Original Contribution

Practice variation in the management for nontraumatic pediatric patients in the ED

Yu-Che Chang MD ^a,b, Chip-Jin Ng MD ^a, Yu-Chuan Chen PhD ^c,

Jih-Chang Chen MD ^a, David Hung Tsang Yen MD, PhD ^b,d,?

^aDepartment of Emergency Medicine, Chang Gung Memorial Hospital and Chang Gung University College of Medicine,

Tao-Yuan, Taiwan, 333

^bInstitute of Emergency and Critical Care Medicine, National Yang-Ming University, Taipei,Taiwan, 112

^cDepartment of Finance, Chihlee Institution of Technology, Taipei, Taiwan, 220

^dDepartment of Emergency Medicine, Taipei Veterans General Hospital, College of Medicine, National Yang-Ming University, Taiwan, 112

Received 6 November 2008; revised 25 November 2008; accepted 28 November 2008

Abstract

Study objective: To improve the management quality and monitoring for common pediatric illnesses in the general emergency department (ED), we examined the effect of physician specialty training on medical resource use and patient outcomes.

Methods: This was a retrospective cohort review of visits by children less than 18 years to the ED of 2 university-affiliated teaching hospitals. Clinical management by 2 groups (emergency physicians [EPs] and pediatricians each working 168 h/wk) was compared with respect to demographics, ED resource use, Short-term outcome, disposition, direct ED costs for each visit, and frequency of radiographic and laboratory test use. The effects of medical decision making on resource use was assessed by comparing costs of radiographic studies, laboratory studies, and medication.

Results: Between-group differences in mean patient age, sex, and triage category were insignificant. Compared to pediatricians, EPs used radiographic and laboratory studies more frequently (respectively, 10.1% and 3.8% higher frequency and 90.5% and 7.6% higher cost) and less medication (12.5% lower cost). Patients managed by EPs had longer ED length of stay , higher admission rates to general wards, and shorter LOS per hospitalization but similar 72-hour revisit rates, needed more frequent referral for medical reasons, and left more frequently against medical advice.

Conclusion: Emergency physicians spent more time and medical resources and admitted patients at a higher rate. Emergency physicians and pediatricians managed critical patients similarly.

(C) 2010

* Corresponding author. Institute of Emergency and Critical Care Medicine, National Yang-Ming University, Taipei, Taiwan, ROC. Tel.: +886 2 2875 7377; fax: +886 2 2875 7842.

E-mail address: [email protected] (D.H.T. Yen).

Introduction

Emergency physicians (EPs) and pediatricians often have significantly different training and practice settings. Their management of pediatric patients with common

0735-6757/$ – see front matter (C) 2010 doi:10.1016/j.ajem.2008.11.021

illnesses can be different [1-3]. These differences affect emergency department (ED) length of stay (LOS), ED cost, use of medical resources, and patient outcomes [4]. Previous studies examining unique diseases or diagnoses have found variations in clinical practice between EPs and pediatricians in pediatric EDs [1-3,5]. However, few articles mention the differences in the management of emergency pediatric patients between EPs and pediatricians on the basis of extensive evaluation of demographics, variable diagnoses, use of medical resources, dispositions, and short-term outcome in the ED. Most nontrauma pediatric patients in Taiwan are evaluated in a separate part of the general ED. In this setting, 2 specialists treat pediatric patients as follows: EPs who treat patients of all ages, mainly adults, and pediatricians, who treat children only. Our previous study demonstrated that treatment by pediatric residents and attending physicians in the ED resulted in similar short-term outcomes, but pediatric residents spent more time and used more medical resources [6]. The role of the EP in providing Emergency medical care to pediatric patients is evaluated in another prior article [7]. In this study, we investigated the effect of specialty training (emergency medicine vs pediatrics) on medical resource use, short-term outcomes, and dispositions stemming from the management of nontraumatic common pediatric ill- nesses in a general ED to find ways to improve the quality of monitoring and care.

Materials and methods

Study design

This was a retrospective cohort review of visits to the ED by pediatric nontrauma patients. Our study protocol was reviewed by our institutional review board and deemed exempt from the requirement to obtain informed consent.

Study setting and population

The settings of this study were 2 university-affiliated teaching hospitals with EDs staffed by board-certified emergency medicine and pediatric attending physicians. At both sites, physicians provided 24-hour management of pediatric short-term nontrauma patients in a separate location within the general ED. Hospital A had an annual ED volume of 65 000 visits (approximately 20% for pediatric non- trauma). Hospital B had an annual volume of 60 000 visits (roughly 20% for pediatric nontrauma). The pediatric emergency patient census of both EDs could be categorized as high volume (ie, N12 000 visits per year) [8]. The study cohort consisted of children less than 18 years of age who presented to these EDs between October 1, 2005, and September 30, 2006.

