- Email: [email protected]
Finding the gaps: An assessment of infection control surveillance needs in British Columbia acute care facilities
Finding the gaps: An assessment of infection control surveillance needs in British Columbia acute care facilities Bruce Gamage, RN, BSN, CIC,a Monali Varia, MHSc,a Margaret Litt, RN, FETP,a Sarah Pugh, MA,a Elizabeth Bryce, MD, FRCPC,b and the PICNet Needs Assessment Working Groupa Vancouver, British Columbia, Canada
Background: This paper reports on an infection prevention and control surveillance survey of acute care facilities (ACFs) performed by the Provincial Infection Control Network of British Columbia. Methods: A surveillance questionnaire was sent to all health care facilities that had access to an infection control professional. The questionnaire incorporated questions on organism-specific, disease-specific, and general surveillance activities. Results: Questionnaires were returned from 47 of 51 (92%) of the ACFs surveyed. Participation in surveillance of methicillinresistant Staphylococcus aureus-, vancomycin-resistant Enterococci-, and Clostridium difficile-associated disease ranged from 97% to 100%, but surveillance methodologies were inconsistent. Surgical-site infection surveillance did not correlate with the most commonly performed operations or with those procedures associated with higher morbidity and mortality from a postoperative infection. Considerable variation in data collection methods and case definitions was also identified. Surveillance for urinary tract infections, bloodstream infections, and ventilator-associated pneumonia was present in 28%, 51%, and 23% of responding ACFs, respectively. Conclusion: The current lack of a standardized surveillance system in British Columbia limits the ability of facilities to set appropriate benchmarks to assist in prioritizing and applying infection control interventions. The survey, however, has assisted in prioritizing implementation of surveillance activities and identifying the resources that would be required. (Am J Infect Control 2008;36:706-10.)
The Provincial Infection Control Network (PICNet) of British Columbia (BC) was formed by the Ministry of Health in 2005 to provide advice and strategic intervention on relevant policy, procedures, and issues relating to infection prevention and control across the continuum of health care. One of the first orders of business for PICNet was to review the scope and nature of surveillance activities for health care-associated infections (HAI) in BC to enable future development of surveillance programs at the local and potentially the provincial
From the Provincial Infection Control Network,a and Vancouver General Hospital,b Vancouver, British Columbia, Canada. Address correspondence to Bruce Gamage, RN, BSN, CIC, Network Manager, Provincial Infection Control Network, 655 W 12th Ave, Vancouver, BC V5Z 4R4. E-mail: [email protected]. PICNet Needs Assessment Working Group: Joanne Archer, Heather Blaus, Ian Connell, Fern Davey, Nicki Gill, Janice DeHeer, Jacqueline Hlagi, Betty Johnson, James Lu, Shelley Myatovic, Sue Pollock, Diane Roscoe, Gayle Shimokura, Annalee Yassi. Conflicts of interest: None of the authors received any financial or material support from any organization that may either gain or lose financially from the results or conclusions of this study. None of the authors received any financial support from a manufacturer or were given any product free of charge. 0196-6553/$34.00 Copyright ª 2008 by the Association for Professionals in Infection Control and Epidemiology, Inc. doi:10.1016/j.ajic.2008.06.004
706
level. Identification of required resources, barriers, and impediments, as well as opportunities for standardization of surveillance, was key to developing a foundation for future successful collaboration. This paper focuses on the results from an infection prevention and control surveillance survey of acute care facilities (ACFs).
METHODS An infection prevention and control surveillance questionnaire was developed by members of the PICNet Needs Assessment Working Group and then validated by senior infection control professionals (ICPs) representing the directly funded (ie, public) health care facilities of the 6 provincial Health Authorities. The questionnaire was sent to a convenience sample of 51 of the 81 ACFs in BC. This convenience sample constituted 63% of ACFs in BC and 93% of the acute care beds. Facilities were selected based on the availability of an ICP to supply the necessary information. Follow-up e-mails and telephone calls were made to clarify and/or request additional information. Additional questionnaires were distributed to targeted facilities in February 2006 to increase the response rate. Data were submitted centrally to the PICNet Management Office for analysis. The questionnaire incorporated (1) organism-specific surveillance (methicillin-resistant Staphylococcus aureus [MRSA], vancomycin-resistant Enterococci
Gamage et al
[VRE], Clostridium difficile-associated diarrhea [CDAD]); (2) disease-specific surveillance (urinary tract infections [UTI], bloodstream infections [BSI], ventilatorassociated pneumonias [VAP], health care-associated pneumonias [HAP], and surgical site infection [SSI] surveillance); and (3) general surveillance activities. Data collected included case definitions used, patient populations surveyed, admission screening, laboratory methods used, and data capture methods. At no time were patient or professional names or site-specific surveillance data requested. The survey form is available by e-mail request to [email protected]. Data were entered into EpiData (EpiData Association, Odense, Denmark) and analyzed in SPSS v.10.1 (SPSS, Inc, Chicago, IL). Data were collected for both acute and long-term care; only surveillance data for ACFs are presented here. In instances in which information received from respondents was incomplete or values missing, the respondents were excluded from the question for the purposes of analysis.
