Reducing surgical site infection rates in colorectal surgery – a quality improvement approach to implementing a comprehensive bundle

Abstract Aim Surgical site infections (SSIs) are a preventable cause of morbidity following surgical procedures. Strategies to reduce rates of SSI must address pre‐, peri‐ and postoperative factors and multiple interventions can be combined into ‘bundles’. Adoption of these measures can reduce SSIs, but this is dependent on high levels of compliance. The aim of this work is to assess the change in rates of SSI in elective colorectal surgery after implementing a colorectal SSI bundle. Method This is a single‐centre prospective cohort study. All elective colorectal procedures from 2011 until 2018 (inclusive) were included. The primary outcome was inpatient SSI. A multimodal bundle was implemented using quality improvement methodology. The bundle was altered during the timeframe of the study to optimize outcomes. Data were analysed by interrupted time series analysis assessing points at which the bundle was altered. Results In the study period, 1075 elective colorectal procedures were performed. Prior to the introduction of the colorectal SSI bundle, the SSI rate was 16.4%. During the implementation period (2013–2015), the overall rate of SSI fell from 15.9% to 9.4%, with the most significant reduction being in superficial SSI, from 8.6% to 4.7%. In the postimplementation period from 2015–2018, there was a further reduction in the overall rate of SSI (5.1%). In 2018, there were 87 consecutive cases without infection. Conclusion A successful reduction in the rate of SSI following elective colorectal surgery can be achieved by adopting a comprehensive perioperative bundle. This is complemented by a process of continuous measurement and evaluation. The current bundle has achieved a significant reduction in superficial SSI.


INTRODUC TI ON
The colon contains approximately 10 14 live luminal bacteria, including multiple human pathogens [1]. Any operation in which the colon is opened is therefore a clean-contaminated or a contaminated procedure, with an inherent risk of infection in the surrounding deep tissue or skin surface. The reported rates of surgical site infection (SSI) following colorectal surgery vary (often 10%-30% [2][3][4]), but are consistently higher than the rates for other general surgical specialities [5]. However, high rates of SSI following colorectal surgery should not be accepted as inevitable or unchangeable.
SSIs are the most common cause of healthcare-associated infections (HAIs) and confer an additional morbidity and mortality to the surgical procedure originally performed. SSIs can contribute to a prolonged hospital stay and increased readmission and intervention rates. They can delay rehabilitation and a return to normal activity, which may have a significant psychological impact [6]. Furthermore, SSIs confer an additional 3% risk of mortality after colorectal resection [3]. This is significant, given that colorectal resection is a common elective procedure in the UK (n = 18,796 in 2019) [7]. As a result, SSIs have a substantial economic burden, with an associated 35% increase in direct healthcare costs [8,9].
Interventions to reduce rates of SSI in colorectal surgery are therefore needed to optimize both patient care and healthcare bud- which have reduced the rates of SSI across all surgical specialities [10,11]. However, considerable heterogeneity remains in the components of colorectal-specific SSI bundles. Despite this heterogeneity, recent meta-analyses have shown that, regardless of these variations, implementing a SSI bundle reduces the risk of infection following colorectal surgery by up to 40% [2,3].
Our unit is a large district hospital serving the Highlands and Islands of Scotland, UK with an estimated population of 330,000.
On average, 120-150 elective colorectal procedures are performed annually by full-time colorectal surgeons in a dedicated theatre.
Prospective surveillance of colorectal SSIs following elective surgery has been ongoing in our unit since 2011. An initial 6 month pilot study showed higher than expected rates of SSI in this patient cohort, prompting the creation of a multidisciplinary working group to translate emerging evidence on SSI bundles into a practical strategy to lower the rates in elective colorectal patients.
The aim of this paper is to describe the dynamic process of implementing a comprehensive perioperative colorectal SSI bundle using quality improvement methodology. We also aim to assess the resultant effect on SSI rates after bundle adoption in this singlecentre cohort study.

Patient selection and data collection
This is a single-centre prospective patient cohort and interrupted time series (ITS) study. The study period was from January 2011 to December 2018. The inclusion criterion was patients aged ≥16 years undergoing elective (planned) colorectal resection at Raigmore Hospital, Inverness, Scotland. Both open and laparoscopic operations were included. SSI was defined as infection of the index surgical procedure occurring in an inpatient, using the internationally recognized Scottish Patient Safety Programme (SPSP) definitions of superficial, deep and organ/space SSI [12]. The exclusion criteria were paediatric patients, noncolorectal procedures and patients undergoing emergency colorectal operations.
A hospital-based surveillance system for colorectal SSIs was developed between February and June 2011. Data were initially collected using a paper proforma kept in the patient's medical notes and then transferred to a digital database. An online data collection tool was subsequently developed to allow accurate and timely recording of data within the operating and recovery areas. Postoperative data collection was carried out by the infection control team (independent of the colorectal team). All SSIs identified in an inpatient, readmissions and reinterventions within 30 days were recorded.
However, because the mechanism of recording involved the secondary sector only, community events were not collected.

