- Email: [email protected]
Validation of National Healthcare Safety Network Laboratory-Identified Event Data
Poster Abstracts / American Journal of Infection Control 45 (2017) S16-S93
Maureen Bolon, MD, MS, Medical Director of Healthcare Epidemiology and Infection Prevention, Northwestern Memorial Hospital; Cynthia Barnard, PhD, MBA, MSJS, VP, Quality, Northwestern Medicine; Teresa Zembower, MD, MPH, Medical Director, Healthcare Epidemiology and Infection Prevention, Northwestern Memorial Hospital BACKGROUND: As healthcare-associated infections (HAIs) drive financial incentives and penalties, reputation and quality improvement, accurate identification is increasingly vital. There is literature suggesting significant inconsistency and potential gaming in identification and reporting of HAIs. Infection Preventionists (IP) are usually responsible for performing infection surveillance at healthcare facilities but no standard for ensuring valid identification and reporting exists. This study evaluates the accuracy of HAI detection, focusing on central line associated bloodstream infections (CLABSIs) across a newly formed healthcare system. METHODS: We conducted an audit of all positive blood cultures not deemed to be CLABSIs across our healthcare system in patients with known central lines. Data were sampled based on test volume from four hospitals encompassing over 1,600 inpatient beds. Positive results were obtained directly from the respective facilities’ electronic surveillance systems. Data were collected from JanuaryFebruary 2015 and March-July 2015. Duplicate cultures were excluded as were cultures from patients without central lines. Positive cultures were independently reviewed according to the 2015 National Healthcare Safety Network (NHSN) definitions to validate the negative predictive value of IPs’ determinations. RESULTS: During this period, 452 cultures met criteria across the system and a random sample of 60 were reviewed. In two of the four hospitals, 30 cultures passed validation and no additional CLABSIs were identified. One facility had one additional CLABSI identified and one had two. CONCLUSIONS: These 3 true CLABSIs had been misidentified as secondary infections based on clinical judgement in conflict with NHSN definitions. The two hospitals without discrepancies have robust orientation, ongoing training and oversight of HAI surveillance and classification. Subsequently, this was implemented at all hospitals. This methodology outlines a relatively simple audit tool to monitor CLABSI surveillance across a system. Ongoing education and oversight of IP surveillance is important to ensure appropriate reporting to NHSN.
Session DSV-054 12:30-1:30 p.m. Validating Nurse Sick Call Data as a Method of Early Warning Outbreak Response in an Academic Teaching Hospital Keenan Williamson, MPH, Infection Preventionist, Oregon Health & Sciences University; Molly Hale, MPH, CIC, FAPIC, Manager, Infection Prevention & Control; Dawn Nolt, MD, MPH, Pediatric Medical Director, Oregon Health & Sciences University BACKGROUND: Early outbreak response is essential in mitigating negative health outcomes from infectious diseases in patients and healthcare workers in a hospital setting. Early outbreak detection research has not been conducted extensively in academic teaching hospitals. In June 2015, a university hospital experienced a norovirus outbreak among the Pediatric ICU nursing staff due to ongoing environmental norovirus transmission. This prompted
S35
establishment of an early outbreak detection and monitoring system for inpatient nurses. METHODS: Mean daily Pediatric ICU nurse sick call data were retrospectively analyzed from 2005-December 2014 and compared to the known PICU norovirus outbreak. The outbreak recorded 32 suspect or lab-confirmed norovirus cases from June 6th-June 24th, 2015 with a total of 19 nursing staff out sick. Comparison used the Exponential Weighted Moving Average (EWMA) with an alpha of 0.03. The EWMA forecast was elaborated by three standard deviations to serve as an Upper Control Limit (UCL). RESULTS: In 2015, the PICU reported 548 nurse sick calls. Moderate increases in nurse sick calls were observed during influenza season and holidays, Jan-Feb and Nov-Dec. Analyzing EWMA control charts revealed 12 days in 2015 above the UCL; including 3 days during the known PICU norovirus outbreak in June. Of the remaining 9 days above the UCL, only 2 were clustered in time-suggesting a missed investigation opportunity. CONCLUSIONS: The establishment of real-time analysis of nurse sick call data can serve as a proxy for early outbreak detection by identifying aberrant increases in nurse sick calls. Infection Prevention & Control staff developed a house-wide nurse sick call monitoring system using the EWMA method. Further analysis will be done to make the process sensitive to clusters in time and include an array of healthcare professionals.
