Professional Documents
Culture Documents
E37 Full
E37 Full
E37 Full
e38 SAITO et al
Downloaded from http://pediatrics.aappublications.org/ by guest on November 15, 2017
ARTICLE
and independent variables included appendectomy for acute abdominal hospital. The 2 groups did not differ in
age, gender, BMI percentile, race/ pain during the study interval. Char- terms of gender, BMI percentile, per-
ethnicity, primary insurance type, du- acteristics of the cohort are summa- centage of neutrophils, and neutrophil
ration of abdominal symptoms, WBC rized in Table 1. Almost all children count. Also, final diagnosis distribution
count, neutrophil count, and percent- underwent laparoscopic appendec- of subjects referred from community
age of neutrophils. Univariate analyses tomy (n = 396, 93.6%), and 9 (2.3% of hospitals did not differ from those
included x 2, Fisher exact, t test, Mann- laparoscopic procedures) underwent primarily evaluated at the children’s
Whitney U test, and multinomial logistic conversion to an open approach. The hospital.
regression. Multivariable analysis was proportion of negative appendectomy Only 28 (6.6%) subjects underwent an
performed with multinomial logistic was low (n = 31, 7.3%). Perforated ap- operation without preoperative imag-
regression to control for confounders pendicitis was found in 23.6% (n = 100) ing, whereas 64 (15.1%) had both CT
and referral bias. Independent varia- (Table 1). Sixteen children diagnosed scans and ultrasound (Table 2). Uni-
bles included in the model had asso- with perforated appendicitis were variate analysis of preoperative imag-
ciation with the dependent variable treated with intravenous antibiotics ing utilization by multinomial logistic
(P , .20) in univariate analysis and/or followed by delayed (interval) appen- regression demonstrated significant
were potential confounders. Level of dectomy (3.8% of entire cohort, 16.0% differences by the site of initial evalu-
significance was set at P , .05. of perforated appendicitis subjects). ation, gender, race/ethnicity, BMI per-
Imaging accuracy was assessed by Subjects who were initially evaluated at centile, symptom duration, and WBC
comparing CT and ultrasound impres- a community hospital were younger count. Of note, among subjects initially
sion regarding the appendiceal ap- (10.7 6 3.6 vs 11.8 6 3.7, mean 6 SD, P = evaluated at community hospitals, 10
pearance(normal/indeterminate, acute, .003), more frequently white (P , .001), underwent CT scan and 30 had ultra-
complicated/perforated) to final di- and insured with Medicaid/government sound at the children’s hospital. Among
agnosis (normal, acute/gangrenous source (P = .045) than those evaluated subjects who were initially evaluated at
appendicitis, or perforated appendici- initially at the children’s hospital. There the children’s hospital, the primary
tis) as the reference standard. In- were trends toward shorter duration physician ordered imaging studies (17
determinate studies were included with of abdominal symptoms (1.7 6 1.4 vs CT scans, 2 ultrasounds) that were
normal findings (ie, not appendicitis). 2.0 6 1.9 days, P = .10) and higher performed at another facility.
The Cohen weighted k statistic was
initial total WBC count (15.8 6 4.9 Multivariable analysis of preoperative
determined for the cohort overall
vs 14.8 6 5.4, P = .051) in children ini- imaging revealed significant differences
and by subgroups, including location
tially evaluated at a community hospi- by site of initial evaluation (Table 3). Age,
where the imaging study was per-
tal in comparison with the children’s gender, race/ethnicity, BMI percentile,
formed, and subject age, gender, and
weight classification by BMI percentile;
95% confidence intervals (CIs) were TABLE 1 Cohort Demographics Overall and by Site of Initial Evaluation for Acute Abdominal Pain
calculated by using normal approxima- Variable Overall, n = 423 Community, n = 218 Children’s, n = 205 P
tion or exact method where appropri- Age, y 11.25 6 3.70 10.74 6 3.61 11.80 6 3.73 .003
Gender
ate. Sensitivities for any appendicitis
Female, n (%) 170 (40.2) 88 (40.4) 82 (40.0) .94
(acute or perforated) and perforated BMI percentile 61.90 6 30.88 62.19 6 30.39 61.57 6 31.48 .85
appendicitis were calculated. Specific- Race/ethnicity, n (%) ,.001
ity, positive predictive value, and nega- White 345 (81.6) 193 (88.5) 152 (74.1)
African American 56 (13.2) 16 (7.3) 40 (19.5)
tive predictive value were not assessed, Other 22 (5.2) 9 (4.1) 13 (6.3)
because the cohort did not include Insurance, n (%) .045
patients who had general abdominal Medicaid/government 127 (30.0) 77 (35.3) 50 (24.4)
Private 283 (66.9) 134 (61.5) 149 (72.7)
pain who underwent imaging to eval- None/self-pay 13 (3.1) 7 (3.2) 6 (2.9)
uate for appendicitis and no operation. Symptom duration, days 1.86 6 1.68 1.73 6 1.43 2.00 6 1.91 .10
WBC count, K/mm3 15.32 6 5.17 15.80 6 4.90 14.81 6 5.42 .051
Final diagnosis, n (%) .64
RESULTS Normal 31 (7.3) 17 (7.8) 14 (6.8)
Acute 292 (69.0) 146 (67.0) 146 (71.2)
We identified 423 children who un- Perforated 100 (23.6) 55 (25.2) 45 (22.0)
derwent an operation for the pre- Age, symptom duration, BMI percentile, and WBC count are expressed as mean 6 SD. Percentage for gender, race/ethnicity,
operative diagnosis of appendicitis or insurance, and final diagnosis represents column percentage.
