Nonoperative management of adhesive small bowel obstruction: what is the break point?

Accepted Manuscript Non-operative Management of Adhesive Small Bowel Obstruction – What is the Break Point? Alexander L. Colonna, MD, FACS, Nickolas R. Byrge, MD, Scott D. Nelson, PharmD, MS, Richard E. Nelson, PhD, Michael C. Hunter, MD, Raminder Nirula, MD, MPH, FACS PII:

S0002-9610(16)30625-0

DOI:

10.1016/j.amjsurg.2016.09.037

Reference:

AJS 12108

To appear in:

The American Journal of Surgery

Received Date: 18 March 2016 Revised Date:

10 September 2016

Accepted Date: 14 September 2016

Please cite this article as: Colonna AL, Byrge NR, Nelson SD, Nelson RE, Hunter MC, Nirula R, Nonoperative Management of Adhesive Small Bowel Obstruction – What is the Break Point?, The American Journal of Surgery (2016), doi: 10.1016/j.amjsurg.2016.09.037. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ABSTRACT

RI PT

Background: The current management paradigm for recurrent adhesive small bowel obstruction (SBO) is nonoperative. Rates of recurrence differ based upon time interval between and number of previous occurrences. Optimal time to intervene has not been determined.

SC

Methods: We constructed a Markov model to evaluate costs and quality of life on a hypothetical cohort of 40 y.o. patients after their first episode of medical management for post-operative SBO. We estimated a relative risk reduction of .55 with surgical intervention and a relative risk increase of 2.1, 2.9, and 5.7 after the medical management of the 2nd, 3rd, and 4th SBO.

M AN U

Results: Surgery performed after earlier episodes of SBO was more costly but also more effective. The cost difference between surgery after the 1st SBO recurrence versus the 2nd SBO recurrence was $1,643, with an increase of 0.135 quality adjusted life years (QALYs), the incremental costeffectiveness ratio (ICER) was $12,170/QALY.

AC C

EP

TE D

Conclusion: Surgery after the first episode of SBO provides a small increase in QALY at a small cost since surgical intervention lowers the risk of recurrence.

ACCEPTED MANUSCRIPT

SUMMARY

AC C

EP

TE D

M AN U

SC

RI PT

This paper uses Markov analysis to examine the cost effectiveness of different strategies used to manage adhesive small bowel obstructions. Earlier intervention produces a quality of life benefit at slightly increased cost.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

KEYWORDS cost effectiveness; small bowel obstruction; adhesions; quality adjusted life years

ACCEPTED MANUSCRIPT

Nickolas R. Byrge, MDA University of Utah Department of Surgery 30 N 1900 East Salt Lake City, UT 84132 [email protected]

SC

Corresponding Author: Alexander L. Colonna, MD, FACSA A University of Utah Department of Surgery 30 N 1900 East Salt Lake City, UT 84132 [email protected] Office: (801) 581-4607 Fax: (801) 587-9370

TE D

EP

Richard E. Nelson, PhDC C University of Utah School of Medicine 30 N 1900 East Salt Lake City, UT 84132 [email protected]

M AN U

A

Scott D. Nelson, PharmD, MSB B VA Salt Lake City Health Care System 500 Foothill Drive Salt Lake City, UT 84148 [email protected]

AC C

Michael C. Hunter, MDA A University of Utah Department of Surgery 30 N 1900 East Salt Lake City, UT 84132 [email protected]

Raminder Nirula, MD, MPH, FACSA A University of Utah Department of Surgery 30 N 1900 East Salt Lake City, UT 84132 [email protected]

RI PT

Title: NON-OPERATIVE MANAGEMENT OF ADHESIVE SMALL BOWEL OBSTRUCTION – WHAT IS THE BREAK POINT?

