Pediatr Surg Int (2013) 29:263–268 DOI 10.1007/s00383-012-3232-1
ORIGINAL ARTICLE
Impact of trauma system development on pediatric injury care David P. Mooney • Ivan M. Gutierrez • Qiaoli Chen • Peter W. Forbes • David Zurakowski
Accepted: 4 December 2012 / Published online: 18 December 2012 Ó Springer-Verlag Berlin Heidelberg 2012
Abstract Purpose Trauma systems improve survival by directing severely injured patients to trauma centers. This study analyzes the impact of trauma systems on pediatric triage and injury mortality rates. Methods Population-based data were collected on injured children less than 15 years who were admitted to any hospital in New England from 1996 to 2006. Data from three trauma system states were compared to three nontrauma system states. The percentages of injured children, severely injured children, and brain-injured children admitted to trauma centers were determined as well as injury hospitalization and death rates. Time trend analysis examined the pace of change between the groups. Results A total of 58,583 injured children were hospitalized during the study period. Injury hospitalization rates were initially similar between the two groups (with and without trauma systems) and decreased over time in both. Rates decreased more rapidly in trauma system states compared to those without, (P = 0.003). Injury death rates decreased over time in both groups with no difference between the groups, (P = 0.20). A higher percentage of injured children were admitted to trauma centers in nontrauma system states throughout the study period, and this percentage increased in both groups of states. A higher percentage of severely injured children and brain-injured D. P. Mooney (&) I. M. Gutierrez D. Zurakowski Trauma Program, Department of Surgery, Children’s Hospital Boston and Harvard Medical School, 300 Longwood Avenue, Fegan 3, Boston, MA 02115, USA e-mail:
[email protected] Q. Chen P. W. Forbes Clinical Research Program, Children’s Hospital Boston, Boston, MA, USA
children were admitted to trauma centers in non-trauma system states and both percentages increased over time. The increase was more rapid in trauma system states for children with severe injuries (P \ 0.001) and children with brain injuries (P \ 0.001). Discussion Trauma systems decreased childhood injury hospitalization rates and increased the percentage of severely injured children and brain-injured children admitted to trauma centers. Mortality and overall triage rates were unaffected. Keywords Trauma Pediatric Triage Outcome assessment
Introduction State and regional trauma systems have evolved in North America in an attempt to improve the outcome of patients with traumatic injuries. Trauma systems are organized public health efforts across a state or region to deliver care to injured patients in a coordinated way. One of their many functions is to direct patients to the most appropriate facility, diverting severely injured patients away from nontrauma centers to trauma centers, where survival has been shown to be significantly better. Several published studies have demonstrated the efficacy of trauma systems in adult patients, and the triage of pediatric patients within trauma systems is based upon the premise that similar benefits are present for children [1–4]. Prior studies on the effectiveness of trauma systems involve comparisons between states or regions with and without a trauma system or between newly instituted trauma systems and historical controls. In the United States, comparisons between states may be affected by other differences between the type of states that
123
264
develop trauma systems and those that do not, unrelated to the presence of a trauma system [3]. In addition, comparisons using historical controls within states negate the influence of time-borne changes in the patterns of patient flow. To minimize the potential impact of these two factors, we utilized a relatively unique situation: the near simultaneous development of trauma systems in three of the six New England states: Maine, New Hampshire, and Connecticut, and the lack of a trauma system in the remaining three New England states: Vermont, Massachusetts, and Rhode Island.
Pediatr Surg Int (2013) 29:263–268
data from the three non-trauma states Massachusetts (MA), Vermont (VT), and Rhode Island (RI). Time series analysis was performed with Prais–Winsten regression estimation and panel-corrected standard errors to assess rates of change over time [5, 6]. The models applied involve a firstorder autocorrelation (AR1) over time (analogous to Yule– Walker estimates) [7]. Comparisons of slopes between the trauma system and no trauma system groups were compared with a group-by-time interaction term in the model, as assessed by the Z test. Two-tailed P \ 0.05 was considered statistically significant. Data analysis was performed using the Stata 12 software package (Stata Corp, College Station, TX).