RESULTS Questionnaires were returned from 47 of 51 (92%) of the ACFs surveyed. Of these, 7 questionnaires combined responses for the ACFs and their associated long-term care facilities (LTCF) and were included in the results below. Six (9%) of the 47 responses from ACFs were completed by teaching hospitals (43% of the teaching hospitals in BC). Participation in organism-specific surveillance and disease-specific monitoring in ACFs is presented in Fig 1. Eight facilities indicated that organismspecific surveillance had existed for at least 10 years. All responding facilities conducting surveillance reported that MRSA, VRE, and CDAD were essential elements of their programs, and 97%, 94%, and 100% (respectively) of facilities conducted surveillance for these organisms on all clinical units (Table 1). Although 80% of facilities reported surveillance of MRSA cases for greater than 5 years, only 15 (68%) ACFs had been performing VRE surveillance and 42% collecting CDAD data for that length of time. Approximately half of the ACFs used a combination of laboratory and clinical information as their data sources, and the majority included both inpatient and outpatient populations in their MRSA, VRE, and CDAD surveillance activities. MRSA and VRE screening was performed on admitted patients meeting the criteria of admission to a health care facility for greater than or equal to 48 hours in the last 90 days and/or known contact with a MRSA or VRE case by all of the responding facilities. Approximately 60% of facilities retained a bank of MRSA and VRE isolates, and 50% characterized isolates by molecular methods. Twenty-five facilities (53%) reported that denominators were not collected to calculate MRSA or VRE rates.
December 2008
707
Percentae of responding facilities conducting surveillance VAP UTI HAP BSI CDAD VRE MRSA 0
20
40
60
80
100
Fig 1. Surveillance of health care-associated infection in responding BC acute care facilities. Table 1. Organism-specific surveillance programs in responding BC acute care facilities* MRSA, n (%) Surveillance on all units 36 (97) Surveillance data Laboratory only 16 (42) Clinical only 3 (8) Both 19 (50) Patients populations surveyed Inpatient 8 (22) Inpatient/outpatient 28 (76) 32 (86) Rates reported to medical advisory committeesy Admission screening 37 (97) Contact screening 38 (100) Organisms saved 23 (64) Organisms typed 20 (57)
VRE, n (%)
CDAD, n (%)
34 (94)
38 (100)
21 (43) 2 (8) 18 (49)
17 (44) 5 (13) 17 (44)
8 (22) 28 (78) 32 (89)
14 (36) 25 (64) 36 (97)
36 (97) 38(100) 21 (60) 16 (47)
9 (23) NAz 2 (6) 5 (15)
NOTE. N 5 38 facilities. *Information from 9 of the responding facilities included missing values. These responses were excluded for the purposes of analysis. y Medical Advisory Committee: the committee included the chief of staff for each medical service in a facility. z Not applicable. Asymptomatic contacts of CDAD patients are not eligible for laboratory testing.
Of those facilities that calculated rates, 13 (28%) used patient-days as a denominator, and 5 (11%) used patient admissions. Definitions for health care-associated MRSA or VRE varied based on the length of time in the facility prior to organism identification. Of the 29 facilities providing definitions, 14 (48%) facilities defined health care-associated cases as acquisition greater than 48 hours after admission, 4 (14%) facilities used greater than 72 hours after admission, 3 (10%) facilities used greater than 1 week after admission, and 8 (28%) sites used the Public Health Agency of Canada (PHAC) definition of 48 to 72 hours postadmission. Community-associated cases of MRSA were defined as no history of hospitalization by 19 (68%) facilities, the PHAC definition of no hospitalization in the past 12
708
Gamage et al
Vol. 36 No. 10
months by 8 (26%) facilities, and based solely on clindamycin and trimethoprim-sulfamethoxazole sensitivity in 2 (6%) facilities. Whereas all facilities reported laboratory detection of toxin A for CDAD diagnosis, culture and antigen detection were infrequently done. Nine (23%) facilities reported admission screening for CDAD. In these facilities, this was based on the use of a gastrointestinal algorithm for admitted, symptomatic patients. Most sites (83%) employed either PHAC or Centers for Disease Control and Prevention (CDC) case definitions. Relapse cases of CDAD (defined as a recurrence of diarrhea) were detected by most (85%) ACFs, and 21 (60%) sites reported inclusion of repeat events. Common case definitions for CDAD were used by all facilities in a given health authority; however, the definition of relapse varied by facility and by health authority. CDAD relapse definitions ranged from recurrence of diarrhea within 6 weeks to 3 months from the previous CDAD episode. UTIs were followed in only 13 (28%) of ACFs, and all participating sites employed PHAC or CDC case definitions. Only 6 of the 13 responding centers (46%) collected denominator data to calculate rates. Of those facilities, 5 (38%) used patient-days for a denominator, and only 1 used number of catheter-days. Surveillance for BSIs was conducted by 24 (51%) of responding ACFs with 2 facilities initiating these programs shortly before the provincial survey. Eleven (40.7%) facilities indicated that BSI surveillance had existed for at least 10 years. Comprehensive surveillance was conducted at 15 (63%) of facilities, and the majority (78%) of sites used a combination of laboratory and clinical data to classify BSIs. All but 1 facility used PHAC or CDC case definitions, and 19 (79%) facilities compiled data as rates. In only 4 (21%) facilities, device-days was used as a denominator. In other cases, patient-days was used to calculate rates. Surveillance for VAP was conducted in 11 (23%) ACFs, and HAP were followed at 16 (34%) sites. Surveillance was based on both laboratory and clinical data in 10 of 13 responding (77%) facilities. All facilities following HAP or VAP used CDC or PHAC case definitions. These definitions differ in the length of time used to define association with mechanical ventilation (48 hours vs 96 hours, respectively). Eleven facilities reported collection of only HAP cases and no denominator data. Similarly, 3 facilities did not collect denominator data for calculation of VAP rates. Of those facilities conducting surveillance for VAP, 4 (36%) used ventilator-days as their denominator for calculating rates, and 4 (36%) used patient-days. Table 2 details the results for SSI surveillance activity. Most of the responding ACFs performed orthopaedic (77%), obstetrical (87%), breast (77%), and
Table 2. Surgical site infection surveillance programs in responding BC acute care hospitals
Surgery
Number (%) of acute care Number (%) of acute facilities conducting care facilities conducting surgery (n 5 47) SSI surveillance*
Orthopaedic Breast Neurosurgical Cardiovascular Obstetrics Renal Gastrointestinal
36 (77) 36 (77) 8 (17) 7 (15) 41 (87) 15 (36) 45 (96)
21/36 (58) 16/36 (44) 5/8 (62) 5/7 (71) 25/41 (61) 10/15 (67) 16/45 (36)
*Of those facilities conducting surgery.