ACCEPTED MANUSCRIPT

Introduction Acute small bowel obstructions (SBOs) caused by adhesions are a significant burden to patients, surgeons, and the health care system overall. A 2011 study estimated the yearly cost

RI PT

attributed to lysis of adhesions at close to $2.3billion a year.1 Approximately 93% of patients will develop adhesions after laparotomy, 1% of all general surgical admission can be attributed to adhesive disease, and 3.3% of all laparotomies performed for obstruction are due to

SC

adhesions.2 Morbidity and mortality have improved significantly over the last half century owing to advancements in fluid and electrolyte management, critical care, and radiologic

M AN U

imaging.3,4 These advancements, including the use of contrast study protocols, have allowed surgeons to more rapidly identify and treat patients whose adhesive bowel obstructions will not likely resolve without intervention.5,6 As 60 to 70% of patients with partial small bowel obstructions due to adhesions will resolve with non-operative management, conventional

TE D

wisdom dictates that intervening in these patients is not indicated.7,8 However, a significant portion of these patients require readmissions, some serially, and the probability that they will have another episode increases with each readmission.9,10

EP

The current management paradigm for adhesive SBO is nonoperative unless there is evidence of bowel compromise or resolution failure. Our group follows this paradigm as long as avoiding

AC C

repeat operation appears to be in the patient’s best interest. Rates of recurrence seem to differ based upon the time interval between occurrences and the number of previous occurrences. Determining the optimal time to operate requires weighing the morbidity, mortality, cost, and quality of life impact of operative versus medical management. To aid in this decision making process, we conducted an economic analysis comparing the costs and outcomes associated with performing surgery after the 1st, 2nd, 3rd, 4th, or 5th SBO recurrence.11 Our hypothesis was that

2

ACCEPTED MANUSCRIPT

after the 2nd admission for small bowel obstruction that resolves medically it becomes cost effective to intervene on the patient, lyse their adhesions, and decrease their risk for recurrence. Methods

RI PT

Overview

We constructed a Markov microsimulation model using TreeAgePro 2014 (TreeAge

Software, Inc., Williamston, MA). The model simulated a hypothetical cohort of patients

SC

(100,000 first-order simulations) 40 years old that have had a laparotomy and are post-medical treatment for an adhesion related SBO. We compared the costs and outcomes between various

M AN U

strategies of performing surgical management of a SBO after the 1st, 2nd, 3rd, 4th, or 5th recurrence, with medical treatment for the other recurrences. We assumed that patients would only undergo surgery once, and that subsequent SBOs would be managed with medical treatment. A selection of the model is shown in Figure 1.

TE D

Patients remained in the stable health state until they experienced a recurrence of a SBO, after which the patient transitioned to surgery or medical treatment for one cycle, incurred the costs and utility decrements, after which the patient transitioned back to stable, as shown in Figure 1.

EP

If a patient transitioned to surgery, there were additional risks of complication such as acute renal injury, deep vein thrombosis (DVT), ileus, myocardial infarction (MI), pulmonary embolism

AC C

(PE), pneumonia, surgical site infections (SSI), and death. Patients continued in the model until they died or reached the end of simulation, whichever came first. The model was run for 10 years in 2-month increments with a half-cycle correction. To test the model results for uncertainty in the parameter inputs, we conducted probabilistic sensitivity analyses where each input was simultaneously varied randomly over a given distribution and range of possible values. (1,000 first-order simulations with 1,000 second-order simulations)

3

ACCEPTED MANUSCRIPT

Input parameters Costs were incurred by medical treatment or surgery, with additional costs from major

reimbursement rates. All costs were adjusted to 2013 US dollars.

RI PT

surgical complications, as shown in Table 1. The costs in the model represent Medicare

Probabilities for events in the model are shown in Table 2. Probabilities of SBO recurrence were derived from a study by Fevang et al.12 Because most of the SBOs reoccurred in the first

SC

year, we used the first year probability of recurrence for the first year after SBO, then the

reported cumulative 10-year probability for years 2-10. Because the population was reported as a

M AN U

heterogeneous mix of patients that received medical or surgical treatment, we adjusted the probability of recurrence using the relative risk reduction of having surgery of 0.55.13 Furthermore, if a patient experienced a SBO recurrence, the relative risk increase of future SBO recurrences was increased as described by Fevang et al.12 Probabilities of surgical complications

