Available online at www.sciencedirect.com
ScienceDirect EJSO 43 (2017) 330e336
Costs and resource use following defunctioning stoma in low anterior resection for cancer e A long-term analysis of a randomized multicenter trial H. Floodeen a,b,*, O. Hallb€ o€ ok c,d, L.A. Hagberg e, P. Matthiessen a,b € € Department of Surgery, Orebro University Hospital, S€odra Grev Rosengatan, 70185 Orebro, Sweden b Department of Surgery, Faculty of Medicine and Health, School of Health and Medical Sciences, € € Orebro University, Fakultetsgatan 1, 70182 Orebro, Sweden c Department of Surgery, Link€oping University Hospital, 58183 Link€oping, Sweden d Department of Clinical and Experimental Medicine, Link€oping University, 58183 Link€oping, Sweden e € € Center for Health Care Science, Orebro County Council, 70182 Orebro, Sweden a
Accepted 6 December 2016 Available online 21 December 2016
Abstract Background: Defunctioning stoma in low anterior resection (LAR) for rectal cancer can prevent major complications, but overall costeffectiveness for the healthcare provider is unknown. This study compared inpatient healthcare resources and costs within 5 years of LAR between two randomized groups of patients undergoing LAR with and without defunctioning stoma. Method: Five-year follow-up of a randomized, multicenter trial on LAR (NCT 00636948) with (stoma; n ¼ 116) or without (no stoma; n ¼ 118) defunctioning stoma comparing inpatient healthcare resources and costs. Unplanned stoma formation, days with stoma, length of hospital stay, reoperations, and total associated inpatient costs were analyzed. Results: Average costs were V 21.663 per patient with defunctioning stoma and V 15.922 per patient without defunctioning stoma within 5 years of LAR, resulting in an average cost-saving of V 5.741. There was no difference between groups regarding the total number of days with any stoma (stoma ¼ 33 398 vs. no stoma ¼ 34 068). The total number of unplanned reoperations were 70 (no stoma) and 32 (stoma); p < 0.001. In the group randomized to no stoma at LAR, 30.5% (36/118) required an unplanned stoma later. Conclusion: Randomization to defunctioning stoma in LAR was more expensive than no stoma, despite the cost-savings associated with a reduced frequency of anastomotic leakage. Both groups required the same total number of days with a stoma within five years of LAR. Ó 2016 Elsevier Ltd, BASO ~ The Association for Cancer Surgery, and the European Society of Surgical Oncology. All rights reserved.
Keywords: Rectal cancer; Defunctioning stoma; Costs and cost analysis; Health resources; Anastomotic leak
Introduction Surgical treatment of low rectal cancer with preserved intestinal continuity often involves the routine formation of a defunctioning stoma (DS). DS has been shown to reduce the risk of anastomotic leakage (AL), one of the most feared postoperative complications following rectal cancer surgery.1 AL often leads to extended hospital stay € * Corresponding author. Department of Surgery, Orebro University Hos€ pital, S€ odra Grev Rosengatan, 70185 Orebro, Sweden. Fax: þ46 19 6021970. E-mail addresses: [email protected]
(H. Floodeen), [email protected]
regionostergotland.se (O. Hallb€o€ok), [email protected]
(L.A. Hagberg), [email protected]
and increased treatment costs for the healthcare provider,2,3 and prevention of this serious postoperative complication by means of a DS has been assumed to be cost-effective. However, temporary stoma formation requires an additional hospital admission and operation for planned stoma reversal, which accounts for a considerable additional cost beyond the initial cancer treatment.4 In times of fast medical advancement and tightened healthcare budgets, healthcare providers may be faced with a necessity for more careful cost analysis and decision-making around different treatment options. The present study aimed at evaluating the associated inpatient healthcare resources necessary during the first 5 postoperative years in patients who underwent low anterior
http://dx.doi.org/10.1016/j.ejso.2016.12.003 0748-7983/Ó 2016 Elsevier Ltd, BASO ~ The Association for Cancer Surgery, and the European Society of Surgical Oncology. All rights reserved.
