Appendiceal rupture rate decreased 7% during the 5 years studied from
42% to 35% (P<.001). The decline in the negative
appendectomy rate was not statistically significant.
Rates shown are the mean rates during the 5 years studied. There was
no correlation between negative appendectomy rate and appendiceal rupture
rate (r = 0.08, P = .65).
The regression line was created using a least squares fit model.
Ponsky TA, Huang ZJ, Kittle K, Eichelberger MR, Gilbert JC, Brody F, Newman KD. Hospital- and Patient-Level Characteristics and the Risk of Appendiceal Rupture and Negative Appendectomy in Children. JAMA. 2004;292(16):1977-1982. doi:10.1001/jama.292.16.1977
Author Affiliations: Departments of Surgery
(Drs Ponsky, Eichelberger, Gilbert, Brody, and Newman) and Biostatistics (Dr
Huang), Children’s National Medical Center and George Washington University
Medical Center, Washington, DC; and Child Health Corporation of America, Overland
Park, Kan (Mr Kittle).
Context The rates of appendiceal rupture and negative appendectomy in children
remain high despite efforts to reduce them. Both outcomes are used as measures
of hospital quality. Little is known about the factors that influence these
Objective To investigate the association between hospital- and patient-level characteristics
and the rates of appendiceal rupture and negative appendectomy in children.
Design, Setting, and Patients Retrospective review using the Pediatric Health Information System database
containing information on 24 411 appendectomies performed on children aged
5 to 17 years at 36 pediatric hospitals in the United States between 1997
Main Outcome Measures Rates of negative appendectomy and appendiceal rupture; the odds ratio
(OR) of negative appendectomy and appendiceal rupture by hospital, patient
age, race, and health insurance status, and hospital fiscal year and appendectomy
volume. Negative appendectomy rate was defined as the number of patients with
appendectomy but without appendicitis divided by the total number of appendectomies.
Results The median negative appendectomy rate was 3.06% (range, 1%-12%) and
the median appendiceal rupture rate was 35.08% (range, 22%-62%). The adjusted
OR for appendiceal rupture was higher in Asian children (1.66; 95% confidence
interval [CI], 1.24-2.23) and black children (1.13; 95% CI, 1.01-1.30) compared
with white children. Children without health insurance and children with public
insurance had increased odds of appendiceal rupture compared with children
who had private health insurance (adjusted OR, 1.36; 95% CI, 1.22-1.53 for
self-insured; adjusted OR, 1.48; 95% CI, 1.34-1.64 for public insurance).
No correlation existed between negative appendectomy rate and race, health
insurance status, or hospital appendiceal rupture rate. The negative appendectomy
rate improved as the hospital appendectomy volume increased.
Conclusion The rate of appendiceal rupture in school-aged children was associated
with race and health insurance status and not with negative appendectomy rate
and therefore is more likely to be associated with prehospitalization factors
such as access to care, quality of care, and patient or physician education.
The primary adverse outcome of appendicitis is appendiceal rupture.
Patients with an appendiceal rupture at the time of surgical exploration have
as high as a 39% chance of having a postsurgical complication, such as intra-abdominal
abscess, wound infection, and postoperative paralytic ileus, compared with
an approximately 8% chance if the appendix is not perforated.1,2 It
has been assumed that the natural history of appendiceal rupture is within
the control of the hospital or physician and that a high rate of rupture reflects
a failure of medical care. As a result, appendiceal rupture rates have been
proposed as a measure of intrinsic hospital quality. Given the difficulty
of diagnosing appendicitis in both children and adults, the traditional approach
by hospitals to decrease the rupture rate has been to encourage early surgical
exploration. In fact, high rates of negative exploration for appendicitis
have been tolerated to lessen the likelihood of appendiceal rupture and its
attendant complications.1,2 In
essence, one complication (a negative exploration) is encouraged to decrease
the incidence of another complication (appendiceal rupture). Negative hospital
appendectomy rates as high as 26% have been reported.3
Despite efforts by hospitals and physicians and the advent of new diagnostic
techniques such as ultrasonography and computed tomography, the appendiceal
rupture rate remains high among children and ranges from 30% to 74%.4 Recently, some have argued that high rupture rates
may be unrelated to hospital-level care and that delay in diagnosis and treatment
due to inadequate access to health care may instead be the major factor.3- 8
In this study, we examined the patterns of diagnosis and care of children
aged 5 to 17 years with appendicitis at 36 major children's hospitals
to assess the contributions of race, health insurance status, age, sex, and
hospital volume on the appendiceal rupture rate. We also evaluated the correlation
between rupture rate and negative appendectomy rate.
