Surgical site infection is a significant cause of patient

Transcription

1 Original Research Using Bundled Interventions to Reduce Surgical Site Infection After Major Gynecologic Cancer Surgery Megan P. Johnson, PA-C, Sharon J. Kim, BA, Carrie L. Langstraat, MD, Sneha Jain, MHA, CSSBB, Elizabeth B. Habermann, PhD, Jean E. Wentink, RN, MPH, Pamela L. Grubbs, MS, APRN, Sharon A. Nehring, RN, BSN, Amy L. Weaver, MS, Michaela E. McGree, BS, Robert R. Cima, MD, Sean C. Dowdy, MD, and Jamie N. Bakkum-Gamez, MD OBJECTIVE: To investigate whether implementing a bundle, defined as a set of evidence-based practices performed collectively, can reduce 30-day surgical site infections. METHODS: Baseline surgical site infection rates were determined retrospectively for cases of open uterine cancer, ovarian cancer without bowel resection, and ovarian cancer with bowel resection between January 1, 2010, and December 31, 2012, at an academic center. A perioperative bundle was prospectively implemented during the intervention period (August 1, 2013, to September 30, 2014). Prior established elements were: patient education, 4% chlorhexidine gluconate shower before surgery, antibiotic administration, 2% chlorhexidine gluconate and 70% isopropyl alcohol coverage of incisional area, and cefazolin redosing 3 4 hours after incision. New elements initiated were: sterile closing tray From the Department of Obstetrics and Gynecology, Division of Gynecologic Surgery, the Division of Healthcare Policy and Research, Infection Prevention and Control, the Department of Nursing, the Surgery Research Office, the Division of Biomedical Statistics and Informatics, and the Department of General Surgery, Division of Colorectal Surgery, Mayo Clinic, and Mayo Medical School, Mayo Clinic, Minnesota. Presented at the American College of Surgeons National Surgical Quality Improvement Program Annual Meeting, July 25 28, 2015, Chicago, Illinois. The authors thank Karen Rucker and Cory Hiatt of the Mayo Clinic Revenue Cycle for their expert technical help with International Classification of Diseases, 9th Revision and Current Procedural Terminology code identification as well as Whitney Bergquist, PharmD, MBA, BCPS, for her assistance with pharmacy measure audits. Corresponding author: Jamie N. Bakkum-Gamez, MD, Department of Obstetrics and Gynecology, Mayo Clinic, Eisenberg Lobby 71, 200 First Street SW, Rochester, MN 55905; bakkum.jamie@mayo.edu. Financial Disclosure The authors did not report any potential conflicts of interest by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved. ISSN: /16 and staff glove change for fascia and skin closure, dressing removal at hours, dismissal with 4% chlorhexidine gluconate, and follow-up nursing phone call. Surgical site infection rates were examined using control charts, compared between periods using x 2 or Fisher exact test, and validated against the American College of Surgeons National Surgical Quality Improvement Program decile ranking. RESULTS: The overall 30-day surgical site infection rate was 38 of 635 (6.0%) among all cases in the preintervention period, with 11 superficial (1.7%), two deep (0.3%), and 25 organ or space infections (3.9%). In the intervention period, the overall rate was 2 of 190 (1.1%), with two organ or space infections (1.1%). Overall, the relative risk reduction in surgical site infection was 82.4% (P5.01). The surgical site infection relative risk reduction was 77.6% among ovarian cancer with bowel resection, 79.3% among ovarian cancer without bowel resection, and 100% among uterine cancer. The American College of Surgeons National Surgical Quality Improvement Program decile ranking improved from the 10th decile to first decile; risk-adjusted odds ratio for surgical site infection decreased from 1.6 (95% confidence interval ) to 0.6 ( ). CONCLUSION: Implementation of an evidence-based surgical site infection reduction bundle was associated with substantial reductions in surgical site infection in high-risk cancer procedures. (Obstet Gynecol 2016;127: ) DOI: /AOG Surgical site infection is a significant cause of patient morbidity and mortality as well as increased health care costs. 1 3 The high-complexity surgery often required to cytoreduce ovarian cancer carries a substantial risk for surgical site infection, which is VOL. 127, NO. 6, JUNE 2016 OBSTETRICS & GYNECOLOGY 1135

