Reduction of adverse outcomes from cesarean section by surgical- site infection prevention care bundles in maternity

Objective: To reduce average surgical- site infection (SSI) rates to less than 7.5%, as well as other complications by incrementally implementing an SSI prevention care bundle in maternity: (1) ChloraPrep; (2) PICO dressings, performing elective cesarean sections in a main theater rather than a labor ward and warming blankets; (3) vaginal cleansing; and (4) Hibiscrub. Methods: In this prospective cohort study, the association between categorical variables was assessed by χ 2 tests, temporal trends in the monthly percentage change of SSI were measured using the Joinpoint Regression Program v4.7.0.0. Results: In all, 1682 women (mean age 33.1 ± 5.2 years) underwent either elective (53.9%) or emergency (46.1%) cesarean section. After a small initial increase (10.0%– 11.8%), SSI progressively declined to 4.4% ( χ 2 = 22.1, P < 0.001), as did sepsis, reoperation or readmission for SSI: from 12.5% to 0.5% ( χ 2 = 90.1, P < 0.001). The rates of SSI fell progressively with the cumulative introduction care bundle components. The average monthly percentage change was −14.0% (95% confidence interval −21.8% to −5.4%, P = 0.004), and the average SSI rate was kept below 7.5% for the last 12 months of the study. Conclusion: The maternal SSI prevention care bundle is simple

USA they account for 15% of all nosocomial infections. 9 More than a decade ago in England, SSI occurred in about 10% after cesarean sections. 5 Previous studies have shown that SSI could be reduced by wellorganized surveillance and infection control programs, 10 and national benchmarking for SSI rates would serve as a powerful driver to achieve this. The Surgical-Site Infection Surveillance Service was therefore set up, including an aim to use local data to follow SSI rates over time and provide a benchmark to guide clinical practice.
The Getting It Right First Time (GIRFT) programme extended this initiative and focused on involving frontline clinicians to collect data. 11 As part of the GIRFT national audit, data from 269 patients in the initial 4-month period (June 1 to September 30, 2019) revealed that the rate of SSI was relatively high in the Ashford and St Peter's Hospitals NHS Foundation Trust (16.3%). We therefore strived to improve outcomes by implementing an SSI prevention care bundle in maternity, in addition to the standard care, from August 1, 2020. Because there was no existing consensus of benchmark targets for SSI after cesarean section, we therefore set an ambitious local target of 7.5% to achieve over a period of 17 months.
This target was about half of the rate of SSI detected in our initial GIRFT audit. The objective of this study was to assess changes in SSI over time following the implementation of the SSI prevention care bundles in maternity.

| Study design, participants, and setting
In this prospective cohort study, data were collected over a period of 17 months in women undergoing cesarean section at a single National Health Service hospital. This centre serves a catchment area of over 400 000 people. The present study was conducted in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study does not require NHS Research Ethics Committee approval because it involves secondary analysis of anonymized data. Informed consent was not required for the present study as it is observational and mandates no deviation from routine medical practice.

| Recruitment
Patients were identified as completely as possible by a number of approaches: (1) informing all colleagues in the Department of Obstetrics and Gynecology of the audit, producing a Patient Information Leaflet about SSI so that patients would present to our triage department rather than their general practitioner, and informing midwives in community of this referral pathway; (2) entering any SSI diagnosed onto the Maternity Badgernet database (routinely used by maternity services in the UK) 12 to be easily identified for the audit; and (3) attempting to contact every patient directly, 6 weeks postnatally, to capture SSI that might have been treated by a general practitioner.

