- Email: [email protected]
Pediatric trauma center criteria: An outcomes analysis
Pediatric Trauma Center Criteria: An Outcomes Analysis By E d w a r d J. Doolin, A n n e M. B r o w n e , and Carla DiScala
Camden, N e w Jersey and Boston, Massachusetts
Background~Purpose:Trauma centers (TC) are certified based on widely accepted criteria. These specific criteria rarely are scrutinized individually. The purpose of this study was to analyze the individual components of a pediatric trauma center for their effect on outcome. Methods: Members of the National Pediatric Trauma Registry were queried about the following: (1) separate pediatric emergency department (ED), (2) pediatric intensive care unit (PICU), (3) pediatric intensivist as PICU director, (4) pediatric surgeon as TC director, (5) in-house attending surgeon, (6) in-house pediatric emergency physician, (7) 24-hour operating room, (8) 24-hour computed tomography (CT) scan. Outcomes analyzed included mortality, length of stay, time in ED, days in PICU, and disability. Victims were stratified based on age (<7 or ->7 years) and severity of injury (ISS -< 16, 17-35, >-36). Results were compared using Student's t test and X2 analysis. Results: A total of 59 of 74 centers responded, 18 were dropped because of low enrollment (mean, 1.6 patients).
HERE HAS BEEN a great deal of development and progress in the institution of trauma systems over the last three decades. Since the development of the first statewide trauma programs, experience and data have shown that there is improved survival rate, lowered disability, and cost effective care when a trauma program is in place, 1-4 There has also been a constant discovery as to what is required for a trauma program. There are many components that come together to create the trauma system. The prehospital program, the trauma center facility, personnel, and rehabilitation programs all constitute the system. In addition, education and research figure in a major way. Many documents recommend the required resources for a trauma center. 5-8 There is no situation in which the controversy is as great as in the area of pediatric trauma centers. Not only is there a question among pediatric trauma systems as to what the requirements should be, but these programs also are being compared with their adult counterparts for content and o u t c o m e ) There is no universal definition or credentialing body for trauma centers. This makes comparisons difficult. 1°-12 In addition, there has been very little in the way of objectively scrutinizing the individual requirements of a trauma center for their value/3-15 A recent survey of pediatric surgeons' activity suggests that the majority are involved
T
Journal of Pediatric Surgery,Vo134, No 5 (May), 1999: pp 885-890
Questions 3, 4, 6, and 7 were eliminated because of skewed data. An in-house surgeon reduced the amount of time a mildly injured patient (ISS -< 16) spent in the ED (210 v 434 minutes), as did the separate pediatric ED (333 v592 minutes) and pediatric emergency physicians (344 v 507 minutes) in younger patients (>-7 years). An in-house surgeon reduced the morality rate in older (>-7) severely injured (ISS >- 36) patients (46.7% v 56.8%; P < .05 for all). No other differences were significant. Conclusions: In-house personnel improved efficiency for the less severely injured, and an in-house attending surgeon reduced mortality in the severely injured older patient. None of the other variables were found to have a significant impact on outcome. J Pediatr Surg 34:885-890. Copyright © 1999 by W.B. Saunders Company.
INDEX WORDS: Trauma, standards, outcomes.
in pediatric trauma care. 16 The settings are as varied as their number. In this project, we did not evaluate different types of trauma centers. Rather, we evaluated the different components of a trauma center for the value it gave the system in terms of efficiency and outcome. The goal o f this type of analysis was to identify how the components of the pediatric trauma center contributed to its success. In addition, the hope was to identify which components were or were not significant in their contribution to the value of the center. MATERIALS A N D METHODS
In 1995, there were 74 Pediatric Trauma Centers that belonged to the National Pediatric Trauma Registry (NPTR). This made up the study
From the Division of Pediatric Surgery, Department of Surgery, Robert Wood Johnson Medical School at Camden, Camden, N J, and the Research and Training Center, Tufts University School of Medicine, Boston, MA. Presented at the 1998 Annual Meeting of the Section on Surgery of the Ameriean Academy of Pediatrics, San Francisco, California, October 16-19, 1998. Address reprint requests to Edward J. Doolin, MD, Department of Surgery, Division of Pediatric Surgery, Three Cooper Plaza, Suite 411, Camden, NJ 08103. Copyright © 1999 by W.B. Saunders Company 0022-3468/99/3405-0043503.00/0
885
886
DOOLIN, BROWNE, AND DISCALA
group for this report. Each center was contacted in writing or by phone to complete a simple questionnaire. Eight components of a trauma center were to be evaluated (Table 1). For the purposes of this evaluation, each component was studied in a singular fashion. Any given center could be a "yes" for one analysis and a "no" for another. The database of the National Pediatric Trauma Registry was then used to measure outcomes. The cohort of patients for the NPTR phase II (ending 1996) was the study group. Five different outcomes were measured for this analysis (Table 2). The patients were stratified to allow comparison. Patients were grouped by age (--> or < 7 years) and by severity of injury (1SS 1-16, 17-35, or >35). For each parameter in the questionnaire the trauma centers were grouped according to their answer. The outcomes between these trauma centers were then compared controlling for age and severity of injury. The significance of the results was then verified using a Student's t test and X2 analysis for each outcome. Significance was determined at P < .05.
RESULTS
Of the 74 trauma centers, 59 were able to be included in the study. Reasons that a center was not included were inability to complete information (n = 15) or the submission of less than 25 patients through phase 2 (n = 25). It was felt that a small volume lent itself to a skewed value for these centers because the centers were not weighted for size. (The mean patient volume for this group was 1.6). The responses from the 59 centers included are listed (Table 1). Based on these answers the center could then be categorized for outcome analysis. Outcomes data were then collected and collated based on the answers to these questions. Questions 2, 3, 7, and 8 had largely affirmative responses (>85%), and it was felt this did not allow for a valid analysis, because the group sizes were not comparable. Each outcome was analyzed by comparing the centers based on their answers to questions 1, 4, 5, and 6. The first outcome compared was days in pediatric intensive care unit (PICU) (Table 3). The presence or absence of any of the four studied criteria did not have an effect on days in the PICU. Table 1. Questions and Responses From Each Trauma Center
Question 1. Does your institution have a separate pediatric emergency department. 2. Does your institution have a separate pediatric intensive care unit? 3. Is the medical director of the PICU a pediatric intensivist? 4. Is there a pediatric surgeon who is the director of the pediatric trauma service? 5. Does your institution have an attending surgeon available in house 24 hours a day? 6. Does your institution have an in-house pediatric emergency room attending surgeon 24 hours a day 7. Does your institution have 24-hour in-house operating room availability? 8. Does your institution have 24-hour in-house CT availability?
Response (n = 59) (% yes)
54 100 97 76 34 59 85 88
Table 2. Outcomes Data Measured a. b. c. d. e.
