Sustained Reduction in Surgical Site Infection after Abdominal Hysterectomy

Abstract

Background:

During a period of five years, the rate of surgical site infection (SSI) after abdominal hysterectomy at our institution was >10%. With the implementation of a multifaceted intervention designed to reduce this, the rate of SSI fell to <2% in the post-intervention period. The pre- and post-intervention periods were compared to determine which of the interventions in the multifaceted array of interventions was most valuable in decreasing SSI.

Methods:

A retrospective chart review was done to identify: (1) Parameters associated with SSI, and (2) parameters that differed in the pre- and post-intervention periods. The intervention included providing departmental SSI rates to the gynecology faculty, re-educating operating room (OR) staff personnel about appropriate perioperative antibiotic choice and timing, and changing the preferred sterile preparation for abdominal surgery from 10% povidone–iodine (PI) to 4% chlorhexidine gluconate (CHG). The preliminary results of our review also led to the suggestion that surgeons use blood products sparingly, although an absolute threshold for transfusion was not specified.

Results:

Twenty-one of 192 patients (10.7%) developed an SSI in the pre-intervention period, whereas 1 of 84 patients (1.2%) developed an SSI in the post-intervention period (p=0.006). Surgical site infection was associated with obesity (a body mass index [BMI] ≥30) (11.5% vs. 4.8%, p=0.04), receipt of a blood transfusion (18.2% vs. 6.6%, p=0.03), and abdominal skin preparation with PI as opposed to CHG (10.1% vs. 2.0%, p=0.07). Chlorhexidine gluconate was used more commonly for abdominal skin preparation in the post- than in the pre-intervention period (6.6% pre-intervention vs. 50.7% post-intervention, p <0.0001).

Conclusions:

A multifaceted intervention decreased dramatically the rate of SSI after abdominal hysterectomy at our institution. No single component of the intervention could be identified as most responsible for the improvement.

Surgical site infection (SSI) is the second most common type of health care–associated infection. The rate of SSI after abdominal hysterectomy is reported to range from 1%–4% [1]. The rate of SSI after abdominal hysterectomy is one of the publicly reported measures that will be linked to reimbursement by the Centers for Medicare and Medicaid Services (CMS) beginning in January 2014 [2]. Using logistic regression modeling and specific patient-level information, the National Healthcare Safety Network (NHSN) will calculate an expected number of SSIs per institution. Institutions with more SSIs than expected will be penalized [2].

Over a period of five years, the rate of SSI after abdominal hysterectomy at our institution was >10%, exceeding the 90th percentile for rates of SSI rates published in the 2009 report of the NHSN [1]. We investigated this rate, [3] instituted a multifaceted intervention to prevent SSI after hysterectomy, and were pleased to see rates of SSI fall to <2% in the post-intervention period. The goal of the present study was to compare formally the rates of SSI in the pre- and post-intervention periods and determine which intervention was most valuable in decreasing the occurrence of SSI.

Methods

Study design

We conducted a retrospective review of all cases of elective abdominal hysterectomy at a single institution between December 31, 2005, and September 27, 2011. The intervention intended to prevent SSI in this procedure was begun in March 2010, and the pre-intervention period was defined as to the period before March 9, 2010, and the post-intervention period as the period after this date. Patients who had an emergent abdominal hysterectomy or operations for tubo-ovarian abscess in the 30 d following abdominal hysterectomy, and those who had no follow-up visits during this period, were excluded. The primary outcome measure in our review was SSI in the 30 d after abdominal hysterectomy. A secondary outcome was adherence to the core measures to prevent SSI set forth by the CMS.

