GENERAL THORACIC

ORIGINAL ARTICLES: GENERAL THORACIC

GENERAL THORACIC SURGERY:

The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/
home. To take the CME activity related to this article, you must have either an STS member or an
individual non-member subscription to the journal.

From Emergency Department Visit to

Readmission After Esophagectomy:
Analysis of Burden and Risk Factors
Biniam Kidane, MD, MSc, Sean Higgins, BSc, Dhruvin H. Hirpara, MD, Suha Kaaki, MD,
Yu Cindy Shen, BSc, Frances Allison, BSc, Thomas K. Waddell, MD, PhD, and
Gail E. Darling, MD
Section of Thoracic Surgery, University of Manitoba, Winnipeg, Manitoba, Canada; Division of Thoracic Surgery, University of Toronto,
Toronto, Ontario, Canada; and Division of Thoracic Surgery, Toronto General Hospital, Toronto, Ontario, Canada

Background. Frequent emergency department (ED) Sensitivity analysis using Poisson regression to model
visits occur after esophagectomy. We aimed to identify conversion as a rate identified that living in regions
the incidence of and risk factors for conversion from ED further away was associated with lower conversion rate to
visit to inpatient admission. admission (risk ratio 0.35; 95% confidence interval, 0.13 to
Methods. A retrospective cohort study was performed 0.94; P [ .04).
of consecutive esophagectomies at a tertiary Canadian Conclusions. Although postesophagectomy ED utili-
center (1999 to 2014). Multivariable regression analyses zation is high, the majority of visits do not convert to
identified factors associated with conversion from ED admission. With each increasing ED visit, likelihood of
visit to admission. converting to admission increases. Anastomotic leakage
Results. There were 520 esophagectomies with 6% was associated with higher odds of conversion to
inhospital mortality (n [ 31). Of those discharged, 29.7% admission, possibly related to development of strictures.
(n [ 145) had one or more emergency visit and 43.4% Access to urgent outpatient endoscopy may help reduce
(n [ 63) of these patients were readmitted to the hospital. the incidence of ED visits and admission. Although living
First-time ED visits resulted in inpatient conversion in regions further away is associated with lower conver-
23.4% of the time (n [ 34); successive ED visits resulted sion rates to admission at the index hospital, that may be
in increasing conversion. On multivariable analysis, due to patients utilizing closer local hospitals.
anastomotic leak (adjusted odds ratio 2.45; 95% confi-
dence interval, 1 to 6.01; P [ .05) was independently (Ann Thorac Surg 2021;112:379-86)
associated with higher odds of conversion to admission. Ó 2021 by The Society of Thoracic Surgeons

S urgery is the primary treatment for operable cancers

of the esophagus and may be used as the sole treat-
ment for cancers or in combination with chemotherapy or
Emergency department (ED) visits following the oper-
ation are common and have consequences in terms of
health resource utilization and quality of life.6-9 We have
chemoradiation therapy.1,2 Although it is an important previously demonstrated that there is a high incidence of
intervention in the treatment of many esophageal can- ED utilization after esophagectomy with nearly one third
cers, esophagectomy is among the most morbid of sur- of patients visiting the ED at least once in the year after
gical procedures and can result in complications that may the procedure and approximately 10% of patients visiting
persist for sustained periods after surgery or arise in a the ED at least three times in the year after the proced-
delayed fashion.3-5 That has implications for the quality of ure.10,11 We also found that feeding tube problems were
life of patients as well resource utilization.6 the most common cause for ED utilization, and that pa-
tients who lived further away from the hospital were less
likely to utilize the ED.10,11 It is unclear, however, what
Accepted for publication Nov 16, 2020. proportion of these patient ED visits convert to true
Presented at the Forty-third Annual Meeting of the Western Thoracic inpatient admissions; it is also unclear what ED visit
Surgical Association, Colorado Springs, CO, Jun 21-24, 2017. reasons are more likely to result in conversion to inpa-
Address correspondence to Dr Kidane, Health Sciences Center, Rm GE- tient admission or risk factors associated with conversion
611, Winnipeg, Manitoba R3A, Canada; email: [email protected]. to inpatient admission are. Characterizing the incidence