TE D

were derived from the literature and the National Surgical Quality Improvement Program (NSQIP),14 and represented the incremental difference between surgery and conservative treatment. Probabilities of death while stable were derived from 2009 US life tables,15 with

EP

probabilities of death during surgery or medical treatment derived from Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS).16

AC C

Utilities for health states and utility decrements from adverse events are shown in Table 3. We assumed that as long as the patient was stable, they had an annual utility value of 1. We also assumed that utility decrement for acute renal injury was experienced for 1 month, DVT for 1 year, ileus for 1 month, MI for 5 years, PE for 1 year, pneumonia for 2 months, and SSI for 2 months. Outcomes

4

ACCEPTED MANUSCRIPT

We compared costs and effectiveness outcomes between the 5 strategies for managing a SBO recurrence. The cost outcomes represent Medicare costs. Effectiveness outcomes were measured as quality-adjusted life years (QALYs). QALY values were derived from the literature and

RI PT

represent the utility weight of various health states from 0 to 1, with 0 representing death and 1 representing perfect health. We then calculated the incremental cost-effectiveness ratio (ICER) by dividing the difference in cost between comparison strategies by the difference in

SC

effectiveness between comparison strategies. The ICER can be interpreted as the additional cost

and utilities were discounted at 3%.

M AN U

necessary to gain 1 additional QALY by using one strategy compared to another. Future costs

We assumed that patients would not undergo subsequent surgeries for SBO recurrence; however, in practice a few patients may have surgery again, depending on the number of recurrences and time between recurrences. We did evaluate this assumption, and found that our

Results

TE D

assumption did not affect the study results.

EP

Table 4 and Figure 2 show the results of the model simulation. Cost difference between surgery after 1st SBO recurrence versus 2nd SBO recurrence is $1,643, with an increase of 0.135

AC C

QALY (0.010 unadjusted life years), resulting in an ICER of $12,170/QALY. Well below the commonly cited $50,000/QALY threshold.33 These results suggest that the sooner surgery is done, the better for the treatment of SBO, as operating after the 1st recurrence was cost effective compared to operating after the 2nd, which was also more cost effective than operating after the 3rd. Figure 3 shows the cost acceptability curve from the probabilistic sensitivity analysis, which suggests that unless a decision maker’s

5

ACCEPTED MANUSCRIPT

willingness to pay is very low (as seen in Figure 3), the sooner the surgery, the better. Figure 4 shows the scatter plot for the ICER comparison between surgery after 1st SBO recurrence versus

M AN U

SC

RI PT

2nd SBO recurrence.

Discussion

The overwhelming surgical paradigm has been to defer surgical intervention in the case of

TE D

adhesive SBO unless clinical circumstances force the surgeons hand. If the patient is clinically stable bowel rest, enteric decompression, fluid and electrolyte management, and an appropriate

EP

period of waiting is successful in 65%-90% of cases.6 However, studies published in the last decade have suggested that patients with recurrent SBO have higher readmission rates and

AC C

progressively lower times to recurrence. In particular, a study published by Rocha et al. in 2009 showed that patients treated medically had a higher readmission rate (24% vs. 9%) and a shorter time to recurrence (39 days vs. 105 days). 34 Fevang et al. published the largest series of patients with recurrent adhesive small bowel disease in 2004. Their study is unique as they had follow up times of up to 40 years. Their group showed that while most recurrences happen in the first 5 years after the initial laparotomy, this risk persists across the lifetime of the patient and increases over time, while the time period between recurrences decreases.12

6

ACCEPTED MANUSCRIPT

Given that recurrent SBO presents as compounding problem we endeavored to challenge the current surgical paradigm of not intervening in these patients by showing that early surgery was actually a cost effective management strategy. We constructed a hypothetical cohort of adult

RI PT

patients that presented with SBO and modeled their outcomes while taking into account

complications of surgery, chances and rates of recurrence, and costs of these health states. To our knowledge our study is the first to examine operative versus medical management of recurrent

SC

small bowel obstructions.

