Flum DR, Morris A, Koepsell T, Dellinger EP. Has Misdiagnosis of Appendicitis Decreased Over Time?A Population-Based Analysis. JAMA. 2001;286(14):1748-1753. doi:10.1001/jama.286.14.1748
Author Affiliations: Robert Wood Johnson Clinical Scholars Program (Drs Flum, Morris, and Koepsell), Department of Surgery (Drs Flum and Dellinger), and Departments of Epidemiology and Health Services (Dr Koepsell), University of Washington, Seattle.
Context Misdiagnosis of presumed appendicitis is an adverse outcome that leads
to unnecessary surgery. Computed tomography, ultrasonography, and laparoscopy
have been suggested for use in patients with equivocal signs of appendicitis
to decrease unnecessary surgery.
Objective To determine if frequency of misdiagnosis preceding appendectomy has
decreased with increased availability of computed tomography, ultrasonography,
Design, Setting, and Patients Retrospective, population-based cohort study of data from a Washington
State hospital discharge database for 85 790 residents assigned International Classification of Diseases, Ninth Revision
procedure codes for appendectomy, and United States Census Bureau data for
Main Outcome Measure Population-based age- and sex-standardized incidence of appendectomy
with acute appendicitis (perforated or not) or with a normal appendix.
Results Among 63 707 nonincidental appendectomy patients, 84.5% had appendicitis
(25.8% with perforation) and 15.5% had no associated diagnosis of appendicitis.
After adjusting for age and sex, the population-based incidence of unnecessary
appendectomy and of appendicitis with perforation did not change significantly
over time. Among women of reproductive age, the population-based incidence
of misdiagnosis increased 1% per year (P = .005).
The incidence of misdiagnosis increased 8% yearly in patients older than 65
years (P<.001) but did not change significantly
in children younger than 5 years (P = .17). The proportion
of patients undergoing laparoscopic appendectomy who were misdiagnosed was
significantly higher than that of open appendectomy patients (19.6% vs 15.5%; P<.001).
Conclusion Contrary to expectation, the frequency of misdiagnosis leading to unnecessary
appendectomy has not changed with the introduction of computed tomography,
ultrasonography, and laparoscopy, nor has the frequency of perforation decreased.
These data suggest that on a population level, diagnosis of appendicitis has
not improved with the availability of advanced diagnostic testing.
Appendectomy is one of the most frequently performed surgical procedures
in the United States and the most common surgical emergency of the abdomen.
Despite a lifetime cumulative incidence of nearly 7%,1
diagnosis of appendicitis remains a challenge. The risk of 2 primary adverse
outcomes must be balanced in management of presumed appendicitis: perforation,
often occurring in the prehospital setting,2
and misdiagnosis, resulting in removal of a normal appendix. Although reduction
of the frequency of appendiceal perforation has received much scrutiny, the
factors leading to misdiagnosis are less understood.
To reduce the incidence of perforation, the surgical community traditionally
accepts that approximately 15% of appendectomies overall and 20% in women
will yield a noninflamed appendix.3,4
The rate of misdiagnosis in certain populations of patients may be as high
as 40%.2,5- 7
This relatively high rate of unnecessary appendectomy is being challenged
in some quarters as an outdated standard, given the dramatic expansion of
diagnostic testing options for appendicitis during the last decade.8 Indeed, many investigators have demonstrated that
in research environments, advanced diagnostic testing using computed tomography
(CT), ultrasonography (US), and laparoscopy decreases the frequency of misdiagnosis.9- 14
These diagnostic tests are often targeted toward populations deemed at increased
risk for misdiagnosis: children, the elderly, and women of reproductive age.
The purpose of this study was to evaluate changes in the frequency of
misdiagnosis among patients undergoing appendectomy during a period coincident
with the growing availability of CT, US, and laparoscopy. The frequency of
misdiagnosis was measured in 2 ways: as a percentage of procedures performed
during a given period and as a population-based incidence rate. Reported rates
of misdiagnosis classically have been based on the total number of appendectomies
performed in a single institution over a fixed period; however, population-based
rates account more closely for the true population at risk for appendectomy.
We hypothesized that misdiagnosis has decreased since the introduction of
advanced diagnostic techniques in the late 1980s. Furthermore, we hypothesized
that the greatest decreases in misdiagnosis rate would be identified in populations
at increased risk for misdiagnosis because they are more likely to undergo
A retrospective cohort study was conducted, using a statewide, population-based
hospital discharge database.
