Application of a framework to mitigate the risk of surgical site infection after exploratory celiotomy in horses: A retrospective study

Abstract

Objective

To describe the methodology used to identify the contributors to a perceived sudden increase in exploratory celiotomy surgical site infections (SSI) and complications at the North Carolina State University Veterinary Teaching Hospital (NCSU VTH) between 2019 and 2020 and evaluate the effect of the designed intervention up to 4 years after its implementation.

Study design

Case–control retrospective study over a five‐year period.

Animals

A total of 448 horses that underwent exploratory celiotomy for the treatment of acute abdominal pain were included.

Methods

Medical records of horses that underwent exploratory celiotomy between 2019–2024 were reviewed from software systems used at the NCSU VTH. A surgical audit was conducted to assess adherence to best practices and identify factors contributing to increased SSI incidence. This led to the development of an evidence‐based intervention to address procedural deficiencies and incorporate preventative perioperative strategies. The approach, resultant protocols, and reduction of SSI incidence are described. Data were analyzed using Fisher's exact test and univariate logistic regression. Statistical significance was set to p < .05.

Results

A significant increase in %SSI was observed from 7.7% in 2019 to 29% in 2020 (p = .0067). Following new protocol implementation, %SSI decreased to 2.3%.

Conclusion

A surgical audit enabled the development of an evidence‐based intervention that significantly reduced SSI incidence after exploratory celiotomy surgery.

Clinical significance

Surgical audits serve as critical quality‐of‐care measure, allowing hospitals to identify procedural deficiencies. There is currently no literature that describes structured processes to manage this kind of problem in veterinary medicine. Surgical audit implementation may help other hospitals faced with similar challenges.

1. INTRODUCTION

Acute abdominal pain is a common condition in horses which may require surgical intervention as the only life‐saving option for severe cases. Each year, up to 17% of horses presented with colic signs may require surgery. ¹ , ² Exploratory celiotomy is a common surgery performed in horses, and like any other invasive procedure carries the risk of postoperative complications which challenge the success of its outcome. Surgical site infection (SSI) is the most common incisional complication following exploratory celiotomy in horses. ³ , ⁴ , ⁵ , ⁶ Its presentation is characterized by fever, heat, edema, presence of discharge (serosanguinous or purulent) and pain at the surgical site. These signs are attributed to bacterial contamination during or after surgery with E. coli spp. Enterococcus spp. and Staphylococcus spp. are most commonly cultured. ⁷ While bacterial contamination plays a key role in SSI development, it has long been recognized that SSI is a multifactorial condition influenced by various patient‐related, surgical, and environmental factors. ⁸

Patient outcomes can be significantly and negatively affected by SSI. Incisional infection delays healing and increases the risk of herniation and/or peritonitis, as well as the cost of hospitalization and the duration of recovery time. Additionally, it may reduce the rate of return to work or competition level. ⁹ The incidence of incisional infection in colic cases ranges from 0% to 43% ¹⁰ ; this variation is likely influenced by hospital postoperative management, definition of SSI, or incomplete reporting of SSI incidence.

Around 80 horses undergo exploratory celiotomy annually at our institution. Historically, the NCSU VTH has had a low rate of SSI at 7.5%. However, in 2020–2021, faculty and residents noted an increase in the rate and severity of SSI. There was a clinical impression that, in addition to an increased prevalence of SSI cases and difficulty of management, infections appeared more rapidly, within 3 to 4 days after surgery instead of the typical 7–14 days. ¹⁰ In severe cases, horses experienced herniation and some even developed peritonitis. This led to a data collection effort to determine both current and past rates of SSI. At that time, SSI after colic surgery was determined to be 29%, well within previously published rates, but a substantial increase compared to the previous rate of 7.5% in 2019. In response, a clinician‐driven audit of SSI rates was initiated to identify contributing factors and improve patient care and outcomes.

Previous studies have investigated both the causes of SSI in colic patients as well as various strategies to reduce the rate of SSI rates in this population, including bandaging techniques, ¹¹ , ¹² suture materials and patterns, ¹³ , ¹⁴ , ¹⁵ surgical experience, ⁵ , ¹⁶ and the use of antibiotics. ¹⁷ Unfortunately, there is still not a validated protocol to completely prevent this complication. Identifying key contributors to SSI development within a hospital is difficult. However, objective methods to create a framework to identify hospital issues have been developed and utilized in human medicine to decrease iatrogenic causes of SSI. ¹⁸ , ¹⁹ , ²⁰ , ²¹ In collaboration with Duke University, surgical audits were coconducted to gather data and objectively assess the factors contributing to the increased SSI rate. The information collected from this audit provided the foundation for detailed analysis using a modified fishbone diagram analysis method, a tool often employed in human medicine to systemically identify potential contributors to iatrogenic complications. ²² , ²³ , ²⁴ This approach helped to develop and establish a new perioperative protocol in exploratory celiotomy surgeries.

