Children and youth receive a substantial amount of care for acute illnesses and injuries in emergency departments (EDs). In 2010, children younger than 15 years accounted for 19.6% of all ED visits, and approximately 5% of these visits by children resulted in hospitalization. Of all ED patients with an injury diagnosis, children younger than 15 years accounted for 21.4% of all patient visits, with most injured children between 5 and 14 years of age.1
The Emergency Medical Services for Children (EMSC) program promotes a more uniform approach to improve statewide pediatric emergency care. State funding from EMSC supports initiatives to improve the quality of pediatric emergency care. One specific performance measure for EMSC state partnership grants is:
- The percentage of hospitals recognized as part of a statewide, territorial, or regional standardized system that are able to stabilize and/or manage pediatric traumatic emergencies.2
This performance measure calls for a pediatric emergency department designation program that is expected to improve the quality of care for injured children. A similar performance measure addresses the care of children with medical emergencies. These performance measures were developed in response to findings that EDs were not well equipped or staffed to manage pediatric emergencies.3,4
California, Illinois, Tennessee, and Oklahoma were the first states to implement a program with a verification process to ensure that health care facilities and care providers meet training and resource requirements for the level of designation sought.5,6 The higher designation level is usually awarded to a hospital with pediatric critical care capability. The lowest designation level requires the hospital ED to have the resources for basic assessment and stabilization and to transfer children with serious health conditions needing more advanced care.
Many hospitals seek trauma center verification using criteria developed by the American College of Surgeons Committee on Trauma (ACS-COT). These criteria describe the specific personnel, resources, and quality improvement requirements for each level of verification, including requirements of pediatric trauma care.7 States also designate trauma centers, and designation is often based on the ACS-COT criteria or trauma center verification by the ACS-COT. Most states have established trauma systems to improve the trauma care access for all residents. Studies have documented that organized systems of trauma care for critically injured pediatric patients improve outcomes, and some sources indicate that pediatric trauma centers produce better outcomes for injured children.8–10
In recent years, many states have initiated a pediatric ED designation process to meet EMSC program performance measure requirements, but the degree to which such a designation process improves care of injured children is unknown. In 2009, the State of Delaware began developing criteria for the Pediatric Emergency Care Facility Recognition (PECFR) program and planned statewide voluntary implementation. The PECFR program standards identified 4 pediatric capability levels that addressed personnel; equipment, medication, and supplies; physician and nurse coordinators for pediatric emergency care; continuing education; and participation in a pediatric ED quality improvement program supported by the state. The levels of ED preparedness for pediatric care include the following:
- Level 1. Capable of providing comprehensive specialized pediatric medical and surgical care to all types of ill children; a regional referral center for specialized pediatric care.
- Level 2. Capable of identifying and stabilizing children who are critically ill or injured and providing ongoing inpatient care or appropriate timely transfer to a Level 1 facility.
- Level 3. Same as Level 2, but principally capable of managing minor pediatric inpatient illnesses.
- Level 4. Same as Level 3, but no pediatric inpatient admission capability.
All 8 of Delaware’s acute care hospitals have an ED and submitted an application to the PECFR program. As of December 31, 2011, all 8 facilities achieved recognition after a verification site visit, resulting in 1 Level 1, 1 Level 2, and 6 Level 3 PECFR hospitals. Delaware also has established a trauma center designation process based on ACS-COT criteria.
Implementation of a pediatric ED designation process is expected to improve the care provided to ill and injured children. The effort and costs associated with the implementation of this process are significant for the state and the hospitals that choose to participate. Determining whether an ED designation process improves patient care is needed.
Goal of this Investigation
The intent of this study was to determine if the hospital requirements to achieve PECFR result in an indirect improvement in the daily processes of care provided to injured children. We hypothesize that implementation of the PECFR program in all Delaware hospitals would improve the processes of emergency care provided to injured children, when compared with injured children treated in control hospitals in another state during a similar time frame (no notable differences between states would be found during the preimplementation phase, but significant care efficiencies would be noted in Delaware PECFR hospitals compared with control hospitals during the postimplementation phase).