Data were obtained from the Washington State Comprehensive Hospital
Abstract Reporting System (CHARS) database. This data set is derived from
all public and private hospitals in Washington State (Veterans Affairs and
US military hospitals excluded) and includes nearly all of the population
of patients undergoing appendectomy in the state of Washington during the
study period. The data set contains demographic variables, admission and discharge
administrative details, payer status, International Classification
of Diseases, Ninth Revision (ICD-9) procedure
and diagnostic codes, and coded hospital identifiers. United States Census
Bureau data for total and age- and sex-specific yearly state population estimates
were used for population-based analyses. Error due to inclusion of patients
in the denominator who had prior appendectomy was assumed to be stable over
time. Error due to missing Veterans Affairs and military hospital patients
was recognized but represented approximately 0.63% of total statewide cases.
This study was exempted from human subjects review by agreement of the
University of Washington Human Subject Review Committee and the Washington
State Department of Health. The data set includes only anonymous data and
is considered to be within the public domain.
All CHARS reports from 1987 through 1998 were searched for ICD-9 procedure codes pertaining to appendectomy
(BOX). This group
was then evaluated based on associated ICD-9 diagnostic
codes that described appendiceal pathologic findings and other relevant variables.
The total number of appendectomies (incidental and nonincidental) performed
were recorded by year.
Nonincidental Appendectomy47 Operations on appendix, excludes incidental
47.0 Appendectomy, excludes incidental
47.01 Laparoscopic appendectomy
47.09 Other appendectomy
Incidental Appendectomy47.1 Incidental appendectomy
47.11 Laparoscopic incidental appendectomy
47.19 Other incidental appendectomy
Appendicitis540 Acute appendicitis
5400 With perforation, peritonitis, rupture*
5401 Abscess with generalized peritonitis*
5409 Without mention of perforation, peritonitis, rupture
541 Appendicitis unqualified
542 Other appendicitis
*These codes are used to define patients with perforation or rupture.
A case of appendicitis was defined as any patient undergoing nonincidental
appendectomy with an associated diagnostic code of appendicitis or related
appendiceal pathologic finding. A case of misdiagnosis was defined as any
patient undergoing nonincidental appendectomy without an associated diagnostic
code of appendicitis or related appendiceal pathologic finding. Perforation
was defined in a similar fashion, using appropriate ICD-9 codes
We calculated yearly procedure-based percentages of all appendectomies
(incidental and nonincidental) for the entire cohort and for women only, using
the total number of procedures as the denominator. In the nonincidental appendectomy
group, frequencies of misdiagnosis, appendicitis, and perforation were computed,
using the number of nonincidental appendectomies as the denominator. Misdiagnosis
frequency was also computed for 3 subpopulations considered at increased risk
for misdiagnosis: children younger than 5 years, patients older than 65 years,
and women of reproductive age (15-45 years).
Because population-based rates are sensitive to changes in the underlying
population at risk for each of these outcomes, yearly population-based rates
were calculated for all appendectomies (incidental and nonincidental). Population-based
rates of misdiagnosis, appendicitis, and rupture were calculated for the entire
nonincidental cohort and for women. Age- and sex-specific population values
were used as appropriate. All data were standardized for sex and age using
the direct method, with the 1990 population of Washington State as the reference
population. All population-based rates are reported as rate per 10 000
person-years. Patients undergoing laparoscopic appendectomy were considered
in a separate analysis as well to compare the overall frequency of use and
misdiagnosis associated with laparoscopic appendectomy in different subpopulations.
Poisson regression was used to test for a significant increase or decrease
over time in the population-adjusted rates of appendectomy, incidental appendectomy,
appendicitis, misdiagnosis, and perforation/rupture. The Poisson regression
model was used to adjust for the sex and age of the patient. A test for trend
(ie, the P value for whether the coefficient for
calendar year was significantly different from 0) was applied to evaluate
significant changes over time in the procedure-based rates of all appendectomies
and incidental appendectomies and frequencies of appendicitis, misdiagnosis,
and perforation. Statistical analysis was performed using STATA statistical
analysis software, version 7 (STATA Corp, College Station, Tex).
During the 12-year study period, 85 790 patients (mean [SD] age,
32.1 [18.6] years; 54.4% female) underwent appendectomy (22.1% incidental).