Ó 2021 by The Society of Thoracic Surgeons 0003-4975/$36.00

Published by Elsevier Inc. https://doi.org/10.1016/j.athoracsur.2020.11.020
 380 KIDANE ET AL Ann Thorac Surg
GENERAL THORACIC

READMISSION AFTER ESOPHAGECTOMY 2021;112:379-86

of conversion to inpatient admission and the risk factors patients and also fit within usual patterns of categoriza-
associated with it serves as valuable information and can tion in the literature. Year of surgery was dichotomized
guide development of local and system-level in- into 1999 to 2006 and 2007 to 2014. Patients were classified
terventions to address preventable causes and risk fac- as having social support if they lived with a partner or an
tors, with the ultimate goal of optimizing health resource adult child (aged 18 years or more).
utilization for the benefit of both our patients and the Competing risk analysis was not used for the
health care system. competing risks of ED visits and mortality after hospital
The primary objectives of this study were to determine discharge. Dying at home after discharge prevents a pa-
the incidence of conversion from ED visit to inpatient tient from presenting to the ED, and our purpose was to
admission and to identify the reasons and risk factors identify burden of resource utilization incurred by ED
associated with conversion to inpatient admission. visits; therefore, competing risk analysis is not necessary
as patients who die at home after discharge would not be
expected to increase the burden of ED visits or resource
Material and Methods
utilization. A two-sided alpha of 0.05 was used for all tests
This study was approved by the University Health of significance. Statistical analyses were performed using
Network Research Ethics Board. Requirement for patient IBM SPSS Statistics 24 (IBM Corp, Armonk, NY).
consent was waived owing to the retrospective nature of The usual postesophagectomy follow-up protocol at
the analysis. Using data from a prospectively collected our center is as follows: (1) immediate follow-up with
esophageal cancer database, we conducted a retrospec- surgeon 2 to 3 weeks after discharge and then 3 to 4 weeks
tive cohort study of consecutive esophagectomies for as needed thereafter until the 3-month mark; (2) radio-
cancer in a high-volume tertiary hospital in Ontario, graphic surveillance with computed tomography scans
Canada, from 1999 to 2014. We identified all ED visits every 3 months for 3 years, then every 6 months to 5
within 1 year of esophagectomy. Reasons for ED visits years; and (3) clinical follow-up every 3 months for 3
were assessed by chart review. The outcome of interest years, then every 6 months to 5 years, with the majority of
was incidence of conversion from ED visit to inpatient patients traveling to our center. Patients living far away
admission. were able to alternate clinical visits with their local phy-
For univariable analysis, normally distributed contin- sicians after the first year and able to get all their radio-
uous data were reported as mean with standard deviation graphic surveillance locally.
and analyzed using independent sample Student’s t tests.
Data that were not normally distributed were reported as
median with interquartile range (IQR) and analyzed us-
Results
ing the Mann-Whitney U test. Fisher’s exact tests were There were 520 esophagectomies performed during the
used for univariate analysis of categoric data. Hierarchical study period with an inhospital mortality of 6% (n ¼ 31). Of
multivariable logistic regression was used to identify the surviving patients discharged from hospital (n ¼ 489),
factors independently associated with higher odds of 145 (29.7%) went on to utilize the ED within 1 year of
conversion from ED visit to inpatient admission. Hierar- esophagectomy and 63 (43.4%) of those patients were
chy was defined by using a forced entry method as the admitted to the hospital as the result of their ED visit;
first step for the following factors, which were identified a these 63 patients were therefore classified as having a
priori as being important: age, sex, income, comorbidity, conversion from ED visit to inpatient admission. Distri-
resection type, use of hybrid/minimally invasive surgery, bution of esophagectomy type among patients who con-
use of chemotherapy or radiation therapy, or both, and verted from ED visit to inpatient admission was as
year of surgery. follows: transhiatal, 17.5% (n ¼ 11); McKeown, 17.5%
In the second step of the hierarchy, we included the (n ¼ 11); Ivor-Lewis, 36.5% (n ¼ 23); left thoracoabdominal,
following factors: social supports, ability to speak English, 25.4% (n ¼ 16); and pharyngolaryngectomy, 3.2% (n ¼ 2).
and specific complication types (respiratory, cardiac, Approximately 25% of these patients (n ¼ 16) had at least
thromboembolic, anastomotic, chylothorax, recurrent a thoracoscopic or laparoscopic component and were
laryngeal nerve palsy) that occurred during the index classified as hybrid/minimally invasive. Use of minimally
operation hospital stay. We also included any factor that invasive approaches increased significantly in the latter
had a P value less than .2 on univariable analysis. We half of the study period, rising from 4.4% of cases to 40%
applied backward stepwise elimination to this second of cases (P < .001). Median length of stay for the index
step of the hierarchical logistic regression analysis. As a operation of esophagectomy was 14 days (IQR, 11 to 23).
sensitivity analysis, we also preformed Poisson regression Figure 1 illustrates the reasons for patients’ first ED
to model conversion as a rate rather than an incidence. visits and the proportion that converted to inpatient
Income was estimated based on the median family in- admission. Overall, first-time ED visits after esoph-
come of the individual’s postal code using Canadian agectomy resulted in inpatient conversion 23.4% of the
Census tract data, which has been shown to be a robust time. Dysphagia and feeding tube problems, which were
and reliable estimation of income.12 Comorbidity status identified as the two most common culprits for first ED
was defined using the Charlson Comorbidity Index (CCI) visits, were also the visit reasons least likely to lead to
and was stratified as low (CCI 0 to 2) and high (CCI
3). inpatient conversion (P ¼ .007). Second-time ED visits
This stratification was based on the distribution of converted to inpatient admission at a rate of 33.3%, an
Ann Thorac Surg KIDANE ET AL 381