The purpose of this study was to describe the methods used to confirm the existence of and address a sudden perceived increase in SSI at the NCSU VTH between March 2020 and February 2021and to evaluate the effect of the established intervention up to 4 years after its implementation. We hypothesized that the implementation of the new protocol would improve the SSI rate of horses that underwent exploratory celiotomy.

2. MATERIALS AND METHODS

2.1. Retrospective SSI and SSC

The clinical case records of horses that presented to NCSU VTH for colic and that underwent abdominal surgery from January 2019 to December 2024 were reviewed. Horses were included if they were presented for evaluation of acute abdominal pain and underwent exploratory laparotomy (including repeat laparotomy during hospitalization). Data collected for each case included signalment, type of colic, and incisional postoperative complications. Surgical site infection, herniation, and septic peritonitis were considered as surgical site complications (SSC). Complications were identified either during hospitalization or, in some cases, by the referring veterinarian after discharge.

Surgical site infection was defined as purulent discharge with or without a positive culture or the presence of serosanguinous discharge starting more than 24 h postoperatively. ²⁵ , ²⁶ , ²⁷ To evaluate the severity of the infections the following signs were considered: presence of purulent material, time between surgery and commencement of incisional discharge, ultrasonographic evaluation, positive culture with or without antibiotic‐resistant bacteria, depth of infection, development of hernia or peritonitis, and euthanasia due to this complication. Antimicrobial resistance was determined by the hospital‐based microbiology service using a broth dilution method; microtiter plates were read on a ThermoFisher Sensititer instrument with interpretations based on established Clinical and Laboratory Standards Institute recommendations.

For the purposes of this retrospective study, horses were considered to have peritonitis if the attending clinician documented the diagnosis in the medical record. The diagnosis of peritonitis was supported by clinical features such as abnormal appearance of fluid, increased peritoneal white blood cell count (WBC > 200 000 cells/μL at day 4 and remaining elevated at 40 000 cells/μL through day 4), increased mean protein concentration (>6 g/dL at day 6 after surgery) and the presence of clinical signs such as mild colic, depression, anorexia, fever, ileus, diarrhea, tachycardia or tachypnea. ²⁸ Hernia was defined as failure of the incisional suture with separation of a portion of the surgical incision.

Severe SSI was classified when at least three of the following were present: purulent material, incisional discharge within the first 5 days after surgery, infection involving the linea alba on ultrasound evaluation, positive culture with antibiotic‐resistant bacteria, secondary hernia or peritonitis, or euthanasia due to SSI. Severe peritonitis was defined as peritonitis with a positive culture that ultimately required euthanasia. Severe herniation was defined as protrusion of abdominal viscera through the linea defect with a clinical recommendation for surgical repair. Severe SSC were defined as those cases with at least one type of complication (SSI, herniation, or septic peritonitis) classified as severe.

No exclusion criteria were determined in this study, including no restrictions of age, breed, or sex of the horses included. Since this study was retrospective in nature, no minimum follow‐up duration was established.

2.2. Internal audit and fishbone diagram development

To address the perceived increase in severity of SSI, key stakeholders, including emergency and soft tissue surgeons, as well as internal medicine clinicians, house officers and microbiology specialists, performed an internal audit. This audit included retrospective evaluation of all the patients which developed SSI, including analysis of infections rates, microbial culture results, and postoperative complications. The goal was to collect data to identify patterns and support the clinicians' suspicion of the potential failure points in the previously established surgical preparation protocol. Additionally, given the complexity of the problem, external input was sought from Duke University. Based upon this collaboration, key themes for developing a successful intervention included group problem solving, staff engagement, data collection and refinement, and continuous monitoring.

As part of the process, stakeholders set out to create a fishbone diagram as a visualization tool to identify the root causes of the increased SSI rates. This diagram sorted potential contributors into categories and bundled different potential interventions to address multiple likely contributors with a single action. Ultimately, the goal was to identify the “top 10” likely contributors and improvements in the preoperative process that could address each one, while considering ease of implementation and its sustainability. Ultimately, the sustainability of the improvement was considered most essential by the group. The key contributory factors were identified and then organized and classified into the main skeleton and subsequent branches of the diagram. The main skeleton included six dimensions: (1) Materials, (2) personnel, (3) procedures, (4) medication, (5) environment, and (6) equipment. For each dimension, potential causes that could influence the outcomes were summarized within skeleton branches. The completed fishbone diagram was employed to design a new protocol to address deficiencies potentially contributing to ongoing issues.