When considered simply as a percentage of all appendectomies performed yearly
and analyzed for trend, incidental appendectomy decreased significantly by
year, with a corresponding increase in nonincidental appendectomy (P<.001). During the study period, 9880 nonincidental appendectomies
were performed without an associated diagnosis of appendicitis, representing
15.5% of all nonincidental appendectomies. The average rate of perforation
was 25.8%. The percentage of misdiagnoses was greater among women than men
(22.8% vs 9.2%, respectively; P<.001). Patients
with misdiagnoses were slightly older compared with those with appropriate
diagnoses (mean, 30.5 vs 28.6 years, respectively; P<.001).
Among patients undergoing nonincidental appendectomy (n = 63 707), the
percentage of misdiagnoses remained stable over time (P = .06), as did the percentage with perforated and nonperforated appendicitis
(P = .08) (Table 1).
Between 1987 and 1998, there was a 20.3% increase in the state's population.
Age- and sex-standardized population-based incidence trends are displayed
in Figure 1. The overall rate of
appendectomy per 10 000 person-years decreased yearly by 3.1% (P<.001). Incidental appendectomy decreased 7.3% yearly,
from 5.52 per 10 000 person-years in 1987 to 2.45 per 10 000 person-years
in 1998 (P<.001). There was essentially no change
in the rate of nonincidental appendectomy per 10 000 person-years (P = .07). Regression analysis indicates a small decrease
(−1.5%) in the yearly rate of appendicitis per 10 000 person-years
(P<.001), but stable rates of misdiagnosis (P = .27) and perforation (P =
.51) over time.
Among women, the decreasing rate of total appendectomies manifested
as a significant trend toward fewer incidental appendectomies over time (P<.001). The frequency of misdiagnosis among women remained
stable over time at an overall rate of 23.2% (P =
.52). Similar findings were also noted in the sex-specific population-based
rates. There was no significant change in the population incidence of nonincidental
appendectomy, overall appendicitis, or rate of misdiagnosis in women. Among
other subpopulations at risk for misdiagnosis, the population incidence of
misdiagnosis actually increased 1% yearly in women of reproductive age (P = .005) and increased 8% per year in patients older than
65 years (P<.001) (Table 2). Misdiagnosis rates did not change significantly in children
younger than 5 years (P = .17).
Performance of laparoscopic appendectomy, first coded for in 1996, has
rapidly increased over time. Laparoscopic appendectomy was performed in 2.8%
of all patients in 1996 and in 15.0% of patients in 1998. Laparoscopic appendectomy
was performed more often in certain groups of patients. For example, during
the 3 years that it was recorded, laparoscopic appendectomy was performed
in 13.8% of women and in 17.3% of women of reproductive age. The procedure-based
frequency of misdiagnosis among patients undergoing laparoscopic appendectomy
was higher than that of patients undergoing open appendectomy during the same
period (19.6% vs 15.5%, respectively; P<.001)
(Table 3). For all women, the
overall frequency of misdiagnosis with laparoscopic appendectomy from 1996
to 1998 was similar to the open appendectomy misdiagnosis rate during the
same period (24.4% vs 22.5%, respectively; P = .17).
Among women of reproductive age, laparoscopic appendectomy was associated
with a higher rate of misdiagnosis (29.1% vs 24.9%, respectively; P = .02).
The availability of CT, US, and laparoscopy to aid in diagnosis of appendicitis
has increased dramatically over the last decade. Controlled clinical trials
have suggested that this technology improves diagnostic accuracy.9- 14
We evaluated the 85 790 patients in Washington State who underwent appendectomy
in 1987-1998, a period concurrent with the increasing availability of this
technology. Contrary to expectation, we found that both procedure-based and
population-based frequencies of misdiagnosis remained stable over time (15.5%
and 1.56 per 10 000 person-years, respectively). Likewise, age- and sex-adjusted
incidences of perforation remained stable over time (25.8% and 2.71 per 10 000
person-years, respectively). Among subpopulations considered at increased
risk, the incidence of misdiagnosis actually increased among women of reproductive
age and patients older than 65 years. A decrease in the overall rate of appendectomy
was also identified but was related to a significant decrease in the rate
of incidental appendectomy.
In their landmark 1990 article, Addiss et al1
first described the method we used to define misdiagnosis and appendicitis
within a large administrative data set. These investigators evaluated population
trends in appendectomy and appendicitis in 1970-1984 (n = 16 547). The
data source for this work, the National Hospital Discharge Summary, incorporated
a 0.5% sample of all patients hospitalized yearly in the United States. While
this previous study was based on national survey data, our study was based
on data from all Washington State hospital discharge summaries during the
years of interest. Both studies defined misdiagnosis as the absence of a diagnosis
of acute appendicitis. Both studies are limited by the absence of pathologically
confirmed diagnosis and the potential that administrative data are subject
to reporting or interpretation errors by personnel who abstract chart data.