GENERAL THORACIC
2021;112:379-86 READMISSION AFTER ESOPHAGECTOMY

Figure 1. Conversion of first emergency department visit to inpatient admission stratified by visit reason. Feeding tube problems and dysphagia
were less likely to lead to conversion to admission (P ¼ .007). Blue bars indicate no conversion to inpatient; green bars indicate conversion to
inpatient. (GI, gastrointestinal; Mets, metastases; NYD, not yet diagnosed; SOB, shortness of breath.)

increase from first-time ED visits. Feeding tube problems occurred and were identified and treated during the in-
were still the most common reason for ED visits and were dex/original hospital admission.
again less likely to lead to inpatient conversion; ED visits As a sensitivity analysis, we also preformed Poisson
for dysphagia, postoperative pain, and incisional hernia regression to model conversion as a rate rather than an
were likely to lead to readmission (Figure 2). Third-time incidence. Sensitivity analysis with Poisson regression
ED visits converted to inpatient admission 24.4% of the identified that living in regions further away was signifi-
time with feeding tube problems again identified as being cantly associated with lower rate of conversion from ED
the most common reason for ED visits and less likely to visit to inpatient admission (risk ratio 0.35; 95% confi-
lead to readmission (Figure 3). Figure 4 illustrates that dence interval, 0.13 to 0.94; P ¼ .04).
fourth-time ED visitors convert to inpatient admission at
a rate of 47.8%. Therefore, the fourth ED visit appears to
Comment
be the inflection point at which patients are more likely to
be readmitted when they utilize the ED this frequently. We aimed to identify the incidence of and risk factors
Although feeding tube problems again were the most for the conversion of ED visits to inpatient admission.
common reason for fourth ED visits, they lead to read- Whereas one third of esophagectomy patients have one
mission 88.9% of the time, a sharp contrast from the trend or more ED visits within 1 year of being discharged
in patients’ first three ED visits, which rarely lead to home, fewer than half these patients (43.4%) are read-
admission. The majority of patients who visited the ED mitted to the hospital as a result of their ED visit.
for a fifth time and beyond were readmitted as inpatients, Feeding tube problems were found to be the most
including all patients who visited the ED a seventh or common reason behind ED visits, but the least likely
eighth time (Figure 5). This finding is consistent with the reason for conversion of ED visit to inpatient admis-
trend of increasing ED visit number being more likely to sion. The likelihood of converting from an ED visit to
result in readmission. an inpatient at the index hospital is increased among
Table 1 depicts the univariable analysis of differences patients who visited the ED multiple times and had
between patients with and without conversion to inpa- anastomotic leaks but lower among patients who live
tient admission from ED visits. On multivariable analysis, further distances away.
anastomotic leak (adjusted odds ratio 2.45; 95% confi- Feeding tube problems mostly related to tube block-
dence interval, 1.00 to 6.01; P ¼ .05) was independently ages or unsubstantiated infection concerns were the most
associated with higher odds of conversion to inpatient common and consistent reason for ED utilization and
admission from ED visit (Table 2). These were leaks that were the most common reason for readmission overall.
 382 KIDANE ET AL Ann Thorac Surg
GENERAL THORACIC