Length of Stay in PICU (d) Mortality rate Initial time in the emergency department (min) Overall length of stay (d) Long-term impairment (>7 too)
The second outcome studied was mortality (Table 4). The presence of an in-house attending surgeon reduced the mortality rate of the severely injured (ISS > 35) older ( > 7 years) patient from 56.8% to 46.7% (P < .05). In addition, the presence of a pediatric emergency department (ED) increased the mortality rate from 44.8% to 61.3% (P < .05). These components did not affect any other group, and no other components had any effect on mortality. The third outcome measured was time in the emergency room (Table 5). The presence of an in-house attending surgeon, a pediatric ED, and a pediatric ED physician all reduced the length of stay for a young (<7 years) mildly injured (ISS < 16) patient. No other differences were found. The overall hospital stay also was examined. No component studied had any effect on the overall length of stay (Table 6). The patient disability was evaluated as the frequency of long-term problems (Table 7). None of the components studied had an effect on the long-term sequelae. DISCUSSION
As the need for cost-effective and efficient care increases, the components of a health care system are going to be scrutinized. In addition, the goal of pediatric trauma systems always will be intervening in the most severe of injuries with a successful outcome. In the current study we chose to compare the components of the product to find if they had a significant effect on health care delivery. Many components were too prevalent to create comparable groups and, therefore, were not studied. The criteria studied did give useful data. It is not surprising that the presence of an in-house attending surgeon decreased the mortality rate in the most severely injured group. This is a small cohort (232 of 41,246 or 0.56%), and the improvement (18 patients or .04%) is even smaller. If not isolated as a separate small group, these differences would never be significant. This group may be useful in that it represents the "extreme injury," which often is considered a "mission" for the trauma center. An in-house surgeon allows for rapid surgical intervention. Although most deaths are from central nervous system injury, the preventable deaths in this group usually are caused by hemorrhage or other "golden hour" mechanisms. To validate this, the difference in diagnoses were compared in victims needing emergency surgery. An analysis of patients operated on in less than 1 hour, was performed
PEDIATRIC T R A U M A CENTER CRITERIA
887
Table 3. Average Days in the PICU <7 yr ISS 1-16 A t t e n d i n g in house, yes A t t e n d i n g in house, no Pediatric ED, yes Pediatric ED, no Pediatric ED physician, yes Pediatric ED, physician, no Director Ped Trauma Ped Surgeon, yes Director Ped Trauma Ped Surgeon, no
<7 yr ISS 17-35
<7 yr ISS 36-75
>7 yr ISS 1-16
>7 yr ISS 17-35
>7 yr ISS 36-75 n = 216
n = 1355
n = 567
n = 47
n = 1974
n = 1155
2 (1-49)
5 (1-76)
16 (1-131)
2 (1-31)
6 (1-42)
10 (1-85)
n = 671
n = 878
n = 88
n = 2024
n = 1187
n = 137
2 (1-112)
5 (1-76)
11 (1-192)
2 (1-33)
6 (1-77)
n = 1503
n = 809
n = 84
n = 1797
n = 1019
2 (1-60)
5 (1-76)
11 (1-192)
3 (1-33)
6 (1-105)
n = 1546
n = 641
n = 51
n = 2228
n = 1338
2 (1-112)
5 (1-72)
15 (1-131)
2 (1-31)
6 (1-142)
10 (1-85)
n = 1947
n = 969
n = 97
n = 2406
n = 1317
n = 193
2 (1-112)
5 (1-76)
12 (1-192)
2 (1-33)
6 (1-105)
n = 1102
n = 481
n = 38
n = 1619
n = 1040
2 (1-60)
5 (1-70)
15 (1-113)
2 (1-27)
6 (1-142)
11 (1-85)
n = 2921
n = 1403
n = 133
n = 3785
n = 2220
n = 337
2 (1-112)
5 (1-76)
12 (1-192)
2 (1-33)
6 (1-142)
n = 128
n = 47
n = 2
n = 240
n = 137
n = 17
2 (1-14)
5 (1-17)
21 (6-36)
2 (1-15)
5 (1-46)
10 (1-34)
9 (1-37) n = 113 8 (1-37) n = 241
8 (1-74) n = 161
9 (1-85)
NOTE. None of the groups studied had any differences in this o u t c o m e that w e r e statistically significant. N u m b e r s in parentheses are ranges.
(Table 8). Those who died predominately had head injuries (35%) with associated other problems. Those who lived predominately had fractures with only 17% head injuries. The abdominal injuries also were fewer in the survival group. This suggests that for the fatal group, the injuries (head) are less treatable than those of the survivors (orthopedic). No clear pattern of injuries identified a need filled by the in-house attending surgeon, suggesting his main role may be largely that of decision making. However, the patients who benefit are few, and one could argue that they are not worth the investment of a n in-house attending surgeon. Paradoxically, the presence of a separate pediatric ED increased mortality rate. This is likely a statistical fallout in that most centers with in-house attending surgeons are probably housed in an
adult trauma center without a separate pediatric ED. Again, this is a small group. Interestingly, the presence of a staff surgeon increased the efficiency of the mild to moderate injured group a great deal as did a pediatric emergency physician. This might be evaluated from a financial point of view. It is possible that the improved efficiency justifies the presence of the attending surgeon financially. However, this reduction in price may not be translated to a reduction in cost because of the capital intensive effort of trauma. Also, there may be other strategies that result in this effect without the need of the in-house surgeon. This study evaluated each component in isolation. The treatment of trauma is multifactorial and so are its outcomes. Some of these components may be interre-
Table 4. Morality Rates <7 yr ISS 1-16
<7 yr ISS 17-35
3/3524 0.08%
Attending in house, no
18/11510 0.16%
23.5%
63.3%
0.08%
14.8%
Pediatric E D, yes
17/11309
257/1082
72/115
10/13934
204/1289
0.15%
23.8%
62.6%
0.07%
15.8%
61.3%
Pediatric ED, no
4/3775
198/730
39/73
6/7082
266/1549
138/308"
0.11%
27.1%
53.4%
0.08%
17.2%
44.8%
16/11276
296/1211
81/136
14/15034
273/1619
147/271
A t t e n d i n g in house, yes
Pediatric ED physicia n, yes Pediatric ED physician, no Director Ped Trauma Ped Surgeon, yes Di rector Ped Trauma Ped Surgeon, no
<7 yr ISS 36-75
>7 yr ISS 1-16
>7 yr ISS 17-35
>7 yr ISS 36-75
182/654
35/68
4/6727
250/1354
126/276"
27.8%
51.8%
0.06%
18.5%
46.7%
272/1153
76/120
12/14201
217/1469
108/190" 56.8% 98/160*
0.14%
24.4%
59.6%
0.09%
16.9%
54.2%
5/3808
159/601
30/52
2/5982
197/1219
89/197
0.013% 21/14622
26.5% 444/1759
57.7% 109/446
0.03% 15/19877
16.2% 444/2669
45.0% 226/446
0.14%
25.0%
59.2%
0.08%
16.6%
50.7%
0/462
11/53
2/4
1/1124
26/169
10/22
0%
20.8%
50.0%
0.09%
15.4%
45.5%
NOTE. In the severely injured patient the presence of an in-house attending significantly reduced the m o r t a l i t y rate. This w a s rare b u t significant. Paradoxically, the m o r t a l i t y rate in the same g r o u p w a s higher w h e n a separate pediatric e m e r g e n c y r o o m existed than w h e n it was not available. * P < .05.