2.3. Culture

To identify potential causative agents as part of the new protocol, two cultures were obtained from the ventral abdominal surface: one prior to creation of the skin incision and the other immediately after body wall closure but before closure of the subcutaneous and skin layers.

Cultures were collected using a sterile swab (BBL CultureSwab; Franklin Lakes, New Jersey) and sent to the microbiology laboratory at NCSU VTH. Results were typically returned within 48 h, accompanied by an antibiogram report if bacterial growth was detected. If the incision started draining after surgery, an additional culture was taken following the same protocol.

2.4. Statistical analysis

GraphPad Prism version 10 was used for all statistical analyses. Data are reported as absolute counts or percentages of horses with and without SSI and SSC before and after the new protocol implementation. Fisher's exact test was used for categorical comparisons, including overall and pairwise analyses between years and percent of horses euthanized. Additionally, univariate logistic regression was performed to evaluate changes in absolute numbers of SSI and SSC incidence from 2021 to 2024. The year 2021 was used as the variable of comparison to all years subsequent to the new protocol implementation. Outlier analysis was not applicable, as the outcomes were categorical. All available cases were included; no missing data were identified. Statistical significance was set to p < .05.

3. RESULTS

3.1. Study population

A total of 448 clinical case records of horses that underwent exploratory celiotomy for the treatment of acute abdominal pain were reviewed. Of the 448, 103 were euthanized intraoperatively and 345 were recovered. No complications were reported in 235 cases. Surgical site complications were reported in 57 cases of which 16 were ultimately euthanized and 41 survived until discharge (Figure 1, Table 1). Horses were classified as severe or not severe SSC, SSI, hernia, and peritonitis according to the predefined criteria (Table 2).

FIGURE 1.

Flow diagram of study population and postoperative outcomes.

TABLE 1.

Signalment of horses that developed SSC following exploratory celiotomy. Distribution of horses with SSI, peritonitis, hernia, or more than one SSC, according to gender, age, bodyweight, and type of colic.

Variable	SSI N = 28 (%)	Peritonitis N = 3 (%)	Hernia N = 9 (%)	>1 SSC N = 17 (%)
Gender
Female	9 (32%)	0 (0%)	0 (0%)	9 (53%)
Male	19 (67%)	3 (100%)	9 (100%)	8 (47%)
Age
<15	18 (64%)	2 (66%)	5 (66%)	12 (70%)
>15	10 (36%)	1 (44%)	4 (44%)	5 (30%)
Mean [range]	12 [1–23]	15 [13–19]	9 [2–26]	11 [3–28]
Weight
<500 kg	11 (39%)	0	4 (33%)	10 (59%)
>500 kg	17 (61%)	3 (100%)	5 (66%)	7 (41%)
Mean [range]	524 kg [323–737]	605 kg [570–678]	509 kg [420–659]	498 kg [88–722]
Type of colic
Small intestine	9 (32%)	1 (33%)	6 (66%)	7 (41%)
Large intestine	18 (64%)	1 (33%)	3 (33%)	8 (47%)
Small and large intestine	1 (3%)	1 (33%)	0 (0%)	2 (12%)

Note: Values are presented as number of cases.

Abbreviations: SSC, surgical site complication; SSI, surgical site infection.

TABLE 2.

Trends in surgical outcomes and complication rates over time (2019–2024). The table shows the proportion of horses that had complication per total of horses recovered from surgery and percentage of cases with complication in parenthesis. The orange vertical line indicates the integration of the new protocol.

	2019	2020	2021	2022	2023	2024
SSC	7/52 (13.5%)	23/59 (39.0%)	11/59 (18.6%)	9/63 (14.3%)	4/58 (6.8%)	3/44 (6.8%)
Not severe	3	12	8	9	4	2
Severe	4	11	3	0	0	1
SSI	4/52 (7.7%)	17/59 (29.0%)	11/59 (18.6%)	8/63 (12.7%)	3/58 (5.2%)	1/44 (2.3%)
Not severe	3	9	8	8	3	0
Severe	1	8	3	0	0	1
Peritonitis	1/52 (1.9%)	6/59 (10.2%)	3/59 (5.1%)	‐	‐	1/44 (2.3%)
Not severe	1	4	3	‐	‐	0
Severe	0	2	0	‐	‐	1
Hernia	3/52 (5.8%)	7/59 (11.8%)	1/59 (1.7%)	2/63 (3.2%)	1/58 (1.7%)	2/44 (4.5%)
Not severe	0	6	1	2	1	0
Severe	3	1	0	0	0	1

Abbreviations: SSC, surgical site complication; SSI, surgical site infection.

