Sage Journals: Discover world-class research

Abstract

Background and Objectives:

Colon surgical site infections (SSIs) impose clinical and financial burden. To guide reimbursement, the Centers for Medicare and Medicaid Services utilize the National Healthcare Safety Network’s (NSHN) Complex 30-Day SSI Model for Colon Procedures, which is used to compare the observed number of SSIs to the expected number of SSIs for all colon surgeries, regardless of underlying etiology. This study assessed the predictive accuracy of the Complex 30-day SSI Model for inpatients undergoing emergent colon operations for trauma.

Patients and Methods:

The Trauma Quality Program Participant Use File was used to identify all patients who underwent emergent colon surgery because of trauma between 2017 and 2021 (excluding 2020). The predicted probability of colon SSI was calculated using the NSHN Complex 30-day Model using the following variables: diabetes mellitus, gender, age, body mass index, closure technique, American Society of Anesthesiology Score, and oncology hospital status. Receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC) was calculated to evaluate the model’s predictive accuracy.

Results:

In total, 292,688 patients were included. Among all patients, an SSI occurred in 0.33% (n = 1,097). For patients with complete data, the rate of SSI was 0.34%. The ROC curve for complete case analysis showed an AUC of 0.523 (95% confidence interval 0.506–0.541).

Conclusion:

The predictive accuracy of the NSHN Complex 30-Day SSI Model after colon injury is poor. The current model may not adequately capture the complexities and risk associated with trauma. An improved model is needed for more accurate risk assessment to improve patient care and accuracy of hospital reimbursement.

Keywords

colon reimbursement risk assessment surgical site infection trauma wound infection

Introduction

Surgical site infection (SSI) can cause devastating consequences for patients and the healthcare system. The U.S. Health and Human Services Agency for Healthcare Research and Quality considered healthcare-associated infection (HAI) the most common serious complication in 2008.^1,2 Of the various types of HAI, SSI is the most common. A large retrospective study using the National Inpatient Sample, which consisted of over 35 million patients admitted to the hospital in 2016, found that 1.8% of all admissions were related to hospital acquired infection. Of these, 47% were SSI.³ In 2016 alone, SSI was responsible for $3.7 billion dollars in U.S. healthcare costs; however, when considering estimated additional costs associated with SSI, this increases to $7.2–$14.9 billion dollars.⁴

In addition to the potential harm to the patient and the staggering cost to the national healthcare system, SSI reporting has significant implications for hospital management. The Hospital Inpatient Quality Reporting Program (HIQRP) was originally mandated in 2003 as part of the Medicare Prescription Drug, Improvement, and Modernization Act.⁵ For receipt of full reimbursement by the Centers for Medicare and Medicaid Services (CMS), acute-care hospitals must report SSI outcomes as part of the Inpatient Prospective Payment System. The American Recovery and Reinvestment Act and the Affordable Care Act modified this policy as to reduce reimbursement by up to 25% if HIQRP requirements are not met.⁵

The Centers for Disease Control and Prevention’s National Healthcare Safety Network (NHSN) is the reporting system used by CMS for hospitals to report and track data. HAIs are reported as a standardized infection ratio (SIR). The SIR is calculated by comparing the observed number of HAIs with the expected number of HAIs based on a predictive model.⁶ Of these models, the Colon Procedure, Complex 30-Day Model is the model used for NHSN reporting for all colon surgeries. This study evaluates the accuracy of the Complex 30-Day SSI Model for Colon Procedures to predict SSI in the setting of traumatic colon surgery. We hypothesized that this model is a poor predictor of SSI and thus poses significant implications for quality reporting for hospitals treating trauma patients. A secondary aim of this study was to determine factors associated with an increased risk of SSI in trauma patients undergoing colon surgery.