READMISSION AFTER ESOPHAGECTOMY 2021;112:379-86

Figure 2. Conversion of second emergency department visit to inpatient admission stratified by visit reason. Feeding tube problems and
abdominal pain were less likely to lead to conversion to admission (P ¼ .05). Blue bars indicate no conversion to inpatient; green bars indicate
conversion to inpatient. (C. Diff, Clostridium difficile; Post-op, postoperative.)

Figure 3. Conversion of third emergency department visit to inpatient admission stratified by visit reason. No visit reason was significantly more
likely to result in conversion to admission (P ¼ .41). Blue bars indicate no conversion to inpatient; green bars indicate conversion to inpatient. (C.
Diff, Clostridium difficile; Post-op, postoperative; NYD, not yet diagnosed; SOB, shortness of breath; SSI, surgical site infection.)
Ann Thorac Surg KIDANE ET AL 383

GENERAL THORACIC
2021;112:379-86 READMISSION AFTER ESOPHAGECTOMY

Figure 4. Conversion of fourth emergency department visit to inpatient admission stratified by visit reason. Feeding tube problems were more
likely to lead to conversion to admission (P ¼ .02). Blue bars indicate no conversion to inpatient; green bars indicate conversion to inpatient. (C.
Diff, Clostridium difficile; NYD, not yet diagnosed.)

Figure 5. Conversion of fifth to eighth emergency department (ED) visits to inpatient admission stratified by visit reason. The majority of patients
who visited the ED for a fifth time and beyond were readmitted as inpatients, including all patients who visited the ED a seventh or eighth time.
Blue bars indicate no conversion to inpatient; green bars indicate conversion to inpatient. (Mets, metastases.)
 384 KIDANE ET AL Ann Thorac Surg
GENERAL THORACIC

READMISSION AFTER ESOPHAGECTOMY 2021;112:379-86

Table 1. Baseline Characteristics Conversion to Admission Versus No Conversion to Admission

Conversion From ED to No Conversion to
Risk Factor Admission (n ¼ 63) Admission (n ¼ 82) P Value

Year of surgery .50

1999-2006 58.7 (37) 52.4% (43)
2007-2014 41.3 (26) 47.6 % (39)
Age at time of surgery, mean (SD) 61.5 (13.4) 62.1 (10.7) .74
Male 74.6 (47) 76.8% (63) .85
Median household income, $ 74,983 (60,299-93,623) 77,137 (64,287-94,142) .46
Social support 85.7 (54) 79.3% (65) .39
English-speaking 90.5 (57) 92.7% (76) .76
Residence distance from index hospital .008
Within greater metropolitan area 47.3 (61) 52.7 (68)
Outside greater metropolitan area 12.5 (2) 87.5 (14)
Esophagectomy type .86
Transhiatal 17.5 (11) 15.8 (13)
McKeown 17.5 (11) 23.2 (19)
Ivor-Lewis 36.5 (23) 34.1 (28)
Left thoracoabdominal 25.4 (16) 25.6 (21)
Other/not specified 3.2 (2) 1.2 (1)
Lower CCI (CCI 0-2 vs CCI >3) 7.9 (5) 11 (9) .58
Chemotherapy 25.4 (16) 25.6 (21) >.999
Radiation therapy 25.4 (16) 22 (18) .69
Video-assisted thoracic surgery 25.4 (16) 32.9 (27) .36
Hospital length of stay, d 15 (11-28) 14 (11-24) .35
Respiratory complications 28.6 (18) 23.2 (19) .57
Anastomotic leak 25.4 (16) 14.6 (12) .14
Laryngeal nerve complication 9.5 (6) 8.5 (7) >.999
Thromboembolic complication 4.8 (3) 1.2 (1) .32
Cardiac complication 11.1 (7) 13.4 (11) .80
Chylothorax 4.8 (3) 6.1 (5) >.999
Time to first ED visit, d 20 (5-42) 17 (11-72) .19

Values are percent (n) or median (interquartile range), unless otherwise indicated.
CCI, Charlson Comorbidity Index; ED, emergency department.