Study protocol

During the study period, the number and experience (in years) of enrolled physicians in the 2 groups were similar. Fifteen EPs and 14 pediatricians were enrolled, and all physicians had less than 4 years of attending experience and had well completed 3 years of residency training and 2 to

3 years of Fellowship training before becoming attending physicians. Because the annual volume of ED visits was high in both hospitals A and B, there were 2 attending physicians responsible for ED patient care in each shift of each hospital. One EP was responsible for managing adult patient care in each shift of either hospital A or B, and the other EP was strictly assigned to care for pediatric patient care alternatively with the pediatrician enrolled in the 2 hospitals. For physician assignment for nontraumatic pediatric patient care in hospital A, EPs provided acute care to pediatric patients from 7:00 AM Sunday to 7:00 AM Wednesday (72 hours), and pediatricians provided acute care from 7:00 AM Wednesday to 7:00 AM Sunday (96 hours). The reverse occurred in hospital B. Thus, the 2 groups of physicians enrolled worked an equal amount of time (168 h/wk) for nontraumatic pediatric patient care in the ED. Pediatricians enrolled in this study were strictly limited to work for the care of pediatric patients only in the ED. The clinic and inpatient service were mainly provided by another 7 pediatricians in hospital A and the other 6 pediatricians in hospital B. The tertiary children’s hospital in the same medical system also sustain clinic service in hospital A and B with some physician specialists. However, these physicians were not responsible for pediatric ED service and were deservedly not enrolled in this study. The residency training program in pediatric emergency medicine was conducted only in the tertiary children’s hospital, and physicians enrolled in this study provided acute care directly in the EDs of both hospitals A and B. The ability of physicians to provide next-day care and follow-up was similar between groups.

On arrival at the ED, patients were triaged into 1 of the following 4 categories: emergent, urgent high, urgent low, and nonurgent. When children were extremely ill, they were usually admitted to the pediatric intensive care unit (PICU) and cared for by specialists. Eligibility for critical care was the decision of mainly physicians enrolled in this study or a consensus decision of both ED and PICU physicians, using established PICU admission guidelines. Clinical character- istics of patients admitted to the PICU and use of medical resources were also compared to investigate variations in clinical practice due to the differences in specialty training. Both hospitals had neither a general pediatric surgery facility nor the ability to provide advanced specialty care in the areas of urology, endocrinology, genetics, hematology, and oncology. Thus, patients suspected of needing surgery or specialized management were immediately taken by ambu- lance to a tertiary children’s hospital with an annual pediatric ED volume of 40 000 visits in the same medical system located, respectively, 40 km and 22 km away from hospital A

and hospital B.

Measurements

The data were divided on the basis of physician specialty. Patients were assigned to group 1 if treated by an EP or to group 2 if treated by a pediatric attending physician. We compared demographics, ED resource use, short-term out- come, and short-term disposition. In a centralized computer database, we stored information including the number of ED patient visits, triage category, sex, diagnosis (as indicated by the International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code), direct ED costs for each visit (not charges), and frequency of any radiographic and laboratory test use. To isolate better the effects of medical decision making on resource use, we compared costs of radiographic studies, laboratory studies, and medication as well as direct ED costs between groups. Other costs (registration fee, consultation fee, nursing fee, treatment fee, costs of ED observation, and others) were defined as ancillary costs. Radiographic and laboratory test use was adjusted for the most frequent Diagnostic categories, as well as for other diagnoses. We assessed the influence of the disease (diagnosis) by analyzing similar items in different diagnostic categories for each group. We then identified the 15 most common diagnoses (ICD-9-CM codes) from the computerized data.

Short-term outcomes (duration of ED visit, number of revisits within 72 hours, LOS per hospitalization, and admission rates) were collected from the computer database and chart records. Dispositions were measured as direct PICU or Ward admissions, interhospital referrals, discharge against medical advice, and death. All data were compared between the 2 groups.

Data analysis

Data were analyzed using SPSS 13.0 for Windows (SPSS Inc, Chicago, Ill). The ?2 test was applied for categorical data, and a 2-tailed t test was used for continuous variables. We calculated 95% confidence intervals (CIs) of the differences using standard methods. For data with a nonnormal distribu- tion (ED lengths of stay and costs) [1,4,9], nonparametric post hoc and Mann-Whitney U tests were used. Because the use of diagnostic laboratory and imaging studies was not indicated for most nontrauma visits to the pediatric ED, cost of medical resources is presented as mean (+-SEM) instead of median (interquartile range). A P b .05 indicated a statistically significant difference.

Results

Characteristics of study subjects

During the study period, the medical data for 27 487 patients were collected. There were 629 patients who were

excluded based upon incompleteness of data. Finally, 26 858 patients were enrolled into this study. Of these patients, 13 966 were directly managed by EPs (group 1) and 12 892 were managed by pediatricians (group 2). Table 1 shows their demographics. There were insignificant between-group differences in mean age, sex, and triage category.