This analysis shows that surgery after the first episode of SBO may be effective in the long

M AN U

run for a small gain in QALY since surgical intervention lowers the risk of recurrence even though earlier intervention is slightly more expensive. As the ICER of $12,170 per QALY is much lower than the generally accepted $50,000 per QALY therefore willingness to pay is not a factor.33

TE D

Our study has several limitations. A complex simulation is inherently limited by the quality of the data inputs and the applicability of the model itself. We performed a broad literature search and used acceptable proxies for data points like utilities that have not specifically been studied in

EP

adhesive small bowel disease. The major assumptions that we made were that the risk of recurrent SBO increased over time, that the time interval between recurrences decreased, and

AC C

that operative intervention reset the clock so to speak, decreasing this risk. This is in line with our interpretation of the current literature. A starting age of 40 years old was chosen specifically to minimize possible age related confounders, such as an 80-year-old on Coumadin. Older patients may have higher complication rates that we did not model with our younger cohort. While we included a robust selection of complications in our model our list only covered what we felt were the most common complications and those most likely to have an impact on QALY.

7

ACCEPTED MANUSCRIPT

We did not model every complication or permutation of pre-operaterative factors, such as location of obstruction, presence of an ostomy, or previous hernia repairs for example. As our list is not exhaustive we may have underestimated the true cost and rate of surgical complications.

RI PT

Despite these limitations inherent in constructing a computer model, our simulation showed that even when taking the cost increase and utility decrement associated with post-surgical complications that operative intervention was still a dominant strategy.

SC

Complication rates were mostly obtained from NSQIP data and as stated may underestimate rates. Therefore, probabilistic sensitivity analysis was used to gain an estimate of the uncertainty

M AN U

of the ICER estimates. The model randomly selects parameter estimates based on distributions and calculates an ICER using those estimates. It then randomly selects new parameter estimates based on distributions and calculates another ICER. This process is repeated a large number of times giving a large number of ICERs. The sensitivity analysis performed showed that our

TE D

results were valid over a wide range of inputs and that earlier surgery was cost-effective unless willingness to pay was low.

In conclusion, our study demonstrates that operative intervention earlier in the course of

EP

recurrent SBO is an effective strategy. The economic burden of recurrent disease is high as described in separate studies by Wilson et al. and Tingstedt et al. examining the burden of small

AC C

bowel adhesive disease in both the British and Swedish medical systems.8,9 Patients with recurrent adhesive small bowel disease are a complex cohort that require careful weighing of available interventions. As with most treatments in surgery, therapies should be tailored to the individual patient. Our research suggests that earlier surgical intervention may decrease costs and increase QALYs for this patient population.

8

ACCEPTED MANUSCRIPT

References

8. 9.

10.

11. 12. 13.

14. 15. 16.

17. 18. 19.

RI PT

SC

7.

M AN U

6.

TE D

3. 4. 5.

EP

2.

