OBJECTIVE. To examine the relationship between pediatric patient visit characteristics and pain score documentation in the emergency department (ED) and determine whether documentation of a pain score is associated with increased analgesic use.
METHODS. A cross-sectional analysis was conducted of ED visits for pediatric patients from the National Hospital Ambulatory Medical Care Survey (1997–2000). Survey weighted regression first was used to assess the association between patient visit characteristics and pain score documentation. The regression then was repeated to determine the association between documentation of a pain score and analgesic use, adjusting for visit characteristics.
RESULTS. A total of 24707 visits were included. Only 44.5% of visits had documented pain scores. In the regression analysis, younger age, self-pay, visits to pediatric facilities, and visits that were not designated as injury related were associated with decreased pain score documentation. Documentation of pain score was associated with increased odds of an analgesic prescription and opioid prescription. When no pain score was documented, the odds of receiving any analgesic was similar to visits with pain documented as mild.
CONCLUSION. ED pain score documentation is suboptimal in the pediatric population. Infants and toddlers are at particular risk for not having a pain score documented. There is a significant association between pain score documentation and the use of any analgesic, particularly opioids. Improvements in pain documentation for acutely ill and injured children are needed to improve pain management.
Pain is one of the most common complaints in the emergency department (ED), with up to 78% of patients evaluated reporting pain.1–3 Despite the high frequency of painful complaints, pain often is not assessed or treated adequately. Oligoanalgesia, the lack of effective pain management, is a problem for both adult and pediatric ED patients.4 Children in particular have been identified to have inappropriate treatment of painful conditions and may needlessly experience pain with minor illnesses and injury.5–12 Inappropriate pain assessment is likely to be an important contributor to the poor pain management for children in the ED.
Standards and recommendations for pain assessment have been developed in an effort to improve pain management.13–17 Recognition of pain and its severity is central to these standards. Documentation and display of a pain report to improve recognition and treatment of pain was integral to the American Pain Society's Quality Improvement Guidelines for the Treatment of Acute Pain, released in 1995.14 In 2001, The Joint Commission on Accreditation of Health care Organizations (JCAHO) mandated the assessment and documentation of pain for all health care encounters. Because the ED is the initial site of evaluation for numerous children with painful complaints, it is important that pain assessment and documentation be optimized.
There is little literature on the assessment of documentation of pain scores in the ED for pediatric patients. Our objectives were to (1) describe national practices of pain score documentation for pediatric ED visits, (2) determine which patient characteristics and specific conditions are associated with differences in pain score documentation, and (3) evaluate the association between pain score documentation and analgesic use in the ED.
This was a cross-sectional study of visits in the National Hospital Ambulatory Medical Care Survey (NHAMCS). The NHAMCS is an annual multistage probability sample of visits to EDs and outpatient departments in the United States conducted by the Ambulatory Care Statistics Branch of the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention. This study is restricted to the ED data set. Findings are based on a national sample of visits to EDs in noninstitutional general (medical, surgical, and children's) and short-stay hospitals in the 50 states and the District of Columbia. The survey uses a 4-stage probability sample design involving geographic primary sampling units, hospitals within those areas, EDs within the hospitals, and patient visits within the EDs. Data collection and documentation are standardized by and monitored closely by the NCHS. Except when mentioned, items are noted to be blank or unknown for <5% of visits. More detailed descriptions of this sampling design and a full report of the methods used in these surveys are available from the NCHS.18,19
We examined data for ED patients who were ≤18 years of age between 1997 and 2000 because pain scores were recorded in NHAMCS only during these years. In an a priori manner, we selected predictor variables that are thought to affect pain score documentation. Data were extracted for patient demographics, site characteristics, method of payment, visit immediacy, reason for visit, physicians' diagnoses, and medication prescription.