Patients and Methods

Patient population and variables

The American College of Surgeons Trauma Quality Programs (TQP) Participant Use File (PUF) was utilized from 2017 to 2021 (excluding 2020) to obtain variables including demographics (age, gender, and race), injury type (blunt vs. penetrating), mechanism of injury, volume (mL) of packed red blood cells (PRBCs) received within 4 hours of emergency department (ED) or hospital arrival, and hospital (HLOS) and intensive care unit length of stay (ICU LOS). Age, PRBCs, and LOS variables are reported as mean ± standard deviation. TQP PUF was also queried to obtain variables associated with colon SSI prediction using the NSHN Complex 30-Day SSI Model. Predictive variables included diabetes mellitus (yes vs. no), gender (male vs. female), age (per 10-year increment), body mass index (BMI) (≥30 kg/m² vs. <30 kg/m²), closure technique (non-primary vs. primary), American Society of Anesthesiologists Physical Status Classification System (ASA) score (ranging from 1 to 6), and oncology hospital status (yes vs. no).

Definitions

Because TQP data did not include closure technique, our institutional non-primary closure technique rate (0.265) was utilized for model prediction. ASA score, a variable also not collected in TQP, was designated as a “4” for the purposes of this investigation, which would infer severe trauma and/or shock in this population of patients requiring operative intervention for their injury. The designation of “Yes” for oncology hospital was chosen because of its association with a higher SSI rate according to the model.

Statistical analysis

For each patient, the predicted probability of an SSI (deep or organ/space) was calculated using the NSHN Complex 30-day model. Receiver operating characteristic (ROC) curves were then generated, and the area under the curve (AUC) was calculated to evaluate the model’s predictive accuracy. Initially, the data were analyzed using listwise depletion (complete case analysis). Subsequently, multiple imputation by fully conditional specification methods was employed to handle missing BMI and gender data. Furthermore, to assess the sensitivity of the results, closure technique was randomly assigned to each patient using a Bernoulli distribution, assuming a probability of non-primary closure of 0.265. Statistical analysis was done using SAS (SAS Institute Inc., Cary, NC, USA). The MIANALYZE procedure in SAS was utilized to combine AUC estimates from multiple imputations. Significance was defined as p <0.05.

Results

A total of 333,139 patients met the inclusion criteria with 292,688 patients (87.9%) having complete data (Table 1). Overall, the study population had a mean age of 36.4 (±15.0) years and 79.5% were male. White (44.4%) and Black (39.3%) patients were the predominant racial groups. Penetrating injury (59.7%) was more common than blunt traumatic injury (39.3%). Of those with a penetrating mechanism, a firearm was used in 170,117 patients (85.6%). Nearly one-third of the patients (31.5%) had a BMI >30 mg/kg² and 4.0% of the patients had a diagnosis of diabetes mellitus. Among the total number of patients, SSI occurred in 0.33% (n = 1,097). For patients with complete data, the rate of colon SSI was 0.34%.

Table 1.

Study Population Summary and Demographics

Variable	Category	Study population (N = 333,139)
Age (years)	Mean (SD)	36.4 (±15)
	Median	32
	25th–75th percentile	25–44
Gender	Male	263,013 (79.5%)
	Female	67,828 (20.5%)
BMI	< 30 kg/m²	201,865 (68.5%)
	≥ 30 kg/m²	92,711 (31.5%)
Race	White	147,771 (44.4%)
	Black	130,977 (39.3%)
	Asian	6,867 (2.1%)
	Pacific Islander	893 (0.3%)
	American Indian	3,113 (0.9%)
	Multiracial	1,012 (0.3%)
	Other	31,667 (9.5%)
	Unknown/not reported	10,839 (3.2%)
Diabetes mellitus	No	319,852 (96.0%)
	Yes	13,287 (4.0%)
Type of injury	Blunt	130,843 (39.3%)
	Penetrating	198,811 (59.7%)
	Other (burn/other)	3,485 (1.0%)
Mechanism of injury
Blunt	Total	130,758
	MVC	94,193 (72.0%)
	MCC	10,302 (7.9%)
	Pedestrian struck	7,939 (6.1%)
	Fall	7,210 (5.5%)
	Other	11,114 (8.5%)
Penetrating	Total	198,785
	Stab	28,668 (14.4%)
	Firearm	170,117 (85.6%)
Blood transfusions within 4 h	N	319,404
	Mean (SD)	741 (±2071)
	Median	0
	25th–75th percentile	0–500
Hospital Length of stay (days)	N	331,692
	Mean (SD)	15.4 (±16.1)
	Median	10
	25th–75th percentile	7–19
Total ICU length of stay (days)	N	226,474
	Mean (SD)	8.3 (±10.6)
	Median	5
	25th–75th percentile	3–10

BMI, body mass index; MCC, motorcycle crash; MVC, motor vehicle crash; SD, standard deviation.