Sikirica V, Bapat B, Candrilli SD, Davis KL, Wilson M, Johns A. The inpatient burden of abdominal and gynecological adhesiolysis in the US. BMC Surg. 2011;11:13. Menzies D, Ellis H. Intestinal obstruction from adhesions--how big is the problem? Ann R Coll Surg Engl. 1990;72(1):60-63. Davis SE, Sperling L. Obstruction of the small intestine. Arch Surg. 1969;99(4):424-426. Pollock AV. Laparotomy. J R Soc Med. 1981;74(7):480-484. Galardi N, Collins J, Friend K. Use of early gastrografin small bowel follow-through in small bowel obstruction management. Am Surg. 2013;79(8):794-796. Sarraf-Yazdi S, Shapiro ML. Small bowel obstruction: the eternal dilemma of when to intervene. Scand J Surg. 2010;99(2):78-80. O'Connor DB, Winter DC. The role of laparoscopy in the management of acute smallbowel obstruction: a review of over 2,000 cases. Surg Endosc. 2012;26(1):12-17. Wilson MS, Hawkswell J, McCloy RF. Natural history of adhesional small bowel obstruction: counting the cost. Br J Surg. 1998;85(9):1294-1298. Tingstedt B, Isaksson J, Andersson R. Long-term follow-up and cost analysis following surgery for small bowel obstruction caused by intra-abdominal adhesions. Br J Surg. 2007;94(6):743-748. Ellis H, Moran BJ, Thompson JN, et al. Adhesion-related hospital readmissions after abdominal and pelvic surgery: a retrospective cohort study. Lancet. 1999;353(9163):1476-1480. Kotsis SV, Chung KC. Fundamental principles of conducting a surgery economic analysis study. Plast Reconstr Surg. 2010;125(2):727-735. Fevang BT, Fevang J, Lie SA, Soreide O, Svanes K, Viste A. Long-term prognosis after operation for adhesive small bowel obstruction. Ann Surg. 2004;240(2):193-201. Nelson SD, Malone D, Lafleur J. Calculating the Baseline Incidence in Patients Without Risk Factors: A Strategy for Economic Evaluation. Pharmacoeconomics. 2015;33(9):887892. National Surgical Quality Improvement Program. 2005; http://www.acsnsqip.org/. Accessed March 2014, 2014. Arias E. United States Life Tables, 2009. NVSR. 2014;7:62. Accessed 2014. Healthcare Cost and Utilization Project (HCUP), Healthcare Cost and Utilization Project (HCUP). Agency for Healthcare Research and Quality; 2007-2009. http://www.hcupus.ahrq.gov/nisoverview.jsp. Janson M, Bjorholt I, Carlsson P, et al. Randomized clinical trial of the costs of open and laparoscopic surgery for colonic cancer. Br J Surg. 2004;91(4):409-417. Healey MA, Shackford SR, Osler TM, Rogers FB, Burns E. Complications in surgical patients. Arch Surg. 2002;137(5):611-617; discussion 617-618. Ara R, Brazier J. Comparing EQ-5D scores for comorbid health conditions estimated using 5 different methods. Med Care. 2012;50(5):452-459.

AC C

1.

9

ACCEPTED MANUSCRIPT

25.

26.

27.

28. 29.

30.

31.

32.

33.

RI PT

SC

24.

M AN U

23.

TE D

22.

EP

21.

Ghatnekar O, Bondesson A, Persson U, Eriksson T. Health economic evaluation of the Lund Integrated Medicines Management Model (LIMM) in elderly patients admitted to hospital. BMJ Open. 2013;3(1). Bijen CB, Vermeulen KM, Mourits MJ, et al. Cost effectiveness of laparoscopy versus laparotomy in early stage endometrial cancer: a randomised trial. Gynecol Oncol. 2011;121(1):76-82. Desai AA, Baras J, Berk BB, et al. Management of acute kidney injury in the intensive care unit: a cost-effectiveness analysis of daily vs alternate-day hemodialysis. Arch Intern Med. 2008;168(16):1761-1767. Erickson KF, Japa S, Owens DK, Chertow GM, Garber AM, Goldhaber-Fiebert JD. Costeffectiveness of statins for primary cardiovascular prevention in chronic kidney disease. J Am Coll Cardiol. 2013;61(12):1250-1258. Duriseti RS, Brandeau ML. Cost-effectiveness of strategies for diagnosing pulmonary embolism among emergency department patients presenting with undifferentiated symptoms. Ann Emerg Med. 2010;56(4):321-332 e310. Mahmoudi M, Sobieraj DM. The cost-effectiveness of oral direct factor Xa inhibitors compared with low-molecular-weight heparin for the prevention of venous thromboembolism prophylaxis in total hip or knee replacement surgery. Pharmacotherapy. 2013;33(12):1333-1340. Gould MK, Dembitzer AD, Doyle RL, Hastie TJ, Garber AM. Low-molecular-weight heparins compared with unfractionated heparin for treatment of acute deep venous thrombosis. A meta-analysis of randomized, controlled trials. Ann Intern Med. 1999;130(10):800-809. Gould MK, Dembitzer AD, Sanders GD, Garber AM. Low-molecular-weight heparins compared with unfractionated heparin for treatment of acute deep venous thrombosis. A cost-effectiveness analysis. Ann Intern Med. 1999;130(10):789-799. Jensen CC, Prasad LM, Abcarian H. Cost-effectiveness of laparoscopic vs open resection for colon and rectal cancer. Dis Colon Rectum. 2012;55(10):1017-1023. Lala A, Berger JS, Sharma G, Hochman JS, Scott Braithwaite R, Ladapo JA. Genetic testing in patients with acute coronary syndrome undergoing percutaneous coronary intervention: a cost-effectiveness analysis. J Thromb Haemost. 2013;11(1):81-91. Ito K, Shrank WH, Avorn J, et al. Comparative cost-effectiveness of interventions to improve medication adherence after myocardial infarction. Health Serv Res. 2012;47(6):2097-2117. Hamel MB, Phillips RS, Davis RB, et al. Outcomes and cost-effectiveness of ventilator support and aggressive care for patients with acute respiratory failure due to pneumonia or acute respiratory distress syndrome. Am J Med. 2000;109(8):614-620. Culligan PJS, C.; Lewis, C.; Abell, T.D. Cost-effectiveness analysis comparing robotic sacrocolpopexy to a vaginal mesh hysteropexy for treatment of uterovaginal prolapse. Open Journal of Obstetrics and Gynecology. 2013;3(8):613-620. Rascati KL. The $64,000 question--what is a quality-adjusted life-year worth? Clin Ther. 2006;28(7):1042-1043.