The electronic medical record at the institution in which we conducted our review encompasses records from all inpatient and outpatient providers. It includes a combination of laboratory values, dictated reports, and scanned, hand-written progress notes. This was used to gather data about inpatient and outpatient medical care, radiology studies, and the results of microbiology tests. Parameters of interest included age, body mass index (BMI), past medical and surgical histories, identity of the operating surgeon, length of time in the operating room, concomitant operative procedures, estimated blood loss (EBL), duration of hospitalization, use of blood transfusion, pre-operative complete blood count (CBC) and serum creatinine concentration, and post-operative hemoglobin concentration and hematocrit. Additional data about the skin preparation product used for hysterectomy and the use of a razor to prepare the operative field were also gathered. Surgical site infection was determined in accord with the definition established by the U.S. Centers for Disease Control and Prevention, through a manual review of patients' charts [4].

Data relating to the CMS core measures were collected. The CMS core measures for preventing SSI after abdominal hysterectomy include the administration of peri-operative antibiotics within 60 min prior to surgical incision, appropriate choice of antibiotic (cefazolin, cefotetan, cefoxitin, cefuroxime, or ampicillin-sulbactam; or clindamycin plus an aminoglycoside, fluoroquinolone, or aztreonam; or metronidazole plus an aminoglycoside or fluoroquinolone), the discontinuation of antibiotic within 24 h of the surgical procedure, appropriate removal of hair, removal of a urinary catheter on post-operative day 2, and prevention of peri-operative hypothermia (i.e., a temperature on arrival in the post-anesthesia care unit (PACU) of ≥36°C) [5].

Intervention

The intervention included providing departmental rates of SSI to the gynecology department faculty, re-educating operating room (OR) staff members about appropriate perioperative antibiotic choice and timing, and changing the preferred abdominal skin preparation product from 10% povidone–iodine (PI) to 4% chlorhexidine gluconate (CHG; Table 1). On the basis of preliminary results following institution of the intervention, it was also suggested that surgeons use blood products sparingly, although an absolute threshold for transfusion was not specified.

Table 1.

Components of Intervention and Incidence in the Pre- and Post-Intervention Periods

Intervention	Pre-intervention period	Post-intervention period	p value
Feedback on SSI rate provided to gynecology department	10.7%	1.2%	0.006
Correct antibiotic choice per CMS criteria	94.3%	94.0%	0.55
Antibiotic administration within 60 min of incision	96.3%	100%	0.64
Blood transfusion within 14 d of surgery	12.6%	10.7%	0.66
Skin preparation with CHG	6.6%	50.7%	<0.0001

CHG=4% chlorhexidine gluconate; CMS=Centers for Medicare and Medicaid Services; SSI=surgical site infection.

Statistical analysis

Descriptive statistics were used to analyze patient demographics and surgical characteristics for the purpose of identifying potential differences between variables in the pre- and post-intervention periods. Associations between categorical variables were determined with χ² tests of independence, and Student t tests of differences in means were used to test for associations between continuous variables and the outcomes of interest. Logistic regression was used to determine whether risk factors identified through bivariate analysis with a value of p≤0.3 were associated independently with the development of SSI. Data were analyzed with SAS version 9.3 (SAS Institute: Cary, NC). The Colorado Multiple Institutional Review Board approved the study.

Results

Six hundred seventy-seven hysterectomies, 311 of which were abdominal and 366 vaginal, were performed at our institution during the study period. Five hundred ten procedures were performed in the pre- and 167 in the post-intervention period. Two hundred twenty-three (56.3%) of the patients who underwent abdominal hysterectomy did so in the pre-intervention period, and 88 patients (47.3%) underwent abdominal hysterectomy in the post-intervention period (p=0.044). One hundred ninety-two patients (86.1%) were included in the pre-intervention group in the study and 84 (95.5%) were included in the post-intervention group.