Study Design and Setting
The research was conducted using a controlled pre-post study design. All 8 Delaware hospitals participated in planning and selecting criteria for the PECFR program, which was finalized in 2010. To reduce the potential impact of care process changes that could be associated with the PECFR program, 2009 was chosen for the preimplementation phase of data collection. Data were also collected in January 2013, 1 full year after the complete implementation of Delaware’s PECFR program. Using a probability proportional to size sampling process, we completed structured medical record abstractions within participating hospitals in both states in 2009 and 2013. Medical record abstractions focused on elements characterizing the routine processes of care for injured children in the ED, such as vital sign assessment, pain assessment and management, and diagnostic use of radiation.
Delaware served as the intervention site. All 8 acute care hospitals participated, including a children’s hospital (Level II trauma center), an adult Level I trauma center, and 6 community hospitals (4 are Level III trauma centers). The Winston-Salem region in North Carolina served as the control site. Thirteen of the 22 North Carolina acute care hospitals in that region participated, including one children’s hospital (Level I trauma center), an adult Level II trauma center, a Level III trauma center, and 10 community hospitals without trauma center designation. Both Delaware and North Carolina have mature statewide trauma systems.
Emergency department nurse managers in the 21 participating hospitals from both states were key partners who obtained hospital approval to participate, facilitated institutional review board approval, and identified data collection abstractors. Institutional review board approval was obtained from all hospitals, uniformly waiving the need for informed consent.
Selection of Participants
Cases eligible for study inclusion were children 14 years and younger with an injury classified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 800 through 959.9 presenting to a participating ED during a specified time/date window. Children transferred to another facility were included in the referring hospital cases but excluded from eligible cases in the receiving hospital. On the specified date, serial eligible cases were included in the sample until the hospital’s designated number of cases was obtained. The number of cases per hospital was predetermined based on a probability proportional to size approach, using hospital pediatric ED census as the sampling unit.11 Hospitals with the smallest ED pediatric volume contributed a minimum of 10 cases, whereas the hospitals with the largest volume (both children’s hospitals) contributed more than 160 cases during each data collection period.
Methods and Measurements
REDCap, an open access Web-based data collection software platform, was used, allowing each abstractor to directly enter data into the study database from their hospital setting.12 The database had fields focused on demographic information, assessment, interventions, timing of care provided, and ICD-9 coding. A data dictionary was developed to guide record abstraction.
Abstractors recruited from each hospital were trained to use the Web-based data entry system, including practice entering data for 3 redacted cases. An interrater reliability assessment was conducted among abstractors shortly after training to ensure consistency and precision in the abstraction of records. Additional instruction was offered to remedy any inconsistencies. When the study data collection was launched, abstractors were given the start date and time for identifying cases for inclusion from their hospital’s medical records.
To estimate data abstraction accuracy, approximately 10% of all cases abstracted during the first data collection period were reabstracted by 2 study investigators (J.W.B., N.D.S.). The degree of concordance between abstractions by the investigators and the abstractors was assessed using an intraclass correlation coefficient. The overall intraclass correlation coefficient for all abstractors across all records was 0.985 (95% confidence interval [CI], 0.982–0.987).
Before the study initiation, investigators identified several routine pediatric care processes for analysis, including documentation of weight and vital signs, pain assessment, pain management, oxygen administration, sedation, use of x-ray and computed tomographic (CT) scans, and procedures performed. For each process of care measure, analyses focused on identifying differences between the 2 state cohorts during each time frame. Subsets of cases experiencing a specific procedure, injury, or pain scale score were also selected for analysis a priori. It was later determined that some care processes, such as sedation use and oxygen administration, had inadequate cases for the planned analyses.
Categorical elements were assessed using a Pearson χ2. Change over time among interval-level elements was assessed using a Mann-Whitney U test. To adjust for multiple comparisons (ie, probability of Familywise errors), a partial Bonferroni correction was applied, based on an average Spearman correlation among procedure and assessment end points of 0.11.13 Statistical significance was accepted at P < 0.012. Nevertheless, 95% CIs associated with percentages are provided to illustrate the range of values plausibly compatible with the data. All statistical analyses were performed using SPSS (version 20.0; SPSS, Inc, Chicago, Ill).
Characteristics of Study Subjects
The study sample includes a total of 1737 children, 709 from Delaware hospitals and 1028 from North Carolina hospitals. More than 97% of cases presented with a mild injury (Injury Severity Score ≤ 9). An analysis of demographic factors representing the cases presenting to intervention and control hospitals during the data collection periods suggests that no important differences exist between the groups by sex, age, mechanism of injury, or discharge disposition (Table 1).
