New York State HOSPITAL-ACQUIRED INFECTION REPORTING SYSTEM Pilot Year - 2007 New York State Department of Health Report to Hospitals – June 30, 2008 HOSPITAL-ACQUIRED INFECTION REPORTING SYSTEM NEW YORK STATE - PILOT YEAR 2007 TABLE OF CONTENTS PAGE NUMBER Executive Summary 3 Background 12 HAI Data System Selection and Training 14 Monitoring Hospital Implementation 19 Data Submission and Analysis 21 Colon Surgical Site Infections 23 Coronary Artery Bypass Graft Surgical Site Infections 26 Post-Discharge Surveillance Survey 28 Adult/Pediatric Central Line Associated Bloodstream Infections 30 Neonatal ICU Central Line Associated Bloodstream Infections 31 Infection Control Resources in New York Hospitals 35 HAI Prevention Projects - FY 2007-2008 36 HAI Prevention Projects - FY 2008-2009 37 Mandatory Reporting of Infections in 2008 38 Conclusions 38 HAI Program Personnel 41 Abbreviations 42 References 44 Appendix A – Public Health Law 45 Appendix B – MOU between NYSDOH and CDC 48 Appendix C – Technical Advisory Workgroup Members 51 Appendix D – Commissioner’s Letter August 2006 52 Appendix E – HAI Letter for Audits – 2007 53 Colon SSI Tables 54 CABG SSI Tables 69 Adult and Pediatric ICU CLABSI Tables 82 Neonatal ICU CLABSI Tables 98 2 HOSPITAL-ACQUIRED INFECTION REPORTING SYSTEM NEW YORK STATE - PILOT YEAR 2007 EXECUTIVE SUMMARY According to the federal Centers for Disease Control and Prevention (CDC), there were an estimated 1.7 million healthcare-associated infections and 99,000 deaths from those infections in 2002. 1 Other investigators have estimated the annual costs associated with these infections to be $4.5 billion to $5.7 billion.2,3 None of these parameters measure the effect of these infections on the patients, their family members, friends and colleagues. Their emotional, physical and personal costs are not quantifiable. The Legislature passed and the Governor signed legislation in July 2005 requiring hospitals to report select hospital-acquired infections (HAIs) to the New York State Department of Health (NYSDOH, DOH or “the Department”). The legislation provided a “pilot phase” to develop the reporting system; train hospitals on its use; standardize definitions, methods of surveillance and reporting; audit and validate the hospitals’ infection data and modify the system to ensure that the hospital-specific infection rates, when released, would be fair, accurate and reliable. The legislation provided for an initial report to hospitals assessing the overall accuracy of the data submitted in the pilot phase and providing guidance for improving the accuracy of hospital-acquired infection reporting. New York’s reporting system utilizes the Centers for Disease Control and Prevention’s National Healthcare Safety Network (NHSN) for HAI reporting, and is the first state to do so. Now, 17 states are committed to using the NHSN and it has become the standard for state reporting. Public Health Law Section 2819 sets forth the responsibilities of the Department and New York State hospitals. The following report summarizes the development and implementation of the HAI reporting system, an assessment of the overall accuracy of the data submitted in the pilot phase, guidance for improving the accuracy of hospital acquired infection reporting, lessons learned and next steps. Italicized wording is the explicit language of the law. Technical Advisory Workgroup (TAW) The commissioner shall consult with technical advisors who have regionally or nationally acknowledged expertise in the prevention and control of hospital acquired infection and infectious disease in order to develop the adjustment for potential differences in risk factors to be used for public reporting.[PHL 2819 5.(b)] The TAW has met five times including the first meeting on May 5, 2006. The list of technical advisors is provided in Appendix C. At the first meeting, the TAW endorsed the following NYSDOH goals for the HAI reporting program which are consistent with the legislation: • Develop and implement a reliable, valid, useful HAI reporting system for the public, the hospitals, and the NYSDOH; • Prevent the selected HAIs; 3 • Use the HAI reporting system to evaluate risk factors and potential interventions; and • Use the data to evaluate the impact of initiatives to improve quality of care. The legislation called for the reporting of HAIs, with the initial starter set of central line- associated bloodstream infections (CLABSIs) and infections associated with surgical procedures in intensive care units (ICUs). The workgroup selected surgical site infections associated with coronary artery bypass procedures and colon surgical procedures due to the frequency of these infections, severity of infection-related complications, potential for risk adjustment and potential for quality improvement. The Department continues to meet with the TAW semi-annually. Their input has been invaluable. Establishment and Training Hospitals on Use of the National Health Care Safety Network (NHSN) The department shall establish guidelines, definitions, criteria, standards and coding for hospital identification, tracking and reporting of hospital acquired infections which shall be consistent with the recommendations of recognized centers of expertise in the identification and prevention of hospital acquired infections including, but not limited to the National Health Care Safety Network of the Centers for Disease Control. [PHL 2819 2.