The Pennsylvania Patient Safety Authority and the Pennsylvania National Surgical Quality Improvement Program (PA- NSQIP) initiated a collaborative project in December of 2010 to reduce surgical site infections (SSIs) among the PA-NSQIP member hospitals and to transfer successful strategies and lessons learned to other Pennsylvania hospitals. Participating hospitals’ SSI data as reported to the NSQIP was used to select colorectal and bariatric surgeries as the collaborative focus and to identify high-performing and outlier hospitals’ SSI rates in these areas. An SSI prevention assessment tool was developed to conduct on-site hospital interviews to assess the level of implementation of specific preventive practices at the pre-, intra-, and post-operative levels and to determine if the high-performing hospitals' lower SSI rate could be attributed to variation in implementation of practices. The on-site interviews conducted in December of 2011 revealed that the hospitals that were more diligent with a standardized approach to implementation of practices were also the high-performing hospitals with the lowest SSI rates. This article reveals multifaceted differences in the implementation of practices and identifies specific interventions for facilities needing improvement to reduce SSIs related to bariatric and colorectal surgery.
According to the Centers for Disease Control and Prevention (CDC), approximately 1 in 20 hospitalized patients acquires a healthcare-associated infection (HAI).1 Surgical site infections (SSIs) are cited as the second most common HAI, accounting for 17% to 22% of all HAIs among hospitalized patients.2,3 It has been estimated that SSIs may result in as many as 3.7 million additional hospitalization days and an annual overall cost of $1.6 billion in the United States.4 The 2007 average attributable per-patient treatment costs are estimated at over $34,000.5 SSIs are serious, life-threatening infections. The Klevens et al. report of HAIs in US hospitals in 2002 estimated the number of SSIs at 290,485, with 8,205 associated deaths, a 2.8% case fatality rate.2 Using estimates from national reports and published studies related to HAIs, the study suggests that implementation of infection control practices in all US hospitals could reduce the number of SSIs and save 2,133 to 4,431 lives annually.
Reducing HAIs also saves money. The annual cost of preventable cases of SSI in the United States is estimated to be $166 million to $345 million.6 The 2008 Society for Healthcare Epidemiology of America (SHEA) and Infectious Diseases Society of America (IDSA) practice recommendations to prevent SSIs in acute care hospitals recommends additional practices for modifiable risk factors in the areas of patient preparation, operative characteristics, and surgical suite characteristics.7 The Joint Commission 2010 National Patient Safety Goal 07.05.01 for preventing SSIs includes elements of performance for the prevention of SSIs in the areas of education, implementation of evidence-based guidelines, risk assessment, and process and outcome measures.8
Effective July 1, 2011, the federal Patient Protection and Affordable Care Act prohibits Medicare and Medicaid payments for hospital-acquired conditions, including bariatric SSIs associated with laparoscopic gastric bypass, gastroenterostomy, and laparoscopic gastric restrictive surgery.9,10 Effective January 1, 2012, hospitals that are paid under the inpatient prospective payment system are required to submit SSI data for colon procedures as part of the Hospital Inpatient Quality Reporting program in order to receive full reimbursement in Medicare’s annual payment update for fiscal year 2014.11 Hospitals not submitting the required measures will receive a 2% reduction in their annual market basket update. Beginning October 2012, Hospital Compare will include new surgical outcomes measures submitted on a voluntary basis by hospitals participating in the American College of Surgeons’ (ACS) National Surgical Quality Improvement Program (NSQIP).12,13
As a result of the magnitude of these types of infections and the largely unchanged scope of this problem, the Pennsylvania Patient Safety Authority and the Pennsylvania NSQIP (PA-NSQIP) hospitals initiated a collaborative improvement project to reduce SSIs in bariatric and colorectal surgical cases. Participating facilities looked to successful hospitals that have complied with and extended the SSI prevention “bundle” practices for lessons about implementation and overcoming barriers to SSI prevention practices.