Data in this study were obtained from the Pediatric Health Information
System (PHIS), an administrative database that contains inpatient and selected
outpatient data from 36 not-for-profit, freestanding US pediatric hospitals
that are affiliated with Child Health Corporation of America (Overland Park,
Kan), a business alliance of children’s hospitals. The database was
created in 1992 and quality and accuracy is a joint effort between participating
hospitals, Child Health Corporation of America, and Solucient LLC (Evanston,
Ill). Solucient is a leading source of health care business intelligence that
maintains the nation’s largest health care database composed of more
than 22.6 million discharges annually from 2900 hospitals, representing 77.5%
of all discharges.9 Solucient manages the data
warehouse function for the PHIS database. The number of hospitals participating
in PHIS has grown from 12 to 36 hospitals. Eighty percent of the participating
hospitals provide daily resource use data that is used in comparing clinical
practice. From a demographic perspective, 17 of the 20 major metropolitan
areas in the United States are represented in the PHIS database. The National
Association of Children's Hospitals and Related Institutions (Alexandria,
Va) estimates that there are 50 to 55 freestanding pediatric general acute
care hospitals in the United States. Seventy percent of these hospitals submit
data to the PHIS database.
Information on 24 411 appendectomies performed on children aged 5 to
17 years was abstracted. Although information on infants and young children
to age 4 years was abstracted and is included in the population overview,
this age group was excluded from detailed analysis because of an appendiceal
rupture rate that was significantly higher than older age groups. Inclusion
of these infants and young children would skew the results. All patients were
discharged between January 1, 1997, and June 30, 2002. The PHIS uses all patient-refined
diagnosis related groups to classify patients. The criterion for inclusion
was any child with a principal diagnosis of appendectomy and a principal International Statistical Classification of Diseases, 9th Revision
(ICD-9) procedure code of 47.01 (laparoscopic appendectomy) or 47.09
(other appendectomy). Children undergoing an incidental appendectomy performed
during another abdominal surgical procedure were excluded. Children with appendicitis
who were treated initially by drainage followed by a subsequent admission
for an interval appendectomy were included.
The results were blinded to the identity of the hospitals, which is
consistent with PHIS policies. This study received an exemption from the institutional
review board at Children’s National Medical Center, Washington, DC.
Patients were stratified into 3 groups: appendiceal rupture (ICD-9 codes 540.0 and 540.1), nonruptured appendicitis (ICD-9 codes 540.9, 541, 542), and negative appendectomy (other ICD-9 codes). Rupture rate was
defined as the total number of patients with appendiceal rupture divided by
the number of patients with acute appendicitis. Negative
appendectomy rate was defined as the number of patients with appendectomy
but without appendicitis divided by the total number of appendectomies. Independent
variables included age (5-12 years and 13-17 years), sex, race (white, black,
Asian, or other), and health insurance status (private insurance, self-insured,
or public insurance). These independent variables were the most complete variables
from the PHIS database that characterize the population. Ninety percent of
the children in the public health insurance category had Medicaid insurance;
the other 10% included title 4 and other government-sponsored insurance, such
as Champus or nontraditional Medicaid (Medicaid health maintenance organization).
Other race includes all individuals who were not Asian, black, or white. If
race was not indicated in the database, it was coded as missing. Race was
included as a variable because it has been cited as an issue in many studies
of disparities in quality and access to health care for children. Hospital-level
independent variables were fiscal year (1997-2002), hospital negative appendectomy
rate, and hospital volume (number of appendectomies performed per year). These
hospital-level characteristics were chosen as independent variables because
they were likely to affect patient outcome.