Methods Administrative hospital data from the New England Pediatric Database were obtained for every inpatient hospitalization in New England from children aged 15 years or less. Data for children with an International Classification of Diseases 9th Edition (ICD-9) diagnosis code from 800 to 959.9 from calendar years 1996 to 2006 were included in our study. The year 1996 was chosen because the three trauma system states had implemented their systems within 3 years of that date: Maine-1993, New Hampshire, and Connecticut-1995. Patients with ICD-9 codes indicative of late effects of injury, iatrogenic injuries, and poisoning were excluded. Injury severity scores (ISS’s) were applied using ICDMAP-90 software (Baltimore, MD, Johns Hopkins University), and those children with an ISS [13 were considered to have suffered a severe injury. Children with an ICD-9 diagnosis code of 800–801.9, 803–804.9, and 850–854.1 were considered to have suffered a traumatic brain injury. Childhood injury death data for each of the 6 New England states were obtained from the Web-based Injury Statistics Query and Reporting System (WISQARS) data system of the US Centers for Disease Control (www. cdc.gov/injury/wisqars/fatal.html). Population estimates per state were obtained (United States Census Region 1; www.census.gov/popest/estimates.html) and were used to calculate population-based statistics. Trauma center status was determined by verification of the American College of Surgeons (ACS) and/or state designation as a level 1 or level 2 adult or pediatric trauma center (www.facs.org/trauma/verified.html), whether or not the state had a trauma system. Those hospitals that obtained and/ or lost ACS verification as a trauma center during the study interval were considered to be trauma centers across the whole time period for the purposes of this analysis. Statistical analysis Data from the three trauma states Connecticut (CT), Maine (ME), and New Hampshire (NH) were averaged as were
123
Results Across the 11-year time period, 58,583 injured children were hospitalized in New England. At the start of the study period, fewer injured children were admitted to hospitals in trauma system states than non-trauma system states. Over the study period, the number of injured children admitted to a hospital decreased for both groups of states, albeit faster in the trauma system states (P \ 0.001, Fig. 1). When population-based, the childhood injury hospitalization rates were similar in the two groups of states at the beginning of the study period and decreased over time in both groups. This rate decreased significantly faster in the states with trauma systems than in states without trauma systems (P = 0.003, Fig. 2). Population-based pediatric injury death rates also decreased over time in both groups of states, and no difference was observed between the trauma system and nontrauma system states (P = 0.20, Fig. 3). Across the study period, there was a significant increase in the percentage of injured children who were admitted to a trauma center versus a non-trauma center in both nontrauma system states (P \ 0.001) and trauma system states (P \ 0.001, Fig. 4). The rate of the increase was not different between the two groups of states (P = 0.38). The percentage of severely injured children who were admitted to a trauma center increased over time in both the non-trauma system states (P \ 0.001) and trauma system states (P \ 0.001, Fig. 5). The trend increased significantly faster in trauma system states than non-trauma system states (P \ 0.001). The percentage of children diagnosed with traumatic brain injuries who were admitted to a trauma center increased in non-trauma system states (P \ 0.001) and trauma system states (P \ 0.001) (Fig. 6). As with severely injured children, the rate of increase was significantly faster in the trauma system states than the non-trauma system states (P \ 0.001, Fig. 6).