Outcome Variable: Pain Score Documentation
Our primary outcome was the presence or absence of documented pain score; a secondary outcome was the pain severity. “Presenting level of pain” was a specific item for collection on the Patient Record Form that was used by the data collection agents. As pain score assessment tools were not standardized across participating hospitals, severity of pain was based on the Clinical Practice Guidelines published by Agency for Health Care Policy and Research, which provides a numerical pain intensity scale as follows: none, 0; mild, 1 to 3; moderate, 4 to 6; and severe, 7 to 10.20 Severity of pain was assigned at the discretion of the data collection agent on the basis of the pain score recorded on the patient's medical record. Pain score was classified as documented when severity was listed as none, mild, moderate, or severe. When listed as “unknown,” it was classified as not documented.
NHAMCS surveys were initiated in the last week of December in each year. The data that were collected in that last week therefore were reported for the following year. Each year's data were assessed individually and then compiled.
Race/ethnicity was determined by study personnel on the basis of their observations, unless hospital policy mandated that patients be asked directly. Race categories include white, black, Asian/Pacific Islander, and American Indian. Ethnicity is categorized as Hispanic or non-Hispanic. Ethnicity was a missing variable in 18.5% of visits.
Method of payment was grouped into 5 categories: private insurance, government source of payment (Medicaid, Medicare, or worker's compensation), self-pay, other (no charge or other), or unknown.
NHAMCS designates the term “visit immediacy” for triage assessment assigned by triage personnel at the time of arrival at the ED and was categorized into 4 levels on the basis of time expected until evaluation: emergent (<15 minutes), urgent (15–60 minutes), semi-urgent (1–2 hours), or nonurgent (2–24 hours). This variable was unknown in 25.5% of cases.
Hospital location was dichotomized into rural or urban on the basis of their metropolitan statistical area as coded by the US Office of Management and Budget. Hospitals were classified as pediatric facilities when 95% or more of visits to that ED were for patients who were younger than 21 years. Hospitals were classified as teaching hospitals when >10% of visits were evaluated by resident physicians in training.
Reason for Visit
Reasons for visit were coded according to “A Reason for Visit Classification for Ambulatory Care.”21 Up to 3 reasons for visit were coded, and only the first was used for analysis. Patients were assigned to the “painful reason for visit” group when the reason for visit included any of the following: crying, including crying too much; fussy; fidgety; irritable (1080.1); ear pain, including ear pain/pulling at ears (1355.0–1355.1, 1365.3); throat pain (1455.1–1455.2); abdominal pain (1545.0–1545.3); headache (1210.0); pelvic pain (1775.0–1775.1, 1775.3); and body pain, including pain, ache, soreness, discomfort, cramps, and contractures and spasms of the leg, knee, ankle, foot, toe, arm, elbow, wrist, hand, or finger (1920.1–1920.2, 1925.1–1925.2, 1930.1–1930.2, 1935.1–1935.2, 1945.1–1945.2, 1950.1–1950.2, 1955.1–1955.2, 1960.1–1960.2). All other patients were classified as “nonpainful reason for visit.”
NHAMCS data set specifically designates a visit to be related to an injury or poison at the discretion of the data abstractor. These visits were considered to be injury visits in this data analysis.
Diagnosis was coded according to the International Classification of Diseases, Ninth Revision, Clinical Modification. Up to 3 diagnoses were coded in sequence, and the first 2 diagnoses were used for analysis. Patients were assigned to the “painful diagnosis” group when the International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code included any of the following: fracture (800.0–829.9); burn (941–949.5); laceration (870.0–897.9); head injury, including head injury, concussion, and intracranial injury (959.01, 850.0–854.19); headache (784.0); abdominal pain (789.0–789.7); pharyngitis (462–463); and otitis media (381–382.99). All other patients were classified as “nonpainful diagnosis.” The American Academy of Pediatrics in its consensus statement for the assessment and management of acute pain in infants, children, and adolescents identified the last 4 diagnoses specifically in the management of acute illness.17 The others were included by consensus of the authors.
Patient medication information was collected, classified, and coded according to a unique classification scheme developed at the NCHS.22 Drugs that were classified specifically as “relief of pain” were selected from the List of National Drug Code Directory Drug Class. A medication was designated to be an analgesic prescription when it was classified as an analgesic/general (1720); analgesic, nonnarcotic (1722); antimigraine (1723); nonsteroidal anti-inflammatory drug (1727); or narcotic (1721). Analyses were done for all analgesic medications and opioid/narcotic analgesia only. Up to 6 drugs were recorded for each visit, and all were included in the analysis.