Weight and Vital Sign Documentation
The documentation of weight among study children improved in both states from 2009 to 2013, but this improvement did not reach statistical significance (Table 2). Initial vital signs were documented for most patients (heart rate, respiratory rate, and temperature); however, blood pressure (BP) was documented less often. Nevertheless, both states demonstrated improvement in BP documentation when comparing patients presenting in 2009 with those in 2013. North Carolina hospitals demonstrated a greater improvement in BP documentation for children younger than 3 years, and Delaware hospitals demonstrated greater improvement among children 3 years and older. Secondary analyses suggest that Level I and II trauma centers were predominantly responsible for the increased BP documentation.
Pain Assessment and Management
Hospitals in both states demonstrated improvement in the documentation of at least 1 pain assessment when comparing the 2 time frames. In addition, North Carolina hospitals exhibited a significant decrease in the median elapsed time interval from ED admission to initial pain assessment comparing 2009 with 2013, whereas Delaware hospitals demonstrated an increase in that interval that did not reach statistical significance. No differences were found between the states regarding the number of children treated with pain medication.
To further review issues of pain assessment and management, the subset of children who underwent extremity immobilization for a potential fracture was selected for secondary analysis. We initially examined the elapsed time interval between ED presentation and first pain assessment for this subset. This interval proved significantly reduced among North Carolina hospitals when comparing 2009 and 2013 admissions, whereas shorter elapsed times to assessment were less pronounced in Delaware hospitals. Elapsed time differences were also examined for the interval between the pain assessment and pain medication administration for this subset of children. Interestingly, neither hospital cohort produced shorter intervals between pain assessment and pain management for the children with extremity immobilization for all pain levels. However, Delaware children with a documented pain score of 5 or greater experienced significantly shorter intervals between assessment and pain medication after PECFR implementation. The interval between pain assessment and pain medication trended slightly higher for the North Carolina hospitals, regardless of pain score.
The percentage of children exposed to x-ray and/or CT imaging radiation was significantly diminished for children treated in Delaware hospitals after PECFR implementation. No such relationship was noted for control hospitals in North Carolina. In an effort to determine if radiation exposure was more appropriately applied, a secondary analysis confirmed that plain film x-rays administered among children with immobilized extremities increased in both states during the later period of data collection but only to a significant degree in Delaware. A secondary analysis also suggested that the prevalence of CT imaging among children presenting to Delaware hospitals also dropped comparing 2009 and 2013 admissions but not significantly. A similar, but less pronounced, downward trend in CT imaging was noted among North Carolina hospitals. The majority (89%) of CT imaging performed in both states was of the head.
This study was designed to determine if implementation of a statewide ED designation process focused on pediatric trauma would result in improved care to children presenting with common injuries. The study incorporated a control group to aid in distinguishing effects potentially due to hospital designation from effects associated with temporal changes in pediatric care. A comparison of cohort demographics suggested that intervention and control patients did not significantly differ; however, baseline values for process and outcome elements among the 2 groups did differ. Thus, the analyses focused on comparing the delta change within each state group before and after the implementation of the hospital designation process in the intervention state.
This selected study sample presented with 97% mild injuries, which is comparable with the injury severity of children treated in EDs nationally, in which only 3.4% of injured children 14 years and younger are admitted to the hospitals.1 The EMSC performance measure addressed by this study is intended to benefit children with any injury severity level, and children with mild injuries are not often addressed in research. Thus, the efforts to determine if pediatric ED facility recognition improves care for this group of children are important.
Overall, the assessment of process and outcome elements thought to potentially improve with ED recognition, a priori, showed little improvement over that which was attributed to general temporal trends in pediatric care. Weight is an essential measurement for children seeking emergency care so that medication dosages and fluid volumes can be calculated. Both states demonstrated positive trends in weight documentation. Both Delaware and North Carolina hospitals exhibited significant increases in BP documentation over time; thus, findings did not support an added effect due to the development of the PECFR program in Delaware. That Level I and II trauma centers in North Carolina contributed the most to increased BP documentation highlights an opportunity for improvement by the region’s community hospitals.
The hospital cohorts demonstrated no significant differences in the interval between pain assessment and medication administration. The significantly increased rate of pain assessment in both hospital cohorts may be associated with the increased focus on pain assessment by the Joint Commission, encouraging hospitals to address pain assessment in all patients.14 Another potential explanation is the increased use of electronic patient records by hospitals in 2013 that flagged missing assessment parameters. The subgroup of children with extremity immobilization for potential fracture was selected for analysis, as it was the only subgroup with enough cases in which physician and nurses would expect the child to have consistent pain and to be more proactive in providing pain medication.