(c)] After selecting the CDC’s NHSN as the reporting mechanism, all hospital CEOs were informed by letter of the reporting requirements and training opportunities. Nine regional training sessions were held throughout the state in late 2006 on the NHSN enrollment procedures, guidelines for surveillance, standard definitions, use of the NHSN and reporting indicators. The Greater New York Hospital Association (GNYHA) videotaped the presentations and has made them accessible as training materials. A NYSDOH web site and an email distribution system of reporting hospitals was developed, a direct email link to the HAI program was established ([email protected]) and program staff have answered more than 2,000 inquiries. Ongoing education has been maintained via telephone, regional training sessions to discuss modifications to the reporting system and selection of indicators for 2008, onsite hospital visits, additions to the HAI web site and circulating an electronic newsletter. Timeliness and Completeness of Reporting The first year of data submission under this section shall be considered the "pilot phase" of the statewide hospital acquired infection reporting system. The purpose of the pilot phase is to ensure, by various means, including any audit process referred to in subdivision seven of this section, the completeness and accuracy of hospital acquired infection reporting by hospitals. [PHL 2819 5.(c)(ii) ] The pilot year for hospital reporting of HAIs was January 1-December 31, 2007. The initial legislation stated that the Department could not require reporting more often than every six months, 60 days after the end of the reporting period. The first six months of data were due by the end of August 2007. As of October 2007, 96 percent of facilities had 4 complied with the 2007 reporting requirements for the initial six-month reporting period. The eight facilities that did not comply were cited by DOH and provided a plan of correction. Although all eight facilities are now reporting, one of these eight facilities had not fulfilled the reporting requirements in time for this report. Timeliness and completeness of reporting surgical site infections (SSIs) were delayed due to the long incubation for some infections and the fact that SSIs were often detected after the initial hospitalization. As per the NHSN definition, SSIs were considered hospital- associated if they occur within 30 days or up to a year after the procedure if there was an implant (including sternal wires). Only 63 percent of colon SSIs were detected during the initial hospital admission, 24 percent were detected upon readmission and 13 percent were detected post-discharge. Only 32 percent of chest SSIs and 28 percent of donor vessel SSIs were detected during the initial hospitalization. 63 percent of chest SSIs and 66 percent of donor vessel SSIs were detected upon readmission to the same hospital. Nineteen (5 percent) chest site infections and nine (6 percent) donor site infections were detected post-discharge, including two organ space infections. The NHSN system did not distinguish between post- discharge infections involving readmission to another hospital and infections treated in the private practice or outpatient setting. Therefore, NYSDOH created a custom data entry field to capture this information. HAI reporting program personnel are continuing to conduct audits to determine the accuracy and completeness of reporting of CLABSIs and SSIs. For SSIs, the primary focus is on superficial SSIs occurring during the initial hospitalization and deep and organ space infections involving re-operation and/or readmission to a hospital (the initial hospital where the surgery is performed or another hospital). HAI Infection Rates 2007 For data reported during the pilot phase, hospital identifiers shall be encrypted by the department in any and all public databases and reports. The department shall provide each hospital with an encryption key for that hospital only to permit access to its own performance data for internal quality improvement purposes. [PHL 2819 5.