H. Floodeen et al. / EJSO 43 (2017) 330e336
resection (LAR) for rectal cancer, dependent on whether patients were randomized to receive a defunctioning stoma or no stoma. Patients and methods
stapler line, recto-vaginal fistula, or pelvic abscess upon clinical assessment. Leakage was verified by digital examination, inspection of the wound, assessment of drain contents, rectoscopy and/or radiological investigations without time limit. Radiologically demonstrated leakage without clinical symptoms was not included.
Study design and population The present study formed a secondary end point of the multicenter trial Rectal Cancer Trial on Defunctioning Stoma (RECTODES, NCT 00636948), and aimed at comparing the necessary healthcare resources during the first 5 postoperative years following LAR for rectal cancer, with regard to whether patients were randomized to a defunctioning stoma (DSþ) or no defunctioning stoma (DS) at initial surgery. The study population consisted of 234 patients who underwent LAR according to the RECTODES trial protocol. Patients were recruited from 21 Swedish hospitals performing rectal cancer surgery during the study period from December 1999 to June 2005. Preoperative inclusion criteria were adenocarcinoma of the rectum, planned LAR, and informed consent. Intraoperative inclusion criteria were anastomosis 7 cm above the anal verge, negative intraoperative air leakage test, intact anastomotic stapler rings, and the absence of major adverse events during the initial surgery, as judged by the surgeon. If no exclusion criteria were present, the patients were intraoperatively randomized to receive a DS or not by means of a sealed envelope. 116 patients were randomized to DSþ, while 118 patients were randomized to DS at initial surgery. Stoma reversal was planned within 4 months of rectal resection. Permission for the study was granted by the regional ethics committee of each participating hospital (Table 1). Definition of anastomotic leakage Anastomotic leakage (AL) was defined as postoperative signs of peritonitis caused by leakage from any surgical
Evaluation of healthcare resources during the first 5 postoperative years All patient charts were assessed concerning the first 5 postoperative years (1825 days) starting on the day of the initial surgery of each patient. All hospital admissions and abdominal operations related to the initial rectal cancer surgery were evaluated. Outpatient contacts were not recorded. All data was recorded prospectively at each of the participating hospitals. The analysis of clinical variables included the total number of days with any stoma in the two randomized groups within 1825 days, number of days with a stoma divided by stoma type, number of hospital admissions and length of stay, requirement of intensive care, and number of planned and unplanned abdominal reoperations. Associated inpatient costs for all patients within 1825 days were analyzed, not including the costs for the initial rectal cancer surgery. Rectal examinations with or without general anesthesia were not included in the cost analysis. Subgroup analyses were conducted for patients with symptomatic anastomotic leakage (ALþ) and without symptomatic anastomotic leakage (AL). Evaluation of costs Electronic cost registration for patients treated at our institution was introduced in 2006, and actual costs for patients treated at the other participating institutions were not available. Thus the true costs associated with the actual patient population could not be evaluated retrospectively. For an approximation of costs, mean costs for hospital admissions based on relevant surgical procedure codes
Table 1 Study population demography and perioperative details.
Age at LAR, median (range) Female gender, % (n) BMI, median (range) ASA score 1 or 2, % (n) Tumor level above anal verge, cm, median (range) Preoperative radiotherapy, % (n) Level of anastomosis, cm, median (range) Operation time, minutes, median (range) Intraoperative bleeding, ml, median (range) Symptomatic anastomotic leakage, % (n) Number of patients alive after 5 years, % (n) a
Reduced numbers due to missing data.
Stoma (n ¼ 116)
No stoma (n ¼ 118)
68 (32e86) 40% (46/116) 25 (19e36) 89% (91/102)a 10 (4e15) 81% (94/116) 5 (2e7) 220 (110e605) 550 (50e4500) 10.3% (12/116) 69.0% (80/116)
67.5 (43e84) 51% (60/118) 25 (21e37) 83% (84/101)a 10 (3e15) 77% (91/118) 5 (2e7) 200 (100e541) 550 (50e2500) 28.0% (33/118) 73.7% (87/118)
NS p¼ NS NS NS NS p¼ p¼ NS p< NS
0.009 0.017 0.001
H. Floodeen et al. / EJSO 43 (2017) 330e336
were obtained from the local patient registry for cost registration starting in 2006. For comparison, mean hospital admission costs for the same surgical procedure codes were retrieved from the national database for the year 2014 of the Swedish Association of Local Authorities and Regions (SALAR, Sveriges Kommuner och Landsting, www.skl.se). Every surgical procedure in the patient population was then assigned a total mean cost representing one hospital admission. Hospital admissions without surgical intervention were calculated based on the average per diem cost of hospital admission, obtained from the local patient registry in 2014. Less frequent abdominal reoperations, where representative mean costs could not be obtained, were assigned the local mean cost for exploratory laparotomy. Costs were adjusted to match the price year 2014, assuming an average annual increase in healthcare costs of 2.57% between 2006 and 2014. All costs were converted from Swedish crowns (SEK) to Euros (V), based on the annual average exchange rate for 2014 (9.0968 SEK/V) published by the Swedish Central Bank (Sveriges Riksbank, www.riksbank.se).