2 associated with worse overall survival. 4 For patients with uterine cancer, staging performed through laparotomy, rather than a minimally invasive approach, increases the risk of surgical site infection 15-fold. 5,6 With the feedback of validated outcomes data provided by the American College of Surgeons National Surgical Quality Improvement Program, the Centers for Medicaid and Medicare Services initiated the Surgical Care Improvement Project to reduce surgical complications. 7 Despite high compliance to guidelines for perioperative antibiotic administration, hair removal, normothermia, and glucose control, 1,7 these interventions alone have not been proven to lower surgical site infection rates, 8 which suggests additional interventions are needed. Surgical site infection risk can be complex with many risk factors that may or may not be modifiable, and there may not be adequate time to address modifiable risk factors. Therefore, perioperative bundles of evidence-based practices performed collectively and designed to reduce surgical site infections have been introduced The effect of bundled interventions has not been reported in high-risk gynecologic cancer surgeries. Our aim was to evaluate whether implementing a bundle of interventions that spans the entire surgical encounter, modeled after Cima et al, 13 could reduce 30-day surgical site infection rates by 50% in high-risk gynecologic cancer operations. MATERIALS AND METHODS The Mayo Clinic, Rochester campus is a large academic center and its Division of Gynecologic Surgery has eight board-certified gynecologic oncology surgeons, including one who retired during the preintervention period. All surgical site infections are captured by Mayo Clinic Infection Prevention and Control, which submits institutional surgical site infection data to the Centers for Medicaid and Medicare Services and the Centers for Disease Control and Prevention s National Healthcare Safety Network. The Mayo Clinic also participates in the multispecialty American College of Surgeons National Surgical Quality Improvement Program. A multidisciplinary team was assembled to retrospectively evaluate baseline surgical site infection outcomes among gynecologic surgery procedures. Despite long-term Surgical Care Improvement Project measure compliance, 13 the greatest proportion of surgical site infections in our specialty occurred among patients undergoing hysterectomy for gynecologic cancer. An intervention bundle previously shown to decrease surgical site infection in colorectal surgery 13 was selected, adapted, and implemented in gynecologic surgery as a prospective practice change using the Define, Measure, Analyze, Improve, and Control approach to data-driven improvement. 14 This project was reviewed by the Mayo Clinic institutional review board and, as a quality improvement project, was deemed exempt for initial implementation. However, institutional review board approval was obtained retrospectively for publication of findings. Only patients with consent to use their medical records, in accordance with the Minnesota Statute for Use of Medical Information in Research, were included in the published analyses. All laparotomies for ovarian cancer, including primary staging or debulking, interval debulking, and surgery for recurrent disease, and all total abdominal hysterectomies for primary uterine cancer performed during the project timeline were identified. Ovarian cancer cohorts encompassed all histologies and included epithelial ovarian, primary peritoneal, and fallopian tube cancers as well as sex cord-stromal tumors, germ cell tumors, carcinosarcoma, and borderline tumors. Individual patient encounters were used, because patients with recurrent ovarian cancer may undergo multiple surgeries over the course of their disease. The uterine cancer cohort also encompassed all histologies: endometrial carcinoma, carcinosarcoma, endometrial stromal sarcoma, and uterine leiomyosarcoma. Cases were categorized into the following three cohorts for analysis: 1) staging laparotomy for uterine cancer, 2) ovarian cancer debulking not requiring bowel resection, and 3) ovarian cancer debulking with bowel resection. Cohorts were identified using Current Procedural Terminology codes and International Statistical Classification of Diseases, 9th Revision diagnostic codes. Current Procedural Terminology codes for ovarian cancer procedures included 58700, 58720, 58940, 58943, 58950, 58951, 58952, 58953, 58954, 58956, 58958, and As a result of system constraints, the following Current Procedural Terminology codes, affecting approximately 1.3% of the preintervention ovarian cancer cohorts, were unavailable for the calendar year 2010: (salpingectomy, complete or partial, unilateral or bilateral [separate procedure]) and (salpingo-oophorectomy, complete or partial, unilateral or bilateral [separate procedure]). International Classification of Diseases, 9th Revision codes for ovarian cancer included 158.8, 158.9, 183.0, 183.2, 197.6, and Current Procedural Terminology codes for uterine cancer procedures included 58150, 58210, and International Classification of Diseases, 9th Revision codes included 182.0, 182.1, 1136 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery OBSTETRICS & GYNECOLOGY