Main results

Costs and medical resources

Direct costs per visit were 9.6% higher in group 1 than group 2, averaging 129.6 New Taiwan (NT) dollars more; however, the difference was not statistically significant. The average cost of radiographic studies per patient was 90.5% (84.1 NT dollars) higher in group 1 than group 2, and the cost of laboratory studies per patient was 7.6% (15.9 NT dollars) higher in group 1. However, medication costs were 12.5% (-27.7 NT dollars/patient) lower for group 1, and ancillary costs were 6.9% higher for group 1 (Table 2).

We then examined the use of medical resources after adjustment for most frequent and other diagnosis. The major diagnoses were indicated by 700 ICD-9-CM codes. Fifteen ICD-9-CM codes covered the diagnoses in 69% of the study cases, each code accounting for 1.3% to 13.8% of study cases. Tables 3 and 4 list the 15 most common diagnoses (which were also common diseases) and the other diagnoses (indicated by 685 ICD-9-CM codes) in 31% of cases. After stratification into groups based on most common and least common diagnosis, analysis was performed to see whether there were between-group differences in resource use and outcome.

Use of radiographic studies was 10.1% higher for group 1 (Table 3). Emergency physicians used radiographic studies more frequently in children who were ultimately diagnosed with 1 of the 15 most common illnesses but not in children with noninfectious gastroenteritis and colitis (558.9) (23.5% for group 1 vs 27.8% for group 2). The difference in use of radiographic studies was at least 10% between the 2 groups for cases of infectious colitis, enteritis, and gastroenteritis (009.0); abdominal pain (789.0); upper respiratory tract infection (465.9); and bronchopneumonia (485). The use of

Table 1 Demographics

Group 1 ^a Group 2 ^b P

n = 13 966 n = 12 892

Mean age, y (+-SD) 5.0 (+-3.8) 5.0 (+-3.8) 1 Male patients, % (+-SD) 56.8 (+-1.3) 56.7 (+-1.4) .892 Triage category, n (%) Emergent 3541 (25.4) 3348 (26.0) .252 Urgent high 2129 (15.2) 1918 (14.9) .403 Urgent low 8281 (59.3) 7616 (59.1) .719 Nonurgent 15 (0.11) 10 (0.07) .548

a Patient treated by ED attending physicians. b Patient treated by pediatric attending physicians.

Mean direct ED cost, 1477.7 (+-12.3) 1348.1 (+-10.2) .989 NT dollars (+-SEM) Mean cost of 177.0 (+-5.4) 92.9 (+-3.0) b.001 radiologic studies Mean cost of 225.8 (+-3.9) 209.9 (+-4.0) b.001 laboratory studies Mean cost of 194.6 (+-6.0) 222.3 (+-6.4) b.001 medications Mean ancillary cost 880.3 (+-5.3) 823.0 (+-3.9) b.001

a Patient treated by ED attending physicians. b Patient treated by pediatric attending physicians.

radiographic studies in asthma (493.9) and gastritis and duodenitis (535) cases was 25.1% and 21.5% higher in group 1, respectively. Similarly, the use of radiographic studies for

Table 3 Frequency of radiographic studies according to diagnoses and physician specialty

children in the other diagnostic categories was 14.6% higher in group 1 (Table 3).

Group 1 used laboratory studies 3.8% more frequently. The largest difference in laboratory study use (37.8% for group 1 vs 21.1% for group 2, P = .001) was in cases of infectious colitis, enteritis, and gastroenteritis (009.0). In addition, ED physicians used more laboratory studies in cases of upper respiratory tract infection (465.9), functional Digestive disorders (564), pyrexia of unknown origin (780.6), and other diagnoses, but these differences were smaller than those for radiographic studies (Table 4).

Table 2 Comparison of ED costs between the emergency physician and pediatric physician groups

Group 1 ^a

n = 13 966

Group 2 ^b

n = 12 892

P

Assessment of outcome and disposition

For group 1, ED LOS was longer, with 23.3% more management time and an average of 23.6 more minutes per patient (95% confidence interval [CI], 12.7-34.5 minutes, P b .001), and admission rates were higher (95% CI, 0.09%- 2.58%, P b .001). By contrast, LOS per hospitalization was

0.6 days shorter, but the 72-hour revisit rate was 0.3% higher (95% CI, -1.24 to 0.64, P = .313). The value of this indicator related to prognosis had no statistical significance (Table 5).