to mean national costs for hospital admissions due to AL using the available, but unfortunately rarely used, surgical procedure code for AL in rectal cancer patients during 2014. Statistical analysis Descriptive statistics were used for the analysis of clinical outcomes of interest. The c2 test and Fisher’s exact test were used for categorical variables, when appropriate. Student’s t test and ManneWhitney U test were used for comparisons between groups regarding continuous variables. To test robustness of data, distribution-independent Bootstrap analysis was employed for comparison of costeffectiveness between groups. P values less than 0.05 were considered significant. Statistix version 10 (Analytical Software, Tallahassee, FL) and SPSS version 21 (SPSS Inc, Chicago, IL) software packages were used for all statistical calculations. Results Hospital admissions and length of stay
Evaluation of costs in patients with anastomotic leakage Estimation of costs for treatment of AL turned out to be complex. There is a lack of consistency in the usage of surgical procedure codes for diagnosis of or reoperation for AL. Furthermore, few patients presently undergoing LAR for rectal cancer do not receive a DS. Thus, abdominal reoperation for AL in patients without DS following LAR is uncommon, and the associated current costs could not be evaluated representatively. As a surrogate, mean hospital admission costs for similar representative scenarios were evaluated. These included local costs for rectal cancer patients undergoing high anterior resection without DS, who later required reoperation for AL, and patients undergoing LAR with DS, who later required reoperation due to AL following reversal of the temporary stoma. All patients were recruited from the local patient registry, regardless of surgical procedure code. Local costs were then compared
All patients combined in DSþ accounted for a total of 33,398 days with any type of stoma within 5 years, as compared with 34,068 days in DS (not significant). When excluding all patients who remained stoma-free during follow-up, the median number of days with any stoma was 196 (range 13e1825) in the DSþ (n ¼ 116), as compared with median 839 (range 98e1820) in DS (n ¼ 36; p < 0.001). DSþ required median 2 hospital admissions, and had a median 22 days of total time in hospital. The whole group accounted for 3070 days in hospital. DS required median 1 hospital admission, and had a median 12 days of total time in hospital. The whole group accounted for 2637 days in hospital (not significant) (Table 2). Overall 30day and 90-day readmission rates were 19.2% and 11.1%, respectively, resulting in a cumulative readmission rate of 30.3% within 90 days of discharge following LAR. There were no planned stoma reversals within 30 days, but 19
Table 2 Details regarding hospital admissions and stoma within 5 years of LAR (n ¼ 234).
Total number of days with stoma, whole group Total number of days in hospital, whole group Total number of days in intensive care unit, (n) Number of hospital admissions, median (range) Number of individuals with any stoma, % (n) Number of individuals with permanent stoma, % (n) Permanent loopileostomy, % (n) Permanent colostomy, % (n) a
Stoma (n ¼ 116)
No stoma (n ¼ 118)
33 398 3070 32 (9/116) 2 (1e5) 100% (116/116) 17.5% (20/114)a 8.8% (10/114)a 8.8% (10/114)a
34 068 2637 25 (9/118) 1 (1e5) 30.5% (36/118) 22.0% (26/118) 9.3% (11/118) 12.7% (15/118)
NS NS NS p < 0.001 p < 0.001 NS NS NS
2 patients randomized to defunctioning ileostomy excluded due to early death on postoperative day 29 and 43.