Session DSV-055 12:30-1:30 p.m. Validation of National Healthcare Safety Network Laboratory-Identified Event Data Jarred Gray, MPH, Epidemiologist, Tennessee Department of Health; Ashley Fell, MPH, Epidemiologist, Tennessee Department of Health; Vicky Reed, RN, CRRN, Public Health Nurse Consultant, TN Department of Health; Rebecca Meyer, MPH, Epidemiologist, TN Department of Health; Marion Kainer, MD, MPH, FRACP, FSHEA, Director Healthcare Associated Infection and Antimicrobial Resistance Program, TN Department of Health BACKGROUND: The state health department (SHD) has required acute care hospitals to report Clostridium difficile (CDI) and Methicillin-Resistant Staphylococcus aureus (MRSA) Laboratoryidentified (LabID) event data through the National Healthcare Safety Network (NHSN) since July 2010. SHD conducted a validation of NHSN LabID event data to assess data quality and to identify common reporting errors. METHODS: SHD used the NHSN external validation toolkit and selected 31 facilities for validation (01/01/2015-06/30/2015); 23 facilities were selected for MRSA and 23 for CDI. Up to 60 positive laboratory specimens were reviewed for each selected event type (CDI and/or MRSA). Patient medical records, including location, and laboratory records were reviewed to determine if the selected specimen met LabID event criteria. RESULTS: SHD identified 947 reportable CDI/MRSA events. Of the reportable CDI/MRSA events, 138 (14.6%) were unreported and 121 (12.8%) were reported with errors. Eighty-one percent (767/947) of the events validated were identified at facilities utilizing clinical decision support software (CDSS) to aid in event reporting. In these facilities, SHD found that 27% of reportable CDI/MRSA events were either unreported or reported incorrectly. Facilities that verified their CDSS identified events, at least monthly, with a laboratory line list had significantly higher percentage of correctly reported events than facilities that validated less frequently (77.75% vs 56.25%, P = .0318).
APIC 44th Annual Educational Conference & International Meeting | Portland, OR | June 14-16, 2017
S36
Poster Abstracts / American Journal of Infection Control 45 (2017) S16-S93
CONCLUSIONS: Validation of NHSN LabID events has informed SHD of significant errors in reporting arising from the use of CDSS. The use of CDSS is best supplemented with monthly verification of results using a laboratory line list until the accuracy of CDSS can be improved. In addition, validation has provided SHD with valuable information about data quality as well as common misconceptions that can be addressed through education and training. Data validation also strengthened SHD’s relationships with infection preventionists, improving communication between SHD and the acute care hospital community.
Session DSV-056 12:30-1:30 p.m. Will Audit and Feedback Drive Compliance with UV Robot Disinfection? Michele Fleming, MSN, RN, CIC, Infection Preventionist, Virginia Commonwealth University Health System; Yvette Major, MBA, MT (ASCP), Molecular Microbiology and Epidemiology Laboratory Specialist, Virginia Commonwealth University Health System; Mark Gryskevicz, Director Environmental Services, Virginia Commonwealth University Health System, Aramark; Jerry Fife, MEd, Safety and Traning Manager for Environmental Services, Virginia Commonwealth University Health System, Aramark; Lisa Hassmer, MBA, Executive Assisstand, Virginia Commonwealth University Health System; Nadia Masroor, BS, Project Coordinator, Virginia Commonweatlh University Health System; Kaila Cooper, MSN, RN, CIC, Director Hospital Infection Prevention Program, Virginia Commonwealth University Health System; Michelle Doll, MD, MPH, Associate Professor of Internal Medicine, Associate Hospital Epidemiologist, Virginia Commonwealth University Health System; Michael Stevens, MD, MPH, FACP, FIDSA, Assistant Professor of Internal Medicine, Associate Hospital Epidemiologist, Director of Antimicrobial Stewardship Program, Associate Program Director for Internal Medicine Residency Program, Global Health & Health Disparities Pathway Director, Virginia Commonwealth University Health System; Gonzalo Bearman, MD, MPH, FACP, FSHEA, FIDSA, Hospital Epidemiologist, Associate Professor of Internal Medicine, Virginia Commonwealth University Health System BACKGROUND: Increasingly literature illustrates the importance of environmental cleaning, especially in rooms that house Clostridium difficile positive patients (CDPP). Ultra Violet Robots (UVR) are capable of inactivating