gastrointestinal (96%) operations. For those that conducted surveillance, the orthopaedic procedures monitored were knee surgery, hip replacement surgery, hip fracture procedures, and spinal surgery. The principal obstetric surgeries followed were cesarean sections and hysterectomies, whereas breast surgery surveillance included mastectomies, breast reductions, and reconstructive surgeries. Gastrointestinal procedures were monitored by 16 (36%) of the facilities; appendectomies, bowel surgeries, and laparoscopic cholecystectomies were the procedures more commonly surveyed. Among the other surgical classes, neurosurgical monitoring included cranial and shunt surgeries, whereas cardiovascular procedures followed were valve replacement and bypass procedures. Incomplete information was received on SSI surveillance methodology from the sites that performed this activity, and results in Table 3 reflect this. Many ACFs reported SSI surveillance programs that had been ongoing for more than 10 years, and the majority of sites captured the number of procedures conducted as denominator data for calculation of SSI rates. All facilities conducting SSI surveillance collected wound class data and indicated that the results were reported to other authorities (eg, Infection Control Committee or Medical Advisory Committee, the committee including the chief of staff for each medical service in a facility). Three centers (10%) indicated that no risk stratification method was used. However, many facilities did not complete this field or the question regarding how surgeries were coded (eg, ICD 9 or 10). Postdischarge surveillance varied by surgery type from 0% for neurosurgical procedures to 10 facilities (56%) conducting postdischarge obstetrical surveillance (cesarean section and hysterectomy). When teaching and nonteaching hospitals were compared for level of surveillance activity, only SSI surveillance for cardiovascular and neurosurgical procedures was conducted in teaching hospitals. Thirty-four (94%) of the 36 facilities that provided information indicated that ICPs were responsible for data entry. Sixteen
Gamage et al
December 2008
709
Table 3. Characteristics of SSI surveillance programs in BC acute care hospitals Procedure (n 5 respondents reporting conducting surveillance on specific surgeries)* Orthopaedic, n 5 8 Breast, n 5 8 Neurosurgery, n 5 2 Cardiovascular, n 5 4 Obstetrics, n 5 18 Renal, n 5 6 Gastrointestinal, n 5 11
Retrospective surveillance
Inpatient surveillance only
Inpatient/ outpatient surveillance
ASA risk score recorded
Wound class recorded
Number of procedures used as denominator
Postdischarge surveillance
Results reported to MACy
5 (62) 4 (50) 0 (0) 1 (50) 4 (22) 4 (67) 5 (45)
4 (50) 2 (25) 2 (100) 1 (50) 7 (39) 1 (17) 4 (36)
4 (50) 6 (75) 0 (0) 1 (50) 11 (61) 5 (83) 7 (64)
6 (86) 6 (86) 2 (100) 1 (50) 17 (94) 4 (67) 8 (73)
8 (100) 8 (100) 2 (100) 4 (100) 18 (100) 6 (100) 11 (100)
7 (87) 8 (100) 1 (50) 4 (100) 18 (100) 6 (100) 11 (100)
4 (50) 3 (37) 0 (0) 1 (50) 10 (56) 2 (33) 5 (45)
7 (87) 8 (100) 1 (100) 4 (100) 18 (100) 6 (100) 10 (91)
ASA, American Society of Anesthesiologists. *The values ‘‘n (%)’’ are based on the facilities that provided details of their SSI surveillance program; missing data from facilities not providing program details were excluded from the percent calculations. y Medical Advisory Committee (MAC): the committee included the chief of staff for each medical service in a facility.
(47%) facilities that provided information reported access to epidemiologic services. These services ranged from access to an individual with master’s/PhD level training in epidemiology (79%) to individuals who had taken a single course in epidemiology. Only 2 regions reported access to degree epidemiologists. The vast majority (94%) of ICPs entered their own surveillance data, an activity that could more appropriately be designated to clerical staff and/or minimized with the use of electronic data capture forms.