The rates of negative appendectomy and appendiceal rupture were computed
for each age, race, health insurance status, and fiscal year group. Intergroup
differences were tested for significance using the χ2 test
at the individual level. The odds of rupture and negative appendectomy by
hospital appendectomy volume group, age, sex, race, and health insurance status
were then modeled with multivariable logistic regressions. The logistic regressions
were adjusted for interhospital correlations using generalized estimating
equation models in PROC GENMOD of SAS statistical software (version 8.02,
SAS Institute Inc, Cary, NC).10,11 In
the multivariable model for rupture, the negative appendectomy rate was also
adjusted as a hospital characteristic. Rates of radiological procedures were
not used as covariates in the analysis because of difficulty interpreting
the results. A Pearson correlation analysis was performed to assess if any
correlation existed between negative appendectomy rate and rupture rate among
A multivariate analysis was also performed to assess the association
between negative appendectomy rate and appendiceal rupture rate. Because appendicitis
cases from the same hospital cannot be considered independent observations,
generalized estimating equation models in PROC GENMOD were used to adjust
for the effect of case clustering by hospital.12 The P value cutoff for significance for this study was .05.
Of children with appendectomies discharged between January 1, 1997,
and June 30, 2002, 10.6% were infants and children to age 4 years; 64.8%,
5 to 12 years; and 24.6%, 13 to 17 years. Infants and children to age 4 years
had a higher appendiceal rupture rate (70.5%) compared with older children
(37.1%) and were excluded from further analysis. Previous studies have shown
similar rates of rupture in this age group.13 Of
those between ages 5 and 17 years, 61% were male and 39% were female. Seventy-seven
percent were white; 12%, black; 2%, Asian; and 9%, other races. Fifty-three
percent had private health insurance; 39% had public health insurance; and
8% were self-insured.
To assess the representativeness of the study population, we compared
the demographic profile of all appendectomy discharges within the PHIS database
from July 1, 2001, to June 30, 2002, with the demographic profile of all pediatric
appendectomy discharges for 5- to 17-year-olds during the same period in the
Solucient database. Sixty percent of the PHIS appendectomy discharges were
male, 40% were female, and 29% had public insurance. Fifty-eight percent of
the Solucient population was male, 42% was female, and 25% had public insurance.9 Because patient race was not available from the Solucient
database, the racial profile of the PHIS database for all diagnoses from July
1, 2001, to June 30, 2002, was compared with the Solucient 2003 population
projections based on the 2000 US Census. Seventy-two percent of the individuals
in the PHIS database were white; 25%, black; and 2%, Asian. The Solucient
population projections were 71% white; 16%, black; and 3.8%, Asian.14
The groups with the highest frequencies of negative appendectomies were
females (aged 13-17 years), blacks, those with private health insurance, and
those who had appendectomies in 1997 (Table 1). The groups with the highest frequencies of appendiceal ruptures
were children aged 5 to 12 years, Asians, those with public health insurance,
and those with appendectomies in 1997.
Surgical exploration was performed in 81% of the children on the day
of presentation to the hospital and 15% on the next calendar day. Because
the PHIS database includes date but not time of admission, some of these 15%
may actually have been operated on within the first 24 hours of presentation.
The probability of having an appendectomy on the first day of presentation
to the hospital was similar in all health insurance groups: private insurance
(83.4%), self-insured (84.3%), and public insurance (82.3%); and in all race
groups: Asian (84.2%), black (82.1%), and white (83.3%).
Overall, 49% of children in this study underwent a radiological procedure
prior to surgery, including computed tomography, ultrasound, fluoroscopy,
or magnetic resonance imaging. The percentage of children who underwent a
radiological procedure was lower for Asian children (41.0%) compared with
blacks (52.9%) and whites (50.1%; P<.001). The
percentage of children who underwent a radiological procedure prior to surgery
varied by health insurance status: 49.0% were self-insured, 44.1% had private
health insurance, and 47.3% had public insurance (P<.001).
However, a higher frequency of radiological tests was not associated with
a lower rupture rate: 43.8% of children with appendiceal rupture had a radiological
test compared with 35.9% of children with nonruptured appendicitis (P<.001).
The median rupture rate was 35.08% (range, 22%-62%) among the 36 hospitals
studied. The quarterly incidence rate of appendiceal rupture decreased 7%
during the 5 years studied from 42% to 35% (P<.001, Figure 1).