Table 5. 33me Spent in ED (rain) <7 yr ISS 1-16 Attending in house, yes
n = 2468* 210
<7 yr ISS 17-35
<7 yr ISS 36-75
n = 455
n = 47
137
n = 4833
>7 yr ISS 17-35
>7 yr ISS 36-75
n = 986
n = 165
119
219
186
135
n = 95*
n = 12367
n = 1240
n = 154
Attending in house, no
n = 9766*
Pediatric ED, yes
n = 9670*
Pediatric ED, no
n = 2608*
Pediatric ED physician, yes
n = 8947*
Pediatric ED physician, no
n = 3331" 507
126
111
195
152
118
Director pod trauma pod surgeon, yes
n = 11873
n = 1309
n - 138
n = 16263
n = 2081
n = 299
Director pod trauma ped surgeon, no
n = 403
434 333 592 344
395 202
n = 893
>7 yr ]SS 1-16
154 n = 829 158 n = 523 134 n = 843 162 n = 509
113 n = 92
296
236
112
n = 12168
n = 1071
n = 133
118 n = 50
306
255
113
n = 5115
n = 1167
n - 187
109 n = 99
198
175
131
n = 11762
n = 1092
n = 139
117 n = 43
148
117
n = 43
n = 4
168
43
312 n = 5521
278
130
n = 1146
n = 181
279
219
n = 1020
n = 157
208
145
125 n = 21 95
NOTE. In the patient group less than 7 years old with an ISS less than 17, the a m o u n t of time in the ED was decreased significantly by the presence of an in-house attending surgeon, a pediatric ED, and a pediatric ED physician. All other groups were w i t h o u t significant differences. * P < .05.
Table 6. Average Length of Stay (d) <7 yr ISS 1-16 Attending in house, yes Attending in house, no Pediatric ED, yes Pediatric ED, no Pediatric ED physician, yes Pediatric ED physician, no Director ped trauma pod surgeon, yes Director pod trauma ped surgeon, no
<7 yr ISS 17-35
<7 yr ISS 36-75
>7 yr IS9 1-16
>7 yr I$S 17-35
n = 3524
n = 654
n = 68
n = 6727
n = 1354
5 (0-90)
11 (0-140)
20 (0-184)
5 (0-169)
5 (0-703)
n = 11510
n = 1153
n = 120
n = 14201
n - 1469
>7 yr ISS 36-75 n = 276 5 (0-524) n = 190
4 (0-185)
14 (0-226)
18 (0-192)
4 (0-180)
16 (0-206)
18 (0-185)
n = 11309
n = 1082
n = 115
n = 13934
n = 1289
n = 160
4 (0-185)
14 (0-226)
19 (0-192)
4 (0-169)
17 (0-206)
17 (0-185)
n = 3775
n = 730
n = 73
n = 7082
n = 1549
n = 308
5 (0-137)
11 (0-140)
18 (0-184)
5 (0-180)
15 (0-703)
21 (0-524)
n = 11276
n = 1211
n = 136
n = 15034
n = 1619
n = 271
4 (0-137)
13 (0-226)
19 (0-192)
4 (0-206)
16 (0-206)
16 (0-185)
n = 3838
n = 601
n = 52
n = 5982
n = 1219
n = 197
4 (0-185)
11 (0-140)
5 (0-184)
5 (0-180)
17 (0-703)
24 (0-524)
n = 14622
n = 1759
n = 184
n = 19892
n = 2669
n = 446
4 (0-185)
13 (0-226)
19 (0-192)
5 (0-180)
16 (0-703)
20 (0-524)
n = 462
n - 53
n = 4
n - 1124
n = 169
n = 22
4 (0-35)
11 (0-48)
20 (0-42)
4 (0-85)
12 (0-71)
13 (0-34)
NOTE. No groups had a significant difference to their counterpart for this outcome. N umbers in parentheses represent ranges.
Table 7. Impairment Lasting Greater than 7 Months (%)
Attending in house, yes
<7 yr ISS 1-16
<7 yr ISS 17-35
41/3524
80/654 12% 127/1153
1% Attending in house, no
97/11510
Pediatric ED, yes
86/11309
Pediatric ED, no
53/3775
Pediatric ED physician, yes
82/I 1276
Pediatric ED physician, no
57/3808
1% 1% 1% 1% 1%
<7 yr ISS 36-75
>7 yr ISS 1 16
>7 yr ISS 17-35
>7 yr ISS 36-75
15/63
94/6727
235/1354
72/276
22%
1%
17%
15/120
150/14201
211/1469
26% 36/190
11%
13%
1%
14%
19%
112/1082
17/115
129/13934
166/1289
24/308
10%
15%
95/'730
13/73
1% 117/7082
13%
15%
283/1549
84/308
13%
18%
2%
18%
27%
125/1211
22/136
159/15034
207/1619
46/271
10%
16%
1%
13%
82/601
8/52
87/5982
242/1219
17% 62/446
14%
15%
1%
20%
23%
203/1759
29/184
238/19892
422/2669
103/446
1%
12%
16%
16%
23%
2/462
4/53
1/4
8%
25%
Director pod trauma pod surgeon, yes
137/14622
Director pod trauma ped surgeon, no
0.4%
1% 8/1124 1%
27/169 16%
5/22 23%
NOTE. A l t h o u g h the trend f o r an i m p a i r m e n t was much higher with the severe injury, that was not a function of the resources available to the patient.
889
PEDIATRIC TRAUMA CENTER CRITERIA
Table 8. Inquiries in Patients Who Went to the Operating Room Within 1 Hour of Admission Diagnoses
Died (%)
Survived (%)
Skull/head/neck/face Spine Bone injuries (fracture dislocation) Contusion, abrasion Thoracic Abdominal Pelvis Other
35 1 6 28 3 22 4 1
17 1 27 98 6 14 4 3
lated. A n in-house surgeon is of no use without in-house operating r o o m and v i c e versa. It was b e y o n d the scope o f our investigation to study the interactive effects that one c o m p o n e n t had on the other. Perhaps " c o m p o n e n t l i n e s " or " p a c k a g e s " that c o m b i n e c o m p l e m e n t a r y services w o u l d be a better approach. Study the surgeon with the in-house operating r o o m and anesthesia. E v a l u a t e radiological i m a g i n g as a separate line. W h a t laboratory support is critical? This m a y be useful in future w o r k to decide what combinations are optimal.
