2. A. Yaghoubian et al. Appendicitis outcomes better at teaching hospitals 811
Some surgeons and patients believe that educating and Table 1 Morbidity and LOH outcomes in patients with
training future surgeons conflicts with providing the best appendicitis managed at teaching and nonteaching
possible care to patients. Yet few data are available as to institutions
whether resident involvement in common surgical proce-
dures adversely affects patient outcomes. In the current era Outcome Teaching Nonteaching P
of a heightened emphasis on patient outcomes, it is impor- Nonperforated appendicitis 2,289 10,385
tant to document the effect of residents as primary surgeons. Wound infection 2.7% 2.3% .30
Thus, the purpose of this study was to determine the out- Abscess drainage .4% 1% .02
Readmission within 30
comes of appendicitis between teaching and nonteaching
days 1.7% 3.5% Ͻ.0001
institutions. LOH (days) 1.7 Ϯ 1.5 1.8 Ϯ 1.6 .002
Laparoscopic
appendectomy 48% 68% Ͻ.0001
Perforated appendicitis:
Methods operative 953 4098
Wound infection 4.8% 7% .03
After obtaining institutional review board approval, a Abscess drainage 4.9% 10% Ͻ.0001
retrospective review of all appendicitis patients aged Ͼ 18 Readmission within 30
years between 1998 and 2007 at 12 hospitals was per- days 4.2% 8.4% Ͻ.0001
LOH (days) 5.0 Ϯ 4.2 5.2 Ϯ 3.1 .30
formed. Two were teaching hospitals (Harbor-UCLA Med-
Laparoscopic
ical Center and Kaiser Permanente Los Angeles Medical appendectomy 23% 42% Ͻ.0001
Center), and 11 were nonteaching Kaiser Permanente hos- Perforated appendicitis:
pitals. At the teaching institutions, the surgical residents nonoperative 10% 6% Ͻ.0001
are actively involved in all aspects of patient care. They are Data are expressed as numbers, percentages, or mean Ϯ SD.
the first to see consultations in the emergency room, serve as
the primary surgeons, direct the postoperative care, and see
patients back in clinic after discharge. Senior resident sur-
geons at the teaching institutions typically serve as teaching perforated appendicitis rates were 29% at the teaching in-
assistants for the appendectomy cases under attending su- stitutions and 28% at the nonteaching institutions (P ϭ .20).
pervision. In contrast, there are no residents at the nonteach- Outcomes data are summarized in the Table 1. For non-
ing institutions, and the attending surgeons perform all perforated appendicitis, there was no difference in the
aspects of patient care. wound infection rate between teaching and nonteaching
Patient factors collected included age, gender, and the institutions. However, there were lower rates of abscess
presence of perforation. Outcome variables were 30-day drainage (.4% vs 1.0%, P ϭ .02) and readmission (1.7% vs
morbidity and length of hospitalization (LOH). Thirty-day 3.5%, P Ͻ .0001) at teaching than at nonteaching institu-
morbidity included wound infection, postoperative abscess tions, respectively. LOH was shorter at the teaching insti-
drainage, and readmission. Outcomes of patients with ap- tutions. The use of laparoscopy for nonperforated appendi-
pendicitis were compared between teaching and nonteach- citis was lower at the teaching institutions.
ing hospitals. For perforated appendicitis, there were also lower rates
All patient data were collected in an Excel spreadsheet of wound infection (4.8% vs 7.0%, P ϭ .03), abscess drain-
(Microsoft Corporation, Redmond, WA) and translated into age (4.9% vs 10.0%, P ϭ .02), and readmission (4.2% vs
native SAS format using DBMS/Copy (Dataflux Corpora- 8.4%, P Ͻ .0001) at teaching than at nonteaching institu-
tion, Cary, NC). Descriptive statistics were calculated for all tions, respectively. The LOH was similar. The use of lapa-
variables. Numerical variables were compared using the roscopy for perforated appendicitis was lower at the teach-
nonparametric Wilcoxon rank-sum test and are reported as ing institutions. However, nonoperative management of
medians with interquartile ranges. Categorical or nominal perforated appendicitis was higher at the teaching institu-
variables were compared using the 2 test or Fisher’s exact tions.
test, as appropriate.