| Implementation of SSI prevention care bundles
Based on the Plan, Do, Study, Act (PDSA) approach, procedures from the SSI prevention care bundle were introduced in a stepwise manner, at least 1 month apart, in order to assess their efficacy. 13

| Demographic and clinical measures
Age, history of diabetes, and ethnicity of patients were documented as routine. Body mass index (BMI; calculated as weight in kg divided by the square of height in meters) was calculated at booking.
Information on SSI after cesarean section, sepsis, reoperation, and readmission to hospital for SSI was recorded using the GIRFT proforma.
Ethnicity of the patients was recorded as Caucasian, South Asian (Indian, Pakistani or Bangladeshi), Black (Caribbean or African), mixed race (biracial or multiracial backgrounds), and others (namely East Asians other than the aforementioned ethnic groups). Elective or emergency cesarean sections indicate planned or unplanned procedures. Diabetes included those who had a history of type 1 or type 2 diabetes, and those who were diagnosed during pregnancy (gestational diabetes mellitus). Diagnostic criteria for sepsis were based on Royal College of Obstetrics and Gynecology guidelines including clinical features (impaired mental state), respiratory distress (tachypnea or desaturation), hypotension, circulatory collapse or dehydration (tachycardia), hypothermia or fever, and changes in skin discoloration. The diagnosis was supported by biochemical abnormality including raised inflammatory markers (C-reactive protein or white cell counts), leukopenia, and high lactate levels, as well as evidence of organ dysfunction. 14 SSI was based on patients who did not acquire an SSI (low-risk group) and all those who had superficial incisional, deep incisional, or organ/space SSI (high-risk group).
A composite variable was created for the remaining outcomes containing a low-risk group (no adverse outcomes) and a high-risk group (any of sepsis, reoperation, or readmission for SSI).

| Statistical analysis
Chi-squared tests (SPSS version 28; IBM) were used to assess differences between adverse outcomes of cesarean section (SSI and the composite variable of sepsis, reoperation, or readmission) and the periods when each procedure of the care bundles was introduced. Temporal trends in SSI over time were identified using the Joinpoint Regression Program 4.7.0.0. 15 This technique detects join points in data sets and calculates the monthly percentage change for individual linear segments (i.e., different slopes) when one or more join points exist, as well as an average monthly percentage change for the entire period of study. 16 It is driven by the data-measuring discontinuities in the data-and not by the introduction of different variables. The null hypothesis was rejected when P values were less than 0.05.  Bivariate (unadjusted) and partial correlations to adjust for age, BMI, and history of diabetes, showed that these factors did not change the relationship between SSI and the duration of study (bivariate correlation: unadjusted r = 0.171, P < 0.001 versus partial correlation: adjusted r = 0.174, P < 0.001).

| RE SULTS
The rates of any of the three adverse outcomes (sepsis, reoperation for SSI and readmission to hospital) also diminished with time after implementation of the SSI prevention care bundle. However, F I G U R E 1 Temporal trends in surgical-site infection (SSI) rates following implementation of the SSI prevention care bundles in maternity. Square symbol represents baseline data preceding the application of care bundles; the blue dotted line indicates the local benchmark target of 7.5%. the number of these outcomes was too low for temporal trend analysis. With the exception of a small initial increase after ChloraPrep inclusion (10.0%-11.8%), there was a progressive decline in SSI: to 11.8%, 7.7%, and 4.4% (χ 2 = 22.1, P < 0.001). In addition, there was a similar decline in any complications (sepsis, reoperation, or readmission for SSI): from 12.5% to 4.4%, 0.5% and 0.5% (χ 2 = 90.1, P < 0.001) corresponding to the addition of each care bundle component ( Figure 2).