Twenty-one patients (10.7%) developed SSIs in the pre-intervention period and one patient (1.2%) developed SSI in the post-intervention period (p=0.006; Fig. 1). The indications for surgery, patient age, and patient medical and surgical histories were similar in the patients who developed SSIs in the two time periods. Surgical site infection was associated with obesity (BMI ≥30; 11.5% vs. 4.8%, p=0.04), receipt of a blood transfusion (18.2% vs. 6.6%, p=0.03), and abdominal skin preparation with PI as opposed to CHG (10.1% vs. 2.0%, p=0.07). By multiple logistic regression analysis, blood transfusion (odds ratio [OR] 3.30, 95% CI 1.13–9.61) and increasing BMI (OR 1.08 for each 1 kg/m² increase, 95% CI 1.02–1.13) were the strongest predictors of SSI. Abdominal skin preparation with PI rather than CHG was associated with a nonsignificant increase in the risk of developing SSI (OR 5.14, 95% CI 0.67–39.66).

FIG. 1.

Types and rates of surgical site infection before and after implementation of a multifaceted intervention designed to prevent such infections.

The prevalence of obesity and blood transfusion did not change from the pre- to the post-intervention periods (Table 1) although obese patients may require higher-dose prophylaxis [6]. Chlorhexidine gluconate was used more commonly for abdominal skin preparation in the post-intervention period (6.6% pre-intervention vs. 50.7% post-intervention, p<0.0001). Compliance with the CMS core measures of use of appropriate antibiotic, and antibiotic administration in the 60 min before surgical incision were respectively >90% and not statistically different in the pre- and post-intervention periods. Patients were less likely in the post- than in the pre-intervention period to have a temperature of <36°C on arrival in the PACU (76.6% pre-intervention vs. 93.8% post-intervention, p=0.001), although this was not associated with SSI.

Discussion

A multifaceted approach, including reporting of rates of SSI to the gynecology department, improving compliance with CMS measures, and changing the skin preparation used for abdominal hysterectomy resulted in a dramatic decrease in the rate of SSI after hysterectomy at our institution. No single intervention appeared to be most responsible for the decrease in SSI.

Active surveillance for SSI has been associated with decrease rates of its occurrence [7 –10]. Intermittent interruptions in active SSI surveillance have been associated with increased rates of SSI.⁷ However, the effect of providing feedback to individual surgeons and departments has not been studied widely, and in the few studies published on this topic, there has been little objective change in individual performance [11,12]. Using the standardized infection ratio, our institution has begun to provide confidential rates of SSI for individual surgeons. We hope to study the effect of individual-level data on surgeon rates of SSI in the near future.

Although both PI and CHG decrease the number of colony-forming units of bacteria, CHG has been shown in both a randomized controlled trial and a systematic review to decrease SSI by approximately 40% below its frequency with PI [13,14]. Our institution has now adopted CHG as the preferred skin preparation for all sterile procedures; skin preparation in procedures that involve incision through a mucus membrane is still done with PI.

Unlike life-threatening medical conditions, in which physicians provide intensive care to patients unless they decline treatment explicitly, surgeons have the autonomy to determine which patients will undergo an operation and which will not [15]. This complex decision making includes the risks of operating and not operating, the chances of operative and non-operative cure, and the estimated likelihood of survival after surgery or without surgery. With the addition of hospital pay-for-performance systems based on rates of SSI after abdominal hysterectomy, surgeons will soon have another variable added to this equation. The reduction in rates of SSI at our institution was made in the absence of “cream-skimming,” or choosing patients at lower risk from surgery. Interestingly, the relative percentage of abdominal hysterectomies as compared to vaginal hysterectomies at our institution decreased between the pre- and post-intervention periods. We hypothesize that gynecologists may have elected the vaginal approach because of its lower risk of SSI.

Our analysis of the pre-intervention period alone indicated that blood transfusion was associated with an increased risk of SSI, especially in patients with an EBL 500 mL [3]. The relationship between SSI and blood transfusion in patients with a lower EBL was also confirmed by our group in a multi-center study [16]. We were surprised that the post-intervention improvement in the rate of SSI occurred in the absence of a change in the proportion of patients who received blood transfusion. We hypothesize that the absolute number of transfusions may have decreased (i.e., fewer transfusions per patient from the pre- to the post-intervention period) or that patients who received transfusions in the post-intervention period had different characteristics than those who had transfusions in the pre-intervention period.