Frequency a	Group 1, n = 13 966		Group 2, n = 12 892	Odds ratio (95% CI) b	P
Unadjusted	5047/13 966	(36.1)	3351/12 892 (26.0)	1.61 (1.53-1.70)	b.001
Adjusted by diagnoses (ICD-9-CM code) c
Infectious and parasitic diseases
Infectious colitis, enteritis, and	27/90	(30.0)	121/831 (14.6)	2.52 (1.54-4.11)	b.001
gastroenteritis (009.0)
Herpangina (074.0)	21/215	(9.8)	6/138 (4.3)	2.38 (0.94-6.06)	.067
Nervous system and sense organs
Acute suppurative Otitis media (382.0)	56/309	(18.1)	14/121 (11.6)	1.69 (0.90-3.17)	.111
respiratory system
Acute pharyngitis (462)	75/559	(13.4)	76/1050 (7.2)	1.99 (1.42-2.78)	b.001
Acute tonsillitis (463)	100/509	(19.6)	93/846 (11.0)	1.98 (1.46-2.69)	b.001
Upper respiratory tract infection	334/2031	(16.4)	108/1677 (6.4)	2.86 (2.28-3.59)	b.001
(465.9)
Acute bronchitis (466.0)	80/235	(34.0)	64/219 (29.2)	1.25 (0.84-1.86)	.313
Bronchopneumonia (485)	653/727	(89.8)	492/635 (77.5)	2.57 (1.89-3.48)	b.001
Asthma (493.9)	134/268	(50.0)	82/329 (24.9)	3.01 (2.13-4.26)	b.001
Digestive system
Gastritis and duodenitis (535)	69/155	(44.5)	73/317 (23.0)	2.68 (1.78-4.04)	b.001
Other and unspecific noninfectious	363/1547	(23.5)	204/733 (27.8)	0.80 (0.65-0.97)	.026
gastroenteritis and colitis (558.9)
Functional digestive disorders (564)	232/327	(70.9)	68/109 (62.4)	1.47 (0.93-2.32)	.097
Genitourinary system
Urinary tract infection (599.0)	47/187	(25.1)	32/178 (18.0)	1.53 (0.92-2.54)	.100
Symptoms, sign, and ill-defined conditions
Pyrexia of unknown origin (780.6)	388/1414	(27.4)	201/1098 (18.3)	1.69 (1.39-2.05)	b.001
Abdominal pain (789.0)	548/609	(90.0)	805/1038 (77.6)	2.60 (1.92-3.52)	b.001
Other diagnoses	1920/4784	(40.1)	912/3573 (25.5)	1.96 (1.78-2.15)	b.001
a Data are given as number of patients receiving image studies/number of patients in each group (%). b Data in parentheses are 95% CIs of the odds ratio. c Data are given as number of patients receiving image studies/number of patients in each diagnosis category (%). All diagnoses were coded using the ICD-9-CM. P b .05, statistically significant.

Frequency a	Group 1, n = 13 966	Group 2, n = 12 892		Odds ratio (95% CI) b	P
Unadjusted	3933/13 966 (28.2)	3140/12 892	(24.4)	1.22 (1.15-1.29)	b.001
Adjusted by diagnoses (ICD-9-CM code) c
Infectious and parasitic diseases
Infectious colitis, enteritis, and gastroenteritis (009.0)	34/90 (37.8)	175/831	(21.1)	2.28 (1.44-3.60)	.001
Herpangina (074.0)	61/215 (28.4)	30/138	(21.7)	1.43 (0.86-2.36)	.073
Nervous system and sense organs
Acute suppurative otitis media (382.0)	41/309 (13.3)	8/121	(6.6)	2.16 (0.98-4.76)	.062
Respiratory system
Acute pharyngitis (462)	73/559 (13.1)	153/1050	(14.6)	0.88 (0.65-1.19)	.451
Acute tonsillitis (463)	163/509 (32.0)	223/846	(26.4)	1.32 (1.03-1.68)	.030
Upper respiratory tract infection (465.9)	184/2031 (9.1)	119/1677	(7.1)	1.30 (1.03-1.66)	.030
Acute bronchitis (466.0)	19/235 (8.1)	22/219	(10.0)	0.79 (0.41-1.50)	.514
Bronchopneumonia (485)	284/727 (39.1)	244/635	(38.4)	1.03 (0.83-1.28)	.824
Asthma (493.9)	44/268 (16.4)	51/329	(15.5)	1.07 (0.69-1.66)	.822
Digestive system
Gastritis and duodenitis (535)	23/155 (14.8)	49/317	(15.5)	0.95 (0.56-1.63)	.893
Other and unspecific noninfectious gastroenteritis	451/1547 (29.2)	201/733	(27.4)	1.09 (0.90-1.33)	.399
and colitis (558.9)
Functional digestive disorders (564)	38/327 (11.6)	4/109	(3.7)	3.45 (1.20-9.91)	.014
Genitourinary system
Urinary tract infection (599.0)	173/187 (92.5)	165/178	(92.7)	0.97 (0.44-2.13)	1.000
Symptoms, signs, and ill-defined conditions
Pyrexia of unknown origin (780.6)	668/1414 (47.2)	452/1098	(41.2)	1.28 (1.09-1.50)	.002
Abdominal pain (789.0)	194/609 (31.9)	285/1038	(27.5)	1.24 (0.99-1.54)	.064
Other diagnoses	1483/4784 (31.0)	959/3573	(26.8)	1.23 (1.11-1.35)	b.001
a Data are given as number of patients receiving laboratory studies/number of patients in each group (%). b Data in parentheses are 95% CIs of the odds ratio. c Data are given as number of patients receiving laboratory studies/number of patients in each diagnosis (%). All diagnoses were coded using the ICD-9-CM. P b .05, statistically significant.