H. Floodeen et al. / EJSO 43 (2017) 330e336
patients (8.1%) were readmitted for stoma reversal within 90 days, which accounted for the majority of all 90-day readmissions. Nine patients in each group required intensive care during the follow-up period accounting for a total of 32 days in DSþ, and 25 days in DS (not significant). Stoma All patients randomized to DSþ (n ¼ 116) had a stoma at some point during the follow-up period, whereas 30.5% of all patients initially randomized to DS had a stoma at some point during the 5 years following LAR (n ¼ 36). Eighty-two patients remained stoma-free during the entire follow-up period. All patients were analyzed on an intention-to-treat basis according to the original trial protocol, although there were 3 cases of violation of the protocol (2 patients randomized to DS received a DS at LAR, and 1 patient randomized to DSþ remained free of stoma). In DSþ, 20 patients ended up with a permanent stoma, 10 of which maintained a loopileostomy as permanent stoma, while 10 received a permanent end colostomy. Two patients randomized to DSþ were excluded from this analysis due to death on day 29 and 43 during the early postoperative period. In DS, 26 patients ended up with a permanent stoma, 11 of which maintained a loopileostomy as permanent stoma, while 15 received a permanent end colostomy. There were no deaths in the early postoperative period in this group (Table 2).
Table 3 illustrates the history of the study population with regard to planned and unplanned stoma formation within the first 5 years of LAR. Eleven different scenarios were identified and grouped into categories. The number of days with a loopileostomy or colostomy present are shown for each separate category. On the group level, DS accounted for a total of 19,430 days with a loopileostomy and 14,638 days with a colostomy. DSþ accounted for 28,070 days with a loopileostomy and 5328 days with a colostomy. Planned and unplanned reoperations A total of 103 patients in DSþ underwent planned reversal of the loopileostomy after a median time of 6 months. There were a further 32 unplanned reoperations in DSþ, most commonly due to ileus (n ¼ 9), AL (n ¼ 5) and unsatisfactory anorectal function (n ¼ 5). DS required a total of 70 unplanned reoperations, most commonly due to AL (n ¼ 37), reversal of an urgently created loopileostomy (n ¼ 13), and ileus (n ¼ 6). Reasons for unplanned abdominal reoperations in both groups are listed in Table 4. Cost analysis The overall cost analysis based on local data revealed a mean cost of V 21.663 per patient with DSþ and V 15.922 per patient in DS within 5 years of LAR, resulting in an
Table 3 Stoma history including stoma type during the first 5 postoperative years for all randomized patients (n ¼ 234). n
days with loop
days with colostomy
Permanent end colostomy
Loop -> reversal
Loop -> becomes permanent loopileostomy
Loop -> conversion to permanent end colostomy
Loop -> reversal -> permanent end colostomy
Loop -> reversal -> new permanent loop
Randomization to no stoma
no stoma group total (36/118 with any stoma)
Randomization to stoma Reversal
Becomes permanent loopileostomy*
Reversal -> permanent end colostomy
Conversion to permanent end colostomy
stoma group total (116/116 patients with any stoma)
*2 patients randomized to defunctioning ileostomy deceased on postop day 29 and 43.
H. Floodeen et al. / EJSO 43 (2017) 330e336
Table 4 Reasons for unplanned abdominal reoperations within 5 years of LAR. Unplanned abdominal reoperation
No stoma n ¼ 118
Stoma n ¼ 116
Anastomotic leakage Reversal of loopileostomy Ileus Unsatisfactory anorectal function Incisional/parastomal hernia Wound rupture/revision Rectal stenosis Recurrent colorectal cancer Other P
37 13 6 5 (4)a 1 1 1 6 70
5 e 9 5 3 3 e 1 6 32
p < 0.001b p < 0.001b NS NS NS NS NS NS NS p ¼ 0.002c
a Hernia repair was performed as a combined procedure with another surgery listed. b Fischer’s exact test. c Chi2 test.