REFERENCES
1. Tepas JJ, Patel SC, DiScala C, et al: Relationship of trauma patient volume to outcome experience: can a relationship be defined? J Trauma 44:827-830, 1998 2. Schwab W, Frankel HL, Rotondo MF, et al: The impact of true partnership between a university Level I trauma center and a community Level II trauma center on patient transfer practices. J Trauma 44:813-819, 1998 3. Johnson DL, Krishnamurthy S: Send severely head-injured children to a pediatric trauma center. Pediatr Neurosurg 23:309-314, 1996 4. Hall JR, Reyes HM, Meller JL, et al: The outcome for children with blunt trauma is best at a pediatric center. J Pediatr Surg 31:72-76, 1996 5. Ramenofsky ML, Moulton SL: The pediatric trauma center. Semin Pediatr Surg 4:128-134, 1995 6. Committee on Trauma American College of Surgeons: Resource document for optimal care of the injured patient. Chicago, IL, American College of Surgeons, 1990, pp 31-34 7. F 1286-90 Standard guide for development and operation of Level I pediau:ic trauma facilities, in Annual Book of ASTM Standards 13:01, Philadelphia, PA, American Society for Testing and Materials, 1990 8. Pennsylvania Trauma Systems Foundation: 1990-1991 pediatric standards for trauma center accrediation, Pennsylvania Trauma, Harrisburg, PA, The Foundation, 1990 9. D'Amelio LF, Hammond JS, Thomasseau J, et al: "Adult" trauma
surgeons with pediatric commitment: A logical solution to the pediatric trauma manpower problem. Am Surg 61:968-74, 1993 10. Waldrop R, Peck CQ, Hutchinson S, et al: Comparison of pediatric hospitalization using the pediatric appropriateness evaluation protocol at three diverse hospitals in Louisiana. J La State Med Soc 150:21 1-7, 1998 11. Van der Sluis CK, Kingma J, Eisma WH, et al: Pediatric polytrauma: Short-term outcomes. J Trauma 43:301-306, 1997 12. Harrison T, Thomas SH, Wedel SK: Interhospital aeromedical transports: Air medical activation intervals in adult and pediatric trauma patients. Am J Emerg Med 15:122-124, 1997 13. Kissoon N, Tepa JJ, Peterson RJ, et al: The evaluation of pediatric trauma care using audit tilters. Pediatr Emerg Care 12:272276, 1996 14. Konvoliaka CW, Copes WS, Sacco WJ: Institution and persurgeon volume versus survival outcome in Pennsylvania's trauma centers. Am J Surg 170:333-340, 1995 15. Ramenofsky M, Luterman QA, Quindlen E, et al: Maximum survival in pediatric trauma: The ideal system. J Trauma 24:818-823, 1984 16. American Pediatric Surgical Association Trauma Committee Survey presented at the Annual Meeting of the American Pediatric Surgical Association, Vancouver, British Columbia, Canada, May 18, 1990
Discussion M. Fallat (Louisville, KY): D r Doolin, I applaud your goal of attempting to analyze individual c o m p o n e n t s of a trauma center that m i g h t affect o u t c o m e and intuitively agree that the results s e e m valid. H o w e v e r , I a m concerned that y o u r m e t h o d o l o g y is too simplistic and w o u l d e n c o u r a g e you to use a statistician to reevaluate the data b e f o r e publishing this work. F o r example, statistical analysis of survival data could be e v a l u a t e d w i t h a log rank test, and logistic regression analysis w o u l d lend itself w e l l to evaluate the potential multifactorial influences on outcome, w h i c h you did not currently address. It is e v e n c o n c e i v a b l e to stratify by trauma center. M y concern is that you h a v e collected an e n o r m o u s a m o u n t of data and done a lot o f work, and the m e t h o d o l o g y you h a v e c h o s e n does not allow accurate interpretation. I h a v e a f e w questions. T h e first is h o w did y o u arrive at the age categorization
o f less than 7 years or greater than or equal to 7 years? E i g h t years is the usual break point for airway considerations. M o s t adult trauma centers admit patients 16 and 17 years old and include t h e m in their data. It m i g h t be better to stratify into different age groups. The second question is w h a t is the average n u m b e r of patients per center per year for the centers y o u h a v e chosen to include? D i d you l o o k at penetrating versus blunt trauma separately? It could be that an urban center w o u l d see m o r e penetrating trauma, and that m i g h t h a v e a higher impact on mortality rate, based purely on m e c h a n i s m . I notice that the length of stay in days was up to 703 days in s o m e o f the children o v e r 7 years old, w h i c h is almost 2 years, and w o n d e r e d if y o u had an explanation for this w h a t seems like an almost inordinate length o f stay for an individual patient? E. Doolin (response): T h e age chosen was a balance I
890
tried to reach. There was a lot of work done in terms of studying the mechanism of injury many years ago that would suggest that 13 is a good break point for mechanism of injury. But then more recent work suggests that age 5 is a point for physiological distinction. I did not want to isolate the greater than 14, because that would have completely eliminated, in my mind, some of the trauma centers that do not admit older children. These data certainly can be restratified, because we have it in the gross form, and that could be looked into to answer any question. Penetrating and blunt trauma were not separated, although one thing I did look into was the presence of the surgeon, and the outcome of any patient that went to the operating room in less than an hour. Of those patients that went into the operating room in less than an hour and died, the overwhelming diagnosis was head injury. In those patients that went to the operating room in less than an hour and survived, the overwhelming diagnosis was extremity injury, hence treatable versus nontreatable. That is a very small cohort. In terms of the outliers, I did not scrutinize the outliers in terms of number of days. I could only speculate that this may have been an earlier time in the study at which relationships with rehabilitation hospitals were not as close. A. Hailer (Baltimore, MD): I think your information about the in-house surgeon is a critical one, because that has been the stickiest wicket for most of us in determining the exact nature of our children's trauma programs and children's trauma centers. So I want to be sure I understood what you said before I carry some of the word. If I understood you correctly, there were 34% of the centers that had in-house surgeons? E. DooIin (response): Correct, surgeons were defined as an attending board-certified surgeon. A. Haller (Baltimore, MD): Is this definition that they are actually geographically in the hospital? Or do they meet the criteria of the college that they are within 20 to 30 minutes of arriving in the hospital at the time of the seriously injured patient? The second question is, I am worried about the statistics that you gave us about the in-house surgeon. As I look at your abstract, there were 46% versus 56% who are included in that over-7 age group. And that does not look like a statistically significant difference to me, even though you have got this huge number. The final question is, on the basis of this information you have given us, do you think you would feel comfort-
DOOLIN, BROWNE, AND DISCALA
able recommending that a pediatric trauma center have a surgeon available outside the hospital within 30 minutes, therefore changing our criteria? Or do you believe that on the basis of this information you would want to have a pediatric trauma center designated with an in-house surgeon in the sense of being physically in the hospital? E. Doolin (response): I made a very special effort not to come up with a recommendation, because these are all numbers, and that was my mission. But I have thoughts on it, which are separate. The in-house surgeon was a board-certified general surgeon who stayed within the walls of the trauma center 24 hours a day. And the reason for the phone calls was to be sure there was no misunderstanding about that. An R-5 did not count, a Fellow did not count unless he was board-certified in general surgery. In terms of the statistics, the 10% drop in mortality rate represents a 20% change in the groups. This was done within that cohort by both the Student's t test and the X2 test, again within that cohort. That statistic, if expanded to the entire cohort, would be meaningless. Because, essentially, out of 46,000 patients, it was nine lives saved. And so in terms of the overall system, that statistic would be meaningless. It is only addressing that cohort. And that information can be used however you choose to as long as you recognize that. In terms of the pediatric surgeon or any surgeon being in the house or not in the house, everybody just has to interpret these data accordingly. My feeling is that the role of the in-house surgeon is so limited by these data that it really does not serve a purpose, although the purpose it does serve is in the dozen kids who got shot in the heart. But outside of that, they do not serve a purpose. You could you argue that the in-house attending surgeon will save some money because of the efficiency of the mildly injured patients. And I kind of like that idea. But the pediatric ED physician and the separate pediatric ED have the same effect, and you do not have to take the 400 or 500 pediatric surgeons and stick them in the house just to get a Glasgow of 13 out of the emergency room. Baltimore clearly is different than Children's Memorial of Chicago, which is four miles from Northwestern, which is different from Washington DC, which although four miles from GW is 400 yards from the Hospital Center. We all have different resources, and to compare them just as categories is not as effective as comparing what they offer on an individual basis.