microorganisms and have a positive impact on Clostridium difficile rates. Evidence of conformity with established protocols for UVR disinfection in rooms that house CDPP does not exist without retrospective review. This study assesses the impact of ongoing audit and feedback to end users to improve compliance with UVR deployment. METHODS: Protocols were established to ensure proper cleaning of patient rooms in an urban, tertiary, 865 bed academic medical center. Enhanced cleaning of rooms that housed patients diagnosed with Clostridium difficile included UVR deployment for room disinfection. On weekdays, environmental services (EVS) was provided a list of all rooms with a CDPP for enhanced room cleaning. A retrospective, manual review of CDPP room assignments was compared to an automated report of UVR deployment. Rooms were identified as compliant if the UVR ran at terminal clean; the compliance goal was 90%. Capture rate was generated by dividing the number of rooms that received UVR disinfection by the number of opportunities for UVR disinfection. EVS leadership was notified
of all non-compliant rooms and the monthly capture rate. Capture rate was also reported at monthly Infection Control Committee meetings and to key stakeholders. RESULTS: Data was analyzed and rate of compliance calculated during a 25 month period, (February 2015 through February 2017), 1,562 opportunities existed for UVR deployment. UVR compliance increased from 20% to 100% during this time period. Capture remained above 80% consecutively for 19 of the 25 months, and above 90% consecutively for the last 5 months. Tracking of daily UVR utilization indicated that as compliance increased, the number of days UVR was not used decreased. CONCLUSIONS: Auditing compliance and providing ongoing feedback of performance can improve compliance with established protocols for the deployment of a UVR for Clostridium difficile room disinfection.
Education and Competencies Session EC-057 12:30-1:30 p.m. Are Adults in Hong Kong Using Face Mask Correctly? Linda Yin King Lee, PhD, RN, RM, Professor, The Open University of Hong Kong; Evangeline Pui Wah Lam, MPH, BScN(Hons), RN, RM, IBCLC, Lecturer, The Open University of Hong Kong; Chin Kiu Chan, Student, The Open University of Hong Kong; Sum Yi Chan, Student, The Open University of Hong Kong; Man Ki Chiu, Student, The Open University of Hong Kong; Wing Hei Chong, Student, The Open University of Hong Kong; Kin Wai Chu, Student, The Open University of Hong Kong; Man Sze Hon, Student, The Open University of Hong Kong; Lok Ki Kwan, Student, The Open University of Hong Kong; Kit Lam Tsang, Student, The Open University of Hong Kong; Siu Lai Tsoi, Student, The Open University of Hong Kong; Chung Wai Wu, Student, The Open University of Hong Kong BACKGROUND: Correct use of face mask is important in decreasing the spread of respiratory infection. Although correct use of face mask implies adopting right practice and performing proper technique, previous evaluations on use of face mask in general population mainly focus on the practice aspect. Existing evidence is unlikely to provide comprehensive reference for promotion of public health. The objective of this study was to assess the practice and technique for using face mask among adults in Hong Kong. METHODS: This descriptive study was conducted in 2016. It recruited a convenient sample of six hundred adults in the public area in Hong Kong. With reference to established guidelines, a questionnaire and an observation checklist were developed to assess subjects’ practice and technique for using face mask respectively. The psychometric properties of these two instruments, including content validity, test-retest reliability and inter-rater reliability, were tested and established. Descriptive statistics, including frequency and percentage, were used to describe subjects’ performance. RESULTS: Subjects’ overall performance was unsatisfactory. Regarding their practice, only 21% of the subjects reported that they always wear face mask when caring for sick people with fever or respiratory infection. Regarding their technique, more than 95% of the subjects did not perform hand hygiene before putting on or taking off the face mask, or after disposing the face mask. On average,
APIC 44th Annual Educational Conference & International Meeting | Portland, OR | June 14-16, 2017