Descriptive statistics were used to summarize the data. The χ2 test was used to assess the association between individual demographic and clinical characteristics and pain score documentation. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Multivariate logistic regression that produce adjusted ORs and 95% CIs then was used to determine the independent association between each predictor variable and pain score documentation. We hypothesized a priori a differential association between painful complaint or painful diagnosis and pain score documentation in different age groups, so appropriate interaction terms were included in the model. We then used multivariate logistic regression to determine the association between pain score documentation and analgesic and opioid prescription, adjusting for the other covariates. All analyses were adjusted to account for survey weighting and clustering (using the svy commands in Stata 8.0; Stata Corp, College Station, TX) to yield national estimates, as recommended by the NCHS.
Our sample included 24707 visits, representing an estimated 20 million ED visits by children who were ≤18 years during the 4-year study period. Mean age was 7.5 years, and median age was 6 years. Table 1 provides demographic characteristics of the study population and their survey-weighted proportions.
Description of Pain Score Documentation for Pediatric ED Visits
Only 44.5% (95% CI: 42.1–47.0) of pediatric visits had documented pain scores. Table 2 shows the proportion of visits for patient characteristics, reasons for visits, and diagnoses. With increasing age, there were incrementally higher proportions of pain scores in preschool (aged 2–5 years) and older children (aged 6–18 years) compared with infants (aged 0–1 year). Visits that were categorized as a “painful reason for visit” and “painful diagnosis” were assessed individually and demonstrated pain score documentation in <60% of visits except those for pelvic pain. Visits for crying were noted to have the lowest proportion of pain scores documented. Pediatric ED visits for injuries that are known to be painful, including burns and fractures, had pain scores for only half of the visits. Acute illnesses including otitis media, pharyngitis, headache, and pelvic pain are identified in the American Academy of Pediatrics policy statement for the assessment and treatment of pain in children as diagnoses for which pain should be addressed.17 Our analysis revealed that pain is documented in <60% of these visits.
Association Between Predictors and Pain Score Documentation
A decreased likelihood of pain score documentation was associated with the following characteristics of visits: infants, preschool children, Asian race, Hispanic ethnicity, Midwest region, urban, pediatric EDs, government insurance and nonurgent visit immediacy (Table 2). Increased likelihood of pain score documentation was associated with painful reasons for visit, injury visits, painful diagnosis and prescription of opioid analgesic. Pain score documentation did not improve on a year over year basis, over the 4-year study period.
As hypothesized, significant associations were found between pain score documentation and age, reason for visit, and diagnosis, with significant interaction (Table 3). Compared with children who were older than 5 years, infants had 52% lower odds of having a pain score when the visit reason or diagnosis was not painful. Pain score documentation was slightly higher in infants with painful reason for visit (OR: 0.53; 95% CI: 0.40–0.71) and painful diagnosis (OR: 0.60; 95% CI: 0.50–0.72) but still significantly lower than that for older children with similar conditions. However, toddlers with a nonpainful reason for visit or diagnosis and painful diagnosis had lower odds of having a pain score (OR: 0.70 [95% CI: 0.60–0.81] and OR: 0.75 [95% CI: 0.63–0.91], respectively). Toddlers with painful reason for visit were found to have similar odds of pain score documentation compared with children who were older than 6 years and had a painful reason for visit. Visits to pediatric facilities were associated with lower odds of pain score documentation than nonpediatric facilities, as were visits for patients with self-payment as the primary payer. Increased odds of pain score documentation were associated with injury visits. Our analysis demonstrated no association between pain score documentation and race, ethnicity, geographic region, metropolitan area, or teaching hospital.