AC C

20.

10

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Rocha FG, Theman TA, Matros E, Ledbetter SM, Zinner MJ, Ferzoco SJ. Nonoperative management of patients with a diagnosis of high-grade small bowel obstruction by computed tomography. Arch Surg. 2009;144(11):1000-1004.

AC C

34.

11

ACCEPTED MANUSCRIPT

Value $5,502 $18,890 $39,626

Note: all costs were adjusted to 2013 USD

First year post treatment Following years (per year) Post surgical treatment First year post treatment Following years (per year)

Source

Fevang et al 2004 12 Adjusted* Fevang et al 2004 12 Adjusted* Fevang et al 2004 12 Adjusted* Fevang et al 2004 12 Adjusted* Fevang et al 2004 12

For multiple recurrences

Fevang et al 2004 12

TE D

Relative risk of recurrence Post surgical treatment

Death Probability of death with medical treatment Probability of death due to surgery

HCUP NIS

16

2009 life tables

EP

Probability of death while stable

HCUP NIS 16

15

Distribution Gamma Gamma Gamma

Value

Range

Distribution

0.0923

+/- 30%

Beta

0.0173

+/- 30%

Beta

0.0508

+/- 30%

Beta

0.0092

+/- 30%

Beta

+/- 30%

Beta

0.0266

+/- 30%

Beta

0.0179 Varies by each year of age

+/- 30%

Beta

+/- 30%

Beta

M AN U

Table 2: Probability inputs Parameter Small bowl obstruction recurrence Post medical treatment

Range +/- 30% +/- 30% +/- 30%

RI PT

Source CMS CMS CMS

SC

Table 1: Cost inputs Parameter Medical treatment Surgery without major complication Surgery with major complication

0.55 After 2nd = 2.1 After 3rd = 2.9 After 4th = 5.7

Surgical complications Acute renal injury Deep vein thrombosis Ileus Myocardial infraction Pulmonary embolism

AC C

NSQIP 14 0.006 +/- 30% 14 NSQIP 0.009 +/- 30% Janson et al 2004 17 0.01 +/- 30% NSQIP 14 0.006 +/- 30% 14 NSQIP 0.009 +/- 30% Healey et al 2002 18 0.0079 to Respiratory (pneumonia) 0.01 and NSQIP 14 0.0129 18 Healey et al 2002 Surgical site infection 0.0189 0.01 to 0.04 and NSQIP 14 13 *Adjusted for the proportion of patients that received surgery, and relative risk reduction of 0.55 from surgery12 Table 3: Utility inputs Parameter

Source

Value

Range

Beta Beta Beta Beta Beta Beta Beta

Distribution

12

ACCEPTED MANUSCRIPT

Table 3: Utility inputs Parameter Health states Stable Medical treatment Surgery Death Surgical complications