3 and Each encounter was manually and independently verified by three authors (M.P.J., S.J.K., and J.N.B.-G.) to ensure criteria for both diagnosis and surgical procedure were met. Patient and cancer demographics and perioperative risk factors were abstracted for all three cohorts. Surgical site infections were defined according to the Centers for Disease Control and Prevention s National Healthcare Safety Network category 1B criteria as superficial, deep, and organ or space. 15 Thirtyday surgical site infection data were obtained from the institution s infection prevention and control division. This division is a separate entity from the division of gynecologic surgery, where a dedicated group of infectious disease specialists independently abstracts, reviews, classifies, and validates all potential surgical site infections through review of microbiology reports, readmission diagnoses, surgical listing, and reports from outside institutions on both inpatient and outpatient procedures (category IB). 15 This project used existing data that are part of an established hospital surveillance program within which reporting methodology did not change throughout its course. The surgical site infection reduction bundle consisted of 15 processes throughout the surgical encounter, including preoperative, intraoperative, postoperative, and postdismissal elements (Box 1). Elements in place during the preintervention period included: patient education on surgical site infection prevention, Hibiclens (4% chlorhexidine gluconate) shower before surgery, prophylactic antibiotic administration, ChloraPrep (2% chlorhexidine gluconate and 70% isopropyl alcohol) coverage of incisional area, and redose of cefazolin within 3 4 hours after incision. New components initiated during the intervention period were: sterile closing tray for fascia and skin closure, staff glove change (and gown if soiled) before fascia closure, dressing removal between 24 and 48 hours, dismissal with 4% chlorhexidine gluconate, and follow-up phone call from nurses. All patients undergoing gynecologic surgery are managed with an enhanced recovery protocol. 16 An established system to maintain normothermia is in place, hair is removed only when indicated, and blood glucose in diabetics is monitored throughout the perioperative period by a diabetes care team. Aside from standardized bundle elements and these institutional protocols, other perioperative care measures were surgeon-dependent. Performance metrics on tracked bundle elements (9/15) were collected during audits of electronic medical record documentation. Random audits of Box 1. Gynecologic Surgery Surgical Site Infection Reduction Bundle Elements Preoperative processes Preventing surgical site infection pamphlet for patient education 4% chlorhexidine gluconate shower night before and day of surgery Chlorhexidine cloths at morning admission Intraoperative processes Surgical Care Improvement Project compliance with antibiotic administration Complete coverage of incisional area with 2% chlorhexidine gluconate and 70% isopropyl alcohol solution Redose of cefazolin within 3 4 hours after incision Sterile closing tray for fascia and skin closure Staff glove change before fascia closure; gown change if soiled Postoperative processes Practice good hand hygiene Hand-cleansing agent readily available Ensure dressing removal within hours Patient shower with 4% chlorhexidine gluconate after dressing removal Patient education on wound care and infection symptoms Postdismissal processes Dismiss patient with 4-oz bottle of 4% chlorhexidine gluconate Follow-up phone call from nurses within hours Adapted from Figure 1 in Cima R, Dankbar E, Lovely J, Pendlimari R, Aronhalt K, Nehring S, et al. Colorectal surgery surgical site infection reduction program: a national surgical quality improvement program driven multidisciplinary single-institution experience. J Am Coll Surg 2013;216:23 33, with permission from Elsevier. the nursing elements were incorporated into routine weekly perioperative nursing care reviews and included providing the patient education pamphlet, hand hygiene, dressing removal, 4% chlorhexidine gluconate showering after dressing removal, dismissal with 4% chlorhexidine gluconate, postoperative patient education on wound care and infection symptoms, and follow-up phone call. Antibiotic compliance and cefazolin redosing were audited by the inpatient pharmacy for all cases. The preintervention period spanned January 1, 2010, through December 31, This period established baseline surgical site infection rates for VOL. 127, NO. 6, JUNE 2016 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery 1137