(c)(ii)] By law, this pilot year report does not provide hospital identified infection rates. Future reports will include comparative hospital specific infection rates. Hospitals that perform very few procedures or have ICUs with very few patients with central lines will usually have infection rates that fluctuate greatly over time. The NHSN uses minimum thresholds to report infections and infection rates. The Department will use the same thresholds: • For surgical site infections, there must be a minimum of 20 patients undergoing a surgical procedure in the specific risk category before infection and rate data will be released. • For CLABSIs and rates in adult and pediatric ICUs, there must be a minimum of 50 central-line days. 5 • For CLABSIs and umbilical catheter-associated blood stream infections (BSIs) in neonatal ICUs, there must be a minimum of 50 central line or umbilical catheter days within a birth weight category. Colon Surgical Site Infection Rates The SSI rate for patients undergoing colon procedures in New York hospitals in 2007 ranged from 4.5 infections per 100 procedures in the lowest-risk patients to 9.4 per 100 procedures in the highest-risk group (Table 1). National colon SSI rates ranged from 4.0 to 11.3, for the lowest- and highest-risk patients, respectively (Table 2). De-identified hospital rates by risk category are provided in Table 3. HAI program staff members have been evaluating facilities with the highest and lowest infection rates, determining if there are surveillance and reporting differences, assessing trends, risk factors and interventions to reduce infections. Tables 4a, 4b, and 8 describe patient and procedure related risk factors associated with colon surgical site infections. Patient-related factors included obesity and male gender. Emergency procedures, especially those involving trauma patients, were the most likely to result in surgical site infections. The department will work with the surgical community to identify possible strategies to reduce infections in this extremely vulnerable group. In addition, the DOH will evaluate the need for further risk adjustment and incorporate these risk factors prior to releasing the report of 2008 infection rates. Coronary Artery Bypass Graft (CABG) Surgical Site Infection Rates CABG surgery most often involves two surgical sites: a chest incision and a separate site to harvest donor vessels. Because infections can occur at either incision, the infection rates are presented separately. New York State (NYS) donor vessel site infection rate was significantly lower than national rates across the majority of risk categories. Chest site infection rates were similar to national rates with the exception of a higher rate of deep incisional chest site infections (Tables 11 and 13). This difference may be due in part to difficulties in appropriately classifying the depth of chest SSIs. When the rates of deep and organ/space infections were combined, the difference was no longer statistically significant. The State is considering combining these rates when reporting 2008 hospital- specific rates due to the difficulty in accurately and reliably distinguishing between deep and organ/space infections. Individual hospital infection rates are provided in Tables 17 and 18 for donor vessel site infections and chest incision site infections, respectively. Individual hospitals reported performing as few as 65 CBGB procedures and as many as 1,065. Half the hospitals reported less than one chest or donor site infection per month. The donor vessel site infection rates ranged from zero to 4.0 percent, and from zero to 5.3 percent for chest incision sites. HAI program staff members have been evaluating facilities with the highest and lowest infection rates, determining if there are surveillance and reporting differences, and assessing trends, risk factors and interventions to reduce infections. Additional information will be collected during the 2008 audit process to systematically evaluate possible prevention practices. 6 Risk factors associated with chest SSIs included female gender, chronic lung disease, diabetes, immunodeficiency, obesity, post-operative renal failure, GI bleeding, bleeding re- operations and emergency procedures (Table 24). The department will work with infection preventionists, surgeons and the cardiac advisory committee to identify possible strategies to reduce infections in these patients. In addition, the DOH will evaluate the need for further risk adjustment due to these factors prior to releasing hospital-specific infection rates with identifiers. Central Line Associated Blood Stream Infections (CLABSIs) in Adult/Pediatric ICUs Table 28 provides the New York CLABSI rates by type of adult or pediatric ICU. The ICU-specific rates vary from a low of 2.0 infections per 1,000 central line (CL) days in cardiothoracic ICU patients to 4.0 infections per 1,000 CL days in pediatric ICU patients. NYS CLABSI rates in coronary and pediatric ICUs were significantly lower than national data but higher in surgical ICUs (Table 29). Within the State, New York City (NYC) facilities had lower CLABSI rates in medical and surgical intensive care units than the rest of the State (Tables 30 and 31). This difference may be attributable to a major regional collaborative to reduce CLABSI rates that began in 2006 in the NYC area, sponsored by GNYHA and United Hospital Fund. This possible explanation is currently being evaluated during 2008 audits. Tables 32-39 provide the individual hospital CLABSI rates by type of ICU. Hospitals