After patients presented to the ED beyond three times, 66 of 100 patients were unnecessarily subjected to a
readmission for feeding tube issues occurred in the ma- risky procedure.15 Furthermore, during those first 8
jority of patients. As such, addressing feeding tube issues postoperative days, they found that only 48% of the
in an upfront and organized outpatient fashion can pro- target nutrition was delivered through the feeding tube
vide an opportunity to decrease both repeated ED utili- owing to feeding tube cessations in response to
zation and eventual inpatient admission. For instance, abdominal fullness, distension, pain, or diarrhea.15
many of these repeat ED visits for feeding tube issues Their findings, in addition to other literature, prompted
were over a short period within the first 2 months of us to engage in selective rather than routine use of
discharge home. feeding jejunostomy tubes in our later experience. How-
Although routine insertion of feeding tubes has been ever, it is important to note that some patients require
a standard component of esophagectomy, reevaluation feeding tubes as a lifeline, and those tend to be the pa-
of this practice may be warranted. Several studies have tients who come in repeatedly and then eventually end
questioned whether routine insertion of feeding tubes up admitted. They are also often the patients who could
after esophagectomy is necessary.13-15 Srinathan and be possibly targeted for selective intraoperative feeding
associates15 documented that a significant number of tube insertion or reactive postoperative placement.
patients neither require nor benefit from feeding tubes Therefore, it is important to note that eliminating routine
and also reported significant complications related to feeding tube placement will certainly not eliminate the
feeding tubes. In their study, they found that 66% of need for one in some patients.
patients were able start an oral full fluid diet by post- As the number of times a patient visited the ED
operative day 8 and that 6% of patients had feeding increased, so too did the likelihood of converting to an
tube-related complications (of which half required inpatient admission, regardless of the reason for ED uti-
operative intervention).15 Therefore, they inferred that lization. The reason for this is difficult to determine.
Ann Thorac Surg KIDANE ET AL 385