4 all three cohorts. The intervention period was from August 1, 2013, through September 30, Bundle implementation planning occurred between January 1, 2013, and July 31, Patient, procedure, and cancer characteristics were compared between the preintervention and intervention periods using the two-sample t test for continuous variables and the x 2 or Fisher exact test for categorical variables. Control charts, specifically p- charts, were created to plot overall, superficial incisional, and organ and space surgical site infection rates over time for each calendar month. Denoting the overall surgical site infection rate per time period as pbar, the upper and lower p confidence limits were calculated as pbar63 3 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðpbar½12pbarš=n i Þ, where n i is the number of encounters per month. Odds ratios and corresponding 95% confidence intervals (CIs) were estimated using logistic regression models for the odds of having a surgical site infection in the intervention period compared with the odds in the preintervention method. Adjusted odds ratios were obtained using stabilized inverse probability weights that were stabilized according to the proportion of procedures in the two time periods. The inverse probability weights were derived from a separate logistic regression model of the probability of being in the intervention period (compared with the preintervention period) that included the patient, procedure, and cancer characteristics listed in Tables 1 and 2 and trimmed at the 99th percentile. The covariate balance between the two periods was evaluated by calculating standardized differences for each of the covariates in both the original and weighted cohorts. 17 Analyses were performed using SAS 9.3. All calculated P values were two-sided and those,.05 were considered statistically significant. Mayo Clinic Rochester gynecologic surgery patients are identified and charts abstracted each month according to the National Surgical Quality Improvement Program s standard sampling methodology. Their semiannual report decile rankings and odds Table 1. Baseline Patient Characteristics in the Preintervention and Intervention Periods Characteristic Total (N5825) January 2010 to December 2012 (n5635) August 2013 to September 2014 (n5190) P* Age at surgery (y) BMI (kg/m 2 ) ASA class (1.7) 12 (1.9) 2 (1.1) (57.5) 357 (56.2) 117 (61.6) (37.6) 245 (38.6) 65 (34.2) 4 5 (0.6) 4 (0.6) 1 (0.5) Unknown 22 (2.7) 17 (2.7) 5 (2.6) Functional capacity,.001 Fewer than 4 METs 22 (2.7) 16 (2.5) 6 (3.2) 4 or more METs 441 (53.5) 377 (59.4) 64 (33.7) Unknown 362 (43.9) 242 (38.1) 120 (63.2) Diabetes 119 (14.4) 99 (15.6) 20 (10.5).08 Smoking status.19 Current smoker 69 (8.4) 57 (9.0) 12 (6.3) Former smoker 222 (26.9) 179 (28.2) 43 (22.6) Never smoker 532 (64.5) 397 (62.5) 135 (71.1) Unknown 2 (0.2) 2 (0.3) 0 (0.0) Cancer,.001 UC 301 (36.5) 253 (39.8) 48 (25.3) OC 524 (63.5) 382 (60.2) 142 (74.7) Procedure.001 OC without BR 369 (44.7) 269 (42.4) 100 (52.6) OC with BR 155 (18.8) 113 (17.8) 42 (22.1) UC (TAH) 301 (36.5) 253 (39.8) 48 (25.3) Intraoperative or postoperative blood 411 (49.8) 319 (50.2) 92 (48.4).66 transfusion Operative time (min) ,.001 BMI, body mass index; ASA, American Society of Anesthesiologists; MET, metabolic equivalent of task; UC, uterine cancer; OC, ovarian cancer; BR, bowel resection; TAH, total abdominal hysterectomy. Data are mean6standard deviation or n (%) unless otherwise specified. * x 2 or Fisher exact P value presented for categorical variables and two-sample t test P value presented for continuous variables Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery OBSTETRICS & GYNECOLOGY

5 ratios (the odds of a surgical site infection at our site compared with the odds at the average site enrolled in the program), adjusted for variables contemporary to each corresponding reporting period, were reported for the project time frame to serve as an additional validation measure. RESULTS A total of 825 cases that met criteria for this study were identified: 635 cases in the preintervention period and 190 cases in the intervention period. Compliance with nursing bundle elements was 97.6%. Surgical Care Improvement Project and cefazolin redosing compliance was 95.3%. Baseline patient characteristics are listed in Table 1. Cohort characteristics for uterine and ovarian cancer are listed in Table 2. Patient demographics were relatively well balanced between the preintervention and intervention periods with no difference in age, body mass index, American Society of Anesthesiologists class, diabetes, or smoking status. There was a higher proportion of ovarian cancer cases and longer operative times during the intervention period (Table 1). Among the uterine cancer cohort, there was a higher proportion of advanced-stage disease, sarcoma, and residual disease within the intervention period. Among those with ovarian cancer, there was a lower proportion of early-stage disease and a higher rate of suboptimal debulking (greater than 1 cm residual disease) during the intervention period (Table 2). The baseline overall surgical site infection rate during the preintervention period was 6.0%. After implementation of the reduction bundle, the overall surgical site infection rate declined to 1.1% (P5.01) (Fig. 1). Among ovarian cancer without bowel