Limitations of this study include its retrospective design. Although we collected information on variables that we suspected were most associated with SSI, there may have been other unmeasured factors that contributed to the sustained decrease in SSI. In particular, post-operative incision care and patient selection are complex variables that we were not able to evaluate fully.

In conclusion, this investigation and intervention were successful in decreasing rates of SSI after hysterectomy, although we were unable to identify a single intervention that was most responsible for the improvement. We encourage other facilities to provide feedback to their staff members about SSI, and to investigate and intervene in the setting of high SSI rates to improve patient care.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Edwards

, Peterson

, Mu

et al. National Healthcare Safety Network (NHSN) report: Data summary for 2006 through 2008, issued December 2009. Am J Infect Control, 2009; 37:783–805.

Medicare Program; Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and the Long-Term Care Hospital Prospective Payment System and FY 2012 Rates; Hospitals' FTE Resident Caps for Graduate Medical Education PaymentIn: CMS, ed. 42 CFR Parts 412, 413, and 476; CMS-1518-F; CMS-1430-FU.S. Department of Health and Human Services, 2011.

Young

, Bliss

, Carey

et al. Beyond core measures: identifying modifiable risk factors for prevention of surgical site infection after elective total abdominal hysterectomy. Surg Infect, 2011; 12:491–496.

Healthcare-Associated Infections (HAIs). 2010(2010 April 17 http://www.cdc.gov/ncidod/dhqp/healthDis.html.)

Specifications Manual for National Hospital Inpatient Quality Measures. Terrace, IL: Joint Commission National Quality Core Measures, 2010.

Stitely

, Sweet

, Slain

et al. Plasma and tissue cefazolin concentrations in obese patients undergoing cesarean delivery and receiving differing preoperative doses of drug. Surg Infect, 2013; 14:455–459.

Brandt

, Sohr

, Behnke

et al. Reduction of surgical site infection rates associated with active surveillance. Infect Control Hosp Epidemiol, 2006; 27:1347–1351.

Sykes

, Brodribb

, McLaws

et al. When continuous surgical site infection surveillance is interrupted: the Royal Hobart Hospital experience. Am J Infect Control, 2005; 33:422–427.

Kaya

, Yetim

, Dervisoglu

et al. Risk factors for and effect of a one-year surveillance program on surgical site infection at a university hospital in Turkey. Surg Infect, 2006; 7:519–526.

10.

Haley

, Culver

, White

et al. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol, 1985; 121:182–205.

11.

Kasatpibal

, Jamulitrat

, Chongsuvivatwong

et al. Impact of surgeon-specific feedback on surgical site infection rates in Thailand. J Hosp Infect, 2006; 63:148–155.

12.

Rodriguez

, Trobo

, Garcia

et al. The effect of performance feedback on wound infection rate in abdominal hysterectomy. Am J Infect Control, 2006; 34:182–187.

13.

Darouiche

, Wall

Jr. , Itani

et al. Chlorhexidine-alcohol versus povidone-i9odine for surgical-site antisepsis. N Engl J Med, 2010; 362:18–26.

14.

Lee

, Agarwal

, Lee

et al. Systematic review and cost analysis comparing use of chlorhexidine with use of iodine for preoperative skin antisepsis to prevent surgical site infection. Infect Control Hosp Epidemiol, 2010; 31:1219–1229.

15.

Manthous

, Ivy

. Why surgeons can say “No”: exploring “Unilateral Withholding.” J Hosp Med, 2012; 7:249–253.

16.

Young

, Berumen

, Knepper

et al. Statewide collaboration to evaluate the effects of blood loss and transfusion on surgical site infection after hysterectomy. Infect Control Hosp Epidemiol, 2012; 33:90–93.