When comparing those in the study population who developed a colon SSI with those who did not, characteristics associated with colon SSI include penetrating mechanism, use of a firearm, and BMI ≥30 kg/m² (Table 2). There was no association in the development of a colon SSI with gender, race, diabetes mellitus diagnosis, or the volume of PRBCs transfused in the first 4 hours after ED/hospital arrival. Although the association with younger age and colon SSI was statistically significant, the differences in age were not clinically relevant (35.2 ± 13.1 yrs vs. 36.4 ± 15.0 yrs). Patients who developed a colon SSI had longer ICU LOS (17.0 ± 16.6 d vs. 8.2 ± 10.5 d; p <0.001) and HLOS (36.7 ± 28.4 d vs. 15.3 ± 16.0 d; p <0.001).

Table 2.

Variables Predictive of Surgical Site Infection

		Surgical Site Infection
Variable	Category	No(N = 332,042)	Yes(N = 1,097)	p
Age (years)*	Mean (SD)	36.4 (±15)	35.2 (13.1)	0.0071*
	Median	32	32
	25th–75th percentile	25–44	25–43
Gender	Male (N = 263,013)	262,119 (99.7%)	894 (0.34%)	0.1008
	Female (N = 67,828)	67,625 (99.7%)	203 (0.3%)
BMI*	< 30 kg/m² (N = 201,865)	201,239 (99.7%)	626 (0.31%)	0.0005*
	≥ 30 kg/m² (N = 92,711)	92,350 (99.6%)	361 (0.39%)
Race	White (N = 147,771)	147,346 (99.7.4%)	425 (0.29%)
	Black (N = 130,977)	130,471 (99.6%)	506 (0.39%)
	Asian (N = 6,867)	6,849 (99.7%)	18 (0.26%)
	Pacific Islander (N = 893)	893 (100%)	0
	American Indian (N = 3,113)	3,106 (99.8%)	7 (0.22%)
	Multiracial (N = 1,012)	1,012 (100%)	0
	Other (N = 31,667)	31,556 (99.6%)	111 (0.35%)
	Unknown/not reported	10,809 (99.7%)	30 (0.28%)
Diabetes mellitus	No (N = 319,852)	318,807 (99.7%)	1,045 (0.33%)	0.2025
	Yes (N = 13,287)	13,235 (99.6%)	52 (0.39%)
Type of injury*	Blunt (N = 130,843)	130,558 (99.8%)	285 (0.22%)	<0.0001*
	Penetrating(N = 198,811)	198,005 (99.6%)	806 (0.41%)
	Other (burn/other) (N = 3,485)	3,479 (99.8%)	6 (0.17%)
Mechanism of injury
Blunt	MVC (N = 94,193)	94,005 (99.8%)	188 (0.20%)	0.2785
	MCC (N = 10,302)	10,274 (99.7%)	28 (0.27%)
	Pedestrian struck (N = 7,939)	7,918 (99.7%)	21 (0.26%)
	Fall (N = 7,210)	7,188 (99.7%)	22 (0.31%)
	Other (N = 11,114)	11,091 (99.8%)	23 (0.21%)
Penetrating*	Stab (N = 28,668)	28,616 (99.8%)	52 (0.18%)	<0.0001*
	Firearm (N = 170,117)	169,363 (99.6%)	754 (0.44%)
Blood transfusions within 4 h	N	318,345	1,059	0.1195
	Mean (SD)	742 (2073)	321 (1452)
	Median	0	3
	25th–75th percentile	0–540	0–10
Hospital length of stay (days)*	N	330,604	1,088	<0.0001*
	Mean (SD)	15.3 (16.0)	36.7 (28.4)
	Median	10	28
	25th–75th percentile	7–19	18–45
Total ICU length of stay (days)*	N	225513	961	<0.0001*
	Mean (SD)	8.2 (10.5)	17.0 (16.6)
	Median	5	12
	25th–75th percentile	3–10	5–23

*p <0.05.

BMI, body mass index; SD, standard deviation.