resection, the overall surgical site infection rate dropped from 4.8% to 1.0% for a relative risk reduction of 79.3% (P5.12). The overall surgical site infection rate Table 2. Uterine and Ovarian Cancer Characteristics in the Preintervention and Intervention Periods Characteristic Total January 2010 to December 2012 August 2013 to September 2014 P* Uterine cancer FIGO stage,.001 I to II 196 (65.1) 176 (69.6) 20 (41.7) III to IV 105 (34.9) 77 (30.4) 28 (58.3) Histology.02 Endometrioid 165 (54.8) 144 (56.9) 21 (43.8) Nonendometrioid 93 (30.9) 79 (31.2) 14 (29.2) Sarcoma 43 (14.3) 30 (11.9) 13 (27.1) Neoadjuvant chemotherapy,.001 No 291 (96.7) 250 (98.8) 41 (85.4) Yes 10 (3.3) 3 (1.2) 7 (14.6) Residual disease.02 Microscopic, none 282 (93.7) 241 (95.3) 41 (85.4) 1 cm or greater 15 (5.0) 10 (4.0) 5 (10.4) Less than 1 cm 4 (1.3) 2 (0.8) 2 (4.2) Ovarian cancer FIGO stage.048 I to II 133 (25.4) 106 (27.7) 27 (19.0) III to IV 328 (62.6) 236 (61.8) 92 (64.8) Recurrence 63 (12.0) 40 (10.5) 23 (16.2) Histology.82 Serous 380 (72.5) 276 (72.3) 104 (73.2) Nonserous 144 (27.5) 106 (27.7) 38 (26.8) Neoadjuvant chemotherapy.20 No 442 (84.4) 327 (85.6) 115 (81.0) Yes 82 (15.6) 55 (14.4) 27 (19.0) Residual disease.02 Microscopic, none 394 (75.2) 290 (75.9) 104 (73.2) 1 cm or greater 104 (19.8) 79 (20.7) 25 (17.6) Less than 1 cm 26 (5.0) 13 (3.4) 13 (9.2) FIGO, International Federation of Gynecology and Obstetrics. Data are n or n (%) unless otherwise specified. * x 2 or Fisher exact P value. VOL. 127, NO. 6, JUNE 2016 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery 1139

6 among ovarian cancer with bowel resection decreased from 10.6% to 2.4% for a relative risk reduction of 77.6% (P5.19). Among hysterectomy for uterine cancer, the overall surgical site infection rate declined from 5.1% to 0.0% (P5.23), despite the fact that patients were more likely to have advanced-stage uterine cancer in the intervention group. The relative risk reduction for superficial incisional surgical site infections was 100% ( % P5.08) (Fig. 2A). The organ and space surgical site infection relative risk reduction was 73.3% ( %; P5.05) (Fig. 2B). Among ovarian cancer without bowel resection, the rate declined from 3.0% to 1.0% (P5.45), ovarian cancer with bowel resection from 8.8% to 2.4% (P5.29), and among uterine cancer from 2.8% to 0.0% (P5.60). There were only two deep incisional surgical site infections within the project, both of which occurred in cases of ovarian cancer without bowel resection in the preintervention group. Overall the odds of having a surgical site infection in the intervention period relative to the preintervention period were 0.17 (95% CI ) (Table 3). The adjusted odds ratio, derived using inverse probability weighting, was similar at 0.13 (95% CI ) Inverse probability weighting was used to create a weighted sample in which the distributions of measured baseline covariates were more similar between the two time periods. The distribution of the stabilized weights was acceptable (mean 0.99, standard deviation 0.38, range ). All of the levels of the covariates in the weighted sample had standardized differences below the recommended threshold of 0.10 with the exception of two (operative time 0.135; stage I and II 0.113). Among all gynecologic surgery cases performed at Mayo Clinic Rochester, the National Surgical Quality Improvement Program surgical site infection decile ranking improved from the 10th decile (preintervention period [January 1, 2012, to December 31, 2012]) to the first decile (April 1, 2013, to March 31, 2014) after implementing the bundle; likewise, the risk-adjusted odds ratio for surgical site infection decreased from 1.6 (95% CI ) to 0.6 (95% CI ) during the same time periods. DISCUSSION Surgical site infection is a key factor in the vitally important measures of patient morbidity and mortality. After bundle implementation, our overall rate of surgical site infection decreased significantly, exceeding our goal and supporting this institutional practice change. Several bundles designed to reduce surgical site infection have been reported with similar results. In 2013, van der Slegt et al 12 observed a 51% surgical site infection reduction in vascular surgery cases on implementation of a four-component bundle, which included hair removal, prophylactic antibiotics, normothermia, and operating room discipline. In 2014, Waits et al 9 attained a 2.0% risk-adjusted rate in colorectal surgery after implementing a bundle of prophylactic antibiotics, normothermia, oral antibiotics with bowel preparation, glycemic control, minimally invasive surgery, and short operative duration. They illustrated an inverse association between number of bundle elements implemented and risk-standardized surgical site infection rate, strengthening the bundle concept. Cima et al 13 then adopted a bundle resulting Fig. 1. Overall monthly surgical site infection rates for patients undergoing laparotomy for ovarian or uterine cancer plotted over time. The baseline rate decreased from 5.9% during the preintervention period to 1.5% after bundle implementation (P5.01). The vertical blue line represents the planning for implementation period (January 1, 2013, to July 31, 2013), the solid black line and points represent the observed values, the dashed gray lines are the average surgical site infection rate for each period, and the top gray line is the upper confidence limit. Johnson. Reducing Surgical Site Infection in Gynecologic Cancer Surgery. Obstet Gynecol Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery OBSTETRICS & GYNECOLOGY Copyright ª by The American College of Obstetricians