Association Between Pain Score and Analgesic Prescription
Only 52.9% of children with pain documented as severe received any analgesic in the ED (Fig 1). An association was found between pain score documentation and analgesic prescription. Adjusting for the other covariates, the odds of an analgesic prescription was 1.19 (95% CI: 1.07–1.31) when a pain score was documented. Odds were higher 1.64 (95% CI: 1.29–2.09) for a narcotic prescription when a pain score was designated. Stratification by pain score severity showed that opioid analgesic prescription has a statistically significant incremental increase with increasing pain severity. The odds of receiving analgesic or opioid prescription when no pain score is documented is similar to the odds of receiving a prescription when the pain is assessed as mild (Table 4).
Nationally, documentation of pain scores for pediatric ED visits is low, occurring in <50% of visits. Similar to our findings, a recent study evaluated pain assessment documentation in children in Illinois EDs and found that 59% of visits assessed pain.23 The low proportion of visits with pain score documentation is inadequate in the face of literature and experience that identify pain as a frequent and significant problem for the majority of patients who present to the ED.1,2 Although our study sample encompasses patient visits before the 2001 JCAHO mandate for pain score documentation, national practice guidelines have supported their use in the management of pain since 1995.14 This analysis identifies the youngest children as those at particularly high risk for poor documentation, infants more so than toddlers.
Whether the low proportion of pain scores that are documented in children is attributable to lack of recognition of pain or inability to quantify pain in these patients is unclear. Recognition is particularly important in the youngest children because they do not articulate their pain. The complaint of “crying” has one of the lowest proportions of pain score documented of all variables assessed. Furthermore, toddlers had a similar proportion of pain score assessments as their older counterparts with painful reasons for visit. This suggests that the toddlers, who indeed express pain in ways similar to older children, are recognized to have pain by providers at a similar proportion to their older counterparts and have similar rates of pain score assessment, but infants do not have pain identified and therefore do not have pain scores assessed as often. Alternatively, inability to quantify pain in young children may lead to low documentation rates. Although there are many validated pain scoring scales for pediatric patients, objective assessment of preverbal children can be difficult.12 The use and the availability of these scales in clinical practice may not be widespread, and inexperience with observational scoring devices may make them time-consuming and impractical for providers in a busy ED. In this case, the pain may be recognized but documentation of a pain score not performed in the younger children.
The low proportion of children with pain score documentation at pediatric facilities is surprising because these sites specialize in the unique care that children require. Pediatric EDs have a higher proportion of children with developmental delay or cognitive disabilities, and it is only in recent literature that pain score assessment tools are being validated for this distinctive population.24,25 However, the small numbers of these children is unlikely to result in this significant difference. This also may be a case of “reverse technology transfer” in that pain assessment research is extensive in the adult literature but only recently has been evaluated in pediatrics. Many nonpediatric facilities may have policies for pain scoring in place and routinely obtain score ratings on all patients evaluated.
Some health care providers may not consider pain scores to be accurate or useful, instead recognizing and treating pain without documenting pain scores. Although pain scales are used in pain research, their clinical practicality is called into question. Whether clinicians use pain scores in decision-making for diagnosis or for treatment plans may affect patterns of documentation. For example, pain score documentation was low for visits with complaints that commonly are considered painful, including ear pain and sore throat. In contrast, a higher proportion of visits had pain assessments for children with abdominal, pelvic, and headache pain. Diagnostic uncertainty in the latter cases might lead to a differential rate of pain score documentation. In these visits, pain assessments guided the clinician in the patient evaluation. Alternatively, the ED is the site of evaluation for many common painful pediatric illnesses and injuries that are not diagnostic dilemmas but are known to be painful. Our analysis shows significantly higher odds of pain score documentation for these visits. In these instances, in contrast, pain assessment was used to guide analgesic prescription.
With the clinical utility of pain scores in question, determining whether pain scores guide analgesic prescription was assessed. As expected, we found a significant association between increasing pain score severity and increasing opioid prescription: the proportion of patients who received opioids doubled with each increase in pain severity. Opioid prescription was significantly higher only in the moderate and severe categories, however, suggesting a threshold effect. There is literature in the adult emergency medicine practice that correlates increased pain score documentation with improved opioid analgesic administration.26,27 In this study, children with no documented pain scores received opioids in a manner similar to children with mild pain documentation, suggesting that undocumented pain may be associated with underestimated pain.