Camden, N e w Jersey and Boston, Massachusetts
Background~Purpose:Trauma centers (TC) are certified based on widely accepted criteria. These specific criteria rarely are scrutinized individually. The purpose of this study was to analyze the individual components of a pediatric trauma center for their effect on outcome. Methods: Members of the National Pediatric Trauma Registry were queried about the following: (1) separate pediatric emergency department (ED), (2) pediatric intensive care unit (PICU), (3) pediatric intensivist as PICU director, (4) pediatric surgeon as TC director, (5) in-house attending surgeon, (6) in-house pediatric emergency physician, (7) 24-hour operating room, (8) 24-hour computed tomography (CT) scan. Outcomes analyzed included mortality, length of stay, time in ED, days in PICU, and disability. Victims were stratified based on age (<7 or ->7 years) and severity of injury (ISS -< 16, 17-35, >-36). Results were compared using Student's t test and X2 analysis. Results: A total of 59 of 74 centers responded, 18 were dropped because of low enrollment (mean, 1.6 patients).
HERE HAS BEEN a great deal of development and progress in the institution of trauma systems over the last three decades. Since the development of the first statewide trauma programs, experience and data have shown that there is improved survival rate, lowered disability, and cost effective care when a trauma program is in place, 1-4 There has also been a constant discovery as to what is required for a trauma program. There are many components that come together to create the trauma system. The prehospital program, the trauma center facility, personnel, and rehabilitation programs all constitute the system. In addition, education and research figure in a major way. Many documents recommend the required resources for a trauma center. 5-8 There is no situation in which the controversy is as great as in the area of pediatric trauma centers. Not only is there a question among pediatric trauma systems as to what the requirements should be, but these programs also are being compared with their adult counterparts for content and o u t c o m e ) There is no universal definition or credentialing body for trauma centers. This makes comparisons difficult. 1°-12 In addition, there has been very little in the way of objectively scrutinizing the individual requirements of a trauma center for their value/3-15 A recent survey of pediatric surgeons' activity suggests that the majority are involved
T
Journal of Pediatric Surgery,Vo134, No 5 (May), 1999: pp 885-890
Questions 3, 4, 6, and 7 were eliminated because of skewed data. An in-house surgeon reduced the amount of time a mildly injured patient (ISS -< 16) spent in the ED (210 v 434 minutes), as did the separate pediatric ED (333 v592 minutes) and pediatric emergency physicians (344 v 507 minutes) in younger patients (>-7 years). An in-house surgeon reduced the morality rate in older (>-7) severely injured (ISS >- 36) patients (46.7% v 56.8%; P < .05 for all). No other differences were significant. Conclusions: In-house personnel improved efficiency for the less severely injured, and an in-house attending surgeon reduced mortality in the severely injured older patient. None of the other variables were found to have a significant impact on outcome. J Pediatr Surg 34:885-890. Copyright © 1999 by W.B. Saunders Company.
INDEX WORDS: Trauma, standards, outcomes.
in pediatric trauma care. 16 The settings are as varied as their number. In this project, we did not evaluate different types of trauma centers. Rather, we evaluated the different components of a trauma center for the value it gave the system in terms of efficiency and outcome. The goal o f this type of analysis was to identify how the components of the pediatric trauma center contributed to its success. In addition, the hope was to identify which components were or were not significant in their contribution to the value of the center. MATERIALS A N D METHODS
In 1995, there were 74 Pediatric Trauma Centers that belonged to the National Pediatric Trauma Registry (NPTR). This made up the study
From the Division of Pediatric Surgery, Department of Surgery, Robert Wood Johnson Medical School at Camden, Camden, N J, and the Research and Training Center, Tufts University School of Medicine, Boston, MA. Presented at the 1998 Annual Meeting of the Section on Surgery of the Ameriean Academy of Pediatrics, San Francisco, California, October 16-19, 1998. Address reprint requests to Edward J. Doolin, MD, Department of Surgery, Division of Pediatric Surgery, Three Cooper Plaza, Suite 411, Camden, NJ 08103. Copyright © 1999 by W.B. Saunders Company 0022-3468/99/3405-0043503.00/0
885
886
DOOLIN, BROWNE, AND DISCALA
group for this report. Each center was contacted in writing or by phone to complete a simple questionnaire. Eight components of a trauma center were to be evaluated (Table 1). For the purposes of this evaluation, each component was studied in a singular fashion. Any given center could be a "yes" for one analysis and a "no" for another. The database of the National Pediatric Trauma Registry was then used to measure outcomes. The cohort of patients for the NPTR phase II (ending 1996) was the study group. Five different outcomes were measured for this analysis (Table 2). The patients were stratified to allow comparison. Patients were grouped by age (--> or < 7 years) and by severity of injury (1SS 1-16, 17-35, or >35). For each parameter in the questionnaire the trauma centers were grouped according to their answer. The outcomes between these trauma centers were then compared controlling for age and severity of injury. The significance of the results was then verified using a Student's t test and X2 analysis for each outcome. Significance was determined at P < .05.
RESULTS
Of the 74 trauma centers, 59 were able to be included in the study. Reasons that a center was not included were inability to complete information (n = 15) or the submission of less than 25 patients through phase 2 (n = 25). It was felt that a small volume lent itself to a skewed value for these centers because the centers were not weighted for size. (The mean patient volume for this group was 1.6). The responses from the 59 centers included are listed (Table 1). Based on these answers the center could then be categorized for outcome analysis. Outcomes data were then collected and collated based on the answers to these questions. Questions 2, 3, 7, and 8 had largely affirmative responses (>85%), and it was felt this did not allow for a valid analysis, because the group sizes were not comparable. Each outcome was analyzed by comparing the centers based on their answers to questions 1, 4, 5, and 6. The first outcome compared was days in pediatric intensive care unit (PICU) (Table 3). The presence or absence of any of the four studied criteria did not have an effect on days in the PICU. Table 1. Questions and Responses From Each Trauma Center
Question 1. Does your institution have a separate pediatric emergency department. 2. Does your institution have a separate pediatric intensive care unit? 3. Is the medical director of the PICU a pediatric intensivist? 4. Is there a pediatric surgeon who is the director of the pediatric trauma service? 5. Does your institution have an attending surgeon available in house 24 hours a day? 6. Does your institution have an in-house pediatric emergency room attending surgeon 24 hours a day 7. Does your institution have 24-hour in-house operating room availability? 8. Does your institution have 24-hour in-house CT availability?
Response (n = 59) (% yes)
54 100 97 76 34 59 85 88
Table 2. Outcomes Data Measured a. b. c. d. e.