7 Fig. 2. Superficial incisional (A) and organ/space (B) surgical site infection rates for patients undergoing laparotomy for ovarian or uterine cancer plotted over time. The vertical blue line represents the planning for implementation period (January 1, 2013, to July 31, 2013), the solid black line and points represent the observed values, the dashed gray lines are the average surgical site infection rate for each period, and the top gray line is the upper confidence limit. Johnson. Reducing Surgical Site Infection in Gynecologic Cancer Surgery. Obstet Gynecol VOL. 127, NO. 6, JUNE 2016 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery 1141 Copyright ª by The American College of Obstetricians

8 Table 3. Comparison of Surgical Site Infection Rates Between the Preintervention and Intervention Periods Type of SSI or Procedure January 2010 to December 2012 August 2013 to September 2014 P* RRR (%) Unadjusted OR Adjusted OR (95% CI) (95% CI) All SSI All procedures 6.0 (38/635) 1.1 (2/190) ( ) 0.13 ( ) OC without BR 4.8 (13/269) 1.0 (1/100) ( ) 0.11 ( ) OC with BR 10.6 (12/113) 2.4 (1/42) ( ) 0.28 ( ) UC (TAH) 5.1 (13/253) 0 (0/48) Superficial Incisional SSI All procedures 1.7 (11/635) 0 (0/190) OC without BR 1.1 (3/269) 0 (0/100) OC with BR 1.8 (2/113) 0 (0/42) UC (TAH) 2.4 (6/253) 0 (0/48) Organ or space SSI All procedures 3.9 (25/635) 1.1 (2/190) ( ) 0.20 ( ) OC without BR 3.0 (8/269) 1.0 (1/100) ( ) 0.18 ( ) OC with BR 8.8 (10/113) 2.4 (1/42) ( ) 0.34 ( ) UC (TAH) 2.8 (7/253) 0 (0/48) Deep SSI All procedures 0.3 (2/635) 0 (0/190) OC without BR 0.7 (2/269) 0 (0/100) OC with BR 0 (0/113) 0 (0/42) 100 UC (TAH) 0 (0/253) 0 (0/48) 100 SSI, surgical site infection; RRR, relative risk reduction; OR, odds ratio; CI, confidence interval; OC, ovarian cancer; BR, bowel resection; UC, uterine cancer; TAH, total abdominal hysterectomy. Data are % (n/n) unless otherwise specified. * x 2 or Fisher exact P value. The OR represents the odds of having an SSI in the intervention period compared with the odds in the preintervention period. The adjusted ORs were obtained using stabilized inverse probability weights. Did not estimate the OR given the lack of events in the intervention period. in a significant decline of overall surgical site infection rate from 9.8% to 4.0% in colorectal surgery. We chose this bundle for implementation because approximately 40% of our ovarian cancer surgeries require bowel resection. 18 We demonstrated statistically significant reductions when comparing overall surgical site infection rates. However, improvements were not significant among subanalyses by procedure or infection classification as a result of small sample sizes. Among the ovarian cancer without bowel resection subgroup, we would have needed an intervention sample size equivalent to the preintervention size of 269 to have had 80% power to declare the observed difference statistically significant. Among ovarian cancer with bowel resection, a sample size of 125 per period would be needed. Nevertheless, we consider the improvement clinically relevant and note it occurred despite the intervention group having a higher rate of ovarian cancer with bowel resection, advanced uterine cancer, longer operative times, and a smaller proportion of stage I disease. Strengths of this quality improvement project include its multidisciplinary approach and the independent identification and reporting of infections by both the infection prevention and control division and National Surgical Quality Improvement Program. The reporting methodologies for those two groups were the same for both periods, and abstractors were blinded to when the intervention period began. This may have limited the Hawthorne effect. 19 Additionally, sustained reduction in surgical site infection rates over the 14 months after bundle implementation suggests that results are attributable to the process rather than reactive health care provider effort. Furthermore, we used high-volume and high-risk cases to determine intervention effectiveness. However, we did not eliminate surgical site infection from our practice. The effect of the bundle on surgical site infection rates among other at-risk procedures and patient populations is needed. An additional limitation was that there were differences in patient characteristics between preintervention and intervention groups. On using inverse probability weighting to adjust for the imbalance in 1142 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery OBSTETRICS & GYNECOLOGY