In analyzing the ability of the Complex 30-Day SSI Model for Colon Procedures to predict SSI in the study population with complete data (292,688 patients), the predicted SSI rate was 9.9% in comparison with the observed SSI rate of 0.34%. The ROC curve for this analysis shows an AUC of 0.523 (95% confidence interval [CI] = 0.506–0.541; Fig. 1). Similar results for the total study population (n = 333,139) were obtained with imputation for non-primary closure technique (AUC = 0.524, 95% CI = 0.507–0.542) and multiple imputations for the missing gender and BMI data (mean AUC = 0.523; 95% CI = 0.506–0.539). Although the total number of SSI observed was low (0.33% of total cases), the accuracy of the analysis of predicting usefulness of the Complex 30-Day Model for Colon Procedures is unaffected by quantity of cases.

FIG. 1.

Receiver operating characteristic (ROC) curve.

Discussion

The results of this study suggest that the ability of the Complex 30-Day SSI Model for Colon Procedures to predict colon SSI in trauma patients is poor. Profoundly negative financial consequences may result with the use of a model having such inaccurate predictive ability. When a hospital has more observed colon SSIs than the number predicted by the Complex 30-Day SSI Model for Colon procedures, that hospital is at risk of having payments by CMS for Medicare beneficiaries reduced.⁷ According to Definitive Healthcare, 775 hospitals received Hospital-Acquired Condition (HAC) Reduction Program penalties in 2021 with the top 25 revenue losses ranging from $1,254,963 to $4,043,719. For the 775 hospitals that faced HAC penalties in 2021, there was $254 million in estimated loss revenue with an average loss of $328,800.⁸

In addition to the severe financial penalties that may be rendered because of the HAC Reduction Program, the information generated by the Complex 30-Day SSI Model for Colon Procedures is publicly reported on sites such as CMS’s Hospital Compare.⁷ Quality indicators such as infection rates are reported in the form of star ratings and report cards. Healthcare consumers utilize this information to choose high-quality providers over low-quality providers. That is not to say that efforts to prevent SSI are not critically important financially and clinically.^1,4,9 Previous studies have reported an increased hospital LOS and increased readmission rates secondary to SSI.^1,10 Our study similarly found a staggering increase in ICU and hospital LOS, that being 1 and 2 weeks longer, respectively. Accurate risk prediction of SSI is paramount, however, given the potential financial and reputational impact on healthcare centers because of using imprecise models.¹¹

There are several reasons that may explain the poor predictive ability of the Complex 30-Day SSI Model for Colon Procedures in trauma patients. Although the model is risk adjusted based on variables found to be statistically significant predictors of SSI, those factors that are specific to trauma are excluded.¹² Trauma patients frequently present in extremis, with coagulopathy, and/or with concomitant injuries.^13–15 One retrospective study reported factors highly prevalent in the trauma setting that significantly increased risk of SSI: penetrating mechanism, injury severity, lactic acidosis, coagulopathy, and damage control surgery.¹⁶ Additionally, there is high potential for fecal contamination in destructive colonic injuries, which may require colostomy creation, thus increasing the risk for SSI.^13,15,17,18 Penetrating injury and use of a firearm were statistically significant risk factors for SSI in this study, thereby demonstrating the increased likelihood of the high-risk factors described above. Although the volume of PRBC transfusion within 4 hours of ED/hospital arrival was not statistically significant in our analysis, other studies have found blood transfusion to be an independent risk factor for SSI.^16,19,20

Trauma patients are substantially different than patients undergoing elective colon surgery, particularly since the widespread adoption of SSI prevention bundles and enhanced recovery after surgery protocols, resulting in a significant reduction in SSI for elective colon surgery patients.^21,22 Patients undergoing elective surgery have the opportunity to adhere to evidence-based guidelines such as smoking cessation, glucose control, nutritional optimization, and bowel preparation.²³ Although proved to be effective in the elective setting, these measures are frequently not feasible in the emergent trauma setting. These factors raise the question of whether the rate of SSI in trauma colon surgeries should be judged similarly to colon procedures that occur in an elective setting with data-proved preventive measures such as preoperative bowel preparation.