Length of Stay in PICU (d) Mortality rate Initial time in the emergency department (min) Overall length of stay (d) Long-term impairment (>7 too)
The second outcome studied was mortality (Table 4). The presence of an in-house attending surgeon reduced the mortality rate of the severely injured (ISS > 35) older ( > 7 years) patient from 56.8% to 46.7% (P < .05). In addition, the presence of a pediatric emergency department (ED) increased the mortality rate from 44.8% to 61.3% (P < .05). These components did not affect any other group, and no other components had any effect on mortality. The third outcome measured was time in the emergency room (Table 5). The presence of an in-house attending surgeon, a pediatric ED, and a pediatric ED physician all reduced the length of stay for a young (<7 years) mildly injured (ISS < 16) patient. No other differences were found. The overall hospital stay also was examined. No component studied had any effect on the overall length of stay (Table 6). The patient disability was evaluated as the frequency of long-term problems (Table 7). None of the components studied had an effect on the long-term sequelae. DISCUSSION
As the need for cost-effective and efficient care increases, the components of a health care system are going to be scrutinized. In addition, the goal of pediatric trauma systems always will be intervening in the most severe of injuries with a successful outcome. In the current study we chose to compare the components of the product to find if they had a significant effect on health care delivery. Many components were too prevalent to create comparable groups and, therefore, were not studied. The criteria studied did give useful data. It is not surprising that the presence of an in-house attending surgeon decreased the mortality rate in the most severely injured group. This is a small cohort (232 of 41,246 or 0.56%), and the improvement (18 patients or .04%) is even smaller. If not isolated as a separate small group, these differences would never be significant. This group may be useful in that it represents the "extreme injury," which often is considered a "mission" for the trauma center. An in-house surgeon allows for rapid surgical intervention. Although most deaths are from central nervous system injury, the preventable deaths in this group usually are caused by hemorrhage or other "golden hour" mechanisms. To validate this, the difference in diagnoses were compared in victims needing emergency surgery. An analysis of patients operated on in less than 1 hour, was performed
PEDIATRIC T R A U M A CENTER CRITERIA
887
Table 3. Average Days in the PICU <7 yr ISS 1-16 A t t e n d i n g in house, yes A t t e n d i n g in house, no Pediatric ED, yes Pediatric ED, no Pediatric ED physician, yes Pediatric ED, physician, no Director Ped Trauma Ped Surgeon, yes Director Ped Trauma Ped Surgeon, no
<7 yr ISS 17-35
<7 yr ISS 36-75
>7 yr ISS 1-16
>7 yr ISS 17-35
>7 yr ISS 36-75 n = 216
n = 1355
n = 567
n = 47
n = 1974
n = 1155
2 (1-49)
5 (1-76)
16 (1-131)
2 (1-31)
6 (1-42)
10 (1-85)
n = 671
n = 878
n = 88
n = 2024
n = 1187
n = 137
2 (1-112)
5 (1-76)
11 (1-192)
2 (1-33)
6 (1-77)
n = 1503
n = 809
n = 84
n = 1797
n = 1019
2 (1-60)
5 (1-76)
11 (1-192)
3 (1-33)
6 (1-105)
n = 1546
n = 641
n = 51
n = 2228
n = 1338
2 (1-112)
5 (1-72)
15 (1-131)
2 (1-31)
6 (1-142)
10 (1-85)
n = 1947
n = 969
n = 97
n = 2406
n = 1317
n = 193
2 (1-112)
5 (1-76)
12 (1-192)
2 (1-33)
6 (1-105)
n = 1102
n = 481
n = 38
n = 1619
n = 1040
2 (1-60)
5 (1-70)
15 (1-113)
2 (1-27)
6 (1-142)
11 (1-85)
n = 2921
n = 1403
n = 133
n = 3785
n = 2220
n = 337
2 (1-112)
5 (1-76)
12 (1-192)
2 (1-33)
6 (1-142)
n = 128
n = 47
n = 2
n = 240
n = 137
n = 17
2 (1-14)
5 (1-17)
21 (6-36)
2 (1-15)
5 (1-46)
10 (1-34)
9 (1-37) n = 113 8 (1-37) n = 241
8 (1-74) n = 161
9 (1-85)
NOTE. None of the groups studied had any differences in this o u t c o m e that w e r e statistically significant. N u m b e r s in parentheses are ranges.
(Table 8). Those who died predominately had head injuries (35%) with associated other problems. Those who lived predominately had fractures with only 17% head injuries. The abdominal injuries also were fewer in the survival group. This suggests that for the fatal group, the injuries (head) are less treatable than those of the survivors (orthopedic). No clear pattern of injuries identified a need filled by the in-house attending surgeon, suggesting his main role may be largely that of decision making. However, the patients who benefit are few, and one could argue that they are not worth the investment of a n in-house attending surgeon. Paradoxically, the presence of a separate pediatric ED increased mortality rate. This is likely a statistical fallout in that most centers with in-house attending surgeons are probably housed in an
adult trauma center without a separate pediatric ED. Again, this is a small group. Interestingly, the presence of a staff surgeon increased the efficiency of the mild to moderate injured group a great deal as did a pediatric emergency physician. This might be evaluated from a financial point of view. It is possible that the improved efficiency justifies the presence of the attending surgeon financially. However, this reduction in price may not be translated to a reduction in cost because of the capital intensive effort of trauma. Also, there may be other strategies that result in this effect without the need of the in-house surgeon. This study evaluated each component in isolation. The treatment of trauma is multifactorial and so are its outcomes. Some of these components may be interre-
Table 4. Morality Rates <7 yr ISS 1-16
<7 yr ISS 17-35
3/3524 0.08%
Attending in house, no
18/11510 0.16%
23.5%
63.3%
0.08%
14.8%
Pediatric E D, yes
17/11309
257/1082
72/115
10/13934
204/1289
0.15%
23.8%
62.6%
0.07%
15.8%
61.3%
Pediatric ED, no
4/3775
198/730
39/73
6/7082
266/1549
138/308"
0.11%
27.1%
53.4%
0.08%
17.2%
44.8%
16/11276
296/1211
81/136
14/15034
273/1619
147/271
A t t e n d i n g in house, yes
Pediatric ED physicia n, yes Pediatric ED physician, no Director Ped Trauma Ped Surgeon, yes Di rector Ped Trauma Ped Surgeon, no
<7 yr ISS 36-75
>7 yr ISS 1-16
>7 yr ISS 17-35
>7 yr ISS 36-75
182/654
35/68
4/6727
250/1354
126/276"
27.8%
51.8%
0.06%
18.5%
46.7%
272/1153
76/120
12/14201
217/1469
108/190" 56.8% 98/160*
0.14%
24.4%
59.6%
0.09%
16.9%
54.2%
5/3808
159/601
30/52
2/5982
197/1219
89/197
0.013% 21/14622
26.5% 444/1759
57.7% 109/446
0.03% 15/19877
16.2% 444/2669
45.0% 226/446
0.14%
25.0%
59.2%
0.08%
16.6%
50.7%
0/462
11/53
2/4
1/1124
26/169
10/22
0%
20.8%
50.0%
0.09%
15.4%
45.5%
NOTE. In the severely injured patient the presence of an in-house attending significantly reduced the m o r t a l i t y rate. This w a s rare b u t significant. Paradoxically, the m o r t a l i t y rate in the same g r o u p w a s higher w h e n a separate pediatric e m e r g e n c y r o o m existed than w h e n it was not available. * P < .05.