9 the measured covariates, the unadjusted and adjusted odds ratios were similar for all procedures combined. The one-time cost to build a resterilizable closing pan was $ However, this quality improvement project did not include a cost analysis, and bundle cost-effectiveness warrants further study. Additional limitations include that we do not know surgical site infection rates for the 7-month planning for implementation period and that compliance of patientdependent elements was not quantified. Further means for improving surgical site infection rates are emerging. Preoperative eradication of methicillin-resistant Staphylococcus aureus colonization has been shown to decrease surgical site infection rates, and we have previously identified prior infection to be an independent risk factor of surgical site infection in endometrial cancer surgery. 6 Prophylactic negative pressure wound therapy has also been shown to reduce abdominal wound surgical site infection rates by up to 70%. 23,24 Finally, oral antibiotic administration has been shown in retrospective studies to decrease surgical site infection rates in colorectal surgery However, less than 50% of patients undergoing surgery for ovarian cancer require a bowel resection, and the risks of oral antibiotics with mechanical bowel preparation need to be weighed against the benefits of surgical outcomes. REFERENCES 1. Anderson DJ, Podgorny K, Berrios-Torres SI, Bratzler DW, Dellinger EP, Greene L, et al. Strategies to prevent surgical site infections in acute care hospitals: 2014 update. Infect Control Hosp Epidemiol 2014;35: Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton DJ. The impact of surgical-site infections in the 1990s: attributable mortality, excess length of hospitalization, and extra costs. Infect Control Hosp Epidemiol 1999;20: Klevens RM, Edwards JR, Richards CL Jr, Horan TC, Gaynes RP, Pollock DA, et al. Estimating health careassociated infections and deaths in U.S. hospitals, Public Health Rep 2007;122: Tran CW, McGree ME, Weaver AL, Martin JR, Lemens MA, Cliby WA, et al. Surgical site infection after primary surgery for epithelial ovarian cancer: predictors and impact on survival. Gynecol Oncol 2015;136: Young H, Berumen C, Knepper B, Miller A, Silverman M, Gilmartin H, et al. Statewide collaboration to evaluate the effects of blood loss and transfusion on surgical site infection after hysterectomy. Infect Control Hosp Epidemiol 2012;33: Bakkum-Gamez JN, Dowdy SC, Borah BJ, Haas LR, Mariani A, Martin JR, et al. Predictors and costs of surgical site infections in patients with endometrial cancer. Gynecol Oncol 2013;130: Bratzler DW, Hunt DR. The surgical infection prevention and surgical care improvement projects: national initiatives to improve outcomes for patients having surgery. Clin Infect Dis 2006;43: Hawn MT, Vick CC, Richman J, Holman W, Deierhoi RJ, Graham LA, et al. Surgical site infection prevention: time to move beyond the surgical care improvement program. Ann Surg 2011;254: Waits SA, Fritze D, Banerjee M, Zhang W, Kubus J, Englesbe MJ, et al. Developing an argument for bundled interventions to reduce surgical site infection in colorectal surgery. Surgery 2014;155: Bull A, Wilson J, Worth LJ, Stuart RL, Gillespie E, Waxman B, et al. A bundle of care to reduce colorectal surgical infections: an Australian experience. J Hosp Infect 2011;78: Revolus T, Tetrokalashvilli M. Implementation of evidencebased innovative bundle checklist for reduction of surgical site infection. Obstet Gynecol 2014;123(suppl 1):32S. 12. van der Slegt J, van der Laan L, Veen EJ, Hendriks Y, Romme J, Kluytmans J. Implementation of a bundle of care to reduce surgical site infections in patients undergoing vascular surgery. PLoS One 2013;8:e Cima R, Dankbar E, Lovely J, Pendlimari R, Aronhalt K, Nehring S, et al. Colorectal surgery surgical site infection reduction program: a national surgical quality improvement program driven multidisciplinary single-institution experience. J Am Coll Surg 2013;216: Stamatis DH. Six sigma fundamentals: a complete introduction to the system, methods, and tools. New York (NY): Productivity Press; Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. Guideline for prevention of surgical site infection, Centers for Disease Control and Prevention (CDC) Hospital Infection Control Practices Advisory Committee. Am J Infect Control 1999;27: Kalogera E, Bakkum-Gamez JN, Jankowski CJ, Trabuco E, Lovely JK, Dhanorker S, et al. Enhanced recovery in gynecologic surgery. Obstet Gynecol 2013;122: Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med 