Asian children had a greater chance of having an appendiceal rupture
compared with white children (adjusted odds ratio [AOR], 1.66; 95% confidence
interval [CI], 1.24-2.23) as did black children (AOR, 1.13; 95% CI, 1.01-1.30)
(Table 2). Children with public health
insurance had a greater chance of having an appendiceal ruptue compared with
children with private health insurance (AOR, 1.48; 95% CI, 1.34-1.64), as
did children who were classified as self-insured (AOR, 1.36; 95% CI, 1.22-1.53).
Children aged 5 to 12 years had an AOR of 1.41 (95% CI, 1.30-1.53) for having
an appendiceal rupture compared with children aged 13 to 17 years. There was
no statistical difference in rupture rate between male and female children.
Hospital experience, defined by the volume of appendectomies performed, was
not associated with appendiceal rupture rate (r =
0.03; P = .86) regardless of adjustments for race,
sex, age, and health insurance status.
The median negative appendectomy rate was 3.03% (range, 1%-12%) in the
36 hospitals studied. The incidence of negative appendectomy did not change
during the 5 years of the study (P = .46; Figure 1).
Race and health insurance status had no statistically significant impact
on negative appendectomy rates (Table 3).
Children aged 5 to 12 years had a lower chance of having a negative appendectomy
compared with children aged 13 to 17 years (AOR, 0.80; 95% CI, 0.69-0.94).
Girls had a 58% greater chance of having a negative appendectomy compared
with boys (AOR, 1.58; 95% CI, 1.30-1.92). A significant interaction was found
between age and sex with regard to negative appendectomy rate. A stratified
analysis showed no association between age and negative appendectomy rate
for boys (2.74% negative appendectomy rate in the 13- to 17-year age group
compared with 2.39% in the 5- to 12-year age group; P =
.22). For girls, however, the older group had a higher negative appendectomy
rate (5.06% compared with 3.58%; P = .001).
The principal diagnoses for the negative appendectomy group were right
lower quadrant abdominal pain (30%); diseases of the appendix not elsewhere
classified (27%); mesenteric lymphadenitis (8%); abdominal pain site not otherwise
specified (7%); noninfectious gastroenteritis (5%); and hyperplasia of the
appendix (3%). These diagnoses accounted for 80% of the negative appendectomies.
Negative appendectomy rate was influenced by hospital volume (Table 3). As the total number of appendectomies
performed at a given hospital increased by 1000, there was a 50% reduction
in the negative appendectomy rate (AOR, 0.50; 95% CI, 0.35-0.71).
There was no correlation between negative appendectomy rate and appendiceal
rupture rate by hospital (r = 0.08; P = .65) (Figure 2) or among
the individual races (P = .65). The rupture rate
was not associated with negative appendectomy rate either before or after
adjustments for race, sex, age, health insurance status, and hospital volume
The data presented herein suggest that hospital-level characteristics
may not be associated with the rate of appendiceal rupture. Hospitals with
increased negative appendectomy rates did not necessarily have lower appendiceal
rupture rates. In addition, no association was found between hospital volume
and appendiceal rupture rate. Evidence supports a relationship between hospital
case volume and quality of outcomes, and in this study, an increased hospital
volume was inversely associated with negative appendectomy rate.15 The
absence of a relationship between hospital volume and appendiceal rupture
rate and the lack of correlation between the negative appendectomy rate and
appendiceal rupture rate suggest that hospital characteristics have little
influence on appendiceal rupture.
The only factors associated with appendiceal rupture in this study were
race, health insurance status, and age. These findings corroborate previous
studies linking appendiceal rupture in both children8 and
adults to certain risk factors including extremes of age, Medicaid insurance,
and nonwhite race.3,5,16,17 Asian
and black children had a significantly higher likelihood of appendiceal rupture
than white children. Moreover, children with public insurance had a 48% greater
chance of having an appendiceal rupture than children with private insurance.
While the racial disparities may represent language barriers, cultural variances,
or a genetic etiology, the association of appendiceal rupture rate with health
insurance status suggests that the incidence of rupture may be related to
medical care access or quality. Finally, the finding of a higher appendiceal
rupture rate in younger children is a well-established phenomenon and correlates
with the difficulty in parental or physician recognition of abdominal symptoms
in this age group.
