patient satisfaction data suggested we were not consistently meeting expectations with regard to pain management in our own institution.
Therefore, we felt it was critically important to understand the epidemiology of acute pain within the current practice environment and the extent to which existing processes and measures common to US children’s hospitals are or are not sufficient for achieving effective pain control in children. We were particularly interested in whether children unable to communicate by standard means, including infants, non-English-speaking patients, and those with severe cognitive impairment, might account for a large proportion of poorly identified pain.17,23,24
We undertook a cross-sectional survey to capture pain experience on a single day in 2011 at a 206-bed, tertiary care, free-standing children’s hospital in Norfolk, Virginia. The hospital had a full breadth of pediatric subspecialties and surgical services, and affiliation with a level I trauma center but no acute pain management service at the time of the audit. Regional pain techniques, such as epidural analgesics, were in limited use during the study period. The Institutional Review Board at Eastern Virginia Medical School approved this study.
All patients aged 0 to 21 years on the medical/surgical wards and PICU and a subset of patients from the NICU who were listed on the hospital census at 8:00 am on the study day were included in the chart audit and were eligible for inclusion in the interview portion of the study. A 25% convenience sample from the 60-bed NICU was selected because of time constraints and to prevent overrepresentation of data from a single unit. Subjects were approached by a member of their own health care team with a flyer requesting their participation in the interview before being consented by the research team. Patients were excluded from the interview if they were not in their hospital room after 3 visits, if they had been discharged, or if a parent was unreachable in person or by telephone to provide consent.
Data were collected from the bedside nursing chart, electronic medical record, and interview by 4 pairs of nurse or physician research team members using a scripted questionnaire modified with permission from Taylor et al (see Supplement).17 Research team members underwent a 2-hour training session before data collection and were unaffiliated with patient care on the day of the audit. A second reviewer independently reviewed each chart to ensure reliability. During interviews, patient responses were sought, with parent responses serving as proxy if the child was incapable of answering due to age, cognition, or clinical status.
Patient age, gender, admitting diagnosis, and service type were collected by bedside chart audit. Subjects were grouped by medical, surgical, PICU, and NICU service types. Patient age was recorded in months and categorized as infant (<1 year old), toddler (1–3 years old), or older child (>3 years old). The inability to communicate in English or presence of significant cognitive delay (defined as a child older than 3 years unable to verbally communicate pain) was elicited from the parent, or if unavailable, the bedside nurse.
Pain Assessment and Treatment
Bedside charts from the preceding 24 hours were reviewed to determine the frequency of nursing pain assessment documentation, pain assessment tool in use for each patient, maximal recorded pain intensity, procedures documented, and methods described for procedural pain reduction. All analgesics and adjuvants prescribed over the preceding 24 hours were recorded from the medication administration record, along with the route, frequency (scheduled around the clock or pro re nata [PRN]), and number of PRN doses administered. Analgesics were categorized as opioid or nonopioid.
Interviewed children or their proxies (hereafter referred to together as “patients”) who answered “yes” to the question of whether they had experienced pain during this hospital stay were asked whether the cause of their pain was surgical, disease-related, procedural, unknown, or other. A patient could select multiple sources of pain. All interviewed patients, including those who did not volunteer procedures as a source of their pain, were asked whether they had undergone any procedures in the preceding 24 hours.
Pain Prevalence and Intensity
Pain intensity scores from each pain scale in use were converted into the categories of no, mild, moderate, or severe pain (Table 1).20,25 Interviewed patients were asked whether they had any pain during this hospital stay, and, if so, the “strongest pain” they had experienced in the last 24 hours, using the 4-point simple Pain Word Rating Scale (PWRS).17,26 Pain intensity ratings collected by interview were evaluated for concordance with nurse-charted maximal pain intensity ratings.
Perceptions of Pain Treatment
Interviewed patients who had experienced pain were asked whether they had received a pain medication within the past 24 hours, whether it was helpful, and whether they would have wanted more pain medication if it could have been prescribed safely. Patients who described procedures within the past 24 hours were asked whether anything (such as creams, positioning, or distraction) had been done to decrease pain during the procedure and whether that intervention was helpful.