The proposed benefits of this collaboration for the participating PA-NSQIP hospitals include the following:
- The reduction of SSIs for a sustained period through the dedication and commitment of PA-NSQIP leadership and teams from each participating facility
- The creation of a collaborative learning network for the prevention of SSIs
- Leveraging existing data already being collected at all hospitals through the national NSQIP
- The establishment of targeted education for facility staff on SSI prevention according to the medical literature, the Authority’s Pennsylvania Patient Safety Advisory article and toolkit database information, and other pertinent educational resources.
- The creation of a shared, password-protected workspace on the Authority’s web-based collaboration site to document all activities in this endeavor
- Guidance, assistance, and technical support provided to the participating facilities
- The collection of follow-up data and the creation of comparison reports to measure progress at the end of the program
Participants and Data Sources
The project design used NSQIP retrospective data to determine baselines and identify high performers and outliers (hospitals with high SSI rates) within the participating PA-NSQIP consortium member hospitals. Each hospital was asked to provide their SSI data from the most recent annual report generated by ACS NSQIP (cases for calendar year 2009) and to complete a brief survey on the observed-versus-expected ratio (O/E ratio) of SSIs from the surgical categories listed in the ACS NSQIP semiannual report: (1) overall general plus vascular surgery, (2) general surgery, (3) colorectal surgery, and (4) vascular surgery. The Authority staff used this information to identify high performers and outliers within the group by analyzing aggregate and facility-level data using meta-analysis (see Tables 1 and 2). General surgery was selected as a focus area due to the high O/E ratio as compared with the other surgical categories.
|Total No. Of Cases||No. Of Cases |
|Observed Rate |
|Expected No. Of Cases||Expected Rate||Observed-Versus-Expected Ratio (95% Ci)||Comment |
0.050 (0.046 to 0.054)
1.010 (0.901 to 1.133)
0.053 (0.049 to 0.058)
1.019 (0.904 to 1.149)
0.133 (0.114 to 0.154)
0.990 (0.790 to 1.239)
0.029 (0.022 to 0.039)
0.904 (0.596 to 1.360)
|Surgery Type||Observed Rate (95% Ci)||I2*|
|Overall (general plus vascular)||0.047 (0.037 to 0.061)||90%|
|General surgery||0.051 (0.040 to 0.065)||88%|
|Colorectal surgery||0.127 (0.093 to 0.171)||74%|
|Vascular surgery||0.029 (0.018 to 0.048)||57%|
Note: These summaries were calculated using meta-analysis, which is a statistical technique that combines data from multiple studies. For this study, data was combined from eight of the nine hospitals that are participants in the Pennsylvania National Surgical Quality Improvement Program, with weighting via inverse standard errors. The specific meta-analytic technique used for these analyses involved logit event rates and used the random-effects approach described by DerSimonian and Laird.† Comprehensive Meta-Analysis Version 2‡ software was used to perform these computations.
* I² "describes the percentage of total variation across studies that is due to heterogeneity rather than chance. . . . A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity." (Source: Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003 Sep 6;327;557-60.)
† Dersimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986 Sep;7(3);177-88.
‡ Developed and available from Biostat Inc., Englewood, New Jersey.