Pediatr Surg Int (2013) 29:263–268
265
Fig. 1 The mean annual number of hospital inpatient injury discharges among children less than 15 years of age in New England for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). The number of children hospitalized per year for an injury decreased across the region, and this decrease was faster in states with a trauma system (Z = -4.45, P \ 0.001)
Fig. 2 The annual mean population-based rate of hospital inpatient injury discharges among children less than 15 years of age in New England for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). When standardized for population, the significant effect of trauma systems on injury admission rate persisted (Z = -2.97, P = 0.003)
Discussion Trauma systems are associated with decreased injury rates and improved mortality in adults [1–4, 8]. Through a variety of means such as pre-hospital triage guidelines, limits on admission to non-trauma centers, and general recognition of trauma center status, trauma systems influence referral patterns. However, few data exist regarding the effect of trauma systems on the outcome of childhood injury [2]. This study analyzes the impact of trauma systems on childhood injury incidence and trauma center admission rates across an 11-year time period. Four states had only one trauma center: two trauma system states (ME and NH) and two non-trauma system states (VT and RI),
and two states had many trauma centers: one trauma system state (CT) and one non-trauma system state (MA). This situation was static during the study period. Our study is unique because it utilizes comparisons of similar states contained within a single geographic region avoiding two major potential sources of confounding that have plagued other similar studies. This study includes information on all injury admissions at every acute care facility in New England and is uniquely well suited to determine changes in childhood injury hospitalization rates in the region. Equally important, the database does not include emergency department visits that result in death in the ED, transfer to another acute care facility, or discharge to home, avoiding double counting. In
123
266
Pediatr Surg Int (2013) 29:263–268
Fig. 3 The annual mean population-based injury death rate among children less than 15 years of age in New England for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). While there was a decrease in the childhood injury death rate across the region, there was no effect found from trauma system development (Z = -1.27, P = 0.20)
Fig. 4 The annual mean percentage of injured children less than 15 years of age in New England children who were discharged from a trauma center for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). While the triage of injured children to trauma centers increased across the region, there was no effect found from trauma system development (Z = -0.87, P = 0.38)
order to capture every childhood injury death and not just those that occur in a hospital, death rates were determined using inclusive statewide childhood injury fatality data from the WISQARS program and not institution-specific information, such as a trauma registry. Each of the parameters analyzed improved significantly across the study time period in states with and without trauma systems. The use of a historical control group alone to conduct this analysis may have falsely attributed these improvements to the presence of trauma systems. These region-wide improvements imply that some factor or factors other than the presence of a trauma system were significantly contributing. This could be related to an ongoing regionalization of all pediatric care, changes in the health
123
care reimbursement system, or other unknown factors, such as improved pre-hospital care, global positions systems, injury prevention, and speeding laws, which appear to be as significant as the effect of trauma system development alone. Trauma system development in New England resulted in an increase in the proportion of severely injured children and brain-injured children triaged to trauma centers, where it is believed that the outcome is better. This significant improvement in triage, which cannot be attributed to market forces, may be enough to justify the dissemination of trauma systems to other states. This occurred despite other strong influences affecting the referral patterns of pediatric patients in these states.
Pediatr Surg Int (2013) 29:263–268
267
Fig. 5 The annual mean percentage of severely injured children (ISS [15) less than 15 years of age in New England who were discharged from a trauma center for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). A greater increase in the triage of severely injured children to trauma centers was noted in the trauma system states (Z = 5.37, P \ 0.001)
Fig. 6 The annual mean percentage of brain-injured children less than 15 years of age in New England children who were discharged from a trauma center for the three states with statewide trauma systems (CT, ME, and NH) compared to the three states without statewide trauma systems (MA, RI, and VT). A greater increase in the triage of brain-injured children to trauma centers was noted in the trauma system states (Z = 4.82, P \ 0.001)
One point of resistance to the development of a trauma system is the potential negative impact on non-trauma hospitals by prioritizing triage of all injured patients to designated trauma centers. Our data reinforce the notion that trauma system development is not, in itself, detrimental to non-trauma centers as the triage of children with minor injuries was not influenced by the presence of a trauma system. Market forces and other factors appear to play a larger role in the increasing diversion of children with minor injuries to trauma centers than does the development of a trauma system. On the other hand, the triage of children with severe injuries and brain injuries to trauma centers was impacted by the presence of a trauma system.