Patient characteristics and pain assessment, treatment, prevalence, intensity, source, and perception of treatment were summarized quantitatively. Confidence intervals for proportions were calculated at 95% with Fisher’s exact method. Sensitivities and specificities of pain assessment tools were calculated by comparison with pain reported by PWRS. Differences in pain detection and pain management across age groups and by cognitive status were analyzed by using χ2 and Fisher exact tests. P values <.05 were considered significant. Odds ratios (ORs) and confidence intervals (CI) for Fisher exact tests are reported as calculated with OpenEpi software, version 3.01 (updated April 6, 2013).
Chart audit was performed on 112 eligible patients the day of the study. Forty-seven subjects returning interest flyers were interviewed (Fig 1).
Patients included in the chart audit were girls (54%) and boys with age ranging from 1 day to 21 years; 44 were infants, 17 toddlers, and 51 older children (Table 2). Twelve (11%) had significant cognitive impairment, and 4 did not speak English. Interviewed subjects were similar in gender distribution (55% girls) and age range (5 weeks–21 years old). Eight of the 47 interviewed patients (17%) “usually” had pain before admission.
Pain assessment documentation frequency varied from 6 times per 24-hour period on medical services to 21 times per 24-hour period in the PICU. The pain assessment tool used was documented in 87% of patient charts and was appropriate to the age and communication ability of the patient in 96% of cases. The Faces, Legs, Activity, Crying, Consolability (FLACC) scale was used most frequently (n = 54) followed by the 0 to 10 Numerical Rating Scale (NRS; n = 24). The Neonatal Infant Pain Scale (NIPS) was used exclusively in the NICU (n = 15). Procedures were inconsistently documented throughout the nursing record, and procedural pain assessment was never documented. Forty-seven patients were also evaluated with the PWRS on the study day (interviewed subgroup).
Chart audit revealed that 76% of all patients had at least 1 analgesic ordered (Table 3). Infants were significantly less likely than toddlers and older children to have analgesics ordered (64% vs 84%, OR 0.34; 95% CI 0.12–0.90) unless they had nurse-documented pain. Children with nurse-documented pain were significantly more likely to have analgesics ordered than those without (97% vs 67%, OR 16.2; 95% CI: 2.8–351.5). Of those children with documented moderate to severe pain, 50% were receiving scheduled analgesics. Children with moderate to severe pain and orders for PRN-only analgesics received a median of 3 doses over the preceding hours, representing 25% (range 12%–75%) of their available analgesia. Of 62 patients with procedures documented in the chart, 38 (61%) had documentation of procedural pain reduction, either pharmacological (24 of 38, 63%) or physical/psychological (16 of 38, 42%) (Table 4).
Thirty-seven interviewed subjects reported on the type of pain experienced during hospitalization. Procedural pain was the most common (n = 20, 54%) followed by disease-related pain (n = 19, 51%) and surgical pain (n = 15, 41%). Thirty-two percent reported >1 source of pain. Of children reporting procedural pain, the majority (18 of 20) experienced needlestick procedures, with an average of 2 per patient on the study day.
Pain Prevalence and Intensity
Chart audit revealed that 30% of children (n = 34 of 112, 95% CI: 22.0–39.34) had pain on the study day, including 25% of medical patients (n = 16), 61% of surgical patients (n = 11), 47% of PICU patients (n = 7), and 0% of NICU patients. Of the 34 with charted pain, maximal documented pain intensity was mild in 18% (n = 6), moderate in 53% (n = 18), and severe in 29% (n = 10).
Pain detection varied significantly by age (P < .001), with older children having the highest documented pain prevalence (51%) compared with 18% of toddlers and 11% of infants (Fig 2). Infants and toddlers had significantly lower odds of pain detection (0.15, 95% CI: 0.056–0.37) compared with older children. Pain detection in children with significant cognitive delay was similar to that in older children (42%, OR 0.62, 95% CI: 0.13–2.72). None of the 4 non-English-speaking patients had documented pain.