Analysis of our disposition data revealed that the percentage of patients discharged after treatment was higher for group 2 (88.1% vs 89.8%, P b .001), with more patients hospitalized after treatment by EPs (11.3% vs 9.9%, P b

Table 4 Frequency of laboratory studies according to diagnoses and physician specialty

Table 5 Comparison of short-term outcomes and dispositions in each group

Group 1, a n = 13 966		Group 2, b n = 12 892	P
Duration of ED visit,	124.8 (+-3.2)	101.2 (+-4.6)	b.001
min (+-SEM)
Patients revisiting	6.7 (+-1.1)	6.4 (+-1.1)	.313
b72 h, % (+-SD)
LOS of hospitalization,	5.1 (+-2.9)	5.7 (+-5.4)	.019
d (+-SD)
Patients admitted c	1574 (11.3)	1282 (9.9)	b.001
Ward c	1462 (10.5)	1167 (9.1)	b.001
ICU c	112 (0.80)	115 (0.89)	.424
Medical referral c	51 (0.43)	13 (0.15)	b.001
Nonmedical referral c	9 (0.06)	6 (0.05)	.612
Leave against	31 (0.22)	13 (0.10)	.016
medical advice c
Death c	1 (0.01)	2 (0.02)	.611
ICU indicates intensive care unit. a Patient treated by ED attending physicians. b Patient treated by pediatric attending physicians. c Values are presented as n (%).

.001). There was no significant between-group difference in percentage of patients admitted to the PICU, but the percentage of patients admitted to general wards was 1.4% higher in group 1 (Table 5). More patients needED referral for medical reasons in group 1. Before referral, 51 patients in

group 1 were initially suspected of needing surgical or

Specialized care. After referral, diseases were confirmed as

13 appendicitis, 2 intussusception, 2 ovarian cysts, 4 nonspecific abdominal pain, 1 mesenteric lymphadenitis, 14 enterocolitis, 3 constipation, and 12 needing specialists. Group 2 had 13 patients initially suspected of needing surgical or specialized care before referral. After referral, the confirmed diagnoses were 7 appendicitis, 3 enterocolitis, 1 constipation, and 2 needing specialists. There was no between-group difference in nonmedical referrals, all being requested by the patients’ families (Table 5). Patients

discharged against medical advice were 31 (0.22%) for group 1 and 13 (0.10%) for group 2.

Characteristic Group 1, a Group 2, b P n = 112 n = 115

Male patients, n (%) 72 (64.3) 63 (54.8) .176 Mean age, y (+-SD) 1.1 (+-2.0) 1.9 (+-3.3) .157 Duration of ED visit, 70.4 (+-10.4) 111.9 (+-46.4) .254 min (+-SEM) LOS of hospitalization, 7.5 (+-6.4) 10.3 (+-13.4) .060 d (+-SD) Mean direct ED cost, 3055.2 (+-167.3) 2858.9 (+-100.7) .800 NT dollars (+-SEM) Mean cost of 324.8 (+-64.1) 170.8 (+-17.5) .003 radiologic studies Mean cost of 790.0 (+-50.4) 991.8 (+-59.8) .008 laboratory studies Mean cost of 127.7 (+-12.2) 130.1 (+-16.9) .001 medications Mean ancillary cost 1812.7 (+-127.0) 1566.2 (+-62.5) .112 Use of diagnostic 75 (67.0) 55 (47.8) .005 radiologic studies, n (%) Use of diagnostic 89 (79.5) 97 (84.3) .390 laboratory tests, n (%)

a Patient treated by ED attending physicians. b Patient treated by pediatric attending physicians.

Clinical characteristics of patients admitted to PICU

Table 6 Clinical characteristics of patients admitted to PICU

Further analysis of data from patients admitted to the PICU (Table 6) found insignificant between-group differ- ences in mean age and sex. Although LOS per hospitaliza- tion was shorter for patients treated by EPs, the average time expended for management was 41.5 minutes less (37.1% reduction) by EPs than by pediatricians. However, these differences were not statistically significant. The average ED cost per patient was similar between the 2 groups. It is worth noting that EPs used more diagnostic radiographic studies but less diagnostic laboratory tests and prescribed fewer medications (Table 6).

Discussion

Overview

It was reported that physicians in the ED may manage pediatric patients differently because of their training [5,10,11]. Moreover, these differences could influence medical resource use [1] and indirectly influence ED LOS [1,12], ED costs [1,2], and even outcomes [2,12,13].

Our study demonstrated that pediatric patients in the ED treated by EPs (when compared to those treated by pediatric physicians) had comparable outcomes, but they also had

longer ED LOS, used more resources more frequently, were admitted to general wards at a higher rate, and received more referrals. However, these differences in the critical care of pediatric patients between EPs and pediatric physicians were small. Nevertheless, an educational program or policy should be implemented to improve the clinical practice of EPs in dealing with pediatric emergencies in the future.