DISCUSSION Surveillance for HAIs is a foundational activity of an infection control program. Continuous monitoring of HAI rates can be used to implement quality improvement activities and programs, assess effectiveness of interventions, and benchmark with comparable facilities. Surveillance data can also be used to identify quickly and confirm outbreaks and provide epidemiologic profiles for clinical and research purposes.1 A convenience sample was used for the distribution of surveys; thus, facilities without access to an ICP were excluded. This limits the ability to compare the surveillance activities between facilities with and without an ICP and likely overestimates the amount and quality of surveillance that is actually done throughout the province. A response rate of 92% was achieved from the convenience sample, with only a few smaller rural centers failing to report. We thought it reasonable to include the 7 facilities that reported data that combined information on LTC and ACF facilities because the methodology used was confirmed to be similar by the responders whether within the LTC or ACF setting. Participation in organism-specific surveillance in ACF was excellent, but activity was less than optimal for disease-specific monitoring. Participation in surveillance of MRSA, VRE, and CDAD was high, but
surveillance methodologies were inconsistent. Of note was the variation in the definition of community-acquired MRSA. This is particularly problematic given the well-documented increase in the community-associated strain of MRSA (USA 300) in disadvantaged populations in our province.2 Over half (56%) of facilities presented raw data rather than rates for MRSA, VRE, and CDAD. Without the calculation of incidence rates, intrafacility trending over time is difficult, and the ability to use the numbers as quality of care measurements to direct interventions could be problematic. Approximately 40% did not characterize or save MRSA or VRE isolates, limiting the ability to provide epidemiologic and molecular information. Most facilities did not save samples or cultures for C difficile, limiting the ability to perform molecular characterization to detect newer more virulent isolates such as the recently documented tcdC deletion or ‘‘Montreal’’ strains.3 SSIs, the third most common nosocomial infections, cause substantial morbidity and mortality and increase hospital costs.4 Surveillance programs have been shown to be an effective measure in reducing SSI rates.5 However, procedures under surveillance did not correlate with the most commonly performed operations. In addition, they did not necessarily correlate to those procedures associated with higher morbidity and mortality from a postoperative infection. Notably, orthopaedic, neurologic, and cardiovascular procedures were followed in less than two thirds of facilities performing these surgeries. Only cesarean sections were followed postoperatively in more than 75% of facilities. Results also found that the intensity of SSI surveillance was not related to facility size or complexity of care. Teaching hospitals did do more surveillance for SSI related to breast, neurologic, cardiovascular, and orthopaedic procedures as would be expected given
710
Gamage et al
Vol. 36 No. 10
the distribution of caseload weighted toward this group. However, there was no difference between teaching and nonteaching facilities for obstetrical or gastrointestinal SSI surveillance. According to National Nosocomial Infection Surveillance System methodology, a minimal SSI surveillance program should include the capture of American Society of Anesthesiologists risk stratification index and denominator data for the calculation of SSI rates.6 Capture of this desirable information ranged from 50% to 100% of facilities for those sites that completed this information. The incompleteness of this data likely indicates a lack of knowledge regarding this risk stratification methodology, a deduction bolstered by subsequent discussion with some of the respondents. Considerable variation in procedure coding, use of American Society of Anesthesiologists scores to stratify procedures by risk, collection of denominator data, inpatient and outpatient data collection, and need to define the methods by which infected cases are identified are just a few of the issues requiring resolution before consistent SSI surveillance across the province could be achieved. The relative lack of surveillance for UTI, BSI, VAP, and HAP likely reflected the intensity of resources and training required for these activities, particularly BSI and VAP. Infection control programs were similarly strikingly under-serviced in terms of access to epidemiologic services and data entry. In 1985, the Study on the Efficacy of Nosocomial Infection Control project identified 3 essential elements of infection surveillance and control programs.5 These included epidemiologic surveillance for HAIs, developing policies and procedures to control these infections based on the surveillance data, and training specific personnel to do the surveillance and coordinate the control activities. The results of this needs assessment demonstrate the great potential and opportunities for provincial standardization of surveillance methodology particularly with regard to MRSA, CDAD, and VRE. Developing a standardized surveillance system to determine the rates and trends of HAI in health care facilities in BC will allow for the setting of appropriate benchmarks as required to assist facilities in applying interventions to improve infection control practice and, ultimately, patient care.
The survey also assisted in prioritizing implementation of surveillance activities and the resources that would be required, particularly with regard to information system (database) support, provision of medical microbiology and epidemiology services, and clerical assistance. The results of this survey can be used by individual facilities and has been used on a provincial basis to prepare business cases for acquiring these resources. The high proportion of facilities participating in this survey and the enthusiasm for the project by the ICP community suggests that development of uniform provincial surveillance system is an achievable goal. With this in mind, PICNet has worked to developed standardized surveillance methodology to collect data on CDAD and will be working with the facilities, province wide, to standardize surveillance methodologies for other health care-acquired infections. The motto of the Provincial Infection Control Network, ‘‘Good Science, Good Will,’’ was exemplified by the commitment of participating ACFs and bodes well for future program development and implementation.
References 1. Schekler WE, Brimhall D, Buck AS, Farr BM, Friedman C, Garabaldi RA, et al. Requirements for infrastructure and essential activities of Infection control and epidemiology in hospitals: a consensus panel report. Am J Infect Control 1998;26:47-60. 2. Al-Rawahi GN, Schreader AG, Porter SD, Roscoe DL, Gustafson R, Bryce EA. Methicillin-resistant Staphylococcus aureus nasal carriage among injection drug users: six years later. J Clin Microbiol 2008;46: 477-9. 3. MacCannell DR, Louie TJ, Gregson DB, Laverdiere M, Labbe A-C, Henwick S, et al. Molecular analysis of Clostridium difficile PCR ribotype 027 isolates from eastern and western Canada. J Clin Microbiol 2006;44:2147-52. 4. Mangram AJ, Horan TC, Silver LC, Jarvis WR. Centers for Disease Control and Prevention (CDC), Hospital Infection Control Practices Advisory Committee. Guideline for prevention of surgical site infection. Am J Infect Control 1999;27:97-132. 5. Haley RW, Culver DH, White JW, Morgan WM, Emori TG, Munn VP, et al. The efficacy of infection surveillance and control programs in preventing nosocomial infection is US hospitals. Am J Epidemiol 1985;121: 182-205. 6. Horan TC, Gaynes R. Surveillance of nosocomial infections. In: Mayhall CG, ed. Hospital epidemiology and infection control. Philadelphia: Lippincott Williams & Wilkins; 2004:1659–1702.