Authority staff conducted on-site visits to the two outlier hospitals to analyze NSQIP monthly data and to identify which surgical procedures in the general surgery category presented opportunities for improvement. Participating hospitals’ 2010 SSI data as reported to the NSQIP was used to select colorectal and bariatric surgeries as the collaborative focus. Hospitals were selected for the on-site practice interview primarily based on their overall NSQIP O/E ratio and secondarily on individual performance in the specific surgical procedure. The specific SSI rates of interest and any risk adjustment that was applied was determined based on the subspecialties in which the most infections occurred at the intervention sites. This data helped identify the general areas with the greatest potential for improvement. Once identified, the two high performers and the two outlier hospitals for colorectal and bariatric procedures were asked to do the following:
- Designate a team leader from each facility as the primary liaison for the collaborative and an interdisciplinary clinical team
- Work with Authority staff to examine the institution’s NSQIP data to identify more specific targets for improvement, based on the hospital’s procedure volume and the clinical services with the highest infection rates
- Convene relevant physicians and other clinical staff in the identified subspecialties for site visits from institutions doing well in those areas
- Provide adequate time for staff to complete surveys, review documents, and attend periodic conference calls and meetings
The Authority staff provided overall coordination, project management, and logistics support for the collaborative and served as independent facilitators to collect, analyze, and report on relevant data to participating hospitals. Of the nine PA-NSQIP hospitals, eight participated in the study. The collaborative formed a steering committee consisting of NSQIP hospital surgeons, NSQIP nurse reviewers, and key Authority staff, including a patient safety liaison and an infection preventionist.
The steering committee developed a unique SSI prevention assessment tool for use during on-site hospital interviews. The tool assessed the level of implementation of specific preventive practices at the pre-, intra-, and postoperative level, and it helped determine if the high-performing hospitals’ lower SSI rate could be attributed to variation in implementation of practices. The tool was based on the ACS comprehensive list of practices consolidated from existing guidelines and those processes thought to be most relevant for the focus areas of bariatric and colorectal surgery. Policies and procedures from the high-performing hospitals for these focus areas were reviewed for evidence of processes and protocols that could be shared with the outlier hospitals. The steering committee developed specific interview questions to elicit detailed information on the levels of implementation of SSI prevention strategies and sorted questions into categories of whether they were related to factors specific to the patient, controlled by the systems, or controlled by the provider. Each category was grouped by provider type in order to help with the flow of the on-site interviews.
Cross-institutional learning was facilitated by arranging site visits and interactions between clinical teams from outlier hospitals and their colleagues from institutions that have achieved and sustained low SSI rates. The collaborative committee elected to visit two high performers and two outlier hospitals in colorectal and bariatric procedures. All four hospital teams were provided with a copy of the SSI prevention assessment tool to perform a self-assessment prior to the visit and to collect supporting documentation such as polices and protocols. Separate survey teams were formed to perform on-site visits at each of the two bariatric-procedure and the two colorectal-procedure high-performing and outlier hospitals.
The collaborative steering committee NSQIP surgeons interviewed anesthesia providers and the hospitals’ surgeons from each surgical specialty. The collaborative steering committee NSQIP nurse reviewers interviewed nursing staff at the pre-, intra-, and postoperative levels. The Authority staff (a patient safety liaison and an infection preventionist) interviewed the directors of the surgical services and quality and the medical-surgical nursing staff. Interviews were conducted with each of these groups in each of the two high-performing and the two outlier hospitals using the SSI prevention assessment tool.
The assessment results were analyzed for a comparison of practice implementation. Analysis of the hospital assessment results was supported by the narrative summary completed for each facility outlining the strategies used to implement each practice. It was evident that the hospitals that were more diligent with a standardized approach to implementation of the items on the assessment tool were also the high-performing hospitals with the lowest rates of SSIs. The comparison found 8 major differences in performance between the high performers and the outliers for bariatric procedures and 16 major differences for colorectal procedures (see Tables 3 and 4).
High-Performer Current Process
Pre-op arterial blood gases
Done on the day of surgery
Hemoglobin A1C and fasting blood sugar
Do not perform surgery
Are involved with surgery
Post-op upper gastrointestinal series
Done on postoperative day 1
Liquid diet for five days prior
Done daily or twice a day in all areas
By registered nurse; bedside report
By technician; report in computer
Clear team roles; everyone aware of others' responsibilities (e.g., "Anesthesia does that, and here is their process.")