ED length of stay and cost

Emergency department LOS is a measure of efficiency of Delivery of care and is an important determinant of patient satisfaction in pediatric emergency medicine [14,15]. The Average ED LOS in clinical practice for pediatric emergency patients is about 2 to 3 hours [2,16]. This ED LOS was also found by survey studies of both parents and physicians [11,17]. The average time for both groups in our study was similar to that in previous studies [1,2]. Emergency physicians on average took 23.6 minutes more per patient (23.3%) than pediatricians to manage pediatric emergencies. The increase in LOS may be attributed to the greater use of radiographic and laboratory studies by EPs than pediatricians (Tables 3 and 4). This finding might demonstrate that EPs have less confidence in clinical diagnostic skills as the basis for decision making in these acute Pediatric cases.

Parent satisfaction is difficult to achieve and assess because of the complex interplay between parental expecta- tions, perceptions, and reality. Direct ED cost was used to assess parental expectations of care received in a tertiary care pediatric ED [17]. It included the cost of medications [17], radiographic studies [17], laboratory studies [9,17], and ancillary costs. Our earlier report on the acute pediatric care in the ED uncovered a tendency toward more expensive evaluation and longer ED LOS by ED residents compared to attending ED physicians [6]. In this study, compared to pediatricians, EPs used diagnostic resources (ie, radiographic and laboratory studies) more frequently but used treatment resources (ie, medication) less frequently. This difference in clinical practice might be attributable to differences in training between EPs and pediatricians, and further study in the future is undoubtedly needed.

Medical resources

Overall, the frequency of radiographic study use was 10.1% higher in group 1 than in group 2. The between-group difference was especially large for children ultimately diagnosed with asthma (493.9) and gastritis and duodenitis

(535). Previous articles have demonstrated that the routine use of chest radiography is not necessary for children with asthma or acute lower respiratory tract infection unless specific clinical symptoms or physical examination findings are present [18-21]. Emergency physicians could be expected to reduce the number of radiographic studies in asthma cases by spending more time correlating clinical symptoms with

examination findings. In this study, EPs also used more radiographic studies in cases of gastritis and duodenitis (535) and other gastrointestinal system diseases such as infectious colitis, enteritis, and gastroenteritis (009.0) and abdominal pain (789.0). However, diagnoses of these diseases actually rely more on clinical course than on findings of diagnostic radiographic studies. For other diagnoses, the between-group difference was still 14.6%. Thus, establishing guidelines for radiographic studies is an important issue for pediatric emergency management in the ED.

About laboratory study, the between-group difference in the frequency of use was smaller for laboratory studies than radiographic studies. This study found that EPs tend to use laboratory studies in cases of common diseases such as infectious colitis, enteritis, and gastroenteritis (009.0); upper respiratory tract infection (465.9); functional digestive disorders (564); pyrexia of unknown origin (780.6); and other diagnoses. Notably, pyrexia of unknown origin (780.6) and upper respiratory tract infection (465.9) were high- volume and also the commonest diseases, representing 9.4% of and 13.8% of our study cases, respectively. McGillivray et al [3] found that attending physicians with clinical experience for 10 years, in particular, prescribed fewer laboratory studies when diagnosing pediatric febrile emer- gencies than did their junior colleagues. This finding implies that clinical experience can reduce dependency on diagnostic tools. On the contrary, a study regarding management of acute gastroenteritis (ICD-9-CM code 009.1 or 558.9) by pediatric EPs showed that Disease progression, admission rate, and family satisfaction were similar regardless of the level of physicians’ dependence on laboratory tools; however, patients treated by physicians with less dependence on laboratory tools had higher ED revisit rates [5]. Although no difference in clinical experience years was noted between physicians in both groups in this study, EPs (compared to pediatricians) spent less time managing common pediatric illnesses because of their workload. Our analysis found that the diagnoses-directed difference was largest (16.7%) for children with infectious colitis, enteritis, and gastroenteritis (009.0). Emergency physicians treating these patients had a higher level of dependence on laboratory studies, and these patients had higher ED revisit rates. Monitoring laboratory resource use of EPs managing this disease might improve quality control of these resources by the pediatric ED.

Assessment of outcome and disposition

To control and improve the quality of clinical care for pediatric emergency patients, short-term outcome, including rate of hospital admissions, 72-hour revisit rate, referral rate, and mortality, should be assessed.

In this study, the overall admission rate was 10.6%, which is similar to previously published rates (3.3%-20%) for pediatric emergency patients [8,22,23]. Johnson et al [4] found that physician specialty affects the admission rate of patients with bronchiolitis. Between-study differences might

be due to differences in the criteria used for patient admission. In our study, EPs (compared to pediatricians) admitted patients to general wards at a 1.42% higher rate, and their patients had 0.6 days shorter hospitalization. Although these differences are small, it is notable that LOS per hospitaliza- tion was also impacted by physician specialty. The absence of a difference in admission rate to the PICU (0.80% vs 0.89%, P = .424) might suggest that decision making regarding PICU admission was similar for EPs and pediatricians.