GENERAL THORACIC
2021;112:379-86 READMISSION AFTER ESOPHAGECTOMY

Table 2. Factors Associated With Conversion of Emergency institution and our health care system and resources.
Department Visit to Readmissiona Patients may require an inpatient admission to facilitate
Risk Factor AOR 95% CI P Value endoscopic interventions. Sometimes, patients also wait
24 hours or more to actually get their urgent endos-
Year of surgery (reference 0.89 (0.42-1.86) .75 copy. Rapid access to urgent outpatient endoscopy
1999-2006) would reduce the need for inpatient admission and
Age at time of surgery, per 0.99 (0.96-1.02) .54 perhaps even need for the ED visit. Such an approach
year increase
would greatly enhance patient quality of life and
Male 1.21 (0.52-2.78) .66
satisfaction; it would also reduce the high and unpre-
Anastomotic leak 2.45 (1.00-6.01) .05
dictable costs associated with repeat ED visits and
Living outside greater 0.13 (0.03-0.63) .01 readmissions. Our findings show no difference in rela-
metropolitan area
tive proportions of esophagectomy type between those
Time to first emergency 1.00 (0.99-1.00) .24
department visit who do and do not have readmissions after ER visits
Presence of social support 2.00 (0.80-4.98) .14
(P ¼ .86; Table 1). Our previous work also showed no
difference in relative proportions of esophagectomy
English speaking 1.03 (0.29-3.66) .96
type between patients who do and patients who do not
a
Results of multivariable regression analyses showed factors have ER visits (P ¼ .98).10
independently associated with conversion of emergency department Living in regions further away from the hospital was
visit to readmission.
associated with a lower rate of conversion to inpatient
AOR, adjusted odds ratio; CI, confidence interval. admission. However, a major caveat is that our current
study only captures ED visits to our hospital and does not
capture local or regional hospital visits. Although we
Although there is little evidence addressing this issue, it is previously showed that living in regions farther away
likely that clinicians are more inclined to readmit patients from the index hospital is associated with lower incidence
who have shown a propensity to repeatedly come back to of ED utilization at the index hospital, we hypothesized
the ED for the same reason; there are likely both medical that these patients may not actually be using the ED less,
and social reasons for this. Previous studies have reported but are instead utilizing their local hospital ED resources.
that ED revisit rates are highest among patients who are We believe that our current finding is also a reflection of
of younger age or lower socioeconomic status.16 Devel- capturing only the ED visits and readmissions at the in-
oping and implementing specific strategies in this patient dex hospital. This hypothesis is supported by Stitzenberg
population to decrease the incidence of repeat ED visits, and colleagues,17 who found that patients who lived
and thus inpatient admission, may be of benefit to the farther away from the index hospital had higher rates of
health care system. Our current results show no differ- both ED visits and readmission overall outside the index
ence between patients with and patients without mini- hospital. Higher overall rates of readmission in this
mally invasive esophagectomy in conversion to population may be related to these patients utilizing ED
readmission from ED visits. Our previous study showed resources at a center different from the index hospital.
that having a minimally invasive component in the That was indeed highlighted by a previous population-
esophagectomy was independently associated with based study of patients in Ontario; when one is able to
higher odds of having any ED visits but not frequent ED track all ED visits across all hospitals in the province,
visits.10 Patients with minimally invasive surgery were rural esophagectomy patients are more likely to have
usually discharged earlier and that may have resulted in higher incidence of ED visits but not necessarily read-
increased ED visits.10 Our current study shows that these missions.11 That results in the fragmentation of care as
visits did not appear to lead to more readmissions. additional providers unfamiliar with the patient, pro-
Having an anastomotic leak was independently cedure, and treatment plan are added to the care team. In
associated with higher odds of conversion from an ED light of this, appropriate patient education regarding
visit to inpatient admission. That could be as a result of postoperative care, telehealth utilization, and continued
delayed presentations of leak or strictures after leaks; or communication between the patient’s surgeon and the
it could also be that patients with leaks may also be patient or new care team members may be an important
patients with higher comorbidity burden who are area of focus in this patient population.
therefore more likely to have more frequent ED visits A limitation of this study is that it is based on retro-
and readmissions. Although our multivariable analyses spective data. Although data were collected prospectively,
do not suggest that to be the case, there is the possi- this study is subject to the potential biases of retrospective
bility of residual confounding. Anastomotic leaks are analysis. A significant limitation is that we only focused
known to increase the risk of the development of on ED visits at our index hospital. Therefore, we likely
strictures and thus dysphagia. This process often results missed ED visits at other hospitals closer to the patients’
in an increased need for clinical visits or endoscopic homes. Wider analyses with the ability to capture ED
dilations, which may ultimately necessitate readmission. visits at other, non-index hospitals are necessary to un-
Although patients requiring urgent endoscopic dilation derstand the true burden of ED use after esophagectomy,
for their dysphagia are likely to present to the ED, their and we are exploring this. In addition, this study does not
admission in our center may be a function of the include visits to primary care physicians or walk-in
 386 KIDANE ET AL Ann Thorac Surg
GENERAL THORACIC