Implementation of change
Before implementation of the bundle, individual clinician preference predominated in relation to aspects of the bundle used for each case.
After implementation, the aim was that each aspect of the bundle be utilized for every case and compliance measured.
A colorectal SSI working group was created in December 2012 comprising colorectal and general surgeons, anaesthetists, microbiologists, scrub nurses, colorectal nurse specialists, ward nurses, hospital infection control team, patient safety and quality improvement advisors and hospital clinical governance advisors. Monthly meetings were held to address predicted issues in the bundle implementation. At these meetings, the constituent parts of the bundle were agreed by the group. Any changes to the bundle components were agreed by this group before they were implemented, based on how easy the interventions were to implement and anticipated compliance rates, in a pragmatic approach. A target of 95% compliance for each bundle component was agreed. Involvement from the Scottish Patient Safety Programme (SPSP) from October 2014 helped to create a standardized data collection tool.

What does this paper add to the literature?
Surgical site infections (SSIs) confer additional morbidity and mortality to elective colorectal surgery. We demonstrate that a coordinated multidisciplinary approach to implementing a 'bundle' of interventions can successfully reduce rates of SSI. We propose that adoption of this bundle is generalizable to other units and could act as a baseline for future work in this field.
Bundle implementation included education of ward and theatre staff to highlight the rationale for the change in practice and to help create a shared sense of responsibility for its development. There was dedicated teaching, which included face to face sessions as well as online modules to augment knowledge. Staff were actively encouraged to challenge nonadherence to the bundle where observed and to record any difficulties arising from factors such as lack of access to equipment. Relevant aspects of the bundle were incorporated into the patient safety brief and surgical pauses (sign in, time out and sign out) to help ensure compliance.
Additional aids, such as a dedicated whiteboard and clock detailing antibiotic timing for the case, were installed in the colorectal theatre.

Compliance and follow-up
A robust evaluation of the bundle through continuous audit was undertaken. SSI incidence and bundle compliance were discussed monthly by the colorectal SSI group. Results were also disseminated electronically in the form of a monthly report. All recorded cases of SSI during the study period had a root cause analysis conducted to identify patient-specific and system-based issues. In instances where bundle compliance was less than 100%, justification for this was sought from key members of staff involved. Common themes, as well as specific issues which arose from these, were used as motivators for further improvement.

Outcome measures
The primary outcomes were overall rates of SSI and rates of superficial SSI during the inpatient stay within secondary care. Readmissions and reinterventions were secondary outcomes.

Statistical analysis
The primary aim of this study was to assess rate of overall, superficial and deep SSI over time in our unit. Generalized estimating equations  It was felt that a single month would be sufficient to implement each bundle change as the specific components of each change were relatively straightforward to undertake and were agreed upon by all stakeholders before their introduction. Results are summarized as preintervention period slope, change in level after each period, change in slope after each period and period slope. Figure S1 in the Supporting Information illustrates these effects. Demographic characteristics were also summarized by year and by period using appropriate summary statistics. All analysis was carried out in Stata 16 (Statacorp).  Compliance data were continually collected and are shown in Figure S2.     (Table 3).