e40 SAITO et al
Downloaded from http://pediatrics.aappublications.org/ by guest on November 15, 2017
ARTICLE
TABLE 4 Accuracy of Preoperative CT characteristics that potentially influ- evaluation of adult acute abdominal
Impression Compared With Final enced the likelihood of interfacility pain.24 Finally, practitioners might have
Diagnosis: Agreement (Weighted k)
transfer. In addition, we found that CT greater confidence in CT scans in
CT Scan k 95% CI n
and ultrasound studies performed at comparison with ultrasound; a pre-
Overall 0.54 0.44–0.64 282 community hospitals in comparison vious survey of North American pedi-
Study location
Community 0.48 0.36–0.60 196 with the children’s hospital had di- atric surgeons in 2004 demonstrated
Children’s 0.69 0.54–0.85 86 minished accuracy for diagnosing ap- preference for CT over ultrasound in
Age, y pendicitis. appendicitis evaluation.25 However, our
1–6 0.35 0.05–0.65 30
7–12 0.56 0.40–0.71 104 Variation in diagnostic imaging use for finding of more frequent ultrasound
13–18 0.58 0.44–0.72 148 pediatric appendicitis by initial evalu- use in the children’s hospital may re-
Weight category
ation location might stem from multiple flect conscious avoidance of ionizing
Underweight/normal 0.58 0.46–0.70 169
Overweight 0.60 0.34–0.86 42 factors, such as availability of imaging radiation exposure. Interestingly, sub-
Obese 0.31 0.02–0.60 44 or the perceived need for diagnosis jects who had both CT scan and ultra-
Gender confirmation. First, compared with ul- sound were more likely to be female
Female 0.60 0.47–0.74 120
Male 0.49 0.35–0.63 162 trasound, the ready availability of and to have lower BMI percentile, lon-
CT scans may account for frequent use ger duration of symptoms, and lower
in community hospitals. CT use for pedi- WBC count. Ultrasound may have
the majority underwent preoperative atric abdominal pain evaluation has served to evaluate for gynecologic pa-
imaging. Diagnostic imaging selection markedly increased over the past decade thology in girls. The longer duration of
and accuracy varied with the site of in emergency departments, particularly symptoms might have increased the
initial evaluation. Controlling for fac- nonpediatric-focused departments.15,16 perceived urgency to establish the di-
tors potentially associated with referral By contrast, ultrasound use over time agnosis of appendicitis, although lower
bias and illness severity, the perfor- has remained constant16; decreased or WBC count would not be expected with
mance of preoperative abdominal-pelvic inconsistent availability of emergent ul- advanced or perforated appendicitis.