Without regard to its definition, pediatric emergency revisit rate ranges from 3.0% to 13.3% in previous studies [24-27]. A large study showed that the 72-hour revisit rate for pediatric emergency patients in Toronto was 5.2% [24]. Our 72-hour revisit rates were similar to those in previous studies [24-27]. After adjusting for the 15 most common diseases, there was a slight between-group difference in 72-hour revisit rate only for children with infectious colitis, enteritis, and gastroenteritis (009.0) (14.4% vs 7.8%, P = .044). Monitoring 72-hour revisit rate of patients with this ICD code might be a simple immediate way to check quality control.

A prior study has shown that years in practice and specialty can influence physicians’ decisions to make referrals [28]. Compared with pediatricians, EPs had a lower percentage of patients in need of surgery or a specialist, but the difference was not statistically significant. In Taiwan, EP training only requires a 4 to 6-month pediatric rotation, similar to training programs in the United States. Although there was little difference in attending staff experience between EP and pediatricians in this study, it was clear that EPs spend less time managing common illnesses of children because of their workload.

Table 5 shows that more group 1 patients left against medical advice. The most common reason for discharge in both groups was the need to wait for ward beds and cancellation of admission due to clinical improvement. Although their EPs suggested a longer stay for observation, these patients chose discharge after arranging for outpatient follow-up. Compared to pediatric physicians, EPs appeared more likely to suggest referral and observation in the ED.

From these study results, it is reasonable to infer an association of physicians’ training with workload. Some studies have examined the influence of malpractice fear or the concept of defensive medicine on the clinical practice of EPs [29,30] and found that EPs might use more medical resources and admit patients more frequently to avoid misdiagnosis, avoid inappropriate discharge, and reduce liability or malpractice resulting in better patient outcomes [30]. However, it is undeniable that pediatric emergency patients receive a great deal of their care in the general ED [8]. It is uncertain whether doctors who admit fewer patients and use fewer tests are better doctors, and it is not known which philosophy is correct or provides the greatest good to patients. However, because medical resources are limited and pediatric ED patient care is important [31], we need to consider not only the cost-effectiveness of medical resources but also ED quality assurance/improvement practices [8].

This study showed only a small difference in ED practice between EPs and pediatric physicians. Moreover, our findings support the previous findings demonstrating the important role that EPs play in delivering pediatric emergency care [7].

Limitations

Our study is retrospective, and the data were collected from a computer database and chart review. Although we made every effort to remain objective, possible errors may have been introduced.

The ICD-9-CM code indicates a specific disease but not the chief complaint or severity of the disease. Compared to previous studies investigating only one disease, we enrolled a sufficient number of cases, and all diseases were analyzed. Moreover, the triage categories were not significantly different between the 2 groups.

This study was conducted in 2 hospitals with similar facilities for managing general pediatric surgery and providing advanced care in some specialties. Furthermore, differences in disease types might be due to regional differences resulting in possible selection bias. However, both hospitals are teaching hospitals in northern Taiwan, and the study samples were quite large. Therefore, selection bias should be limited.

Finally, the reasons for disposition could only be determined by a review of the medical records. Conse- quently, we did not know the physician’s thinking at the time. The medical records are unclear as to whether some patients were discharged against medical advice or left on their own.

Conclusions

Compared to pediatricians, EPs spent more time, used more medical resources to treat pediatric patients in the ED, and admitted patients at a higher rate. Revisit rates within 72 hours were similar for patients in both groups. Emergency physicians and pediatricians managed critical patients similarly. More detailed study on the investigation of factors influencing the clinical care of pediatric emergency patients by physicians with different specialties should be imple- mented in the future.