These findings are consistent with previous articles suggesting that
appendiceal rupture typically occurs prior to hospital presentation. Hale
et al18 reported that 68% of all ruptures occur
prior to surgical evaluation. These authors found that a delay in outpatient
management or diagnosis resulted in a doubling of the appendiceal rupture
rate. Furthermore, the number of appendiceal ruptures related to in-hospital
delay was not statistically significant.18 There
are several reports that document that prehospital delay increases the rate
of complicated appendicitis.4,7,11,19 In
their prospective analysis of 5755 children and adults, Pittman-Waller et
al19 determined that the time from the onset
of symptoms to first seeking medical attention is a significant predictor
of complicated appendicitis (39.8 vs 16.5 hours for acute appendicitis), whereas
the time from surgical evaluation to operative intervention was significantly
shorter for complicated appendicitis (3.8 vs 4.7 hours for acute appendicitis).
Unlike the appendiceal rupture rate, the negative appendectomy rate
was related to hospital-level characteristics. The negative appendectomy rate
was not associated with race or health insurance status but did improve as
hospital volume increased. The mean negative appendectomy rate among the 36
hospitals was 3%, which was much lower than that previously reported.4 The beneficial effects of imaging advances such as
computed tomography and ultrasound probably contributed to the low rate.16,20
The higher negative appendectomy rate in girls and adolescent females
is most likely related to the gynecologic sources of pain after puberty that
often mimic appendicitis. This theory is supported by the intergroup comparisons
showing that adolescent girls but not adolescent boys had a higher negative
If a relationship between negative appendectomy rate and appendiceal
rupture rate exists, a decline in the negative appendectomy rate over time
should result in an increased rupture rate. However, several recent studies6,16,20- 27 performed
in the era of improved diagnostic imaging show stable or declining appendiceal
rupture rates in the face of declining negative appendectomy rates in both
children and young adults. Some reports have noted no change in either rate
over time.28,29 The data presented
herein demonstrate that the appendiceal rupture rate did decline by 7% during
the years studied without a statistically significant change in negative appendectomy
The use of an administrative database created several limitations. For
example, there is a lack of data concerning the prehospital experience of
the patients, including whether patients were transferred from other hospitals.
Additionally, race was provided subjectively by either an admitting clerk
or the patient's family. Ethnicity is not captured in the database and
the Hispanic patient population, therefore, cannot be evaluated separately.
The PHIS database slightly overrepresented blacks, but appears to be similar
to the national pediatric demographic profile for sex and health insurance
status. Studies using different data sets may further elucidate the racial,
ethnic, and health insurance status disparities. This study also was limited
by the inability to analyze physician-specific practice variation. The negative
appendectomy rates and appendiceal rupture rates may not be representative
of all hospitals because all of the institutions were children’s hospitals.
These findings present a dual challenge for improving the outcomes of
children with appendicitis. The low negative appendectomy rates and the relationship
between hospital volume and negative appendectomy rate suggest potential opportunities
for improvement at the hospital level. Efforts to reduce the incidence of
appendiceal rupture should focus on prehospital care. The findings of disparate
care by race and health insurance status are troubling. A public health paradigm
with concentration on access to care and quality-of-care issues as well as
family and physician education might facilitate earlier diagnosis and intervention.
The excessively high rates of appendiceal rupture in children should no longer
Corresponding Author: Kurt D. Newman, MD,
Department of Surgery, Children’s National Medical Center, 111 Michigan
Ave, West Wing, 4th Floor, Suite 200, Washington, DC 20010 (firstname.lastname@example.org).
Author Contributions: Dr Newman had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Ponsky, Kittle, Gilbert,
Acquisition of data: Ponsky, Kittle, Gilbert,
Analysis and interpretation of data: Ponsky,
Huang, Eichelberger, Gilbert, Brody, Newman.
Drafting of the manuscript: Ponsky, Gilbert,
Critical revision of the manuscript for important
intellectual content: Ponsky, Huang, Kittle, Eichelberger, Gilbert,
Statistical analysis: Ponsky, Huang.
Administrative, technical, or material support:
Ponsky, Kittle, Gilbert, Newman.
Study supervision: Eichelberger, Gilbert, Brody,
Acknowledgment: We acknowledge the support
of Jill Joseph, MD, and the Children’s Research Institute.