3.2. Critical deficiencies identified by fish bone analysis

The fish bone diagram which resulted from group problem solving (Figure 2) identified six key dimensions and possible contributors that provided actionable targets for modifying the current practices. Derived from the six key dimensions of materials, personnel, procedures, medication, environment, and equipment, the following key contributors were identified, and corresponding improvements were implemented to address them:

1. Organic debris remaining on the skin during initial preparation was addressed by introducing the routine use of detergent soap for the “dirty scrub.”

2. Use of contaminated or reused clipper blades posed a risk of introducing bacteria to the surgical site; this was addressed by performing clipping prior to entering the operating room and using a new blade for each use.

3. Inconsistencies in skin disinfection were corrected by adopting a standardized two‐step abdominal scrub using non‐sterile detergent‐based scrub followed by chlorhexidine/alcohol scrub until clean.

4. The risk of cross‐contamination between the sheath and the abdominal surgical field was addressed by assigning different individuals to scrub the different areas.

5. Surgical site antisepsis was improved by applying Iodine povacrylex and isopropyl alcohol solution (Duraprep, 3M Health Care, Maplewood, Minnesota) prior to incision.

6. A topical antimicrobial was applied to the incision prior to closure as a precautionary measure to reduce the risk of infection.

7. To minimize incision contamination during recovery, an abdominal stent bandage was routinely placed.

8. The potential for bacterial contamination from uncleaned vacuums and filters between horses was addressed by establishing a cleaning protocol to disinfect the vacuum and changing the filters.

9. Inconsistencies in postoperative abdominal bandaging and incision care were addressed by unifying the protocol.

10. The variability among surgical personnel was addressed by enhancing staff training, safe culture and communication protocols.

FIGURE 2.

Fishbone diagram highlighting the six key dimensions and bundled contributors.

Of these 10 items, three (Nos. 1–3 above) originated from the preoperative preparation period and five (Nos. 4–8 above) were related to intraoperative management. The final two elements, routine postoperative bandaging and monitoring, and training and communication, encompassed all parts of any intervention.

3.3. New protocol derived from fishbone analysis

Based on the key points identified on the fishbone diagram, a new protocol for perioperative management of exploratory celiotomy was developed with critical changes from the “old protocol” (detailed in Table 3). The “old protocol” served as a baseline for the modifications, and “new protocol” was designed following the surgical audits. Over the subsequent years, the protocol evolved in response to clinical outcomes, workflow, and staff feedback. The current version (current protocol) has several components of the “new protocol”; however, certain elements were changed and are described in Table 3.

TABLE 3.

Comparison of the key changes in pre‐, intra‐, and postoperative management following fishbone analysis.

		Old protocol	New protocol	Current protocol
Preoperative/Prior to induction	Grooming	Brushed	Brushed/feet cleaning	Brushed ± feet cleaning
	Clipping	Before or in surgery new or clean clipper blade	Before surgery with new clipper blade	Before or in surgery new or clean clipper blade
	Abdominocentesis	Dirty scrub: non‐sterile Chlorhexidine gluconate 4% w/v with alcohol rinse	Dirty scrub: non‐sterile 5 min detergent soap, non‐sterile Chlorhexidine gluconate 4% w/v with alcohol rinse	Unchanged from New protocol
	Standing two step abdominal scrub for abdominocentesis	Not performed or not consistent (only 1 step sterile prep performed)	Dirty scrub: non‐sterile 5 min detergent soap, non‐sterile Chlorhexidine gluconate 4% w/v with alcohol rinse	Unchanged from Old Protocol
Intraoperative	Clipping	Before or in surgery new or clean clipper blade	Before surgery with new clipper blade	Before or in surgery new or clean clipper blade
	Two step abdominal scrub	Not consistent	Yes: Detergent soap for 2 min followed by 5‐minutes Chlorhexidine gluconate 4% w/v scrub, rinse with sterile water and dried with a towel. Then, the sterile preparation, using sterile gloves and gauzes, 5 min chlorhexidine scrub and alcohol rinse until clean.	Unchanged from New protocol
	Sheath cleaning	Technician or HO	HO – penis preparation with povidone iodine and rinsed with sterile water. Place urinary catheter and the sheath closed horizontally with #2‐0 nylon suture.	Unchanged from New protocol
	Duraprep	No	Yes	No
	Skin culture prior to incision	No	Yes	Unchanged from New protocol
	Topical applied on linea	No	Amikacin pluronic gel	Manuka honey
	Stent placement	Clinician preference	Yes	Unchanged from New protocol
	Antibiotic redose	No	Every 80 min	No
	Vacuum cleaning	Yes	Yes	Unchanged from New protocol
	Filter change	Every 6 months	Once a month	Unchanged from New protocol
Postoperative	Supportive abdominal bandage	Clinician preference	Yes	Unchanged from New protocol
	Abdominal bandage and incision cleaning after surgery	Inconsistent	Inconsistent	Yes
	48 h bandage change	Inconsistent	Yes	Unchanged from New protocol
	Training	Inconsistent	Yes	Unchanged from New protocol

Abbreviation: HO, house officer.