Distribution

1 0.96 0.81 0

0.7 to 1 0.7 to 1 +/- 30%

Beta Beta Beta

-0.217

-0.04 to -0.3

Beta

-0.1

-0.069 to -0.18

Beta

-0.3

+/- 30%

Beta

Myocardial infraction (for 5 years)

-0.3

-0.15 to -0.4

Beta

Pulmonary embolism (for 1 year)

-0.19

-0.145 to -0.349

Beta

Respiratory (pneumonia) (for 2 months) Surgical site infection (for 2 months)

-0.08 -0.15

+/- 30% +/- 30%

Beta Beta

Table 4: Cost-effectiveness results QALY

Life years

TE D

Cost

Incremental Cost

Incremental QALY

Incremental life years

SC

Ileus (for 1 month)

Desai et al. 2008,22 and Erickson et al. 2013 23 Duriseti et al 2013,24 Mahmoudi et al 2013,25 and Gould et al 1999 26,27 Jensen et al 2012 28 Lala et al 2012 29 and Ito et al 2012 30 Duriseti et al 2013 24 and Mahmoudi et al 2013 25 Hamel et al 2000 31 Culligan et al 2013 32

M AN U

Deep vein thrombosis (for 1 year)

ICER (QALY)

5.872

9.615

-

-

-

-

$6,351

5.873

9.616

$7

0.001

0.001

$4,573

$6,387

5.879

9.616

$35

0.005

0.001

$6,581

$6,558

5.900

9.619

$171

0.022

0.004

$7,914

6.035

9.629

$1,643

0.135

0.014

$12,172

EP

$6,345

AC C

Surgery after 5th SBO recurrence Surgery after 4th SBO recurrence Surgery after 3rd SBO recurrence Surgery after 2nd SBO recurrence Surgery after 1st SBO recurrence

Range

Ara et al. 2012 19 Ghatnekar et al. 2013 20 Bijen et al. 201121 Ara et al. 2012 19

Acute renal injury (for 1 month)

Strategy Name

Value

RI PT

Source

$8,201

QALY, Quality adjusted life years; ICER, Incremental cost-effectiveness ratio; SBO, small bowel obstruction

13

RI PT

Figure 1

AC C

EP

TE

D

M AN US C

Markov model

Note: Complications included acute renal injury, deep vein thrombosis, ileus, myocardial infarction, pulmonary embolism, pneumonia, and skin and suture infections.

Figure 2 revised

AC C

EP

TE

D

M AN US C

RI PT

Cost-effectiveness of the different strategies

QALY, Quality adjusted life years; SBO, small bowel obstruction

SBO, small bowel obstruction

AC C

EP

TE

D

M AN US C

Cost-effectiveness acceptability curve (Probabilistic sensitivity analysis)

RI PT

Figure 3

Figure 4

AC C

EP

TE

D

M AN US C

RI PT

Scatter plot comparing ICERs between surgery after 1st SBO recurrence versus 2nd SBO recurrence

QALY, Quality adjusted life years; ICER, Incremental cost-effectiveness ratio; SBO, small bowel obstruction

ACCEPTED MANUSCRIPT

DISCUSSION: DISCUSSANT:

DR. RONALD STEWART (San Antonio, TX):

this interesting paper.

RI PT

Thanks to the Program Committee for the invitation to discuss I think you probably overestimated my

familiarity with the Markov model, but thanks to the authors for

SC

getting the paper to the Journal well in advance of the meeting. Dr. Colonna and team, as described, set up a Markov

M AN U

microsimulation model to evaluate the timing of operation. Their hypothesis was that after the second admission for small bowel obstruction that resolved medically, it becomes cost‑effective to intervene.

All research is prone to bias, and

TE D

certainly normal clinical research bias is a big problem, but in simulation, the bias is particularly acute since it's not just the data that are being analyzed.

It's actually the entire

EP

experiment which is under complete control of the investigator. This leads to my questions.

AC C

Prior to the project, what was your and what was the senior author's approach to adhesive bowel obstruction?