Maureen Bolon, MD, MS, Medical Director of Healthcare Epidemiology and Infection Prevention, Northwestern Memorial Hospital; Cynthia Barnard, PhD, MBA, MSJS, VP, Quality, Northwestern Medicine; Teresa Zembower, MD, MPH, Medical Director, Healthcare Epidemiology and Infection Prevention, Northwestern Memorial Hospital BACKGROUND: As healthcare-associated infections (HAIs) drive financial incentives and penalties, reputation and quality improvement, accurate identification is increasingly vital. There is literature suggesting significant inconsistency and potential gaming in identification and reporting of HAIs. Infection Preventionists (IP) are usually responsible for performing infection surveillance at healthcare facilities but no standard for ensuring valid identification and reporting exists. This study evaluates the accuracy of HAI detection, focusing on central line associated bloodstream infections (CLABSIs) across a newly formed healthcare system. METHODS: We conducted an audit of all positive blood cultures not deemed to be CLABSIs across our healthcare system in patients with known central lines. Data were sampled based on test volume from four hospitals encompassing over 1,600 inpatient beds. Positive results were obtained directly from the respective facilities’ electronic surveillance systems. Data were collected from JanuaryFebruary 2015 and March-July 2015. Duplicate cultures were excluded as were cultures from patients without central lines. Positive cultures were independently reviewed according to the 2015 National Healthcare Safety Network (NHSN) definitions to validate the negative predictive value of IPs’ determinations. RESULTS: During this period, 452 cultures met criteria across the system and a random sample of 60 were reviewed. In two of the four hospitals, 30 cultures passed validation and no additional CLABSIs were identified. One facility had one additional CLABSI identified and one had two. CONCLUSIONS: These 3 true CLABSIs had been misidentified as secondary infections based on clinical judgement in conflict with NHSN definitions. The two hospitals without discrepancies have robust orientation, ongoing training and oversight of HAI surveillance and classification. Subsequently, this was implemented at all hospitals. This methodology outlines a relatively simple audit tool to monitor CLABSI surveillance across a system. Ongoing education and oversight of IP surveillance is important to ensure appropriate reporting to NHSN.
Session DSV-054 12:30-1:30 p.m. Validating Nurse Sick Call Data as a Method of Early Warning Outbreak Response in an Academic Teaching Hospital Keenan Williamson, MPH, Infection Preventionist, Oregon Health & Sciences University; Molly Hale, MPH, CIC, FAPIC, Manager, Infection Prevention & Control; Dawn Nolt, MD, MPH, Pediatric Medical Director, Oregon Health & Sciences University BACKGROUND: Early outbreak response is essential in mitigating negative health outcomes from infectious diseases in patients and healthcare workers in a hospital setting. Early outbreak detection research has not been conducted extensively in academic teaching hospitals. In June 2015, a university hospital experienced a norovirus outbreak among the Pediatric ICU nursing staff due to ongoing environmental norovirus transmission. This prompted
S35
establishment of an early outbreak detection and monitoring system for inpatient nurses. METHODS: Mean daily Pediatric ICU nurse sick call data were retrospectively analyzed from 2005-December 2014 and compared to the known PICU norovirus outbreak. The outbreak recorded 32 suspect or lab-confirmed norovirus cases from June 6th-June 24th, 2015 with a total of 19 nursing staff out sick. Comparison used the Exponential Weighted Moving Average (EWMA) with an alpha of 0.03. The EWMA forecast was elaborated by three standard deviations to serve as an Upper Control Limit (UCL). RESULTS: In 2015, the PICU reported 548 nurse sick calls. Moderate increases in nurse sick calls were observed during influenza season and holidays, Jan-Feb and Nov-Dec. Analyzing EWMA control charts revealed 12 days in 2015 above the UCL; including 3 days during the known PICU norovirus outbreak in June. Of the remaining 9 days above the UCL, only 2 were clustered in time-suggesting a missed investigation opportunity. CONCLUSIONS: The establishment of real-time analysis of nurse sick call data can serve as a proxy for early outbreak detection by identifying aberrant increases in nurse sick calls. Infection Prevention & Control staff developed a house-wide nurse sick call monitoring system using the EWMA method. Further analysis will be done to make the process sensitive to clusters in time and include an array of healthcare professionals.