3.4. Preoperative changes

Resultant preoperative preparation changes focused on patient management, scrub techniques and clipping. Patients were brushed and had their feet cleaned while in the stocks or stall, with abdominal clipping performed when possible. New clipper blades were employed to minimize microtrauma and contamination. The scrubbing procedure was modified to a two‐step preparation process for all invasive procedures performed including abdominocentesis and abdomen preparation for surgery. To ensure consistency, a training video demonstrating the abdominal preparation intraoperatively was shared and a checklist was implemented on every surgery before the incision was made.

3.5. Intraoperative changes

Additional modifications were made to the intraoperative period, particularly in surgical site preparation and protocol after closure (Table 3). A significant point of intervention identified through the analysis was adjustment of the roles during presurgical scrub, and the implementation of a two‐step scrub. The roles during the presurgical scrub were a technician who was in charge of the two‐step scrub while the house officer cleaned the sheath. Sheath preparation consisted of povidone iodine scrub of the glans penis which was then rinsed with sterile water, in addition to placement of a urinary catheter and gauze to maintain the penis within the prepuce. The sheath was then temporarily closed horizontally with 2–0 nylon suture. The two‐step scrub included a 2‐min scrub with detergent soap followed by a 5‐min scrub with chlorhexidine 4% v/w (E‐Z Scrub Red Impregnated scrub brush, Medline Industries), rinse with sterile water and dry with a non‐sterile surgical towel. This was followed by the sterile preparation using sterile gloves and gauzes which included 5 min of chlorhexidine scrub and alcohol rinse until clean. As a final antiseptic step prior to incision, an iodine povacrylex and isopropyl alcohol solution (Duraprep, 3M Health Care) was applied before a swab for microbial culture.

As demonstrated by fishbone analysis, no standardized protocol interventions were deemed necessary for the surgery itself until the linea closure. At that point, the new protocol dictated application of amikacin in pluronic gel at a concentration of 50 mg/mL to the linea after closing the body wall followed by coverage with 1–2 non‐adherent dressing pads (Telfa, Covidien, Mansfield, Massachusetts), a sterile surgical towel secured with No.2 nylon sutures in an interrupted Cushing pattern, and finally a iodophor‐impreganed adhesive drape (Ioban, Solvetum corporation, Bangaluru, India). A standardized systemic antibiotic protocol included potassium penicillin (22 000 IU/kg IV) and gentamicin (6.6 mg/kg IV) or enrofloxacin (7.5 mg/kg IV, if azotemic) administered prior to induction. Potassium penicillin was redosed intraoperatively every 80 min.

Finally, a new practice for cleaning vacuums was also established including brushing to remove organic debris, soaking in 0.5% hydrogen peroxide disinfectant (Rescue, Virox Technologies Inc., Ontario, Canada), and monthly changes of the vacuum filter instead of every 6 months.

3.6. Postoperative changes

Postoperatively, a protocol for incision management was established. Once the horse was standing, the iodine impregnated drape (Ioban) and stent bandage were removed. The incision was gently cleaned with 4% chlorhexidine gluconate followed by isopropyl alcohol by a veterinarian wearing sterile gloves. A sterile, absorbant cotton gauze pad was placed over the incision and covered with a reusable, fabric abdominal bandage (Kruuse, Langeskov, Denmark). The bandage was replaced if dirty or soiled before 48 h, and a new bandage was placed over the incision. The new protocol discouraged palpation of the incision unless directly indicated by concerns such as fever, discharge, or significant edema. If palpation was necessary, sterile gloves were mandatory. When an incision required ultrasound evaluation, additional precautions including scrubbing with betadine and alcohol, use of sterile gloves by the ultrasonographer, coverage of the probe by a sterile cover, and cleaning of the ultrasound and cord with 5% hydrogen peroxide disinfectant (Rescue, Virox Technologies Inc.).

Beyond technical modifications, the success of the protocol relied heavily on improvements in communication and team training, implementing a video training and intraoperative checklist. Over time, updates have been made to the protocol established in 2021 based on new literature published and clinical experience such as eliminating the exclusive use of clipper blades for each horse, not using Duraprep before midline incision, no redosing antibiotics during surgery and replacing amikacin with Manuka honey subcutaneously prior to skin closure. ²⁹ , ³⁰ Additionally, as complications reduced, the culture collection step was omitted.