Pain detection also varied significantly according to the assessment tool used (P < .001) (Fig 3). Assuming the PWRS as the gold standard for pain detection, the NRS had 75% sensitivity and the FLACC 33% sensitivity for the detection of pain (specificities 100%). Insufficient numbers of interviewed patients were evaluated with the NIPS or FACES tools to determine their sensitivities. Replacing children’s charted pain assessment scores with their PWRS scores increased pain prevalence in the overall sample to 46%. When verbal pain intensity reports from the PWRS were compared with the same patients’ nursing documentation, 51% reported more intense pain, 43% reported the same pain intensity, and 6% reported lower pain intensity compared with their charted scores (Table 5).
Perceptions of Pain Treatment
Most (24 of 25) interviewed children who received an analgesic within the preceding 24 hours recalled receiving it and reported that pain medication was a little (25%) or very helpful (75%) in making the pain less bothersome. Interventions to reduce procedural pain were perceived as effective by 85% of the 13 children who reported receiving them. Thirty-eight percent of interviewed patients would have wanted more pain medication if it could have been given safely.
This study adds to the growing literature describing the epidemiology and treatment of pain in children’s hospitals across North America. Pediatric pain is prevalent, but typical assessment tools capture only part of a child’s inpatient pain experience.27,28 In our study, analgesic treatment patterns followed pain detection. The low rates of pain detection in infants correlated with lower analgesic prescription rates than for toddlers and older children, reaffirming effective pain assessment as a requisite first step to good pain control.
Nursing assessments determined that 30% of children were in pain on the study day. By comparison, 72% of interviewed patients reported pain during the same time period. Although it is possible that interviewed patients represented a subset more likely to report pain, the discrepancy persisted when interviewed patients’ self-report PWRS scores were substituted for nursing assessments in the overall sample. The high rate of nursing compliance with performing frequent assessments using the correct tool make poor process adherence an unlikely factor in explaining this difference.
The detection of nearly double the pain prevalence by interview compared with that documented in clinical records is consistent with previous pediatric literature, where pain prevalence rates up to 94% are recorded when patient report is used.1,20,29 This finding is concerning because it suggests that the current process of nursing pain assessment underestimates our children’s pain experience. Patient recall of pain over the preceding 24 hours is an established method of monitoring pain prevalence,1,17,26,30 as is nurse assessment of pain and its intensity at frequent intervals.31,32 However, pediatric self-report and observational pain assessment tools may be prone to influences not encountered in adult self-report, even when performed assiduously.1,23,28
We hypothesized that children unable to communicate due to age or cognitive impairment might account for a large proportion of underrecognized pain.33,34 We were reassured to see that the detection of pain in cognitively impaired children, primarily using the FLACC tool, was not significantly different from that in their unimpaired counterparts. However, detection of pain was inversely correlated with age: no neonate in the intensive care nursery had pain detected, and pain detection rates in infants and toddlers were only 11% and 18%, respectively. Given the breadth of diagnoses as well as the frequency of painful procedures typical for NICU patients, infants are at least as likely to have pain as are other age groups.35 Taylor reported high pain prevalence rates for NICU patients (>90%), and both Birnie and Groenewald found infants to be a high-risk group for significant pain.1,17,18 Many pain assessment tools are available for use on neonates and premature infants; it is plausible that use of different tools in a variety of settings is at least partially responsible for the variance in pain detection.
Both the discrepancy between nursing- and self-report of pain and the suboptimal detection of pain in infants may be attributable to a lack of focus on procedural pain with current assessment practices.1 The universality of needlestick procedures in hospitalized children, as well as the significant fear and pain accompanying them, is well established.6,10,30,35 Procedural pain was the most frequent pain source reported by interviewed children, yet none of the 112 bedside charts reviewed had space dedicated for procedural pain documentation. In the neonatal population, positioning, sucrose, nonnutritive sucking, and kangaroo care are frequently used, but our documentation did not reflect these practices, making judgments about the adequacy of procedural comfort unfeasible. Zhu et al demonstrate that prioritizing the inclusion of such documentation is a key step in facilitating the measurement and subsequent reduction of procedural pain.36
For children in whom pain was detected, our study confirms low rates of analgesic administration when medications are ordered as PRN only rather than scheduled.18,20,37 Analgesics ordered PRN only are typically given reactively rather than preemptively.38 For patients in whom pain detection is difficult or inadequate, as we demonstrate here, PRN dosing likely contributes to suboptimal pain treatment and prevention. Scheduling of nonopioid analgesics for children in acute postsurgical or disease-related pain improves pain scores without significantly increasing side effects,39 is opioid-sparing, and is advocated by pain management experts as one of the most effective ways to improve pain management.2,40 In our study population of interviewed patients, more than one-third would have wanted more pain medication if it could have been safely given.