Team roles defined; staff not aware of others' responsibilities (e.g., "Anesthesia does that; you will have to ask them.")
|Practices||High-Performer Current Process||Outlier Difference|
Standard is to assess every patient; dietary supplements are provided for low albumin; diversion performed if albumin <3.0 g/dL; Child-Pugh score calculated if alcoholic
No standard; no albumin check; no Child-Pugh score calculation
No standard; staff request better communication from anesthesia for timing of next antibiotic dose
At time of data analysis, used antimicrobials and citrate 1 day pre-operatively; Betadine wash used for rectal anastomosis—stopped, and trend now is back to using bowel prep
No bowel prep
Checklists and safety briefings
Policy and/or manual are on chart; attending surgeon must be present (structured process); anesthesia provides report and time for questions
No policy; pathway not on chart; any surgeon may perform safety briefings; pre-, intra-, and post-op checklists are used
Secondary closure for gross-contaminated skin opening
Secondary closure only for emergent situations
Communication and structure
Team roles are clear; everyone is aware of others' responsibilities (e.g., "Anesthesia does that, and here is their process."); communication processes are standardized and structured, with clear accountability
Team roles are defined; not everyone is aware of others' responsibilities (e.g., "Anesthesia does that; you will have to ask them."); processes are not standardized; accountability is unclear
Diversion or colostomy
Policies and standards call for diversion if post-chemo or post-radiation or if the patient is malnourished, has <3.0 g/dL albumin, or has blood loss of more than 400 cc
No policy or standard; diversion is performed in cases of radiation or sepsis
Standardized handoff tool, protocol, policy, and procedure; postanesthesia care unit registered nurse brings patient to unit; in-person handoff performed at bedside
No protocol or policy; check-off process is verbal and not at bedside
History of steroid use
If steroid use within the last six months, then at time of operation, give stress dose steroid (regimen hydrocortisone 50 mg IV Q8 hours for 48 hours; BID for 48 hours; then once a day). If patient has active ulcerative colitis, then the steroid taper will be slowed to a longer time period. Also, any patient on steroid recently will get two-stage operation. If on steroids and infliximab, then three-stage operation.
Operating room (OR) cleaning
Process is more structured, with accountability; training, competency evaluation, and compliance monitoring performed by staff from the OR management team and by the lead anesthesia technician
Training and competency evaluation performed by staff educator
Surgeon does prep; standardized training is provided on patient prep
Nurse or resident does prep; process is not structured
Standardized policy is to redose with Levaquin® or Flagyl®
Redose by anesthesia if remembered
Written policy requires surgeon to perform site prep
Site prep by any resident or fellow; no education
Skin edge/ wound protection
Use for both open and laparotomy cases
Use only for laparotomy cases
Structured accountability, policies, and procedures; assistant OR director or circulating nurse manages observations; observers allowed by appointment
No policy or procedure; entire team is charged with being the traffic police; staff report that this is not well managed
Policy or procedure is that OR has its own beds, which are returned to the OR after transport
No policy or procedure; beds from unit are brought to the OR
Outliers in both the bariatric and the colorectal groups reported variation in implementing methods for bowel preparation, communication, safety briefings, and transport. While bowel preparation is a provider-specific item, communication, safety briefings, and transport are system- or hospital-controlled factors. At both high-performing hospitals, the transport was performed by registered nurses rather than by patient care technicians; safety briefings were conducted in a structured and specific manner; and communication channels were open and encouraged.
The bariatric outlier interviews identified differences with preoperative measurement of arterial blood gas and of hemoglobin A1c (HbA1c) indicating glycemic control, postoperative upper gastrointestinal studies, and the involvement of resident-level physicians. The colorectal outlier interviews revealed differences in a history of steroid use, albumin checks, patient and site preps, decisions of diversion versus colostomy, wound protection and closure methods, and antibiotic timing associated with prolonged procedures, as well as OR cleaning, traffic control, and handoffs. All levels of staff interviewed at the high-performer level for both bariatric and colorectal procedures were able to identify responsibilities of all team members and policies and procedures related to OR practices, as well as protocols for the focus-area procedure. Outlier hospitals reported fewer standardized protocols, and staff were often uninformed as to the workflow of other members of the team. For example, they understood who was responsible for a task but not the specifics of how the task was done.