Table 5. 33me Spent in ED (rain) <7 yr ISS 1-16 Attending in house, yes
n = 2468* 210
<7 yr ISS 17-35
<7 yr ISS 36-75
n = 455
n = 47
137
n = 4833
>7 yr ISS 17-35
>7 yr ISS 36-75
n = 986
n = 165
119
219
186
135
n = 95*
n = 12367
n = 1240
n = 154
Attending in house, no
n = 9766*
Pediatric ED, yes
n = 9670*
Pediatric ED, no
n = 2608*
Pediatric ED physician, yes
n = 8947*
Pediatric ED physician, no
n = 3331" 507
126
111
195
152
118
Director pod trauma pod surgeon, yes
n = 11873
n = 1309
n - 138
n = 16263
n = 2081
n = 299
Director pod trauma ped surgeon, no
n = 403
434 333 592 344
395 202
n = 893
>7 yr ]SS 1-16
154 n = 829 158 n = 523 134 n = 843 162 n = 509
113 n = 92
296
236
112
n = 12168
n = 1071
n = 133
118 n = 50
306
255
113
n = 5115
n = 1167
n - 187
109 n = 99
198
175
131
n = 11762
n = 1092
n = 139
117 n = 43
148
117
n = 43
n = 4
168
43
312 n = 5521
278
130
n = 1146
n = 181
279
219
n = 1020
n = 157
208
145
125 n = 21 95
NOTE. In the patient group less than 7 years old with an ISS less than 17, the a m o u n t of time in the ED was decreased significantly by the presence of an in-house attending surgeon, a pediatric ED, and a pediatric ED physician. All other groups were w i t h o u t significant differences. * P < .05.
Table 6. Average Length of Stay (d) <7 yr ISS 1-16 Attending in house, yes Attending in house, no Pediatric ED, yes Pediatric ED, no Pediatric ED physician, yes Pediatric ED physician, no Director ped trauma pod surgeon, yes Director pod trauma ped surgeon, no
<7 yr ISS 17-35
<7 yr ISS 36-75
>7 yr IS9 1-16
>7 yr I$S 17-35
n = 3524
n = 654
n = 68
n = 6727
n = 1354
5 (0-90)
11 (0-140)
20 (0-184)
5 (0-169)
5 (0-703)
n = 11510
n = 1153
n = 120
n = 14201
n - 1469
>7 yr ISS 36-75 n = 276 5 (0-524) n = 190
4 (0-185)
14 (0-226)
18 (0-192)
4 (0-180)
16 (0-206)
18 (0-185)
n = 11309
n = 1082
n = 115
n = 13934
n = 1289
n = 160
4 (0-185)
14 (0-226)
19 (0-192)
4 (0-169)
17 (0-206)
17 (0-185)
n = 3775
n = 730
n = 73
n = 7082
n = 1549
n = 308
5 (0-137)
11 (0-140)
18 (0-184)
5 (0-180)
15 (0-703)
21 (0-524)
n = 11276
n = 1211
n = 136
n = 15034
n = 1619
n = 271
4 (0-137)
13 (0-226)
19 (0-192)
4 (0-206)
16 (0-206)
16 (0-185)
n = 3838
n = 601
n = 52
n = 5982
n = 1219
n = 197
4 (0-185)
11 (0-140)
5 (0-184)
5 (0-180)
17 (0-703)
24 (0-524)
n = 14622
n = 1759
n = 184
n = 19892
n = 2669
n = 446
4 (0-185)
13 (0-226)
19 (0-192)
5 (0-180)
16 (0-703)
20 (0-524)
n = 462
n - 53
n = 4
n - 1124
n = 169
n = 22
4 (0-35)
11 (0-48)
20 (0-42)
4 (0-85)
12 (0-71)
13 (0-34)
NOTE. No groups had a significant difference to their counterpart for this outcome. N umbers in parentheses represent ranges.
Table 7. Impairment Lasting Greater than 7 Months (%)
Attending in house, yes
<7 yr ISS 1-16
<7 yr ISS 17-35
41/3524
80/654 12% 127/1153
1% Attending in house, no
97/11510
Pediatric ED, yes
86/11309
Pediatric ED, no
53/3775
Pediatric ED physician, yes
82/I 1276
Pediatric ED physician, no
57/3808
1% 1% 1% 1% 1%
<7 yr ISS 36-75
>7 yr ISS 1 16
>7 yr ISS 17-35
>7 yr ISS 36-75
15/63
94/6727
235/1354
72/276
22%
1%
17%
15/120
150/14201
211/1469
26% 36/190
11%
13%
1%
14%
19%
112/1082
17/115
129/13934
166/1289
24/308
10%
15%
95/'730
13/73
1% 117/7082
13%
15%
283/1549
84/308
13%
18%
2%
18%
27%
125/1211
22/136
159/15034
207/1619
46/271
10%
16%
1%
13%
82/601
8/52
87/5982
242/1219
17% 62/446
14%
15%
1%
20%
23%
203/1759
29/184
238/19892
422/2669
103/446
1%
12%
16%
16%
23%
2/462
4/53
1/4
8%
25%
Director pod trauma pod surgeon, yes
137/14622
Director pod trauma ped surgeon, no
0.4%
1% 8/1124 1%
27/169 16%
5/22 23%
NOTE. A l t h o u g h the trend f o r an i m p a i r m e n t was much higher with the severe injury, that was not a function of the resources available to the patient.
889
PEDIATRIC TRAUMA CENTER CRITERIA
Table 8. Inquiries in Patients Who Went to the Operating Room Within 1 Hour of Admission Diagnoses
Died (%)
Survived (%)
Skull/head/neck/face Spine Bone injuries (fracture dislocation) Contusion, abrasion Thoracic Abdominal Pelvis Other
35 1 6 28 3 22 4 1
17 1 27 98 6 14 4 3
lated. A n in-house surgeon is of no use without in-house operating r o o m and v i c e versa. It was b e y o n d the scope o f our investigation to study the interactive effects that one c o m p o n e n t had on the other. Perhaps " c o m p o n e n t l i n e s " or " p a c k a g e s " that c o m b i n e c o m p l e m e n t a r y services w o u l d be a better approach. Study the surgeon with the in-house operating r o o m and anesthesia. E v a l u a t e radiological i m a g i n g as a separate line. W h a t laboratory support is critical? This m a y be useful in future w o r k to decide what combinations are optimal.