With further investigation into specific risk factors, a more accurate prediction model may assist with the development of mitigating factors to prevent SSI in the trauma patient. For example, Wei et al. generated a new predictive model that includes age, gender, race, BMI, mechanism of injury (penetrating vs. blunt), first intra-operative base excess, first intra-operative temperature, units of PRBCs transfused, open small bowel injury, small bowel resection, open colon injury, colon resection, performance of damage control laparotomy, wound class, and surgery duration; this model had improved AUC of 0.78.²³ A more accurate predictive model would allow for better stratification of SSI risk in this highly prone population.^11,24,25

The authors of this study initially hypothesized that the Complex 30-Day Model for Colon Procedures would underestimate the rate of SSI for the reasons described previously. The total number of SSI identified in this study was surprisingly low. This is likely because of a variety of reasons. The Complex 30-Day Model for Colon Procedures only includes deep incisional and organ space infections. These are the colon SSIs that are publicly reported by CMS. There are likely a larger number of superficial SSIs that were not accounted for in this study.

Another likely significant factor in the low number of observed infections is the exclusion of infection present at the time of surgical procedure (PATOS). Since 2017, PATOS SSI was excluded from the NHSN colon SSI SIR. However, there is currently no standardized definition for PATOS for trauma patients. There is currently debate as to how long fecal contamination must be present to be considered PATOS. Furthermore, there is room for interpretation in regard to the term “feculent peritonitis” when classifying a “dirty wound” versus “infected wound” at the time of surgical procedure.⁶ There is a considerable amount of subjectivity in determining PATOS and thus omitting an SSI from reporting as a result.^26,27 PATOS is not a variable that can be extracted from the TQP database, and thus, we were unable to control for this.

Lastly, the use of TQP PUF contributes to the low number of colon SSI infections realized in the current investigation. Unlike the National Surgical Quality Improvement Project, which surveys patients after discharge to capture SSIs,²⁸ TQP PUF is limited to data from the index hospitalization. The contributory impact of the absence of post-discharge information on the SSI rate in this study is unknown. Hübner et al. suggested in their study that the diagnosis of a colon SSI was established after discharge in 22.7% of their patients.²⁹

The authors recognize several limitations of this study. Primarily, the use of a large database poses limitations because of the accuracy of data reporting. Data imported into the TQP are highly dependent on manual coding and are thus subject to error. Specifically, in regard to TQP PUF, some elements used in the Complex 30-Day Model, including ASA classification and non-primary closure method, are not included in this database and had to be estimated. The exclusion of PATOS in this model is another limitation. The lack of consistency in reporting PATOS likely leads to error in SSI reporting and may bias the results. Additionally, a large database study limits the ability to draw cause and effect conclusions. Although there are strong associations with various factors, such as penetrating trauma, with SSI, we cannot definitively state that penetrating trauma causes SSI.

Although this study has known limitations, there are also many strengths of this investigation that should be noted. This study used a large sample size with nationally representative data specific to trauma patients. Furthermore, this study was able to redemonstrate the known independent risk factors proved to be associated with SSI in the trauma population. Despite the low SSI rate following traumatic colon injury revealed in this study, this rate does not affect the accuracy of the ROC curve analysis, which demonstrated an AUC of 0.523 and 0.524 for the study population with complete data and the total study population, respectively, suggesting poor discriminative ability. Our findings contribute to the sparse body of literature that exists, which discusses the issues of the current predictive model used for colon SSI in trauma patients. This study emphasizes the need for a separate model for SSI prediction to improve the equity of reimbursement for acute care and trauma hospitals.

Conclusion

On the basis of the findings of the current study, the predictive accuracy of the current NSHN Complex 30-day SSI Model after penetrating colon injury is poor. These results have significant implications for clinical practice. It suggests that the current model may not adequately capture the complexities and specific risk factors associated with trauma patients. Therefore, there is a critical need for the development of improved models that incorporate additional risk factors specific to trauma patients to enhance the predictive accuracy of colon SSI. Such advancements would have profound implications for patient care, enabling more accurate risk assessment and targeted interventions to reduce the incidence of SSIs in this vulnerable population. With the consequences of a colon SSI being so substantial, a tool to predict the occurrence of this complication needs to be more accurate than a coin flip.