To date, institutions achieving improvements in children’s pain outcomes have used hospital-wide, comprehensive process improvement strategies.41,42 We describe specific areas in which focused process improvements could be leveraged to measurably improve care. For the youngest patients who cannot self-report pain, observational measurement should be undertaken with the most accurate tools available for the clinical setting, by individuals educated in their use. Introducing purposeful conversation or a patient/parent comment card to elicit feedback on pain intensity and adequacy of treatment at the time of the encounter would be a useful adjunct to current assessments. Procedural pain should be specifically assessed and managed because it represents a “hidden” source of pain in a child’s hospital experience. Modification of nursing documentation to facilitate recording of procedural pain assessments and interventions is desirable. Adapting order entry systems to support scheduled nonopioid analgesics where clinically appropriate may also improve children’s pain experience without concerns for opioid overuse.
This study has several limitations. Our patient sample size was smaller than that of previously published studies. Those interviewed represented less than half of our overall study population, which may limit the significance and generalizability of our interview findings. Patients willing to participate in the interview may have had more pain than patients who decided not to participate, although our findings were congruent with those of previously published studies. We limited our NICU population to a convenience sample of 15 patients, who may have differed from the larger population. However, an unpublished performance improvement audit of our entire NICU population after the conclusion of this survey confirmed a <2% pain detection rate. Our audit was limited to a single day in our institution. Audits at different times of year, with differences in surgical, medical, and ICU volumes, may change proportions but would be unlikely to influence overall results or prescribing patterns.
In this study of pain prevalence in a US children’s hospital, neonates and infants had the lowest pain detection and treatment rates compared with older or cognitively impaired children. Inadequate procedural pain measurement contributed to low pain detection. Because pain that is detected is more likely to be treated, increasing focus on infant pain detection and detection of pain during procedures should lead to improved pain outcomes for children.
We thank Meaghann Weaver for adapting the questionnaire and David Austin and Wendy Davis for assistance with data collection.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
Dr Shomaker conceptualized and designed the study, coordinated data collection, carried out the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Ms Dutton and Dr Mark assisted with the initial study design, performed data collection, contributed to data interpretation, drafted portions of the initial manuscript, and critically reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.
- ↵American Academy of Pediatrics. Committee on Psychosocial Aspects of Child and Family HealthTask Force on Pain in Infants, Children, and Adolescents. The assessment and management of acute pain in infants, children, and adolescents. Pediatrics. 2001;108(3):793–797
- Taddio A,
- Appleton M,
- Bortolussi R,
- et al
- ↵Delegates to the International Pain Summit of the International Association for the Study of Pain (IASP). Declaration of Montreal. 2010. Accessed June 4, 2014, 2014.
- Institute of Medicine. Relieving Pain in America: A Blueprint for Transforming Prevention, Care, Education, and Research. Washington, DC: National Academies Press; 2011
- Walco GA
- Wells N,
- Pasero C,
- McCaffery M
- Portenoy R
- Macintyre PE,
- Schug SA,
- Scott DA,
- Visser EJ,
- Walker SM
- Schechter N,
- Berde C,
- Yaster M
- Friedrichsdorf SJ,
- Postier A,
- Eull D,
- et al
- Oakes L
- Gaffney A,
- McGrath P,
- Dick B
- Wells N,
- Pasero C,
- McCaffery M
- Organization TWH
- Twycross A,
- Dowden S
- Copyright © 2015 by the American Academy of Pediatrics