Unnecessary radiation exposure of children, either by x-ray or CT imaging, is receiving increased attention. Research published beginning in 2009 provides specific guidelines for CT imaging among children at low risk of clinically significant brain injury or abdominal injury after trauma.15,16 In our project, Delaware hospitals demonstrated a reduction in radiation exposure, when compared with control hospitals after the implementation of the PECFR program. Secondary analyses support the idea that this reduction in radiation use may be more associated with the prevalence of plain film use versus CT use. Interestingly, the Delaware PECFR program provided no specific training regarding radiation exposure. Thus, additional conservatism in the use of radiation in Delaware hospitals may be more closely associated with factors not measured in this study or is an artifact associated with the presence of the PECFR program in the Delaware hospitals.
Although the focus on children with minor injuries is important, it is possible that additional or different findings associated with the PECFR program could have resulted if the sample had included a larger proportion of children with severe injuries. Additional studies are needed to assess the effect of a PECFR program on the care of children with an acute illness. It would also be important to determine if identified changes in this study are sustained over time.
It is also important to note that this study was not a randomized controlled trial. Thus, our choice of North Carolina control hospitals may not represent the experience of other hospitals across the nation that are not participating in an ED designation process for acute pediatric trauma or illness. Nevertheless, the addition of a control patient sample lessens the likelihood of bias that can be associated with this type a preintervention versus postintervention study.
In conclusion, findings of improvements in care for children associated with the PECFR program was limited to the interval between a pain assessment of 5 or greater and pain medication for children with extremity immobilization. However, postimplementation data were collected only 1 year after the PECFR program implementation, and care processes may improve as the PECFR program matures. Delaware initiated a quality improvement program for PECFR hospitals beginning in 2012 and selected 2 indicators for quality improvement monitoring, neither associated with significant study findings. These pediatric quality improvement meetings involve representatives from all hospitals, and meeting discussions could have potentially focused attention on other processes of care. For example, the practice guideline for the use of CT for brain injury was disseminated to all Delaware hospitals during a quality improvement meeting. This study will help inform Delaware and North Carolina hospitals about opportunities for improvement and indicators to consider in future quality improvement activities.
3. Middleton KR, Burt CW. Availability of pediatric services and equipment in emergency departments: United States, 2002-03. Adv Data
. 2006; 367: 1–16.
4. Gausche-Hill M, Schmitz C, Lewis RJ. Pediatric preparedness of US emergency departments: a 2003 survey. Pediatrics
. 2007; 120: 1229–1237.
5. Cichon ME, Fuchs S, Lyons E, et al. A statewide model program to improve emergency department readiness for pediatric care. Ann Emerg Med
. 2009; 54: 198–204.
6. Hohenhaus SM, Lyons E, Phillippi RG. Emergency departments and pediatric categorization, approval and recognition: a review of two states. J Emerg Nurs
. 2008; 34: 236–237.
7. Resources for Optimal Care of the Injured Patient 2006
, American College of Surgeons Chicago, IL: American College of Surgeons Committee on Trauma, 2006.
8. Hulka F, Mullins RJ, Mann NC, et al. Influence of a statewide trauma system on pediatric hospitalization and outcome. J Trauma
. 1997; 42: 514–519.
9. Esposito TJ, Sanddal ND, Dean JM, et al. Analysis of preventable pediatric trauma deaths and inappropriate trauma care in Montana. J Trauma
. 1999; 47: 243–251; discussion 251–253.
10. Petrosyan M, Guner YS, Emami CN, et al. Disparities in the delivery of pediatric trauma care. J Trauma
. 2009; 67: S114–S119.
11. Yates F, Grundy PM. Selection without replacement from within strata with probability proportional to size. J R Stat Soc Series B Stat Methodol
. 1953; 15: 253–261.
13. Gong J, Pinheiro JC, DeMets DL. Estimating significance level and power comparisons for testing multiple endpoints in clinical trials. Clin Trials
2000: 21: 313–329.
15. Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet
. 2009; 374: 1160–1170.
16. Gorelick MH, Atabaki SM, Hoyle J, et al.; for the Pediatric Emergency Care Applied Research Network (PECARN). Interobserver agreement in assessment of clinical variables in children with blunt head trauma. Acad Emerg Med
. 2008; 15: 795–883.