3.7. Reduction of surgical site complication following new protocol implementation

In 2019, 52 of 66 horses who underwent exploratory celiotomy were recovered, and seven out of 52 horses (13.5%) developed incisional complications: four SSI (7.7%), three hernia (5.8%) and one peritonitis (1.9%). Of these seven, one had more than one complication, including SSI and a hernia at the same time. One of the SSC cases was diagnosed after discharge. In 2020, 59 of 76 horses were recovered, 23 out of 59 (39.0%) developed incisional complications: 17 SSI (29%), seven hernia (11.8%) and six peritonitis (10.2%). Of these 23 horses, eight had more than one concurrent complication: four with SSI and peritonitis and four with SSI and hernia. Four of the SSC cases were diagnosed after discharge. During the old protocol period, five horses with SSI progressed to peritonitis and required euthanasia due to uncontrolled infection and associated complications during hospitalization.

In January 2021, when the SOP was established, 59 of 82 were recovered, 11 out of 59 horses (18.6%) developed incisional complications: 11 SSI (18.6%), one hernia (1.7%) and three peritonitis (5.1%). Of these 11 horses, four developed multiple concurrent complications: three SSI and peritonitis and one SSI and hernia. In 2022, 63 of 84 were recovered, nine (14.3%) developed incisional complications: eight SSI (12.7%) and two hernia (3.2%). Of these nine horses, only one developed multiple concurrent complications; SSI and hernia. One SSC case was diagnosed after discharge.

Two years following implementation of the new protocol, in 2023, 58 of 82 horses were recovered from colic surgery, four (6.8%) developed incisional complications: three SSI (5.2%) and one hernia (1.7%). One SSC case was diagnosed after discharge. The following year in 2024, 44 of 58 were recovered, three horses (6.8%) developed incisional complications: one SSI (2.3%), one peritonitis (2.3%) and two hernia (4.5%). Of these three horses, one developed concurrent complications; SSI and hernia (Figure 3 and Table 2).

FIGURE 3.

Comparison of old to new protocol on the incidence of surgical site infection (SSI) and surgical site complication (SSC) over time. (A) The percentage of horses that developed SSI before and after implementation of the new protocol (orange dash line) was compared using Fisher's exact test. A significant reduction in SSI was observed following protocol implementation (p = .0382). A significant increase in SSI was identified between 2019 and 2020 (p = .0067). No significant difference between subsequent individual years was identified. (B) The percentage of horses that developed SSC before and after implementation of the new protocol (orange dash line) was compared using Fisher's exact test. A significant reduction in SSC was observed following protocol implementation (p = .0010). Significant yearly differences were identified between 2019 and 2020 (p = .0028) and 2020–2021 (p = .0246). No additional yearly differences were significant.

No horses required euthanasia due to SSC after implementation of the new protocol (0/34) compared to 5/35 (14.3%) before the protocol (p = .0536).

Overall, the incidence of SSC significantly decreased following implementation of the new protocol from 27% to 12.05% (p = .0010). A significant reduction in SSI was also observed after protocol implementation with a reduction from 19% to 10% (p = .0382) (Figure 3). Pairwise yearly comparisons identified significant differences between 2019–2020 for both SSI (p = .0067) and SSC (p = .0028), and between 2020–2021 for SSC (p = .0246). In the years following the implementation of the new protocol in 2021, horses were less likely to develop SSI (OR 0.48; 95% CI: 0.29–0.70, p = .0014; Table 4). Similarly, horses were less likely to develop SSC (OR 0.64, 95% CI: 0.42–0.94, p = .0237; Table 4).

TABLE 4.

Univariate logistic regression evaluating temporal incidence of SSC and SSI from implementation of new protocol in 2021.

	Variable (protocol year)	B	Wald statistics	OR	95% CI	p‐value
SSC	2021	−0.72	10.23	0.48	0.29–0.76	.0014
SSI	2021	−0.45	5.119	0.64	0.42–0.94	.0237

Abbreviations: B, regression coefficient; SSC, surgical site complication; SSI, surgical site infection.