Each outlier facility participant took the comparison back to his or her hospital, and the hospital teams selected the following process measures for implementation.
Selected bariatric process measures included the following based on the number of bariatric procedures per month:
- Number of patients who have glycosylated HbA1c drawn prior to surgery
- Number of patients with an HbA1c level over 8% who had surgery
- Number of patients who received chlorhexidine gluconate wipes on the morning of the procedure
- Number of patients who received a Peridex swish on the morning of the procedure
Selected colorectal process measures included the following based on the number of colectomy procedures per month:
- Number of patients who have documentation that the surgical bundle was fully implemented
- Number of patients who have skin edge protection used during surgery
Selected colorectal process measures included the following based on those with a procedure time greater than four hours:
- Number of patients who have an antibiotic redosed
Potential Barriers to Implementation
Bariatric and colorectal outlier hospitals in the collaborative are currently developing methods to implement the processes that will be measured. Standardization of protocols, team member roles, and communication tools were much more prevalent in the high-performing hospitals. These reflect assessment tool questions regarding factors that were identified as controlled by the systems or hospitals rather than as provider- or patient-specific. Implementation of these items will require additional resources when compared with a provider-controlled element due to the complexity of the hospitals themselves. Often, changes to protocols require buy-in from several groups and must be accepted and approved by multiple committees. It takes time, resources, and leadership support to see these types of changes through to adoption. Culture within a facility, as evidenced by communication structures and teamwork within and across units, will also play a role in the length of time and success of implementation.
There are limitations in the study design. The PA-NSQIP consortium consists of only eight hospitals. Of those, two were identified as outliers. This is a small sample size, reflective of only the hospitals that have committed to investing resources to the NSQIP and that may already have bias toward reducing SSIs. The data is limited to NSQIP data. These are not reflective of all procedures performed within a facility. Due to the nature of SSI data collection, real-time data is nearly impossible to collect, causing several-month lag times between implementation of process changes and reflection of the changes in the data. Maintaining momentum and associating changes in process to changes in data is difficult.
Several limitations exist with the tool as well. While literature- and evidence-based items were included on the assessment tool, many of the questions were identified by reviewing the high-performer hospitals’ documents. These items are not necessarily better processes; they are simply identified processes that would need to be explored during the site visits. The tool also contains bias in that the more standardized a process, the more likely it was to correlate with positive outcomes. There was also no weighting of the answers, so any “yes” response was considered equal to any other “yes” response, even though not all risk factors and interventions correlate equally to SSI prevention.
The goal of the collaboration is to strengthen and improve patient safety by reducing the incidence of SSI among surgical patients in participating hospitals and demonstrating a collaborative learning approach to SSI reduction. The assessment highlights the results of on-site visits with hospitals identified as having high or low infection rates in bariatric and colorectal surgery as compared with the expected infection rates in those categories and reveals multifaceted differences in implementation of practices for the purpose of identifying specific interventions for facilities needing improvement to reduce SSIs in relevant areas. The assessment tool provided a detailed comparison of the levels and methods of implementation of SSI prevention practices of high-performing and outlier hospitals with respect to the prevention of postoperative bariatric and colorectal SSI. It also highlighted the significant role of culture and standardization in the prevention of SSIs.