with the highest CLABSI rates have been notified by the department, possible explanations are being evaluated and if the problem is continuing, recommendations have been made and enhanced monitoring is being conducted. Many of the hospitals with the highest rates had already recognized the higher rates, implemented interventions and reduced their rates. Central Line Associated Bloodstream Infections (CLABSIs) in Neonatal ICUs As reported in the literature, neonates in the lowest birth weight categories had the highest CLABSI rates. Neonates born under 750 grams had 7.5 infections per 1,000 CL days whereas neonates weighing more than 2,500 grams had 4.0 infections per 1,000 CL days. State rates are summarized in Table 46 were higher than the national rates (Table 47) but this difference was only statistically significant in one birth weight category (751-1000 grams). Similar trends were seen for neonates with umbilical catheters. Infants weighing less than 750 grams had the highest umbilical catheter-associated BSI rates (12.2 infections per 1,000 umbilical catheter days). The lowest rates were detected in infants born between 1501-2,500 grams (1.7) and more than 2,500 grams (2.2/1,000 umbilical catheter days). State rates are summarized in Table 50. Table 51 provides the most recent national comparison data from CDC. State rates were higher than national rates in the highest and lowest birth weight categories. Hospitals with the highest CLABSI rates have been notified, possible explanations are being evaluated and if the problem is continuing, recommendations have been made and enhanced monitoring is being conducted. In addition, the department is working with neonatologists across the State on a collaborative to reduce CLABSI rates in neonatal intensive care units. 7 Accuracy of Reporting To assure the accuracy of the self-reported hospital acquired infection data and to assure that public reporting fairly reflects what actually is occurring in each hospital, the department shall develop and implement an audit process. [PHL 2819 7.] The NYSDOH HAI reporting program generates bi-weekly reports by region and by hospital to detect data entry errors. These reports are reviewed by the regional HAI program staff members, and hospitals are given the opportunity to verify and/or correct the data. Audits of a sample of medical records were conducted by the department to assess compliance with reporting requirements. Onsite visits were conducted by HAI program staff in 95 percent (183) of the hospitals between July 2007 and January 2008. Data submitted to NHSN for the first quarter of 2007 were used to select medical records for review. The purposes of the audit were to: • Determine the reliability and consistency in applying the surveillance definitions; • Evaluate the adequacy of surveillance methods to detect infections; • Evaluate current risk adjustment methods and determine whether additional factors need to be considered for public reporting purposes; and • Evaluate intervention strategies designed to reduce or eliminate specific infections. If data inconsistencies were identified, hospitals were contacted, the discrepancies were discussed and if the records needed to be modified, the hospitals edited the data. Ongoing monitoring, education and training have been and continue to be provided to ensure the integrity of the data. Strengths and Weaknesses – Use of NHSN for Mandatory Reporting A major objective of the pilot phase was to evaluate the strengths and weaknesses of using the NHSN for mandatory reporting purposes, determining whether the State should continue to use the NHSN reporting system and recommend changes or modifications for 2008. The major strengths of using the NHSN were: • Standard definitions had been developed and could be applied consistently; • These definitions are used throughout the United States and in other countries; • CDC served as a valued partner, was available to assist and support the Department, clarified the interpretation of data elements and definitions, and provided information technology support; • Hospitals could immediately use the information they reported, calculate trends over time and compare their infection rates with national rates; and • Hospitals began to use the system for collaborative intervention initiatives to reduce HAIs. 8 The major weaknesses of using the NHSN were: • Due to confidentiality agreements, hospitals had to take additional steps to confer rights to grant the State permission to view and analyze their data. These steps could have been averted or minimized if the department had been able to make this modification internally. • To make system changes or collect additional information, the department had to request that all hospitals create the same customized data entry fields in the same way. • NYSDOH could not modify definitions unilaterally; CDC had to make these changes. This may not necessarily be a weakness, because any New York-specific modification or change affects the ability