However, the rate of this type of error may be assumed to be stable over time,
allowing for reliable comparison between years.
Most studies of appendicitis use the number of appendectomies performed,
often in a single institution, as the denominator for their calculations.
Evaluation of population-based data, however, allows for analysis of trends
in the true incidence of the process of interest while accounting for changes
in the population at risk. Addiss et al,1 for
example, used a population-based technique and found a declining rate of incidental
appendectomy between 1979 (year of first incidental coding) and 1984, coincident
with a gradually decreasing rate of appendicitis. These researchers found
an overall procedure-based rate of misdiagnosis of 14.7% (8.8% in men and
21.4% in women) and that diagnostic accuracy had improved over time, from
74% to 83% in women and from 86% to 92% in men. However, the authors did not
calculate a population-based rate of misdiagnosis. This earlier cohort may
be considered a historical control for studying the impact of available technology
on the diagnosis of appendicitis. In the 1987-1998 Washington State cohort,
CT, US, and laparoscopy were widely available to assist in diagnosis of appendicitis
(albeit at variable rates), yet we found no improvement in the procedure-based
misdiagnosis rate (15.5%) compared with that found in previous work (14.7%).1
Our data do not support the hypothesis that the rate of misdiagnosis
has improved with the widespread availability of CT, US, and laparoscopy.
While state licensing data demonstrate the increasing availability of CT during
the study period, this study was limited by our inability to determine the
rates of CT and US use with this database. Despite published reports that
advocate increased use of advanced diagnostic testing,9- 14
and the informal impression that such testing has increased dramatically,
we were unable to identify any direct population-level evidence of increasing
CT use in diagnosis of appendicitis. Furthermore, other interventions were
advocated during this period to improve diagnosis and management of appendicitis.
These included clinical pathways,15 increased
senior surgeon involvement in diagnosis,16
and better data management tools.17 We were
also unable to determine the extent to which these available interventions
were used on a population level. As a result, several explanations for our
findings may be proposed. Computed tomography and US may not be performed
frequently enough or in the appropriate subpopulations to affect the rate
of misdiagnosis. Alternatively, diagnostic tests may be less accurate in a
typical clinical environment than in the research setting.10,18,19
Conversely, these tests may be accurate and performed routinely but may be
overruled or not reported rapidly enough to influence decision making. Indeterminate
readings and delayed interpretation clearly reduce the potential benefit of
such tests. A more clinically detailed data set containing reliable radiological
and histological data will be required to clarify this relationship.
Some authors maintain that laparoscopy can decrease the rate of diagnostic
error in patients with presumed appendicitis by identifying alternative pathologic
findings.20 After inserting a laparoscope into
the abdomen, the surgeon can directly observe the appendix to determine if
inflammation is present. This study suggests that the development of laparoscopic
appendectomy has not improved diagnostic accuracy in patients with presumed
appendicitis. In part, this may be explained by selection bias; patients undergoing
laparoscopic appendectomy may be at higher risk for misdiagnosis. However,
in our study, even among women of reproductive age, the rate of misdiagnosis
was not lower in those undergoing laparoscopic appendectomy. Many surgeons
remove a noninflamed appendix when performing laparoscopy, even in the presence
of other pathologic findings. For some, acknowledging that the gross appearance
of the appendix corresponds poorly to the histological changes of appendicitis,21 this practice may not be classified as an incidental
appendectomy. There is little consensus regarding the appropriate laparoscopic
management of the normal-appearing appendix. In 1 survey, the surgeon's assessment
of acute inflammation was correct in 120 of 132 cases, a positive predictive
value of 91%.21 Conversely, when surgeons assessed
the appendix as being noninflamed, they were correct in only 11 of 43 cases,
yielding a negative predictive value of only 26%. These authors suggest that
during a laparoscopic operation for presumed appendicitis, the normal-appearing
appendix should be removed.21
In conclusion, although controlled studies have demonstrated a decrease
in the frequency of misdiagnosis with CT, US, and laparoscopy, we identified
no change in the misdiagnosis of appendicitis on a population level concurrent
with the increasing availability of this diagnostic technology. We recommend
further investigation of the pragmatic relationship between diagnostic tests
for presumed appendicitis and diagnostic accuracy at appendectomy. Finally,
this study indicates that the general population has not realized one suggested
benefit of the era of advanced diagnostic testing: the reduction of unnecessary