Results Comments
Overall, 3,242 patients with appendicitis were treated at There has always been controversy regarding the quality
the teaching institutions and 14,483 at the nonteaching in- of care delivered at teaching hospitals. Academic surgeons
stitutions. The mean ages were 41 years at the teaching have the dual responsibility of ensuring the best possible
institutions and 38 years at the nonteaching institutions. care to patients while simultaneously training and educating
Sixty-one percent of patients were male at the teaching future generations of surgeons. Thus, the ability to ade-
institutions and 54% at the nonteaching institutions. The quately train surgical residents without compromising the
3. 812 The American Journal of Surgery, Vol 200, No 6, December 2010
quality of care is a concern for all faculty members involved appendicitis and a similar LOH for perforated appendicitis.
in surgical education. The LOH at teaching hospitals may have been even shorter
Previous studies examining other procedures have shown than stated because the rate of laparoscopic appendectomy
that the quality of care at teaching and nonteaching institu- was also lower at teaching hospitals. We have recently
tions is similar.1–5 A prior pilot study by our institution, shown that LOH was shorter in patients undergoing lapa-
which compared 1 teaching and 1 nonteaching hospital, roscopic appendectomy versus open appendectomy.9 Al-
found comparable quality of care when outcomes for pedi- though we did not perform a formal cost analysis, we
atric appendicitis were analyzed.6 Unlike our previous believe that cost at the teaching institutions was lower
study, this multi-institutional study demonstrated lower because both morbidity was lower and LOH was shorter
morbidity for both adult perforated and nonperforated ap- compared with nonteaching institutions.
pendicitis at teaching hospitals with respect to readmission In the present study, we also found that patients with
rate, wound infection, and postoperative abscess drainage. perforated appendicitis were more likely to be managed
Meguid et al,7 in a retrospective study of outcomes of lung nonoperatively at the teaching institutions. A potential
cancer resections, demonstrated a lower mortality rate at explanation for this finding is that teaching institutions
teaching hospitals. Similarly, a study by the same group were more willing to implement a fairly novel technique
found a lower mortality rate for abdominal aortic aneurysm of nonoperative management on the basis of research
repair at teaching hospitals although the length of stay was studies.10 –12 By contrast, the nonteaching institutions
longer.8 They attributed their improved outcomes to the performed a much higher rate of laparoscopic appendec-
presence of specialty training and increased volume. The tomies compared with the nonteaching institutions. The
authors hypothesized that the increase in length of stay was reason behind this observation was because one of the
due to sicker patients at the teaching hospitals. The limita- teaching institutions transitioned into performing laparo-
tion of the prior mentioned studies is that they evaluated the scopic appendectomy at a much later date. This was due
outcomes of complex procedures in which residents typi- mainly to resistance from the attending surgeons to
cally play the role of assistants rather than primary sur- change in techniques, and it was the residents’ influence
that ultimately led to the change in practice. Currently,
geons. Looking at more routine operations, Hwang et al1
nearly all patients with appendicitis will undergo laparo-
performed an analysis of outcomes of bowel resection, lapa-
scopic appendectomy at both institutions.
roscopic cholecystectomy, hernia, mastectomy, and appen-
There were several limitations of this study. First, this is
dectomy of 4 attending surgeons who worked with residents
a retrospective analysis and by nature had its limitations.
versus 4 who did not. Comparing all procedures together,
We did not evaluate the length of the operation, which may
there were no differences in complications between the
understandably take longer at the academic institutions and
groups, although there was greater mortality, a greater du-
contribute to increased costs of the operating room, as
ration of stay, and higher costs in the resident group. When
demonstrated by other studies. Furthermore, the overall low
comparing the 5 most common procedures individually,
morbidity and mortality of appendicitis may offset the in-
there was no difference in complications or mortality, al-
creased risk for resident-related errors in patient manage-
though a greater length of stay and higher costs were seen in ment. We also did not review the number of cases that
the teaching group. began laparoscopically and required conversion to open
In general, similar or improved outcomes at teaching appendectomy between the 2 types of institutions. Last,
institutions have come at the expense of increase LOH and teaching hospitals may be inherently different from non-
thus higher cost. A study by Hutter et al2 in the evaluation teaching hospitals independent of the presence of residents
of pancreatic resections also demonstrated improved out- or students.
comes at the teaching hospitals, but they found a longer Our data demonstrate that teaching hospital care of
length of stay and attributed this to their care of uninsured patients with appendicitis is associated with improved
patients. In a subset analysis, the uninsured patients had the outcomes. In addition, patients at teaching institutions
longest length of stay (27 vs 20 days). Unlike previous with perforated appendicitis were more likely to be man-
studies, in our pilot study comparing outcomes of pediatric aged nonoperatively at teaching institutions compared
appendicitis, we found a decrease in the LOH for children with nonteaching institutions. In general, our data refute
with nonperforated appendicitis.6 Because the decrease in the notion that there is a conflict between training future
LOH was only .4 days, one can argue that this may not be surgeons and simultaneously providing the best possible
clinically significant. One possible explanation for this find- patient care.
ing is that more patients at the teaching institution may have
been discharged in the evening rather than the next morning
given the availability of the resident team. If this were
References
indeed the case, then this would have led to substantial cost
savings in this group of patients, making this finding clini- 1. Hwang C, Pagano CR, Wichterman KA. Resident versus no resident:
cally significant. In this current multi-institutional study, we a single institutional study on operative complications, mortality, and
also found a slight decrease in LOH for nonperforated cost. Surgery 2008;144:339 – 44.