Session DSV-055 12:30-1:30 p.m. Validation of National Healthcare Safety Network Laboratory-Identified Event Data Jarred Gray, MPH, Epidemiologist, Tennessee Department of Health; Ashley Fell, MPH, Epidemiologist, Tennessee Department of Health; Vicky Reed, RN, CRRN, Public Health Nurse Consultant, TN Department of Health; Rebecca Meyer, MPH, Epidemiologist, TN Department of Health; Marion Kainer, MD, MPH, FRACP, FSHEA, Director Healthcare Associated Infection and Antimicrobial Resistance Program, TN Department of Health BACKGROUND: The state health department (SHD) has required acute care hospitals to report Clostridium difficile (CDI) and Methicillin-Resistant Staphylococcus aureus (MRSA) Laboratoryidentified (LabID) event data through the National Healthcare Safety Network (NHSN) since July 2010. SHD conducted a validation of NHSN LabID event data to assess data quality and to identify common reporting errors. METHODS: SHD used the NHSN external validation toolkit and selected 31 facilities for validation (01/01/2015-06/30/2015); 23 facilities were selected for MRSA and 23 for CDI. Up to 60 positive laboratory specimens were reviewed for each selected event type (CDI and/or MRSA). Patient medical records, including location, and laboratory records were reviewed to determine if the selected specimen met LabID event criteria. RESULTS: SHD identified 947 reportable CDI/MRSA events. Of the reportable CDI/MRSA events, 138 (14.6%) were unreported and 121 (12.8%) were reported with errors. Eighty-one percent (767/947) of the events validated were identified at facilities utilizing clinical decision support software (CDSS) to aid in event reporting. In these facilities, SHD found that 27% of reportable CDI/MRSA events were either unreported or reported incorrectly. Facilities that verified their CDSS identified events, at least monthly, with a laboratory line list had significantly higher percentage of correctly reported events than facilities that validated less frequently (77.75% vs 56.25%, P = .0318).
APIC 44th Annual Educational Conference & International Meeting | Portland, OR | June 14-16, 2017
S36
Poster Abstracts / American Journal of Infection Control 45 (2017) S16-S93
CONCLUSIONS: Validation of NHSN LabID events has informed SHD of significant errors in reporting arising from the use of CDSS. The use of CDSS is best supplemented with monthly verification of results using a laboratory line list until the accuracy of CDSS can be improved. In addition, validation has provided SHD with valuable information about data quality as well as common misconceptions that can be addressed through education and training. Data validation also strengthened SHD’s relationships with infection preventionists, improving communication between SHD and the acute care hospital community.
Session DSV-056 12:30-1:30 p.m. Will Audit and Feedback Drive Compliance with UV Robot Disinfection? Michele Fleming, MSN, RN, CIC, Infection Preventionist, Virginia Commonwealth University Health System; Yvette Major, MBA, MT (ASCP), Molecular Microbiology and Epidemiology Laboratory Specialist, Virginia Commonwealth University Health System; Mark Gryskevicz, Director Environmental Services, Virginia Commonwealth University Health System, Aramark; Jerry Fife, MEd, Safety and Traning Manager for Environmental Services, Virginia Commonwealth University Health System, Aramark; Lisa Hassmer, MBA, Executive Assisstand, Virginia Commonwealth University Health System; Nadia Masroor, BS, Project Coordinator, Virginia Commonweatlh University Health System; Kaila Cooper, MSN, RN, CIC, Director Hospital Infection Prevention Program, Virginia Commonwealth University Health System; Michelle Doll, MD, MPH, Associate Professor of Internal Medicine, Associate Hospital Epidemiologist, Virginia Commonwealth University Health System; Michael Stevens, MD, MPH, FACP, FIDSA, Assistant Professor of Internal Medicine, Associate Hospital Epidemiologist, Director of Antimicrobial Stewardship Program, Associate Program Director for Internal Medicine Residency Program, Global Health & Health Disparities Pathway Director, Virginia Commonwealth University Health System; Gonzalo Bearman, MD, MPH, FACP, FSHEA, FIDSA, Hospital Epidemiologist, Associate Professor of Internal Medicine, Virginia Commonwealth University Health System BACKGROUND: Increasingly literature illustrates the importance of environmental cleaning, especially in rooms that house Clostridium