Background: This paper reports on an infection prevention and control surveillance survey of acute care facilities (ACFs) performed by the Provincial Infection Control Network of British Columbia. Methods: A surveillance questionnaire was sent to all health care facilities that had access to an infection control professional. The questionnaire incorporated questions on organism-specific, disease-specific, and general surveillance activities. Results: Questionnaires were returned from 47 of 51 (92%) of the ACFs surveyed. Participation in surveillance of methicillinresistant Staphylococcus aureus-, vancomycin-resistant Enterococci-, and Clostridium difficile-associated disease ranged from 97% to 100%, but surveillance methodologies were inconsistent. Surgical-site infection surveillance did not correlate with the most commonly performed operations or with those procedures associated with higher morbidity and mortality from a postoperative infection. Considerable variation in data collection methods and case definitions was also identified. Surveillance for urinary tract infections, bloodstream infections, and ventilator-associated pneumonia was present in 28%, 51%, and 23% of responding ACFs, respectively. Conclusion: The current lack of a standardized surveillance system in British Columbia limits the ability of facilities to set appropriate benchmarks to assist in prioritizing and applying infection control interventions. The survey, however, has assisted in prioritizing implementation of surveillance activities and identifying the resources that would be required. (Am J Infect Control 2008;36:706-10.)
The Provincial Infection Control Network (PICNet) of British Columbia (BC) was formed by the Ministry of Health in 2005 to provide advice and strategic intervention on relevant policy, procedures, and issues relating to infection prevention and control across the continuum of health care. One of the first orders of business for PICNet was to review the scope and nature of surveillance activities for health care-associated infections (HAI) in BC to enable future development of surveillance programs at the local and potentially the provincial
From the Provincial Infection Control Network,a and Vancouver General Hospital,b Vancouver, British Columbia, Canada. Address correspondence to Bruce Gamage, RN, BSN, CIC, Network Manager, Provincial Infection Control Network, 655 W 12th Ave, Vancouver, BC V5Z 4R4. E-mail: [email protected]. PICNet Needs Assessment Working Group: Joanne Archer, Heather Blaus, Ian Connell, Fern Davey, Nicki Gill, Janice DeHeer, Jacqueline Hlagi, Betty Johnson, James Lu, Shelley Myatovic, Sue Pollock, Diane Roscoe, Gayle Shimokura, Annalee Yassi. Conflicts of interest: None of the authors received any financial or material support from any organization that may either gain or lose financially from the results or conclusions of this study. None of the authors received any financial support from a manufacturer or were given any product free of charge. 0196-6553/$34.00 Copyright ª 2008 by the Association for Professionals in Infection Control and Epidemiology, Inc. doi:10.1016/j.ajic.2008.06.004
706
level. Identification of required resources, barriers, and impediments, as well as opportunities for standardization of surveillance, was key to developing a foundation for future successful collaboration. This paper focuses on the results from an infection prevention and control surveillance survey of acute care facilities (ACFs).
METHODS An infection prevention and control surveillance questionnaire was developed by members of the PICNet Needs Assessment Working Group and then validated by senior infection control professionals (ICPs) representing the directly funded (ie, public) health care facilities of the 6 provincial Health Authorities. The questionnaire was sent to a convenience sample of 51 of the 81 ACFs in BC. This convenience sample constituted 63% of ACFs in BC and 93% of the acute care beds. Facilities were selected based on the availability of an ICP to supply the necessary information. Follow-up e-mails and telephone calls were made to clarify and/or request additional information. Additional questionnaires were distributed to targeted facilities in February 2006 to increase the response rate. Data were submitted centrally to the PICNet Management Office for analysis. The questionnaire incorporated (1) organism-specific surveillance (methicillin-resistant Staphylococcus aureus [MRSA], vancomycin-resistant Enterococci
Gamage et al
[VRE], Clostridium difficile-associated diarrhea [CDAD]); (2) disease-specific surveillance (urinary tract infections [UTI], bloodstream infections [BSI], ventilatorassociated pneumonias [VAP], health care-associated pneumonias [HAP], and surgical site infection [SSI] surveillance); and (3) general surveillance activities. Data collected included case definitions used, patient populations surveyed, admission screening, laboratory methods used, and data capture methods. At no time were patient or professional names or site-specific surveillance data requested. The survey form is available by e-mail request to [email protected]. Data were entered into EpiData (EpiData Association, Odense, Denmark) and analyzed in SPSS v.10.1 (SPSS, Inc, Chicago, IL). Data were collected for both acute and long-term care; only surveillance data for ACFs are presented here. In instances in which information received from respondents was incomplete or values missing, the respondents were excluded from the question for the purposes of analysis.