average cost-saving of V 5.741 per patient in DS (p ¼ 0.003; CI 95%: V 1.942e9.541). Two sensitivity analyses to test robustness of the above data were performed using Student’s t test and nonparametric Bootstrap analysis, which confirmed the reported findings. The first sensitivity analysis based on the available national data with a higher total cost for patients with AL revealed an average cost-saving of V 3.696 per patient in DS (p ¼ 0.098; CI 95%: V 693e8.084). Bootstrapping confirmed the expected cost-savings in DS with a probability of 96%. A second sensitivity analysis excluding all patients with a complex clinical course (n ¼ 22), where a retrospective cost assessment was difficult and implied a relative uncertainty revealed an average cost-saving of V 5.287 per patient in DS (p ¼ 0.002; CI 95%: V 1.930e8.644). Bootstrapping confirmed the expected cost-savings in DS with a probability of 100%. Mean local costs for hospital admissions based on representative surgical procedure codes revealed an average cost of V 23.878 (range V 15.129e38.448) and a median hospital admission time of 16 days (range 12e32) per patient with AL. Mean local costs for a hospital admission for planned reversal of a DS with and without intestinal resection were V 9.912 and V 9.712, respectively. Average per diem costs for hospital admission based on local data were V 528 (Table 5).
Cost analysis in patients with symptomatic anastomotic leakage There were 45 patients with symptomatic AL in the present study, 12 from the group initially randomized to DSþ, and 33 from DS. All patients combined with AL accounted for 34% of the total number of days in hospital. Patients with AL had a significantly longer hospital stay, both at initial admission and at all admissions combined within 5 years of LAR. Patients with AL who were randomized to DSþ had a shorter initial hospital stay by median 6 days, as compared with the group randomized to DS. Cost analysis revealed an average total cost of V 34.724 per patient with AL and DS, compared with V 30.120 per patient with AL and DSþ within 5 years of LAR. This difference of V 4.604 per patient was non-significant with the given sample size (p ¼ 0.37; CI 95%: V 5849e15.057). Sensitivity analysis using available national data was nearly significant and revealed a cost-saving of V 11.570 in patients with AL and DSþ, as compared with AL and DS (p ¼ 0.057; CI 95%: V 395e23.534). Discussion The overall cost analysis in this study revealed significantly higher treatment costs for patients randomized to DSþ, regardless of the possible cost-savings associated with the reduced frequency of symptomatic AL in these patients. Total cost-savings per patient in DS amounted to V 5.741 within a 5-year period. AL proved to be an important factor influencing overall costs for the healthcare provider. Patients with AL in this study accounted for one third of the total hospital stay, and half of all unplanned reoperations. Within the subgroup of patients with AL, the cost-savings associated with DSþ were non-significant given the present sample size. Considering the results of the sensitivity analysis, it is reasonable to assume that DSþ entails significant cost-savings in patients with AL when evaluating a larger sample size. The DS has been shown to decrease the rate of symptomatic AL, and to mitigate the effects of symptomatic AL when it occurs.5,6,1 AL is considered to be the most expensive complication of LAR with regard to necessary healthcare
Table 5 Mean hospital admission costs and length of stay associated with surgical procedure codes (total costs in V, median length of stay in days).
Reoperation for anastomotic leakage Reversal of loopileostomy without intestinal resection Reversal of loopileostomy with intestinal resection Ileus Exploratory laparotomy Mean per diem cost of hospital admission (2014) a
All costs inflation-adjusted for the year 2014.