2015;34: Bakkum-Gamez JN, Langstraat CL, Martin JR, Lemens MA, Weaver AL, Allensworth S, et al. Incidence of and risk factors for postoperative ileus in women undergoing primary staging and debulking for epithelial ovarian carcinoma. Gynecol Oncol 2012;125: McCarney R, Warner J, Iliffe S, van Haselen R, Griffin M, Fisher P. The Hawthorne Effect: a randomised, controlled trial. BMC Med Res Methodol 2007;7: Pofahl WE, Goettler CE, Ramsey KM, Cochran MK, Nobles DL, Rotondo MF. Active surveillance screening of MRSA and eradication of the carrier state decreases surgical-site infections caused by MRSA. J Am Coll Surg 2009;208: Kim DH, Spencer M, Davidson SM, Li L, Shaw JD, Gulczynski D, et al. Institutional prescreening for detection and eradication of methicillin-resistant Staphylococcus aureus in patients undergoing elective orthopaedic surgery. J Bone Joint Surg Am 2010;92: Schweizer ML, Chiang H, Septimus E, Moody J, Braun B, Hafner J, et al. Association of a bundled intervention with surgical site infections among patients undergoing cardiac, hip, or knee surgery. JAMA 2015;313: Swift SH, Zimmerman MB, Hardy-Fairbanks AJ. Effect of Single-Use Negative Pressure Wound Therapy on Postcesarean Infections and Wound Complications for High-Risk Patients. Reprod Med 2015;60: VOL. 127, NO. 6, JUNE 2016 Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery 1143

10 24. Soares KC, Baltodano PA, Hicks CW, Cooney CM, Olorundare IO, Cornell P, et al. Novel wound management system reduction of surgical site morbidity after ventral hernia repairs: a critical analysis. Am J Surg 2015;209: Cannon JA, Altom LK, Deierhoi RJ, Morris M, Richman JS, Vick CC, et al. Preoperative oral antibiotics reduce surgical site infection following elective colorectal resections. Dis Colon Rectum 2012;55: Toneva GD, Deierhoi RJ, Morris M, Richman J, Cannon JA, Altom LK, et al. Oral antibiotic bowel preparation reduces length of stay and readmissions after colorectal surgery. J Am Coll Surg 2013;216: Kiran RP, Murray AC, Chiuzan C, Estrada D, Forde K. Combined preoperative mechanical bowel preparation with oral antibiotics significantly reduces surgical site infection, anastomotic leak, and ileus after colorectal surgery. Ann Surg 2015; 262: ACOG Fellows and Journal Subscribers: Get Obstetrics & Gynecology for Free on your ipad The Green Journal app is available for the ipad from Apple s App Store SM. To view journal content, ACOG Fellows and journal subscribers must first register and activate an account at If You Have Not Registered at GreenJournal.org: 1. On click on the gear box at the top right corner of the screen and select Register. 2. On the registration screen, choose a username and password and enter your address. (Usernames must be at least 6 characters in length and contain no spaces or symbols; passwords must be at least 8 characters in length and contain at least one number and one letter.) 3. Click Continue to go to the next step of user registration. 4. On the next screen, enter your name and address and click Continue. 5. The next registration screen asks for additional information about you and your practice to help us recommend articles and rich media that suit your area of specialty. After entering this information, indicate your acceptance of the End User License Agreement and click Complete Registration. 6. After you complete the registration, you will receive an from the site asking you to confirm your registration. Click on the link in the within 48 hours. 7. The link in the leads to a web page where you will be asked if you want to activate your online subscription. Click on Yes! I am a subscriber and want to activate my online subscription. 8. Enter your ACOG Member ID or subscriber ID in the field at the bottom of the next screen. Be sure to enter all characters; note that the last digit for ACOG members is the letter I, not the number 1. Then click on Activate Subscription. If You Have Registered at GreenJournal.org But Not Activated Your Subscription: 1. On click on the gear box at the top right corner of the screen and select Register. 2. Click on Activate Your Subscription at the top right of the page. 3. Enter your ACOG Member ID or subscriber ID in the field at the bottom of the next screen. Be sure to enter all characters; note that the last digit for ACOG members is the letter I, not the number 1. Then click on Activate Subscription. You are now registered for full-text access on both the app and the journal web site! When you access the Green Journal app on your ipad, you will be asked to enter the username and password that you created at Need Help? Contact the Editorial Office at or obgyn@greenjournal.org. Apple and ipad are trademarks of Apple Inc., registered in the U. S. and other countries. App Store is a service mark of Apple Inc. rev 2/ Johnson et al Reducing Surgical Site Infection in Gynecologic Cancer Surgery OBSTETRICS & GYNECOLOGY