RESULTS Questionnaires were returned from 47 of 51 (92%) of the ACFs surveyed. Of these, 7 questionnaires combined responses for the ACFs and their associated long-term care facilities (LTCF) and were included in the results below. Six (9%) of the 47 responses from ACFs were completed by teaching hospitals (43% of the teaching hospitals in BC). Participation in organism-specific surveillance and disease-specific monitoring in ACFs is presented in Fig 1. Eight facilities indicated that organismspecific surveillance had existed for at least 10 years. All responding facilities conducting surveillance reported that MRSA, VRE, and CDAD were essential elements of their programs, and 97%, 94%, and 100% (respectively) of facilities conducted surveillance for these organisms on all clinical units (Table 1). Although 80% of facilities reported surveillance of MRSA cases for greater than 5 years, only 15 (68%) ACFs had been performing VRE surveillance and 42% collecting CDAD data for that length of time. Approximately half of the ACFs used a combination of laboratory and clinical information as their data sources, and the majority included both inpatient and outpatient populations in their MRSA, VRE, and CDAD surveillance activities. MRSA and VRE screening was performed on admitted patients meeting the criteria of admission to a health care facility for greater than or equal to 48 hours in the last 90 days and/or known contact with a MRSA or VRE case by all of the responding facilities. Approximately 60% of facilities retained a bank of MRSA and VRE isolates, and 50% characterized isolates by molecular methods. Twenty-five facilities (53%) reported that denominators were not collected to calculate MRSA or VRE rates.
December 2008
707
Percentae of responding facilities conducting surveillance VAP UTI HAP BSI CDAD VRE MRSA 0
20
40
60
80
100
Fig 1. Surveillance of health care-associated infection in responding BC acute care facilities. Table 1. Organism-specific surveillance programs in responding BC acute care facilities* MRSA, n (%) Surveillance on all units 36 (97) Surveillance data Laboratory only 16 (42) Clinical only 3 (8) Both 19 (50) Patients populations surveyed Inpatient 8 (22) Inpatient/outpatient 28 (76) 32 (86) Rates reported to medical advisory committeesy Admission screening 37 (97) Contact screening 38 (100) Organisms saved 23 (64) Organisms typed 20 (57)
VRE, n (%)
CDAD, n (%)
34 (94)
38 (100)
21 (43) 2 (8) 18 (49)
17 (44) 5 (13) 17 (44)
8 (22) 28 (78) 32 (89)
14 (36) 25 (64) 36 (97)
36 (97) 38(100) 21 (60) 16 (47)
9 (23) NAz 2 (6) 5 (15)
NOTE. N 5 38 facilities. *Information from 9 of the responding facilities included missing values. These responses were excluded for the purposes of analysis. y Medical Advisory Committee: the committee included the chief of staff for each medical service in a facility. z Not applicable. Asymptomatic contacts of CDAD patients are not eligible for laboratory testing.
Of those facilities that calculated rates, 13 (28%) used patient-days as a denominator, and 5 (11%) used patient admissions. Definitions for health care-associated MRSA or VRE varied based on the length of time in the facility prior to organism identification. Of the 29 facilities providing definitions, 14 (48%) facilities defined health care-associated cases as acquisition greater than 48 hours after admission, 4 (14%) facilities used greater than 72 hours after admission, 3 (10%) facilities used greater than 1 week after admission, and 8 (28%) sites used the Public Health Agency of Canada (PHAC) definition of 48 to 72 hours postadmission. Community-associated cases of MRSA were defined as no history of hospitalization by 19 (68%) facilities, the PHAC definition of no hospitalization in the past 12
708
Gamage et al
Vol. 36 No. 10
months by 8 (26%) facilities, and based solely on clindamycin and trimethoprim-sulfamethoxazole sensitivity in 2 (6%) facilities. Whereas all facilities reported laboratory detection of toxin A for CDAD diagnosis, culture and antigen detection were infrequently done. Nine (23%) facilities reported admission screening for CDAD. In these facilities, this was based on the use of a gastrointestinal algorithm for admitted, symptomatic patients. Most sites (83%) employed either PHAC or Centers for Disease Control and Prevention (CDC) case definitions. Relapse cases of CDAD (defined as a recurrence of diarrhea) were detected by most (85%) ACFs, and 21 (60%) sites reported inclusion of repeat events. Common case definitions for CDAD were used by all facilities in a given health authority; however, the definition of relapse varied by facility and by health authority. CDAD relapse definitions ranged from recurrence of diarrhea within 6 weeks to 3 months from the previous CDAD episode. UTIs were followed in only 13 (28%) of ACFs, and all participating sites employed PHAC or CDC case definitions. Only 6 of the 13 responding centers (46%) collected denominator data to calculate rates. Of those facilities, 5 (38%) used patient-days for a denominator, and only 1 used number of catheter-days. Surveillance for BSIs was conducted by 24 (51%) of responding ACFs with 2 facilities initiating these programs shortly before the provincial survey. Eleven (40.7%) facilities indicated that BSI surveillance had existed for at least 10 years. Comprehensive surveillance was conducted at 15 (63%) of facilities, and the majority (78%) of sites used a combination of laboratory and clinical data to classify BSIs. All but 1 facility used PHAC or CDC case definitions, and 19 (79%) facilities compiled data as rates. In only 4 (21%) facilities, device-days was used as a denominator. In other cases, patient-days was used to calculate rates. Surveillance for VAP was conducted in 11 (23%) ACFs, and HAP were followed at 16 (34%) sites. Surveillance was based on both laboratory and clinical data in 10 of 13 responding (77%) facilities. All facilities following HAP or VAP used CDC or PHAC case definitions. These definitions differ in the length of time used to define association with mechanical ventilation (48 hours vs 96 hours, respectively). Eleven facilities reported collection of only HAP cases and no denominator data. Similarly, 3 facilities did not collect denominator data for calculation of VAP rates. Of those facilities conducting surveillance for VAP, 4 (36%) used ventilator-days as their denominator for calculating rates, and 4 (36%) used patient-days. Table 2 details the results for SSI surveillance activity. Most of the responding ACFs performed orthopaedic (77%), obstetrical (87%), breast (77%), and
Table 2. Surgical site infection surveillance programs in responding BC acute care hospitals
Surgery
Number (%) of acute care Number (%) of acute facilities conducting care facilities conducting surgery (n 5 47) SSI surveillance*
Orthopaedic Breast Neurosurgical Cardiovascular Obstetrics Renal Gastrointestinal
36 (77) 36 (77) 8 (17) 7 (15) 41 (87) 15 (36) 45 (96)
21/36 (58) 16/36 (44) 5/8 (62) 5/7 (71) 25/41 (61) 10/15 (67) 16/45 (36)
*Of those facilities conducting surgery.