For unknown reasons, many of the trauma measures analyzed were better in the states without a trauma system at the initial time period analyzed, creating a gap between the two groups of states. Through the study time period, the initial differences in death rate and trauma center admission rate between the two groups of states were not affected. However, the initial gap in the trauma center admission rates between the two groups of states for children with severe injuries and brain injuries narrowed significantly over the study period. In addition, the decline in injury hospitalization rate noted across the study time period was accelerated in the trauma system states compared to the non-trauma system states. No differences in mortality rates were noted between the two groups of
123
268
states. Low baseline childhood injury mortality rates and the occurrence of pre-hospital injury death in over half of cases [9] may make it more difficult to disclose an effect. Mann et al. [10, 11] in 1998, outlined in detail the necessary components of a trauma system. According to Mann’s criteria, each of the trauma systems included in this analysis are incomplete in major ways, which also could have affected the results. For example, none of the trauma systems included a significant injury prevention component, each had insufficient funding, and none placed a limit on the number of trauma centers. Despite these incomplete efforts, the trauma systems implemented by these state governments were effective in diminishing the childhood injury admission rate and improving the triage of severely injured children and brain-injured children to trauma centers. One potential limitation of our study involves a nontrauma system state, Massachusetts, which made at least two attempts to develop a statewide trauma system: in 1994 and 2000, but did not designate its first trauma center until 2006. There may have been a misperception that a system was already in place in Massachusetts, therefore influencing triage behavior. However, if this were true, the influence of trauma system development noted above would have been blunted, not falsely augmented. Also, fully onehalf of the ACS-verified trauma centers were in non-trauma system states and acted as de facto identified trauma centers, regardless of the lack of a trauma system. Again, however, this effect would have blunted any identified effect of trauma system development. Another limitation is the geography of New England, in which the states are in close proximity with overlap in hospital tertiary service areas. It is possible that the development of a system in one state could have influenced the triage pattern in neighboring states. These concerns, however, would again have blunted any of the noted beneficial effects of trauma system development. In New England across the study period, the childhood injury mortality rate decreased and the percentage of injured children admitted to trauma centers increased. However, these trends appeared to be unaffected by the
123
Pediatr Surg Int (2013) 29:263–268
presence or absence of trauma systems. Trauma systems were associated with an accelerated reduction in childhood injury rates and an accelerated increase in the percentage of severely injured children and brain-injured children who were admitted to trauma centers. Further studies in other geographic regions are warranted to evaluate the effectiveness of trauma systems in pediatric trauma.
References 1. Celso B, Tepas J, Langland-Orban B, Pracht E, Papa L, Lottenberg L, Flint L (2006) A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma 60(2): 371–378; discussion 378. doi:10.1097/01.ta.0000197916.99629.eb 2. Nathens AB, Jurkovich GJ, Rivara FP, Maier RV (2000) Effectiveness of state trauma systems in reducing injury-related mortality: a national evaluation. J Trauma 48(1):25–30 discussion 30–21 3. Lansink KW, Leenen LP (2007) Do designated trauma systems improve outcome? Curr Opin Crit Care 13(6):686–690. doi: 10.1097/MCC.0b013e3282f1e7a4 4. Hulka F, Mullins RJ, Mann NC, Hedges JR, Rowland D, Worrall WH, Sandoval RD, Zechnich A, Trunkey DD (1997) Influence of a statewide trauma system on pediatric hospitalization and outcome. J Trauma 42(3):514–519 5. Kmenta J (1997) Elements of econometrics, 2nd edn. University of Michigan Press, Ann Arbor 6. Judge GGGW, Hill RC, Lutkepohl H, Lee TC (1985) The theory and practice of econometrics, 2nd edn. Wiley, New York 7. Brockwell PJDR (2002) Introduction to time series and forecasting, 2nd edn. Springer, New York 8. Mullins RJ, Mann NC (1999) Population-based research assessing the effectiveness of trauma systems. J Trauma 47(3 Suppl):S59–S66 9. Vane DW, Shackford SR (1995) Epidemiology of rural traumatic death in children: a population-based study. J Trauma 38(6): 867–870 10. MacKenzie EJ, Rivara FP, Jurkovich GJ, Nathens AB, Frey KP, Egleston BL, Salkever DS, Scharfstein DO (2006) A national evaluation of the effect of trauma-center care on mortality. N Engl J med 354(4):366–378. doi:10.1056/NEJMsa052049 11. Mann NC, Mullins RJ, MacKenzie EJ, Jurkovich GJ, Mock CN (1999) Systematic review of published evidence regarding trauma system effectiveness. J Trauma 47(3 Suppl):S25–S33