3.8. Bacterial sampling

Of the 44 horses that developed SSI, 37 cultures were performed and 11 had multiple bacteria growing in the same culture. A total of 11 (11/37) were positive for Escherichia coli, 10 of which were resistant to multiple antibiotics. A total of 10 (10/37) were positive for Enterococcus spp., with 9/10 isolates showing resistance to different antibiotics. Nine (9/37) were positive for Staphylococcus spp., with 2/9 resistant to antibiotics. Eight (8/37) were positive for Enterobacter spp., and all were resistant to multiple antibiotics. Seven (7/37) were positive for Klebsiella spp., all of which exhibited multiple antibiotic resistance. One (1/37) was positive for Streptococcus spp. and resistant to one antibiotic whereas another (1/37) was positive for Pseudocitrobacter faecalis sp. and resistant to multiple different antibiotics. Lastly, one (1/37) was positive for Morganella morganii sp. with no resistance profile. The resistance profile of the different bacteria species is specified in Table 5.

TABLE 5.

Bacterial taxa isolated and their antimicrobial resistance profiles. Ranges indicates the minimum and maximum number of antibiotics to which isolates from each bacterial species were resistant.

Bacterial taxa	Number of cases	Number of resistant cases	Resistance profile (number of antibiotics with resistant characterization)
Escherichia coli sp.	11	10	7–11
Enterococcus spp.	10	9
Enterococcus gallinarum			4–6
Enterococcus faecalis			1–10
Enterococcus faecium			10
Staphylococcus spp.	9	2
Staphylococcus epidermidis			1
Staphylococcus haemolyticus			1
Enterobacter spp.	8	8
Enterobacter Hormachei			0–10
Enterobacter cloacae			5–8
Klebsiella spp.	7	7
Klebsiella pneumoniae			2–11
Klebsiella oxytoca			2
Klebsiella aerogenes			3
Streptococcus spp.	1	1	1
Pseudocitrobacter faecalis sp.	1	1	8
Morganella morganii sp.	1	0	0

Of the 37 horses that were cultured during hospitalization, 14 horses were not sampled before surgery because they were managed under the old protocol. The remaining 23 had sequential sampling, before incision and after incisional closure. Only one of the 23 horses sampled presurgery had a positive culture. At closure, nine of the 23 horses had a positive culture, eight developed SSI and the remaining horse was euthanized due to postoperative colic.

4. DISCUSSION

This study focused on a targeted intervention implemented at the NCSU VTH in 2021 to reduce SSI rate following exploratory celiotomy. It found that a targeted intervention, focusing on safety culture, group problem solving, staff engagement, data collection refinement, and continuous monitoring, significantly improved patient outcomes without additional expensive technology or additional ICU resources. Active stakeholder participation during internal audits and the use of a fishbone diagram were essential in identifying possibly overlooked protocol aspects, ultimately leading to meaningful improvements in surgical site management.

This retrospective analysis showed a steady decline in frequency of infection‐related complications following implementation of the new protocol. Although subcutaneous infections rarely extend into the peritoneal cavity, ³¹ an increased number of cases that initially presented with SSI progressed to peritonitis. These cases raised significant concerns affecting patient care and surgical decision‐making. Ultimately, five horses required euthanasia due to the inability to control the infection and its complications, serving as sentinels that highlighted the severity of the issue.

Collaboration with Duke University emphasized a holistic approach and introduced structured tools such as the fishbone diagram to guide the intervention design. This collaboration was critical because human medicine has established quality‐improvement frameworks to address challenges like those discussed here which veterinary medicine currently lacks. The fishbone diagram provided a reference for identifying the most influential contributors to the problem. This, combined with a root cause analysis, allowed the team to interrogate key categories to pinpoint the possible underlying causes and to find issues in our processes. In this analysis, as is the case for most quality control efforts, a singular cause for a problem was not identified. However, the most likely contributors were accurately prioritized. One key issue identified was inconsistency in the use of detergent soap (Ivory Ultra Concentrated Dish Soap, Procter & Gamble, Cincinnati, Ohio) for dirty scrub, especially after the retirement of a key surgical technician. The primary rationale for using ivory soap was its effectiveness in degreasing the skin and removing surface contaminants, thereby enhancing the efficacy of subsequent antiseptic agents. Other potentially impactful changes included the application of an antibacterial product before incision closure, use of stent and ioban dressing for recovery, and a post‐recovery care protocol. To evaluate the impact of interventions, clear data collection protocols were implemented with defined inclusion and exclusion criteria. As part of this process, we adopted a more inclusive SSI definition based on the National Healthcare Safety Network ³² and recent veterinary literature. ¹⁶ , ³³ , ³⁴ Although this may have increased the reported prevalence of SSI, it avoided overlooking any infected incision. ²⁵ , ²⁶ , ²⁷

The sustained and progressive reduction in SSI rates over multiple years likely reflects a combination of early protocol adjustments, data‐driven monitoring, and long‐term shifts in safety culture. ³⁵ , ³⁶ Establishing a safety‐focused culture created an environment where the entire care team felt supported in raising concerns and actively participating in process improvements. ³⁷ Structured onboarding, targeted training and real‐time feedback helped reinforce consistency in the protocol and built confidence between the team members. This confidence resulted in improvements in early infection recognition, team communication, continuous monitoring, regular internal audits and adherence to core practices, which appeared to have had a more durable impact than any single procedural detail. Notably, the reduction persisted even after some aspects of the SOP were relaxed, suggesting the importance of these foundational cultural changes, or that the relaxed measures were not significant contributors to the decrease in SSI.