difficile positive patients (CDPP). Ultra Violet Robots (UVR) are capable of inactivating microorganisms and have a positive impact on Clostridium difficile rates. Evidence of conformity with established protocols for UVR disinfection in rooms that house CDPP does not exist without retrospective review. This study assesses the impact of ongoing audit and feedback to end users to improve compliance with UVR deployment. METHODS: Protocols were established to ensure proper cleaning of patient rooms in an urban, tertiary, 865 bed academic medical center. Enhanced cleaning of rooms that housed patients diagnosed with Clostridium difficile included UVR deployment for room disinfection. On weekdays, environmental services (EVS) was provided a list of all rooms with a CDPP for enhanced room cleaning. A retrospective, manual review of CDPP room assignments was compared to an automated report of UVR deployment. Rooms were identified as compliant if the UVR ran at terminal clean; the compliance goal was 90%. Capture rate was generated by dividing the number of rooms that received UVR disinfection by the number of opportunities for UVR disinfection. EVS leadership was notified
of all non-compliant rooms and the monthly capture rate. Capture rate was also reported at monthly Infection Control Committee meetings and to key stakeholders. RESULTS: Data was analyzed and rate of compliance calculated during a 25 month period, (February 2015 through February 2017), 1,562 opportunities existed for UVR deployment. UVR compliance increased from 20% to 100% during this time period. Capture remained above 80% consecutively for 19 of the 25 months, and above 90% consecutively for the last 5 months. Tracking of daily UVR utilization indicated that as compliance increased, the number of days UVR was not used decreased. CONCLUSIONS: Auditing compliance and providing ongoing feedback of performance can improve compliance with established protocols for the deployment of a UVR for Clostridium difficile room disinfection.
Education and Competencies Session EC-057 12:30-1:30 p.m. Are Adults in Hong Kong Using Face Mask Correctly? Linda Yin King Lee, PhD, RN, RM, Professor, The Open University of Hong Kong; Evangeline Pui Wah Lam, MPH, BScN(Hons), RN, RM, IBCLC, Lecturer, The Open University of Hong Kong; Chin Kiu Chan, Student, The Open University of Hong Kong; Sum Yi Chan, Student, The Open University of Hong Kong; Man Ki Chiu, Student, The Open University of Hong Kong; Wing Hei Chong, Student, The Open University of Hong Kong; Kin Wai Chu, Student, The Open University of Hong Kong; Man Sze Hon, Student, The Open University of Hong Kong; Lok Ki Kwan, Student, The Open University of Hong Kong; Kit Lam Tsang, Student, The Open University of Hong Kong; Siu Lai Tsoi, Student, The Open University of Hong Kong; Chung Wai Wu, Student, The Open University of Hong Kong BACKGROUND: Correct use of face mask is important in decreasing the spread of respiratory infection. Although correct use of face mask implies adopting right practice and performing proper technique, previous evaluations on use of face mask in general population mainly focus on the practice aspect. Existing evidence is unlikely to provide comprehensive reference for promotion of public health. The objective of this study was to assess the practice and technique for using face mask among adults in Hong Kong. METHODS: This descriptive study was conducted in 2016. It recruited a convenient sample of six hundred adults in the public area in Hong Kong. With reference to established guidelines, a questionnaire and an observation checklist were developed to assess subjects’ practice and technique for using face mask respectively. The psychometric properties of these two instruments, including content validity, test-retest reliability and inter-rater reliability, were tested and established. Descriptive statistics, including frequency and percentage, were used to describe subjects’ performance. RESULTS: Subjects’ overall performance was unsatisfactory. Regarding their practice, only 21% of the subjects reported that they always wear face mask when caring for sick people with fever or respiratory infection. Regarding their technique, more than 95% of the subjects did not perform hand hygiene before putting on or taking off the face mask, or after disposing the face mask. On average,
APIC 44th Annual Educational Conference & International Meeting | Portland, OR | June 14-16, 2017