| DISCUSS ION
The present study showed that SSI and three adverse events (sepsis, reoperation, and readmission after cesarean section) were reduced by an SSI prevention care bundle. The agents used in the care bundles are inexpensive and require minimal training for their application by new staff members. This is the first study, as far as we are aware, to examine cumulative effects of the successive introduction of components of an SSI prevention care bundle. Use of a Joinpoint regression analysis, revealed changes with the addition of each procedure so that all four components of the care bundle contributed to SSI reduction.
Perioperative bundles have been introduced into other surgical disciplines to reduce SSI, such as colorectal surgery 17,18 and gynecologic cancer surgery. 19 However, they have not been used routinely in obstetric surgery. A survey of 197 academic centers in the USA showed that 55.3% provided appropriate antibiotic dosing for patients with obesity, 77.2% used chlorhexidine-based skin antisepsis, but only 12.7% applied vaginal antisepsis. 20 In the last 12 months of our study, average SSI rates were kept below an average local target rate of SSI (7.5%), indicating the beneficial effects of the care bundles. Our findings support the routine application of an SSI prevention care bundle in maternity care.
The initial rates of SSI of 16.3% before the introduction of the care bundles were at the upper end of published figures in high-income countries such as the USA 10 and within the UK. 4-6 Measurement of the rate of SSI in our study was obtained by thorough contact tracing of almost all patients after hospital discharge. However, it is well recognized that the rates of SSI are underreported in the published literature. 21 This is because the majority of SSI (up to 84%-89%) occur after hospital discharge. 4,8 Many of these, particularly the less complex cases such as superficial incisional SSI, are treated in the community and therefore may be missed by the hospital, leading to such an underestimation. 7 The other factor contributing to wide variation in reporting is due to differences in the criteria used to define an SSI (e.g., exclusion of superficial incisional SSI or the duration of surveillance). [6][7][8] Furthermore, the definition of SSI for cesarean section tends to vary between reports, which makes interstudy comparison difficult. 6 We adopted the generally accepted WHO definition where SSI occurring within a month following cesarean section are classified into three types: superficial incisional, deep incisional, and organ/space. 22 Cesarean section is one of the most performed obstetric surgical procedures. 23 Global rates of cesarean section have been increasing progressively over the past few decades, but vary widely between countries; from 15.5% of births in Finland to 53.1% in Turkey in 2015, 24 whereas rates are reported as 32% in the USA 25 and Australia, 26 and 41% in China. 27 In England, cesarean section rates were just 9% of deliveries in 1980, but more than tripled to 30% in 2018-2019. 28 Given the high rates of SSI among the global high numbers of cesarean deliveries, the burden of care and financial costs are significant. 29 The strengths of this study lie in its completeness of patient recruitment in one of the largest catchment areas of England; with the characteristics of patients being representative of the UK population.
The protocol for SSI prevention care bundles was followed strictly and the data were gathered using standard proforma. Although the F I G U R E 2 Rates of adverse events after cesarean section in response to stepwise addition of each procedure of the surgical-site infection (SSI) prevention care bundles. Group differences for SSI: χ 2 = 22.1, P < 0.001, and for any of sepsis, reoperation or readmission: χ 2 = 90.1, P < 0.001. sample size was large and adequate for temporal trend analysis for SSI, this type of analysis was not sufficient to analyze other adverse events including sepsis, reoperation, and readmission because of their exceptionally low event rates. However, an alternative analysis using χ 2 tests also showed significant reduction in the proportion of patients with adverse events and therefore strengthened findings from the temporal trend analysis. In addition, the duration of our study was long enough to demonstrate the effects of separate components of the care bundle to reduce SSI. However, a longerterm study involving other centers would be necessary to establish a national benchmark for SSI prevention. Certain limits arise from this study including the lack of control group for the period of study; therefore, bias may have been introduced such as improvement in the overall patient care because all members of the department were alerted of the national audit. Certain additional factors may also have confounded the results, such as improvement of skill and experience in clinical management by staff. However, a seasonal effect is unlikely as the duration of the study was beyond 1 year. Information on patient nutritional status and living environment was not collected in our study, which may potentially have some bearing on our findings.
However, we found that other health-related factors (age, BMI, and history of diabetes) did not influence the association between SSI and the 17-month study period, and is likely to reflect relatively constant patient characteristics throughout the period of study.
In conclusion, SSI prevention care bundles in maternity using simple, inexpensive procedures are effective to reduce SSI after a cesarean section, and should be offered routinely to all women undergoing this procedure.

AUTH O R CO NTR I B UTI O N S
Matthew Erritty, Joann Hale, James Thomas, Anna Thompson, Ria Wright, Anna Low, Megan Carr, Richard George, Lisa Williams, Alexandra Dumitrescu, Jacqui Rees, and Shashi Irukulla were involved in patient care and data collection. Thang S. Han was responsible for data analysis, the writing of the first draft and revision of the paper. All authors contributed to the development of the manuscript, checked, interpreted results, and approved the final version.
All authors comply with all of the ICMJE authorship criteria.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.