Limitations of this study include those that are typical for large cross-sectional databases. We are limited by the variables collected and the accuracy with which they were collected. The potential exists for residual confounding from variables that were not included in the multivariate analysis. Our designation of painful diagnosis and painful reason for visit cannot be considered inclusive of all visits that are painful to the ED, only representative. This categorization has little bearing on the key results of the study. Patients may have had pain assessed but not documented by health care providers. Although all other variables had excellent reporting, visit immediacy was missing in 25% of visits, so its inclusion into the model reduced the number of visits analyzed. However, the regression was performed with and without this variable, and little effect was found in our outcomes. These data were collected before the JCAHO mandate for the assessment and documentation of pain for all health care encounters. This analysis may not represent current practice, but it provides historical standards and a benchmark for comparison. Many of the medications that are prescribed for pain relief also can be used for antipyresis; however, the unique designation in the data set of a medication category for “relief of pain” should limit misclassification.
The JCAHO mandate for routine pain score documentation was a step toward improving documentation. This analysis provides critical information about deficiencies in documentation for the youngest patients and patients at pediatric facilities. Recognition of pain expression in pediatric patients and observational scoring techniques should be emphasized in interventions for health care providers. Integration of these pain assessments into the clinician's approach to all patients should improve pain control. Our study supports others who have shown that patients with higher pain scores are more likely to receive analgesia; however, demonstrating the association between pain score documentation itself and increased analgesic and opioid prescription provides new evidence of the clinical importance of pain scores. Visits with no pain scores documented receive similar treatment to those with mild pain, perhaps underestimating the pain experience for the child. Therefore, lack of pain score documentation may account for the low proportion children's receiving analgesics in the ED.
ED pain score documentation is suboptimal in the pediatric population. Many obviously painful complaints and diagnoses have pain documentation rates of <50%. Infants and toddlers are at particular risk for not having a pain score documented. There is a significant association between pain score documentation and the use of any analgesic, particularly opioids. Improvements in pain documentation for acutely ill and injured children are needed to improve pain management and outcomes.
We thank Steven J. Weisman, MD, for critical review of the manuscript.
- Accepted October 18, 2005.
- Address correspondence to Amy L. Drendel, DO, MS, Pediatric Emergency Medicine, 9000 W Wisconsin Ave, MS 677, Milwaukee, WI 53226. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- Petrack EM, Christopher NC, Kriwinski J. Pain management in the emergency department: patterns of analgesic utilization. Pediatrics.1997;99 :711– 714
- Schecter NL, Allen DA, Hanson K. Status of pediatric pain control: a comparison of hospital analgesic usage in children and adults. Pediatrics.1986;77 :11– 15
- ↵Joint Commission on Accreditation of Healthcare Organizations. The Measurement Mandate: On the Road to Performance Improvement in Health Care. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations; 1993
- Joint Commission on Accreditation of Healthcare Organizations. Examples of Compliance: Pain Assessment and Management. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations; 2002
- ↵Committee on Psychosocial Aspects of Child and Family Health. The assessment and management of acute pain in infants, children, and adolescents. Pediatrics.2001;108 :793– 797
- ↵Ly N, McCaig LF. National Hospital Ambulatory Medical Care Survey: emergency department summary. Adv Data.2002;(327) :1– 27
- ↵Centers for Disease Control, National Center for Health Statistics, National Hospital Ambulatory Medical Care Survey. Available at: www.cdc.gov/nchs/about/major/ahcd/nhamcsds.htm. Accessed October 1, 2005
- ↵US Department of Health and Human Services. Acute Pain Management: Operative or Medical Procedures and Trauma. Clinical Practice Guideline No 1. Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, US Department of Health and Human Services; 1992. AHCPR publication 92-0032
- ↵Schneider D, Appleton L, McLemore T. A reason for visit classification for ambulatory care. Vital Health Stat 2.1979;(78) :i– iv, 1–63
- ↵Koch H, Campbell W. The collection and processing of drug information: National Ambulatory Medical Care Survey—United States, 1980. Vital Health Stat 2.1982;(90) :1– 90
- Copyright © 2006 by the American Academy of Pediatrics