DISCUSS ION
In this prospective, single-centre cohort study, we observed a sig-  [15]. In this regard, the ROSSINI trial failed to demonstrate that use of wound edge protectors as a single intervention that could significantly reduce the rate of SSI following laparotomy [15].
However, subsequent meta-analyses have demonstrated decreasing odds of SSI by their use [16]. In short, it is challenging to analyse individual interventions for reduction of SSI in the complex aetiology of wound infections in colorectal surgery.
A comprehensive approach to modifying risk factors for SSI appears to be required, with targeted pre-, intra-and postoperative interventions. Comparable single-centre studies show that the introduction of various colorectal SSI bundles can have a significant impact on the incidence of SSIs. In one Australian study of 408 patients, the crude SSI rate reduced from 15% to 7% following introduction of a multi-intervention SSI bundle, despite variable compliance with individual bundle elements [4]. A larger study of 5120 from the USA with similar rates of open versus laparoscopic colorectal surgery was able to demonstrate a reduction in SSI rates from 9.8% to 4% following implementation of a multifactor SSI bundle [17]. These findings are corroborated by two recent meta-analyses, comprising 8515 patients and 17,551 patients, respectively, which confirm a significant reduction in colorectal SSI rates after implementation of a dedicated bundle of care [2,3]. Our study adds to the literature by demonstrating the positive impact on infection rates of dynamic changes in a SSI bundle in line with changes in published best practice.
The evidence-based bundles outlined above will have a limited effect without effective strategies for implementation. Compliance with SSI bundles in published studies is highly variable, ranging from 19% to 99% [3,4,17]. Unsurprisingly, those with the lowest rates of overall compliance following bundle introduction often report little or no change in rates of SSI [14]. Compliance with individual bundle elements also has an additive effect in lowering the rate of SSI [17]. The staff commitment required to fully implement a successful bundle should not be underestimated. Indeed, it has been suggested that staff participation is such a vital factor that it should be incorporated in the SSI bundle as a specific intervention [14]. Each member of the team, from ward staff to theatre, needs to appreciate and feel valued in their role. Shared ownership of bundle interventions can help to move away from viewing SSIs as a purely 'surgical problem' [18]. By ensuring buy-in and agreement by all stakeholders throughout this process, our compliance with each aspect of the bundle has remained high throughout.
One of the strengths of our approach was staff engagement during the process of bundle design, and implementation is crucial.
A multidisciplinary work pattern appears to be most successful in reducing SSI rates [17,19] [20,21]. Visual reminders (such as documentation of timings for subsequent antibiotic doses) was another simple strategy which we found to be effective. Limiting factors, such as a lack of forcing air warmers and other equipment, must be addressed at an organizational level if bundle implementation is to be successful [4]. Staff engagement is also key to building a sustainable system.
Establishing the infrastructure to monitor the consistency of practice over time, as well as identifying developing issues, builds longerterm success. Many studies demonstrate success in reducing SSI rates over 12 months or less, but our study provides more substantive longitudinal data. Documenting and sharing this success has been vital in maintaining staff motivation. In addition, collection of follow-up data by an independent auditor or via an electronic system helps to maintain the integrity of the data [22]. Perhaps more importantly, rigorous interrogation of cases in which a SSI occurred or where compliance was poor is integral to improvement.
We acknowledge that this study has several limitations. Firstly, it is a single-centre cohort study. However, the challenges faced in this work are likely to be familiar to most other units. We therefore think that this approach would be effective in other colorectal departments The HAI report published by NHS England highlighted that colorectal surgery had the biggest variability in SSI risk between participating hospitals [5]. Thus, sharing best practices from single centres helps to establish a framework for reduction of colorectal SSI nationally, allowing all patients to benefit.
Other limitations include a lack of information on comorbidity, smoking status or immunosuppressive medications or conditions, all of which can influence the rate of SSI. BMI was also not collected for the final 2 years of this study, which may have influenced some of the SSI rates. Length of operation was also not recorded in this work.
Furthermore, our analysis only included elective patients. Although transferable to the emergency setting, this has not yet been universally established in our unit. Due to the methods of data collection, we were only able to analyse inpatient events, readmissions and reinterventions. This will therefore be an underestimation of the overall rate of SSI. However, we feel that the severe SSIs will be captured in this work and we had a consistent approach throughout this study.
A further limitation is that we have not undertaken a cost analysis.
Work to evaluate cost-effectiveness is ongoing. We also acknowledge that the SSI trend was decreasing slightly before the start of this implementation, and therefore other factors not measured in this study may have influenced the improvements we observed.
A further limitation is that only in-hospital SSIs were identified in this assessment, missing those that will have been detected in the community.
Our experience shows that a significant and sustained reduc-  Colorectal units must therefore consider how best to translate the available evidence into a strategy to lower rates of postoperative SSI [23]. Successful implementation of these strategies relies on 'buy-in' from the wider clinical team. In addition, a robust system of audit and evaluation is needed to ensure both consistent implementation and to facilitate critical appraisal when a SSI does occur. This allows the bundle to change and evolve over time, promoting further falls in SSI rates and achieving parity with those reported for many 'cleaner' specialities with traditionally lower rates of postoperative infection [5].

ACK N OWLED G EM ENTS
This paper has been written and published on behalf of the NHS Highland Colorectal SSI Working Group and the named authors would like to formally acknowledge the contribution of all those involved in this collaboration.

CO N FLI C T O F I NTE R E S T S
There are no conflicts of interest or anything else to disclose.

E TH I C A L A PPROVA L
The study was approved by the NHS Highland clinical governance team. Formal ethics review was not deemed to be required.

AUTH O R CO NTR I B UTI O N
AJMW conceived the project and edited the manuscript. The NHS Highland SSI group co-designed the programme and collated the data. JH performed statistical analysis and co-authored the manuscript. RF performed data analysis and co-authored the manuscript.
GR performed data analysis and co-authored the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.