Was it early

operation or continued attempts at nonoperative management? Two, you allude to it in the discussion and allude to it in your nicely presented paper, but what happens if the assumption of a significant risk reduction following surgery is not present in the model?

I.e., what if surgery leads to an acute cure but the

ACCEPTED MANUSCRIPT

risk of adhesive bowel obstruction is the same or even increased, what happens to the model in that setting? The risk of surgical complications were obtained from the NSQIP

complications.

RI PT

sample, which I believe is probably the best data for But in reality, NSQIP may actually

underrepresent the risk of complications when compared to the

that participate in NSQIP.

SC

overall U.S. cohort because of a selection bias of hospitals So what happens in the model if

M AN U

there is a higher surgical complication rate or, for that matter, even a lower surgical complication rate, what happens to the conclusions?

Lastly, with respect to the model, what happens if the cohort is I think 40 is probably a reasonable

TE D

significantly older?

starting point, but probably a lot of people take care of a significantly older cohort with bowel obstruction.

EP

Finally, as really an editorial comment, this is a very well‑done manuscript, very nicely written.

But I think the power of this

AC C

approach is not in reaching a conclusion about the timing of operation but rather a description of the factors that might lead to the surgeon or the patient choosing one approach versus another depending upon what the variables are with respect to that particular patient. discuss this paper.

Thank you again for the opportunity to

ACCEPTED MANUSCRIPT

DR. ALEXANDER COLONNA (Salt Lake City, UT): for your comments.

Thank you so much

I'll start with the editorial comment first.

Yes, I absolutely agree with you.

I think these economic

way of looking at the literature.

RI PT

analysis models are much like meta‑analyses where it's a novel We all look at literature.

We all try to synthesize it and present that data to our

SC

patients for our informed consent discussions, and I think this is another way of looking at the data that's out there and using

M AN U

that to talk to our patients.

For your first question, our bias, how do we treat patients? Both myself and senior author Ram Nirula, we basically wait until we have to operate on people, like most of us in the room.

TE D

You are correct that the patients that we generally operate on are a little bit older and have several comorbidities.

We

picked the 40‑year‑old patient to try to simplify some of that,

EP

but we tend to wait until we're forced to operate, like I think most of us in general surgery here.

AC C

Your second question, yes, you are correct.

If that significant

risk reduction is not present, then the model kind of falls apart.

There was a well‑done paper by Fevang, and that's what we

based our assumptions.

We did model in the sensitivity analysis

the variation in some of these risks, and through all of our different sensitivity analyses, the model still held up.

But if

ACCEPTED MANUSCRIPT

you disagree with the basic assumption, then the model as we constructed it does not work. The risk percentages from NSQIP, you're correct.

It can

RI PT

underrepresent the risks, but, again, we try to put in the mean standard deviations and the distributions in our probabilistic

DR. JAMES DAVIS (Fresno, CA):

SC

sensitivity analysis and the model still held up.

I would like to congratulate you

M AN U

on what I think is a fascinating presentation on simulation.

I

have a couple of questions for you that maybe echo some of Dr. Stewart's.

One of your assumptions is that surgery resets the risk factor

obstruction.

TE D

rather than perhaps increases your risk for adhesive bowel I think Ronnie addressed that adequately. What are your plans from here?

How are you going

to test this?

Is it going to be a multicentered trial?

What's

EP

a simulation.

This is

Thank you.

AC C

going forward?

DR. ALEXANDER COLONNA (Salt Lake City, UT):

So, again, if you

don't think that surgery helps, then the idea that operating and lysing the adhesions in the absence of a hard clinical sign ‑‑ that was our assumption.

If you disagree with it, then the

model falls apart, like I said. Moving forward, we don't have any specific plans for studying

ACCEPTED MANUSCRIPT

this more rigorously.

I think if I were to redesign the model,

I would change my cohort of patients that I start out with. Again, I think most of us that do this, we see typically older So I think it would be a more

RI PT

patients with more comorbidities.

interesting next way to model this if we selected older patients with comorbidities to start off with that may change our

AC C

EP

TE D

M AN U

SC

decision to operate.