of hospitals to compare themselves with other hospitals across the nation. To deal with these weaknesses, CDC and NYSDOH worked together to make changes to the NHSN or the department developed custom data entry fields to collect additional information. Legislative Changes (Chapter Amendments) In July 2007, the Governor signed chapter amendments to Public Health Law Section 2819 to ensure appropriate (1.), complete (2.) and timely (3.) reporting of HAIs by hospitals in New York. These revisions include: 1. An HAI was redefined as an infection that was: “not found to be present or incubating at the time of admission unless the infection was related to a previous admission to the same setting.” [PHL 2819 1(b)] 2. “For hospital acquired infections for which the department requires tracking and reporting as permitted in this section, hospitals shall be required to report a suspected or confirmed hospital-acquired infection associated with another hospital to the originating hospital. Documentation of reporting should be maintained for a minimum of six years.” [PHL 2819 2( e)] 3. Each hospital shall regularly report to the department the hospital infection data it has collected. The department shall establish data collection and analytical methodologies that meet accepted standards for validity and reliability. The frequency of reporting shall be monthly, and reports shall be submitted not more than sixty days after the close of the reporting period. [PHL 2819 3] Lessons Learned During the pilot year, the Department and hospitals learned the following important lessons regarding HAI reporting: 1. Strict adherence to the surveillance definitions is critical to provide consistency and comparability of data across hospitals. Clinical findings are appropriate for treatment decisions but are not appropriate for mandatory reporting purposes since there is significant variability between providers and different institutions. 2. Additional risk factors were identified and need to be further assessed to determine if they affect the hospital-specific infection rates. 9 3. Post-discharge surveillance methods are highly variable, dependent upon allocated resources and integration of information systems. In addition, the majority of severe infections were detected during the initial hospitalization or upon readmission. Therefore, NYSDOH is not mandating a uniform post-discharge methodology but will continue to monitor the impact of these efforts. 4. The original legislative language prohibited the department from receiving timely, actionable data from the hospitals. The laws were amended in 2007 to require HAI reporting within 60 days of the end of the surveillance month. 5. Timely and complete data submission was often affected by infection control staffing turnover, prolonged vacancies and the need for education and training to comply with the legislative mandate. Hospitals need to provide back-up personnel to ensure compliance with reporting requirements and patient safety. 6. Very few facilities made use of electronic data transfer and therefore relied on cumbersome manual data collection and entry. Hospitals need to integrate information systems to support infection prevention and reporting efforts. Next Steps The Department will work to improve HAI reporting and infection prevention efforts including taking the following actions: 1. Continue to monitor the accuracy and timeliness of data being submitted, discuss findings and ensure corrective action is taken. 2. Conduct onsite audits to evaluate surveillance methods, interpretation of surveillance definitions, and completeness of reporting. 3. Continue to evaluate the effectiveness of various post-discharge methods. 4. In conjunction with the TAW, evaluate the need for further risk adjustment and if deemed necessary, integrate into the public reports. 5. Develop methods and format for public reporting of identified hospital infection rates in collaboration with the TAW. 6. Conduct surveys or additional audits to evaluate the effectiveness of prevention strategies to reduce HAIs. 7. Identify and evaluate hospitals with the lowest and highest infection rates to determine if reported data are reliable and if the data are reliable, attempt to identify reasons for the differences. 8. Monitor infection control resources to evaluate the impact of public reporting on other infection prevention and control responsibilities. 9. Collaborate with other department staff to investigate outbreaks, evaluate emerging trends and/or provide regulatory action for non-compliance with the legislative mandates. 10. Consult with infection preventionists, hospital epidemiologists, surgeons and the Cardiac Advisory Committee to identify possible strategies to reduce HAIs. 11. Monitor HAI prevention projects for compliance with program objectives, fiscal responsibility and potential applicability to other hospitals or healthcare settings. 12. Continue to provide education, training and ongoing support to hospital infection reporting staff. 10