Local patient registrya 2006e2014
National data for 2014
Length of stay (median)
Length of stay (median)
23.878 9.712 9.912 12.704 21.597 528
16 6 7 9 16 1
32.762 8.262 9.963 21.717 18.387 e
27 3 3 14 12 e
H. Floodeen et al. / EJSO 43 (2017) 330e336
resources.7,2,8 The rationale behind the use of the DS in this setting is lowering these costs and improving the associated morbidity and mortality for patients with AL. However, the DS entails its own morbidity and mortality, and thus the necessity of further healthcare resources.9,4 All patients with a DS who underwent stoma reversal required a second hospital admission and planned reoperation for stoma reversal. This group of patients, who did not suffer from AL, accounted for almost half of the total hospital stay and half of the necessary reoperations in this study. Additionally, stoma creation is an independent risk factor for unplanned readmission.10e12 The present randomized trial demonstrated that the two groups of patients required a similar total number of days with a stoma within 5 years of LAR, which the authors found to be an unexpected finding. Understandably, all patients in the group randomized to DSþ shared the burden of stoma treatment, which was reflected in the median time with a stoma of 196 days across the group. In the group randomized to DS, only 30% of all patients accounted for the total number of days with a stoma, which resulted in a longer median time with a stoma of 839 days. The other 70% of the patients remained stoma-free during the entire follow-up time. Although the Health-related Quality of Life (QoL) between individual patients may have differed considerably, the associated costs for stoma care from the point of view of the healthcare provider during the first 5 postoperative years can be assumed to be similar across the two randomized groups. There was a difference observed between the total number of days spent with a loopileostomy and a colostomy, respectively, in the two randomized groups. In the group randomized to DS, days with a colostomy accounted for 43% of all days spent with a stoma within 5 years of LAR. In DSþ, this number was 16%. As the clinical manifestation of AL is generally more dramatic in patients without a DS, the observed difference can be explained by the necessity of definitive measures in the face of this feared complication. Surgeons treating a patient with AL, and without the protective DS, may have chosen to create a permanent colostomy more readily during the clinical course. Days with a planned loopileostomy accounted for the majority of days with a stoma in the group randomized to DSþ. There is no consensus in the current literature as to the optimal timing of ileostomy reversal, but reversal is generally considered safe around 8e12 weeks postoperatively.13e15 Median time to reversal of the DS in the present study was 6 months. Additionally to possibly enhancing the individual patient’s QoL, there are large potential cost-savings in reducing the amount of time spent with a DS by planning and performing stoma reversal as early as feasible after LAR.16e18 Current practice in Sweden involves the routine formation of a DS in LAR in more than 90% of cases.19 The present cost analysis implies that such a high percentage of DS in LAR cannot be assumed to be cost-effective. The
challenge lies in identifying those patients at greatest risk to develop AL, and further research is necessary to determine which patients will benefit most from a DS. An issue of interest in this context is the Health-related QoL in patients with a stoma. The current literature is ambivalent concerning the impact of temporary and permanent stomas on QoL,20e22 and there is a lack of reliable data regarding long-term QoL in patients with AL. According to current policy in Sweden, a threshold of SEK 500.000 (wV 55.000) per gained QALY is employed for decisions about subsidized medicine.23 This is in line with current accepted threshold ranges in the UK (£ 20.000e30.000 per gained QALY) and the US (US$ 50.000e100.000 per gained QALY).24,25 Thus, an average gained 0.1 QALYs per patient would be necessary to compensate for the higher treatment costs associated with the routine formation of a DS in LAR. Strengths of the present trial were the randomized study design, regardless of patient predisposition to develop postoperative complications, the long and exactly defined follow-up time, and the size of the study population. To our knowledge, there is no other published cost analysis regarding DS in LAR in a randomized study population. A weakness of the study was the lack of available cost data for the actual patient population, necessitating the estimation of representative treatment costs based on local and national cost data for 2014. To balance this uncertainty, sensitivity analysis was performed using more conservative cost estimates, which yielded similar results. Conclusion Randomization to defunctioning stoma following low anterior resection was overall more expensive than no defunctioning stoma, despite the cost-savings associated with a reduced frequency of anastomotic leakage. Conflict of interest statement None declared.
Acknowledgments € This work was supported by a grant from the Orebro € County Council, Orebro, Sweden (OLL-326961, OLL500101).
References 1. Matthiessen P, Hallbook O, Rutegard J, et al. Defunctioning stoma reduces symptomatic anastomotic leakage after low anterior resection of the rectum for cancer: a randomized multicenter trial. Ann Surg 2007; 246(2):207–14. 2. Ashraf SQ, Burns EM, Jani A, et al. The economic impact of anastomotic leakage after anterior resections in English NHS hospitals: are we adequately remunerating them? Colorectal Dis 2013;15(4):e190–8.