REFERENCES
1. Tepas JJ, Patel SC, DiScala C, et al: Relationship of trauma patient volume to outcome experience: can a relationship be defined? J Trauma 44:827-830, 1998 2. Schwab W, Frankel HL, Rotondo MF, et al: The impact of true partnership between a university Level I trauma center and a community Level II trauma center on patient transfer practices. J Trauma 44:813-819, 1998 3. Johnson DL, Krishnamurthy S: Send severely head-injured children to a pediatric trauma center. Pediatr Neurosurg 23:309-314, 1996 4. Hall JR, Reyes HM, Meller JL, et al: The outcome for children with blunt trauma is best at a pediatric center. J Pediatr Surg 31:72-76, 1996 5. Ramenofsky ML, Moulton SL: The pediatric trauma center. Semin Pediatr Surg 4:128-134, 1995 6. Committee on Trauma American College of Surgeons: Resource document for optimal care of the injured patient. Chicago, IL, American College of Surgeons, 1990, pp 31-34 7. F 1286-90 Standard guide for development and operation of Level I pediau:ic trauma facilities, in Annual Book of ASTM Standards 13:01, Philadelphia, PA, American Society for Testing and Materials, 1990 8. Pennsylvania Trauma Systems Foundation: 1990-1991 pediatric standards for trauma center accrediation, Pennsylvania Trauma, Harrisburg, PA, The Foundation, 1990 9. D'Amelio LF, Hammond JS, Thomasseau J, et al: "Adult" trauma
surgeons with pediatric commitment: A logical solution to the pediatric trauma manpower problem. Am Surg 61:968-74, 1993 10. Waldrop R, Peck CQ, Hutchinson S, et al: Comparison of pediatric hospitalization using the pediatric appropriateness evaluation protocol at three diverse hospitals in Louisiana. J La State Med Soc 150:21 1-7, 1998 11. Van der Sluis CK, Kingma J, Eisma WH, et al: Pediatric polytrauma: Short-term outcomes. J Trauma 43:301-306, 1997 12. Harrison T, Thomas SH, Wedel SK: Interhospital aeromedical transports: Air medical activation intervals in adult and pediatric trauma patients. Am J Emerg Med 15:122-124, 1997 13. Kissoon N, Tepa JJ, Peterson RJ, et al: The evaluation of pediatric trauma care using audit tilters. Pediatr Emerg Care 12:272276, 1996 14. Konvoliaka CW, Copes WS, Sacco WJ: Institution and persurgeon volume versus survival outcome in Pennsylvania's trauma centers. Am J Surg 170:333-340, 1995 15. Ramenofsky M, Luterman QA, Quindlen E, et al: Maximum survival in pediatric trauma: The ideal system. J Trauma 24:818-823, 1984 16. American Pediatric Surgical Association Trauma Committee Survey presented at the Annual Meeting of the American Pediatric Surgical Association, Vancouver, British Columbia, Canada, May 18, 1990
Discussion M. Fallat (Louisville, KY): D r Doolin, I applaud your goal of attempting to analyze individual c o m p o n e n t s of a trauma center that m i g h t affect o u t c o m e and intuitively agree that the results s e e m valid. H o w e v e r , I a m concerned that y o u r m e t h o d o l o g y is too simplistic and w o u l d e n c o u r a g e you to use a statistician to reevaluate the data b e f o r e publishing this work. F o r example, statistical analysis of survival data could be e v a l u a t e d w i t h a log rank test, and logistic regression analysis w o u l d lend itself w e l l to evaluate the potential multifactorial influences on outcome, w h i c h you did not currently address. It is e v e n c o n c e i v a b l e to stratify by trauma center. M y concern is that you h a v e collected an e n o r m o u s a m o u n t of data and done a lot o f work, and the m e t h o d o l o g y you h a v e c h o s e n does not allow accurate interpretation. I h a v e a f e w questions. T h e first is h o w did y o u arrive at the age categorization
o f less than 7 years or greater than or equal to 7 years? E i g h t years is the usual break point for airway considerations. M o s t adult trauma centers admit patients 16 and 17 years old and include t h e m in their data. It m i g h t be better to stratify into different age groups. The second question is w h a t is the average n u m b e r of patients per center per year for the centers y o u h a v e chosen to include? D i d you l o o k at penetrating versus blunt trauma separately? It could be that an urban center w o u l d see m o r e penetrating trauma, and that m i g h t h a v e a higher impact on mortality rate, based purely on m e c h a n i s m . I notice that the length of stay in days was up to 703 days in s o m e o f the children o v e r 7 years old, w h i c h is almost 2 years, and w o n d e r e d if y o u had an explanation for this w h a t seems like an almost inordinate length o f stay for an individual patient? E. Doolin (response): T h e age chosen was a balance I
890
tried to reach. There was a lot of work done in terms of studying the mechanism of injury many years ago that would suggest that 13 is a good break point for mechanism of injury. But then more recent work suggests that age 5 is a point for physiological distinction. I did not want to isolate the greater than 14, because that would have completely eliminated, in my mind, some of the trauma centers that do not admit older children. These data certainly can be restratified, because we have it in the gross form, and that could be looked into to answer any question. Penetrating and blunt trauma were not separated, although one thing I did look into was the presence of the surgeon, and the outcome of any patient that went to the operating room in less than an hour. Of those patients that went into the operating room in less than an hour and died, the overwhelming diagnosis was head injury. In those patients that went to the operating room in less than an hour and survived, the overwhelming diagnosis was extremity injury, hence treatable versus nontreatable. That is a very small cohort. In terms of the outliers, I did not scrutinize the outliers in terms of number of days. I could only speculate that this may have been an earlier time in the study at which relationships with rehabilitation hospitals were not as close. A. Hailer (Baltimore, MD): I think your information about the in-house surgeon is a critical one, because that has been the stickiest wicket for most of us in determining the exact nature of our children's trauma programs and children's trauma centers. So I want to be sure I understood what you said before I carry some of the word. If I understood you correctly, there were 34% of the centers that had in-house surgeons? E. DooIin (response): Correct, surgeons were defined as an attending board-certified surgeon. A. Haller (Baltimore, MD): Is this definition that they are actually geographically in the hospital? Or do they meet the criteria of the college that they are within 20 to 30 minutes of arriving in the hospital at the time of the seriously injured patient? The second question is, I am worried about the statistics that you gave us about the in-house surgeon. As I look at your abstract, there were 46% versus 56% who are included in that over-7 age group. And that does not look like a statistically significant difference to me, even though you have got this huge number. The final question is, on the basis of this information you have given us, do you think you would feel comfort-
DOOLIN, BROWNE, AND DISCALA
able recommending that a pediatric trauma center have a surgeon available outside the hospital within 30 minutes, therefore changing our criteria? Or do you believe that on the basis of this information you would want to have a pediatric trauma center designated with an in-house surgeon in the sense of being physically in the hospital? E. Doolin (response): I made a very special effort not to come up with a recommendation, because these are all numbers, and that was my mission. But I have thoughts on it, which are separate. The in-house surgeon was a board-certified general surgeon who stayed within the walls of the trauma center 24 hours a day. And the reason for the phone calls was to be sure there was no misunderstanding about that. An R-5 did not count, a Fellow did not count unless he was board-certified in general surgery. In terms of the statistics, the 10% drop in mortality rate represents a 20% change in the groups. This was done within that cohort by both the Student's t test and the X2 test, again within that cohort. That statistic, if expanded to the entire cohort, would be meaningless. Because, essentially, out of 46,000 patients, it was nine lives saved. And so in terms of the overall system, that statistic would be meaningless. It is only addressing that cohort. And that information can be used however you choose to as long as you recognize that. In terms of the pediatric surgeon or any surgeon being in the house or not in the house, everybody just has to interpret these data accordingly. My feeling is that the role of the in-house surgeon is so limited by these data that it really does not serve a purpose, although the purpose it does serve is in the dozen kids who got shot in the heart. But outside of that, they do not serve a purpose. You could you argue that the in-house attending surgeon will save some money because of the efficiency of the mildly injured patients. And I kind of like that idea. But the pediatric ED physician and the separate pediatric ED have the same effect, and you do not have to take the 400 or 500 pediatric surgeons and stick them in the house just to get a Glasgow of 13 out of the emergency room. Baltimore clearly is different than Children's Memorial of Chicago, which is four miles from Northwestern, which is different from Washington DC, which although four miles from GW is 400 yards from the Hospital Center. We all have different resources, and to compare them just as categories is not as effective as comparing what they offer on an individual basis.