Authors’ Contributions

A.B.: Visualization, writing—original draft, and writing—review and editing. W.I.: Methodology, formal analysis, investigation, and writing—review and editing. M.H.: Conceptualization, data curation, investigation, methodology, writing—original draft, and writing—review and editing. M.N.: Conceptualization, data curation, investigation, supervision, writing—original draft, writing—review and editing, and Guarantor.

Footnotes

Author Disclosure Statement

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding Information

The authors received no financial support for the research, authorship, and/or publication of this article.

Ethical Approval

The East Carolina University Institutional Review Board waived the requirements for IRB approval, as all data are publicly available.

References

Shepard

, Ward

, Milstone

, et al. Financial impact of surgical site infections on hospitals: The hospital management perspective. JAMA Surg, 2013; 148(10):907–914; doi: 10.1001/jamasurg.2013.2246

Turner

, Migaly

. Surgical site infection: The clinical and economic impact. Clin Colon Rectal Surg, 2019; 32(3):157–165; doi: 10.1055/s-0038-1677002

Forrester

, Maggio

, Tennakoon

. Cost of health care-associated infections in the United States. J Patient Saf, 2022; 18(2):e477–e479; doi: 10.1097/PTS.0000000000000845

Broex

ECJ

, van Asselt

ADI

, Bruggeman

, et al. Surgical site infections: How high are the costs? J Hosp Infect, 2009; 72(3):193–201; doi: 10.1016/j.jhin.2009.03.020

Hospital inpatient quality reporting program | CMS. CMS.gov Web site. Updated 2023. Available from: https://www.cms.gov/medicare/quality/initiatives/hospital-quality-initiative/inpatient-reporting-program [Last accessed: Jul 22, 2024].

THE NHSN standardized infection ratio (SIR): A guide to the SIR. CDC.gov: National Healthcare Safety Network Web site. Updated 2024. Available from: https://www.cdc.gov/nhsn/pdfs/ps-analysis-resources/nhsn-sir-guide.pdf [Last accessed: Jul 22, 2024].

Hospital quality initiative public reporting | CMS. www.cms.gov Web site. Updated 2024. Available from: https://www.cms.gov/medicare/quality/initiatives/hospital-quality-initiative/hospital-compare [Last accessed: Nov 20, 2024].

25 hospitals with the greatest revenue loss from HAC penalties. Definitive Healthcare Web site. Updated 2022. Available from: https://www.definitivehc.com/resources/healthcare-insights/hospitals-highest-hac-penalty [Last accessed: Dec 11, 2024].

Hsu

, Wang

, Broadwell

, et al. Association between federal value-based incentive programs and health care-associated infection rates in safety-net and non-safety-net hospitals. JAMA Netw Open, 2020; 3(7):e209700; doi: 10.1001/jamanetworkopen.2020.9700

10.

Mujagic

, Marti

, Coslovsky

, et al. Associations of hospital length of stay with surgical site infections. World J Surg, 2018; 42(12):3888–3896; doi: 10.1007/s00268-018-4733-4

11.

Fornek

, Ata

, Jimenez

, et al. Reportable infections following colon surgery in a large public healthcare system in New York City: The consequences of being a level 1 trauma center. Infect Control Hosp Epidemiol, 2024; 45(1):21–26; doi: 10.1017/ice.2023.150

12.

Boston

, Ellsworth

, Thomas

, et al. Impact of penetrating trauma on surgical site infection standardized infection ratio (SIR) for colon procedures. Open Forum Infect Dis, 2021; 8(Suppl 1) MPH, CIC, CPHQ, FAPIC1 DO MPH, CIC, CSSGB3 MS, BSN, RN, CIC4 MD.

13.

O’Neill

, Kirton

, Dresner

, et al. Analysis of 162 colon injuries in patients with penetrating abdominal trauma: Concomitant stomach injury results in a higher rate of infection. J Trauma, 2004; 56(2):304–313; doi: 10.1097/01.TA.0000109856.25273.07

14.

Chammas

, Abdul Jawad

, Pust

, et al. Association between fecal contamination and outcomes after emergent general surgery colorectal resection: A post hoc analysis of an eastern association for the surgery of trauma (EAST) multicenter study. Surg Infect (Larchmt), 2023; 24(6):561–565; doi: 10.1089/sur.2022.256

15.