Additionally, another invaluable component of the protocol was the routine collection of subcutaneous cultures from horses undergoing colic surgery. This practice was implemented to determine whether the source of infection could be related to inadequate surgical preparation or intraoperative contamination. This was particularly important because before the new protocol culturing had been inconsistent and performed only in response to complications, making it difficult to trace infection sources. Routine culture collection allowed us to confirm that surgical preparation was adequate and not a major contributor to infection. We also questioned whether the primary cause was antibiotic resistance or if other factors contributed to the problem. Culturing of the drainage has been shown in the literature to frequently identify potential hospital‐associated antibiotic resistance pathogen as it was in this case. ³⁸ While not all resistant bacteria were associated with severe cases, these findings emphasized the multifactorial nature of the SSI development. The most frequently isolated organisms were: E. coli, Enterococcus spp. and Staphylococcus spp. (Table 5), in agreement with existing literature. ⁷ Although the positive culture was not essential to consider an incision infected, we used the culture results as one of the severity indicators, based on the antibiotic resistance of the bacteria found. One of the latest changes of the protocol was the replacement of amikacin pluronic gel to medical‐grade Manuka honey in the incision. Manuka honey provides broad‐spectrum antimicrobial effects without contributing to antibiotic resistance. ²⁹ , ³⁰ This was guided by antimicrobial stewardship principles, aiming to reduce reliance on topical antibiotics while maintaining effective infection control.

The study has several limitations. Multiple changes were implemented at the same time, making it challenging to isolate the effect of individual components. Nonetheless, the reduction in SSI, hernia formation, peritonitis, and the absence of mortality due to SSC supports a strong association between the protocol and improved outcomes. There is potential underreporting of SSI and complications, as horses are typically discharged 7–10 days post‐surgery and subsequently cared for by their owners or referring veterinarians. It is possible that some owners did not realize that the horse had an incisional infection, or that they sought care from their primary veterinarian without contacting NCSU VTH. While minor infections may go unnoticed, communication with referring veterinarians and owners is generally reliable, as cases are reported even if managed at home. Finally, initial observations were based on clinical impressions rather than data collection; however, this study provides empirical support for those observations.

The reduction in the number of SSI following the implementation of the new perioperative protocol highlights the importance of analyzing events holistically and not as a specific single cause. Beyond the direct impact on infection rates, this study underscores the value of safety culture, where every team member feels responsible for patient care quality. Due to the ease of implementation and the minimal financial investment required, this protocol could be implemented by other equine veterinary hospitals, promoting standardized SSI control practices across different settings. The results from this intervention serve as a foundation for future studies and protocol refinements, reinforcing the importance of continuous evaluation and adaptation in veterinary hospitals.

AUTHOR CONTRIBUTIONS

Lopez Cruz C, DVM: Contributed to the design of the study, identified suitable medical records, compiled all data, interpreted data, wrote the draft and completed all revisions of the manuscript. Gonzalez LM, DVM, PhD, DACVS (Large Animal): Contributed to the design of the study, interpreted data, contribution to the initial drafting of the manuscript and extensively revised the manuscript. Hepworth‐Warren KL, DVM, DACVIM (Large Animal): Contributed to data collection and interpretation, elaborated and revised the manuscript. Roessner HA, DVM, DACVS (Large Animal): Contributed to data collection and interpretation, elaborated and revised the manuscript. Burke M, DVM, DACVS (Large Animal), DACVECC (Large Animal): Contributed to data collection and interpretation, elaborated and revised the manuscript. McKinney‐Aguirre CA, DVM, MPH, PhD, DACVIM (Large Animal): Contributed to the design of the study, interpreted data, contribution to the initial drafting of the manuscript and extensively revised the manuscript. Fogle CA, DVM, DACVS (Large Animal): Contributed substantially to the conception and the design of the study, acquisition, interpretation of the data, and extensively revised the manuscript. All the author provided a critical review of the manuscript and endorse the original version. All authors are aware of their respective contributions and have confidence in the integrity of all contributors.

CONFLICT OF INTEREST STATEMENT

No conflicts of interest to declare.