gastrointestinal (96%) operations. For those that conducted surveillance, the orthopaedic procedures monitored were knee surgery, hip replacement surgery, hip fracture procedures, and spinal surgery. The principal obstetric surgeries followed were cesarean sections and hysterectomies, whereas breast surgery surveillance included mastectomies, breast reductions, and reconstructive surgeries. Gastrointestinal procedures were monitored by 16 (36%) of the facilities; appendectomies, bowel surgeries, and laparoscopic cholecystectomies were the procedures more commonly surveyed. Among the other surgical classes, neurosurgical monitoring included cranial and shunt surgeries, whereas cardiovascular procedures followed were valve replacement and bypass procedures. Incomplete information was received on SSI surveillance methodology from the sites that performed this activity, and results in Table 3 reflect this. Many ACFs reported SSI surveillance programs that had been ongoing for more than 10 years, and the majority of sites captured the number of procedures conducted as denominator data for calculation of SSI rates. All facilities conducting SSI surveillance collected wound class data and indicated that the results were reported to other authorities (eg, Infection Control Committee or Medical Advisory Committee, the committee including the chief of staff for each medical service in a facility). Three centers (10%) indicated that no risk stratification method was used. However, many facilities did not complete this field or the question regarding how surgeries were coded (eg, ICD 9 or 10). Postdischarge surveillance varied by surgery type from 0% for neurosurgical procedures to 10 facilities (56%) conducting postdischarge obstetrical surveillance (cesarean section and hysterectomy). When teaching and nonteaching hospitals were compared for level of surveillance activity, only SSI surveillance for cardiovascular and neurosurgical procedures was conducted in teaching hospitals. Thirty-four (94%) of the 36 facilities that provided information indicated that ICPs were responsible for data entry. Sixteen
Gamage et al
December 2008
709
Table 3. Characteristics of SSI surveillance programs in BC acute care hospitals Procedure (n 5 respondents reporting conducting surveillance on specific surgeries)* Orthopaedic, n 5 8 Breast, n 5 8 Neurosurgery, n 5 2 Cardiovascular, n 5 4 Obstetrics, n 5 18 Renal, n 5 6 Gastrointestinal, n 5 11
Retrospective surveillance
Inpatient surveillance only
Inpatient/ outpatient surveillance
ASA risk score recorded
Wound class recorded
Number of procedures used as denominator
Postdischarge surveillance
Results reported to MACy
5 (62) 4 (50) 0 (0) 1 (50) 4 (22) 4 (67) 5 (45)
4 (50) 2 (25) 2 (100) 1 (50) 7 (39) 1 (17) 4 (36)
4 (50) 6 (75) 0 (0) 1 (50) 11 (61) 5 (83) 7 (64)
6 (86) 6 (86) 2 (100) 1 (50) 17 (94) 4 (67) 8 (73)
8 (100) 8 (100) 2 (100) 4 (100) 18 (100) 6 (100) 11 (100)
7 (87) 8 (100) 1 (50) 4 (100) 18 (100) 6 (100) 11 (100)
4 (50) 3 (37) 0 (0) 1 (50) 10 (56) 2 (33) 5 (45)
7 (87) 8 (100) 1 (100) 4 (100) 18 (100) 6 (100) 10 (91)
ASA, American Society of Anesthesiologists. *The values ‘‘n (%)’’ are based on the facilities that provided details of their SSI surveillance program; missing data from facilities not providing program details were excluded from the percent calculations. y Medical Advisory Committee (MAC): the committee included the chief of staff for each medical service in a facility.
(47%) facilities that provided information reported access to epidemiologic services. These services ranged from access to an individual with master’s/PhD level training in epidemiology (79%) to individuals who had taken a single course in epidemiology. Only 2 regions reported access to degree epidemiologists. The vast majority (94%) of ICPs entered their own surveillance data, an activity that could more appropriately be designated to clerical staff and/or minimized with the use of electronic data capture forms.