CT scan was significantly associated trasound within community hospitals Rather, imaging with both CT and ul-
with initial evaluation at community might contribute to this pattern.10,11 trasound might have been obtained in
hospitals, whereas abdominal ultra- Second, concern over diagnostic errors clinically confusing cases, the identity
sound was more likely obtained with might prompt CT use. Appendicitis is of which cannot be discerned in ret-
initial evaluation at the children’s among the leading diagnoses associated rospect. In many instances, CT scans
hospital. Variation in CT use by hos- with pediatric diagnostic errors22 and followed nondiagnostic ultrasound,
pital type has been reported,20,21 and, malpractice claims.23 Low physician as recommended in several previous
here, we extend the observation by risk tolerance among emergency studies.26–28
examining ultrasound in combination medicine physicians has been associ-
Despite frequent use, CT accuracy was
with CT use and adjusting for subject ated with more frequent CT use for
reduced when performed in the com-
munity setting. Although overall CT
TABLE 5 Accuracy of Preoperative CT Impression Compared With Final Diagnosis: Sensitivities for sensitivity for any appendicitis was
Any Appendicitis and Perforated Appendicitis
similar to previous reports,7 CT scans
CT Scan Any Appendicitis Perforated Appendicitis
performed at the children’s hospital
Sensitivity P Sensitivity P were somewhat more sensitive than
Overall 249/262 95.0% 42/73 57.5% at referring institutions. For perforated
Study location .07 .045
Community 169/181 93.4% 24/49 49.0%
appendicitis, CT studies from the
Children’s 80/81 98.8% 18/24 75.0% children’s hospital had significantly
Age, y .42 .77 higher sensitivity. One potential reason
1–6 26/28 92.9% 6/11 54.5%
7–12 95/98 96.9% 16/30 53.3%
for diminished accuracy is that multi-
13–18 128/136 94.1% 20/32 62.5% detector CT, which is used at the child-
Weight category .23 .17 ren’s hospital, might be less available
Underweight/normal 145/155 93.5% 30/46 65.2%
at referring community hospitals. Multi-
Overweight/obese 80/82 97.6% 9/20 45.0%
Gender .78 .35 detector CT offers the advantages
Female 102/108 94.4% 20/31 64.5% of improved resolution through thinner
Male 147/154 95.5% 22/42 52.4%
sections and coronal reconstructions
e42 SAITO et al
Downloaded from http://pediatrics.aappublications.org/ by guest on November 15, 2017
ARTICLE
such as patient age, gender, body habitus, subjects were not recorded with suffi- within subgroups precluded multivari-
symptoms, potential alternate diagnoses, cient consistency to permit a detailed able analysis of CT and ultrasound
accuracy of imaging modality for patient analysis of imaging utilization with accuracy.
subtype, and specific hospital resources. regard to clinical presentation. Conse-
The value of diagnostic confirmation in quently, the few subjects who did not
avoiding unnecessary interfacility trans- have imaging during initial evaluation CONCLUSIONS
fer, hospital admission, operations, and may have had more obvious clinical The near universal use of preoperative
treatment delays must be balanced evidence of appendicitis. Nonetheless, imaging was associated with a low
against the harm of radiation exposure the high imaging utilization implies that proportion of negative appendectomy in
from CT, and costs to maintain ultrasound at least some CT scans and ultrasounds children who underwent operation for
technical proficiency and to provide pe- were confirmatory rather than essen- presumed appendicitis. Preoperative CT
diatric expertise. tial. The selection criteria for the cohort use was significantly higher in children
The retrospective and single-center were chosen to capture negative ap- initially evaluated at community hospi-
study structure presents several limi- pendectomies; despite this, the pro- tals in comparison with the children’s
tations. Additional similar analyses in portion of operations performed for a hospital, whereas ultrasound use was
other sites will ascertain the generaliz- normal appendix may be underesti- significantly lower. Potential targets to
ability of our findings. We cannot address mated if the appendix was not removed. streamline the evaluation for pediatric
what specific impact imaging had in the Whether community physiciansobtain appendicitis include algorithm de-
evaluation of children with possible ap- imaging in children with suspected velopment with broad validity to de-
pendicitis; the value of normal imaging in appendicitis routinely or selectively crease reliance on preoperative imaging
preventing an unnecessary operation or to confirm the diagnosis before in- and radiation exposure while avoid-
hospital transfer could not be assessed terfacility transfer for operative care ing unnecessary hospital transfers,
with this cohort. The initial symptoms cannot be determined from this study. admissions, operations, and missed
and physical examination findings of Finally, the limited number of subjects diagnoses.