H. Floodeen et al. / EJSO 43 (2017) 330e336
3. Frye J, Bokey EL, Chapuis PH, et al. Anastomotic leakage after resection of colorectal cancer generates prodigious use of hospital resources. Colorectal Dis 2009;11(9):917–20. 4. Macafee DA, West J, Scholefield JH, et al. Hospital costs of colorectal cancer care. Clin Med Oncol 2009;3:27–37. 5. Dehni N, Schlegel RD, Cunningham C, et al. Influence of a defunctioning stoma on leakage rates after low colorectal anastomosis and colonic J pouch-anal anastomosis. Br J Surg 1998;85(8):1114–7. 6. Alberts JC, Parvaiz A, Moran BJ. Predicting risk and diminishing the consequences of anastomotic dehiscence following rectal resection. Colorectal Dis 2003;5(5):478–82. 7. Koperna T. Cost-effectiveness of defunctioning stomas in low anterior resections for rectal cancer: a call for benchmarking. Arch Surg 2003; 138(12). 1334e1338; discussion 9. 8. Hammond J, Lim S, Wan Y, et al. The burden of gastrointestinal anastomotic leaks: an evaluation of clinical and economic outcomes. J Gastrointest Surg 2014;18(6):1176–85. 9. Chow A, Tilney HS, Paraskeva P, et al. The morbidity surrounding reversal of defunctioning ileostomies: a systematic review of 48 studies including 6,107 cases. Int J Colorectal Dis 2009;24(6):711–23. 10. Wick EC, Shore AD, Hirose K, et al. Readmission rates and cost following colorectal surgery. Dis Colon Rectum 2011;54(12):1475–9. 11. Damle RN, Cherng NB, Flahive JM, et al. Clinical and financial impact of hospital readmissions after colorectal resection: predictors, outcomes, and costs. Dis Colon Rectum 2014;57(12):1421–9. 12. Damle RN, Alavi K. Risk factors for 30-d readmission after colorectal surgery: a systematic review. J Surg Res 2016 Jan;200(1):200–7. 13. Perez RO, Habr-Gama A, Seid VE, et al. Loop ileostomy morbidity: timing of closure matters. Dis Colon Rectum 2006;49(10):1539–45. 14. den Dulk M, Smit M, Peeters KC, et al. A multivariate analysis of limiting factors for stoma reversal in patients with rectal cancer entered into the total mesorectal excision (TME) trial: a retrospective study. Lancet Oncol 2007;8(4):297–303.
15. Sier MF, van Gelder L, Ubbink DT, et al. Factors affecting timing of closure and non-reversal of temporary ileostomies. Int J Colorectal Dis 2015;30(9):1185–92. 16. Yau T, Watkins D, Cunningham D, et al. Longitudinal assessment of quality of life in rectal cancer patients with or without stomas following primary resection. Dis Colon Rectum 2009;52(4):669–77. 17. Floodeen H, Lindgren R, Matthiessen P. When are defunctioning stomas in rectal cancer surgery really reversed? Results from a population-based single center experience. Scand J Surg 2013;102(4):246–50. 18. Robertson J, Linkhorn H, Vather R, et al. Cost analysis of early versus delayed loop ileostomy closure: a case-matched study. Dig Surg 2015; 32(3):166–72. 19. Swedish Rectal Cancer Registry (Nationellt kvalitetsregister f€or Cancer recti 2012). The National Board of Health and Welfare. (www.oc. umu.se, Accessed 22 March 2016). 20. Tsunoda A, Tsunoda Y, Narita K, et al. Quality of life after low anterior resection and temporary loop ileostomy. Dis Colon Rectum 2008; 51(2):218–22. 21. Bloemen JG, Visschers RG, Truin W, et al. Long-term quality of life in patients with rectal cancer: association with severe postoperative complications and presence of a stoma. Dis Colon Rectum 2009;52(7):1251–8. 22. Pachler J, Wille-Jorgensen P. Quality of life after rectal resection for cancer, with or without permanent colostomy. Cochrane Database Syst Rev 2012;12:Cd004323. 23. The Pharmaceutical Benefits Agency: Allm€anna r ad om ekonomiska utv€arderingar (The Pharmaceutical Benefits Agency General Guidelines for Economic Evaluations); 2003. Stockholm, May. 24. Shiroiwa T, Sung YK, Fukuda T, et al. International survey on willingness-to-pay (WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ 2010;19(4):422–37. 25. Hagberg LA, Brekke HK, Bertz F, et al. Cost-utility analysis of a randomized controlled weight loss trial among lactating overweight/obese women. BMC Public Health 2014;14:38.