DISCUSSION Surveillance for HAIs is a foundational activity of an infection control program. Continuous monitoring of HAI rates can be used to implement quality improvement activities and programs, assess effectiveness of interventions, and benchmark with comparable facilities. Surveillance data can also be used to identify quickly and confirm outbreaks and provide epidemiologic profiles for clinical and research purposes.1 A convenience sample was used for the distribution of surveys; thus, facilities without access to an ICP were excluded. This limits the ability to compare the surveillance activities between facilities with and without an ICP and likely overestimates the amount and quality of surveillance that is actually done throughout the province. A response rate of 92% was achieved from the convenience sample, with only a few smaller rural centers failing to report. We thought it reasonable to include the 7 facilities that reported data that combined information on LTC and ACF facilities because the methodology used was confirmed to be similar by the responders whether within the LTC or ACF setting. Participation in organism-specific surveillance in ACF was excellent, but activity was less than optimal for disease-specific monitoring. Participation in surveillance of MRSA, VRE, and CDAD was high, but
surveillance methodologies were inconsistent. Of note was the variation in the definition of community-acquired MRSA. This is particularly problematic given the well-documented increase in the community-associated strain of MRSA (USA 300) in disadvantaged populations in our province.2 Over half (56%) of facilities presented raw data rather than rates for MRSA, VRE, and CDAD. Without the calculation of incidence rates, intrafacility trending over time is difficult, and the ability to use the numbers as quality of care measurements to direct interventions could be problematic. Approximately 40% did not characterize or save MRSA or VRE isolates, limiting the ability to provide epidemiologic and molecular information. Most facilities did not save samples or cultures for C difficile, limiting the ability to perform molecular characterization to detect newer more virulent isolates such as the recently documented tcdC deletion or ‘‘Montreal’’ strains.3 SSIs, the third most common nosocomial infections, cause substantial morbidity and mortality and increase hospital costs.4 Surveillance programs have been shown to be an effective measure in reducing SSI rates.5 However, procedures under surveillance did not correlate with the most commonly performed operations. In addition, they did not necessarily correlate to those procedures associated with higher morbidity and mortality from a postoperative infection. Notably, orthopaedic, neurologic, and cardiovascular procedures were followed in less than two thirds of facilities performing these surgeries. Only cesarean sections were followed postoperatively in more than 75% of facilities. Results also found that the intensity of SSI surveillance was not related to facility size or complexity of care. Teaching hospitals did do more surveillance for SSI related to breast, neurologic, cardiovascular, and orthopaedic procedures as would be expected given
710
Gamage et al
Vol. 36 No. 10
the distribution of caseload weighted toward this group. However, there was no difference between teaching and nonteaching facilities for obstetrical or gastrointestinal SSI surveillance. According to National Nosocomial Infection Surveillance System methodology, a minimal SSI surveillance program should include the capture of American Society of Anesthesiologists risk stratification index and denominator data for the calculation of SSI rates.6 Capture of this desirable information ranged from 50% to 100% of facilities for those sites that completed this information. The incompleteness of this data likely indicates a lack of knowledge regarding this risk stratification methodology, a deduction bolstered by subsequent discussion with some of the respondents. Considerable variation in procedure coding, use of American Society of Anesthesiologists scores to stratify procedures by risk, collection of denominator data, inpatient and outpatient data collection, and need to define the methods by which infected cases are identified are just a few of the issues requiring resolution before consistent SSI surveillance across the province could be achieved. The relative lack of surveillance for UTI, BSI, VAP, and HAP likely reflected the intensity of resources and training required for these activities, particularly BSI and VAP. Infection control programs were similarly strikingly under-serviced in terms of access to epidemiologic services and data entry. In 1985, the Study on the Efficacy of Nosocomial Infection Control project identified 3 essential elements of infection surveillance and control programs.5 These included epidemiologic surveillance for HAIs, developing policies and procedures to control these infections based on the surveillance data, and training specific personnel to do the surveillance and coordinate the control activities. The results of this needs assessment demonstrate the great potential and opportunities for provincial standardization of surveillance methodology particularly with regard to MRSA, CDAD, and VRE. Developing a standardized surveillance system to determine the rates and trends of HAI in health care facilities in BC will allow for the setting of appropriate benchmarks as required to assist facilities in applying interventions to improve infection control practice and, ultimately, patient care.
The survey also assisted in prioritizing implementation of surveillance activities and the resources that would be required, particularly with regard to information system (database) support, provision of medical microbiology and epidemiology services, and clerical assistance. The results of this survey can be used by individual facilities and has been used on a provincial basis to prepare business cases for acquiring these resources. The high proportion of facilities participating in this survey and the enthusiasm for the project by the ICP community suggests that development of uniform provincial surveillance system is an achievable goal. With this in mind, PICNet has worked to developed standardized surveillance methodology to collect data on CDAD and will be working with the facilities, province wide, to standardize surveillance methodologies for other health care-acquired infections. The motto of the Provincial Infection Control Network, ‘‘Good Science, Good Will,’’ was exemplified by the commitment of participating ACFs and bodes well for future program development and implementation.
References 1. Schekler WE, Brimhall D, Buck AS, Farr BM, Friedman C, Garabaldi RA, et al. Requirements for infrastructure and essential activities of Infection control and epidemiology in hospitals: a consensus panel report. Am J Infect Control 1998;26:47-60. 2. Al-Rawahi GN, Schreader AG, Porter SD, Roscoe DL, Gustafson R, Bryce EA. Methicillin-resistant Staphylococcus aureus nasal carriage among injection drug users: six years later. J Clin Microbiol 2008;46: 477-9. 3. MacCannell DR, Louie TJ, Gregson DB, Laverdiere M, Labbe A-C, Henwick S, et al. Molecular analysis of Clostridium difficile PCR ribotype 027 isolates from eastern and western Canada. J Clin Microbiol 2006;44:2147-52. 4. Mangram AJ, Horan TC, Silver LC, Jarvis WR. Centers for Disease Control and Prevention (CDC), Hospital Infection Control Practices Advisory Committee. Guideline for prevention of surgical site infection. Am J Infect Control 1999;27:97-132. 5. Haley RW, Culver DH, White JW, Morgan WM, Emori TG, Munn VP, et al. The efficacy of infection surveillance and control programs in preventing nosocomial infection is US hospitals. Am J Epidemiol 1985;121: 182-205. 6. Horan TC, Gaynes R. Surveillance of nosocomial infections. In: Mayhall CG, ed. Hospital epidemiology and infection control. Philadelphia: Lippincott Williams & Wilkins; 2004:1659–1702.