REFERENCES
1. Kids’ Inpatient Database (KID) 2009. 7. Doria AS, Moineddin R, Kellenberger CJ, survivors exposed in utero or as young
Healthcare Cost and Utilization Project et al. US or CT for diagnosis of appendicitis children. J Natl Cancer Inst. 2008;100(6):
(HCUP), Agency for Healthcare Research in children and adults? A meta-analysis. 428–436
and Quality. Available at: http://hcupnet. Radiology. 2006;241(1):83–94 13. image gently: The Alliance for Radiation
ahrq.gov/. Accessed May 31, 2012 8. Brenner DJ, Hall EJ. Computed tomography Safety in Pediatric Imaging. Available at: www.
2. Becker T, Kharbanda A, Bachur R. Atypical —an increasing source of radiation expo- pedrad.org/associations/5364/ig/. Accessed
clinical features of pediatric appendicitis. sure. N Engl J Med. 2007;357(22):2277–2284 April 28, 2011
Acad Emerg Med. 2007;14(2):124–129 9. Abo A, Shannon M, Taylor G, Bachur R. The 14. Paterson A, Frush DP, Donnelly LF. Helical CT
3. Sung T, Callahan MJ, Taylor GA. Clinical and influence of body mass index on the ac- of the body: are settings adjusted for pe-
imaging mimickers of acute appendicitis in curacy of ultrasound and computed to- diatric patients? AJR Am J Roentgenol.
the pediatric population. AJR Am J Roent- mography in diagnosing appendicitis in
2001;176(2):297–301
genol. 2006;186(1):67–74 children. Pediatr Emerg Care. 2011;27(8):
15. Larson DB, Johnson LW, Schnell BM, Goske
4. Kaiser S, Jorulf H, Söderman E, Frenckner 731–736
MJ, Salisbury SR, Forman HP. Rising use of
B. Impact of radiologic imaging on the 10. Moore CL, Molina AA, Lin H. Ultrasonogra-
CT in child visits to the emergency de-
surgical decision-making process in sus- phy in community emergency departments
partment in the United States, 1995–2008.
pected appendicitis in children. Acad in the United States: access to ultraso-
Radiology. 2011;259(3):793–801
Radiol. 2004;11(9):971–979 nography performed by consultants and
status of emergency physician-performed 16. Hryhorczuk AL, Mannix RC, Taylor GA. Pedi-
5. Applegate KE, Sivit CJ, Salvator AE, et al.
Effect of cross-sectional imaging on neg- ultrasonography. Ann Emerg Med. 2006;47 atric abdominal pain: use of imaging in the
ative appendectomy and perforation (2):147–153 emergency department in the United
rates in children. Radiology. 2001;220(1): 11. Burr A, Renaud EJ, Manno M, et al. Glowing States from 1999 to 2007. Radiology. 2012;
103–107 in the dark: time of day as a determinant of 263(3):778–785
6. Warner BW, Kulick RM, Stoops MM, Mehta S, radiographic imaging in the evaluation of 17. Harris PA, Taylor R, Thielke R, Payne J,
Stephan M, Kotagal UR. An evidenced-based abdominal pain in children. J Pediatr Surg. Gonzalez N, Conde JG. Research electronic
clinical pathway for acute appendicitis 2011;46(1):188–191 data capture (REDCap)—a metadata-driven
decreases hospital duration and cost. J 12. Preston DL, Cullings H, Suyama A, et al. methodology and workflow process for
Pediatr Surg. 1998;33(9):1371–1375 Solid cancer incidence in atomic bomb providing translational research informatics
e44 SAITO et al
Downloaded from http://pediatrics.aappublications.org/ by guest on November 15, 2017
Use and Accuracy of Diagnostic Imaging by Hospital Type in Pediatric
Appendicitis
Jacqueline M. Saito, Yan Yan, Thomas W. Evashwick, Brad W. Warner and Phillip I.
Tarr
Pediatrics 2013;131;e37
DOI: 10.1542/peds.2012-1665 originally published online December 24, 2012;
Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/131/1/e37
References This article cites 37 articles, 6 of which you can access for free at:
http://pediatrics.aappublications.org/content/131/1/e37.full#ref-list-1
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Radiology
http://classic.pediatrics.aappublications.org/cgi/collection/radiology_
sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
https://shop.aap.org/licensing-permissions/
Reprints Information about ordering reprints can be found online:
http://classic.pediatrics.aappublications.org/content/reprints
Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since . Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2013 by the American Academy of Pediatrics. All rights reserved. Print
ISSN: .
The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/131/1/e37
Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since . Pediatrics is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2013 by the American Academy of Pediatrics. All rights reserved. Print
ISSN: .