Adesanya

, Ekanem

. A ten-year study of penetrating injuries of the colon. Dis Colon Rectum, 2004; 47(12):2169–2177; doi: 10.1007/s10350-004-0726-5

16.

Isbell

, Hatton

, Wei

, et al. Risk stratification for superficial surgical site infection after emergency trauma laparotomy. Surg Infect (Larchmt), 2021; 22(7):697–704; doi: 10.1089/sur.2020.242

17.

Cullinane

DCM

, Jawa

RSM

, Como

JJM

, et al. Management of penetrating intraperitoneal colon injuries: A meta-analysis and practice management guideline from the eastern association for the surgery of trauma. J Trauma Acute Care Surg, 2019; 86(3):505–515; doi: 10.1097/TA.0000000000002146

18.

Yamamoto

, Morimoto

, Kita

, et al. The preventive surgical site infection bundle in patients with colorectal perforation. BMC Surg, 2015; 15:128; doi: 10.1186/s12893-015-0115-0

19.

Kirton

, O’Neill

, Kestner

, et al. Perioperative antibiotic use in high-risk penetrating hollow viscus injury: A prospective randomized, double-blind, placebo-control trial of 24 hours versus 5 days. J Trauma, 2000; 49(5):822–832; doi: 10.1097/00005373-200011000-00006

20.

Egbe

, Iranya

, Dimala

, et al. Effect of perioperative blood transfusion on complications following emergency non-trauma laparotomy in Mulago hospital: A prospective cohort study. Cureus, 2024; 16(7):e65759; doi: 10.7759/cureus.65759

21.

Cavallaro

, Bordeianou

. Implementation of an ERAS pathway in colorectal surgery. Clin Colon Rectal Surg, 2019; 32(2):102–108; doi: 10.1055/s-0038-1676474

22.

Varadhan

, Neal

, Dejong

CHC

, et al. The enhanced recovery after surgery (ERAS) pathway for patients undergoing major elective open colorectal surgery: A meta-analysis of randomized controlled trials. Clin Nutr, 2010; 29(4):434–440; doi: 10.1016/j.clnu.2010.01.004

23.

Wei

, Green

, Kao

, et al. Accurate risk stratification for development of organ/space surgical site infections after emergent trauma laparotomy. J Trauma Acute Care Surg, 2019; 86(2):226–231; doi: 10.1097/TA.0000000000002143

24.

Eckmann

, Kramer

, Assadian

, et al. Clinical and economic burden of surgical site infections in inpatient care in Germany: A retrospective, cross-sectional analysis from 79 hospitals. PLoS One, 2022; 17(12):e0275970; doi: 10.1371/journal.pone.0275970

25.

Caroff

, Wang

, Zhang

, et al. The limited utility of ranking hospitals based on their colon surgery infection rates. Clin Infect Dis, 2021; 72(1):90–98; doi: 10.1093/cid/ciaa012

26.

Seidelman

, Ge

, Baker

, et al. Colon surgical-site infections and the impact of “present at the time of surgery (PATOS)” in a large network of community hospitals. Infect Control Hosp Epidemiol, 2023; 44(8):1255–1260; doi: 10.1017/ice.2022.236

27.

Konnor

, Russo

, Leaptrot

, et al. How infection present at time of surgery (PATOS) data impacts your surgical site infection (SSI) standardized infection ratios (SIR), with focus on the complex 30-day SSI SIR model. Am J Infect Control, 2021; 49(11):1423–1426; doi: 10.1016/j.ajic.2021.05.002

28.

ACS National Surgical Quality Improvement Program. Available from: https://www.facs.org/quality-programs/data-and-registries/acs-nsqip/ [Last accessed: Sep 19, 2025].

29.

Hübner

, Diana

, Zanetti

, et al. Surgical site infections in colon surgery: The patient, the procedure, the hospital, and the surgeon. Arch Surg, 2011; 146(11):1240–1245; doi: 10.1001/archsurg.2011.176

Predictive Accuracy of the Complex 30-Day Model for Colon Surgical Site Infections after Traumatic Colon Injury: Any Better than a Coin Flip?