References

1. Hampers LC, Faries SG. Practice variation in the emergency management of croup. Pediatrics 2002;109:505-8.
2. Isaacman DJ, Kaminer K, Veligeti H, Jones M, Davis P, Mason JD. Comparative practice of emergency medicine physicians and pediatric emergency physicians managing fever in young children. Pediatrics 2001;108:354-8.
3. McGillivray DL, Roberts-Brauer R, Kramer MS. Diagnostic test ordering in the evaluation of febrile children. Physician and Environmental factors. Am J Dis Child 1993;147:870-4.
4. Johnson DW, Adair C, Brant R, Holmwood J, Mitchell I. Differences in admission rates of children with bronchiolitis by pediatric and general emergency departments. Pediatrics 2002;110:e49.
5. Powell EC, Hampers LC. Physician variation in test ordering in the management of gastroenteritis in children. Arch Pediatr Adolesc Med 2003;157:978-83.
6. Chang YC, Lo HC, Tzeng YM, Yen DH, Jeng MJ, Huang CI. Comparative clinical practice of residents and attending physicians who care pediatric patients in the emergency department. Pediatr Emerg Care 2008;24:364-9.
7. Gausche-Hill M, Johnson RW, Warden CR, Brennan JA. American College of Emergency Physicians’ Pediatric Emergency Committee. The role of the emergency physician in emergency medical services for children. Ann Emerg Med 2003;42:206-15.
8. Gausche M, Rutherford M, Lewis RL. Emergency department quality assurance/improvement practices for the pediatric patient. Ann Emerg Med 1995;25:804-8.
9. McNulty JE, Hampers LC, Krug SE. Primary care and emergency department decision making. Arch Pediatr Adolesc Med 2001;155: 1266-70.
10. Cimpello LB, Khine H, Avner JR. Practice patterns of pediatric versus general emergency physicians for pain management of fractures in pediatric patients. Pediatr Emerg Care 2004;20:228-32.
11. Schweich PJ, Smith KM, Dowd MD, Walkley EI. Pediatric emergency medicine practice patterns: a comparison of pediatric and general emergency physicians. Pediatr Emerg Care 1998;14:89-94.
12. Seow VK, Lin AC, Lin IY, Chen CC, Chen KC, Wang TL, et al. Comparing different patterns for managing febrile children in the ED between emergency and pediatric physicians: impact on patient outcome. Am J Emerg Med 2007;25:1004-8.
13. Bourgeois FT, Shannon MW. Emergency care for children in pediatric and general emergency departments. Pediatr Emerg Care 2007;23: 94-102.
14. Arendt KW, Sadosty AT, Weaver AL, Brent CR, Boie ET. The left- without-being-seen patients: what would keep them from leaving. Ann Emerg Med 2003;42:317-23.
15. Magaret ND, Clark TA, Warden CR, Magnusson AR, Hedges JR. Patient satisfaction in the emergency department–a survey of pediatric patients and their parents. Acad Emerg Med 2002;9:1379-88.
16. Gorelick MH, Yen K, Yun HJ. The effect of in-room registration on emergency department length of stay. Ann Emerg Med 2005;45:128-33.
17. Harris N, Hostetler MA. Parental expectations of care and charges in a tertiary care pediatric ED. Am J Emerg Med 2002;20:601-3.
18. Bordley WC, Viswanathan M, King VJ, Sutton SF, Jackman AM, Sterling L, et al. Diagnosis and testing in bronchiolitis: a systematic review. Arch Pediatr Adolesc Med 2004;158:119-26.
19. Swingler GH, Hussey GD, Zwarenstein M. Randomised controlled trial of clinical outcome after chest radiograph in ambulatory acute lower-respiratory infection in children. Lancet 1998;351:404-8.
20. Gershel JC, Goldman HS, Stein RE, Shelov SP, Ziprkowski M. The usefulness of chest radiographs in first asthma attacks. N Engl J Med 1983;309:336-9.
21. Rushton AR. The role of the chest radiograph in the management of childhood asthma. Clin Pediatr (Phila) 1982;21:325-8.
22. Losek JD. Characteristics, workload, and Job satisfaction of attending physicians from pediatric emergency medicine fellowship programs. Pediatric Emergency Medicine Collaborative Research Committee. Pediatr Emerg Care 1994;10:256-9.
23. Pershad J, Whitley M, Herman M. Visit-level acuity and resource- based relative value unit utilization in a pediatric emergency department. Pediatric Emergency Care 2006;22:423-5.
24. Goldman RD, Ong M, Macpherson A. UnschedulED return visits to the pediatric emergency department-one-year experience. Pediatr Emerg Care 2006;22:545-9.
25. Zimmerman DR, McCarten-Gibbs KA, DeNoble DH, et al. Repeat pediatric visits to a general emergency department. Ann Emerg Med 1996;28:467-73.
26. Mintegui Raso S, Benito Fernandez J, Vazquez Ronco MA, Ortiz Andres A, Capape Zache S, Fernandez Landaluce A. Children’s unschedulED return visits to an emergency department. An Esp Pediatr 2000;52:542-7.
27. Alessandrini EA, Lavelle JM, Grenfell SM, Jacobstein CR, Shaw KN. Return visits to a pediatric emergency department. Pediatr Emerg Care 2004;20:166-71.
28. Franks P, Williams GC, Zwanziger J, Mooney C, Sorbero M. Why do physicians vary so widely in their referral rates? J Gen Intern Med 2000;15:163-8.
29. Rodriguez RM, Anglin D, Hankin A, Hayden SR, Phelps M, McCollough L, et al. A longitudinal study of emergency medicine residents’ malpractice fear and defensive medicine. Acad Emerg Med 2007;14:569-73.
30. Katz DA, Williams GC, Brown RL, Aufderheide TP, Bogner M, Rahko PS, et al. Emergency physicians’ fear of malpractice in evaluating patients with possible Acute cardiac ischemia. Ann Emerg Med 2005;46:525-33.
31. Wiley 2nd JF, Fuchs S, Brotherton SE, et al. A comparison of pediatric emergency medicine and general emergency medicine physicians’ practice patterns: results from the Future of Pediatric Education II Survey of Sections Project. Pediatr Emerg Care 2002;18: 153-8.