READMISSION AFTER ESOPHAGECTOMY 2021;112:379-86

clinics. However, despite this limitation, our current 3. Atkins BZ, Shah AS, Hutcheson KA, et al. Reducing hospital
study allows us to achieve our more important objective, morbidity and mortality following esophagectomy. Ann
Thorac Surg. 2004;78:1170-1176 [discussion: 1170-1176].
which is to understand the causes of ED use with 4. Alanezi K, Urschel JD. Mortality secondary to esophageal
appropriate granularity so that we may design in- anastomotic leak. Ann Thorac Cardiovasc Surg. 2004;10:71-75.
terventions and solutions to reduce them. 5. Blewett CJ, Miller JD, Young JE, Bennett WF, Urschel JD.
In conclusion, although a large proportion of esoph- Anastomotic leaks after esophagectomy for esophageal
cancer: a comparison of thoracic and cervical anastomoses.
agectomy patients utilize the ED after their procedure, the
Ann Thorac Cardiovasc Surg. 2001;7:75-78.
majority do not convert to inpatient admission. Although 6. Trudel JG, Sulman J, Atenafu EG, Kidane B, Darling GE.
feeding tube problems are the most common reason for Longitudinal evaluation of trial outcome index scores in
ED visits, they are among the least likely reason for patients with esophageal cancer. Ann Thorac Surg. 2016;102:
conversion of ED visit to inpatient admission. Addressing 269-275.
7. Abbott DE, Gaitonde SG, Hanseman DJ, et al. Resource
feeding tube management and considering whether the utilization in esophagectomy: when higher costs are associ-
utilization of feeding tubes in certain patients is appro- ated with worse outcomes. Ann Surg Oncol. 2014;112(suppl):
priate appears to provide an opportunity to decrease both S34.
repeated ED utilization and eventual inpatient admission. 8. Chen SY, Molena D, Stem M, Mungo B, Lidor AO. Post-
discharge complications after esophagectomy account for high
As the number of times a patient visits the ED increases,
readmission rates. World J Gastroenterol. 2016;22:5246-5253.
so, too, does the likelihood of converting to an inpatient 9. Shah SP, Xu T, Hooker CM, et al. Why are patients being
admission, regardless of the reason for ED utilization. readmitted after surgery for esophageal cancer? J Thorac
Having an anastomotic leak was associated with higher Cardiovasc Surg. 2015;149:1384-1391.
odds of conversion to inpatient, and that was likely 10. Kidane B, Kaaki S, Hirpara DH, et al. Emergency department
use is high after esophagectomy and feeding tube problems
related to the development of strictures that cause are the biggest culprit. J Thorac Cardiovasc Surg. 2018;156:
dysphagia or delayed presentations of leak. Rapid access 2340-2348.
to urgent outpatient endoscopy may help to reduce the 11. Kidane B, Jacob B, Gupta V, et al. Medium and long-term
incidence of ED visits and conversion to admission in this emergency department utilization after oesophagectomy: a
population-based analysis. Eur J Cardiothorac Surg. 2018;54:
patient population. Living in regions further away from
683-688.
the index hospital is associated with a lower rate of 12. Dunlop S, Coyte PC, McIsaac W. Socio-economic status and
inpatient admission to the index hospital but is likely due the utilisation of physicians’ services: results from the Ca-
to patients utilizing closer local resources. Improving nadian National Population Health Survey. Soc Sci Med.
postoperative education and implementing strategies that 2000;51:123-133.
13. Fenton JR, Berferon EJ, Coello M, Welsh RJ,
allow for patient and care team communication with the
Chmielewski GW. Feeding jejunostomy tubes placed during
surgeon may improve outcomes in this population. esophagectomy: are they necessary? Ann Thorac Surg. 2011;92:
504-511.
14. Wheble GA, Benson RA, Khan OA. Is routine postoperative
This work was supported by the Kress Family Chair in Esopha- enteral feeding after oesophagectomy worthwhile? Interact
geal Cancer. Cardiovasc Thorac Surg. 2012;15:709-712.
15. Srinathan SK, Hamin T, Walter S, Tan AL, Unruh HW,
Guyatt G. Jejunostomy tube feeding in patients undergoing
esophagectomy. Can J Surg. 2013;56:409-414.
16. Steiner C, Barrett M, Hunter K. Hospital readmissions and
References multiple emergency department visits, in selected states,
2006-2007: statistical brief #90. In: Healthcare Cost and Uti-
1. Ajani JA, D’Amico TA, Almhanna K, et al. Esophageal and lization Project (HCUP) Statistical Briefs. Rockville, MD:
esophagogastric junction cancers, version 1.2015. J Natl Agency for Healthcare Research and Quality; 2006:1-10.
Compr Cancer Netwk. 2015;13:194-227. Available at: https://pubmed.ncbi.nlm.nih.gov/21413203/.
2. Kidane B, Coughlin S, Vogt K, Malthaner R. Preoperative 17. Stitzenberg KB, Chang YK, Smith AB, Nielsen ME. Exploring
chemotherapy for resectable thoracic esophageal cancer. the burden of inpatient readmissions after major cancer
Cochrane Database Syst Rev. 2015(5):CD001556. surgery. J Clin Oncol. 2015;33:455-464.