In the short term, the principal outcome measures that will indicate the success of this first project of the PA-NSQIP consortium is a reduction in the SSI rates at the institutions selected for the initial intervention. Secondary measures include process metrics identified through the project and thought to have an impact on SSI reduction in bariatric and colorectal surgery (see Tables 2 and 3). Over the longer term, the consortium will demonstrate improvement by reducing the SSI O/E ratio based on risk-adjusted data published by ACS NSQIP. The consortium will track these outcomes prospectively for all participating facilities. The success of this collaboration requires a high level of commitment from PA-NSQIP leaders at each facility and from interdisciplinary clinical teams at each hospital. PA-NSQIP teams and the Authority will work to transfer knowledge gained to other hospitals throughout Pennsylvania.
PA-NSQIP consortium committee members Pat Toselli, DO, MMM, FACOS; Herbert E. Cohn, MD; James Reilly, MD; Virginia Wesner, MPA, CIP; Randi E Altmark, RN, BSN, CNOR; Kelly Gemmill, BA, RN; Lisa Baro, RN, CPHQ; Susan Diehl, RN, BSN; Patricia Fisher, BS, RN, CNOR; Cynthia L. Brophy, RN, BSN, CPAN; and Cathy Webber, RN, MSHA, contributed to the development of this project and to on-site hospital assessments. William M. Marella, MBA, Pennsylvania Patient Safety Authority, contributed to project development and statistical testing for this article.
- Centers for Disease Control and Prevention. Healthcare-associated infections (HAIs): the burden [online]. [cited 2012 Jul 9]. http://www.cdc.gov/HAI/burden.html.
- Klevens RM, Edwards JR, et al. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep 2007 Mar-Apr;122(2):160-6.
- Stone PW, Braccia D, Larson E. Systematic review of economic analyses of health care-associated infections. Am J Infect Control 2005 Nov;33(9):501-9.
- Martone WJ, Nichols RL. Recognition, prevention, surveillance, and management of surgical site infections: introduction to the problem and symposium overview. Clin Infect Dis 2001 Sep 1;33(Suppl 2):S67-8.
- Scott RD. The direct medical costs of healthcare-associated infections in U.S. hospitals and the benefits of prevention [report online]. 2009 Mar [cited 2012 Jul 9]. http://www.cdc.gov/hai/pdfs/hai/scott_costpaper.pdf.
- Umscheid CA, Mitchell MD, Doshi JA, et al. Estimating the proportion of healthcare-associated infections that are reasonably preventable and the related mortality and costs. Infect Control Hosp Epidemiol 2011 Feb;32(2):101-14. Also available at http://www.jstor.org/stable/10.1086/657912.
- Anderson DJ, Kaye KS, Classen D, et al. Strategies to prevent surgical site infections in acute care hospitals [2008 SHEA/IDSA practice recommendation]. Infect Control Hosp Epidemiol 2008 Oct;29(Suppl 1):S51-61.
- Joint Commission. National Patient Safety Goals effective January 1, 2012 [online]. 2012 [cited 2012 Jul 9]. http://www.jointcommission.org/assets/1/6/NPSG_Chapter_Jan2012_HAP.pdf
- US Department of Health and Human Services. Medicaid program; payment adjustment for provider-preventable conditions including health care-acquired conditions; final rule. Fed Regist 2011 Jun 6;76(108):32816-38. Also available at http://www.gpo.gov/fdsys/pkg/FR-2011-06-06/pdf/2011-13819.pdf.
- Centers for Medicare and Medicaid Services. Hospital-acquired conditions [online]. [cited 2012 Jul 26]. http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HospitalAcqCond/Hospital-Acquired_Conditions.html.
- Centers for Medicare and Medicaid Services. Hospital inpatient quality reporting program [online]. [cited 2012 Jul 26]. http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/HospitalRHQDAPU.html.
- US Department of Health and Human Services (HHS). Hospital Compare [website]. [cited 2012 Jul 26]. Washington (DC): HHS. http://www.hospitalcompare.hhs.gov.
- American College of Surgeons. Be a quality leader with ACS NSQIP [online]. [cited 2012 Jul 26]. http://www.facs.org/hospitalcompare/index.html.