Victor D. Rosenthal, MD, a Dennis G. Maki, MD, b and Nicholas Graves, MA, PhD c Buenos Aires, Argentina; Madison, Wisconsin; and Brisbane, Australia

Transcription

1 The International Nosocomial Infection Control Consortium (INICC): Goals an objectives, escription of surveillance methos, an operational activities Victor D. Rosenthal, MD, a Dennis G. Maki, MD, b an Nicholas Graves, MA, PhD c Buenos Aires, Argentina; Maison, Wisconsin; an Brisbane, Australia We have shown that intensive care units (ICUs) in countries with limite resources have rates of evice-associate health care-associate infection (HAI), incluing central line relate bloostream infection (CLAB), ventilator-associate pneumonia (VAP), an catheter-associate urinary tract infection (CAUTI), 3 to 5 times higher than rates reporte from North American, Western European, an Australian ICUs. The International Nosocomial Infection Control Consortium (INICC) is an international ongoing collaborative HAI control program with a surveillance system base on that of the US National Healthcare Safety Network. The INICC was foune 10 years ago to promote evience-base infection control in hospitals in limite-resource countries an in hospitals of evelope countries without sufficient experience in HAI surveillance an control, through the analysis an feeback of surveillance ata collecte voluntarily by the member hospitals. It evelope from a hanful of South American hospitals in 1998 to a ynamic network of 98 ICUs in 18 countries, an is the only source of aggregate stanarize international ata on HAI epiemiology. Herein we report the criteria an mechanisms for gaining membership in INICC; the training of personnel in INICC hospitals; the INICC protocol for outcome surveillance of CLABs, VAPs, an CAUTIs in ICUs, microorganism profiles, bacterial resistance, antibiotic use, extra length of stay, extra costs, extra mortality, an risk factor analysis, an for process surveillance, incluing compliance rates for han hygiene, vascular catheter care, urinary catheter care, an measures for prevention of VAP; an the use of surveillance ata feeback as a powerful weapon for control of HAIs. The INICC will continue to evolve in its quest to fin more effective an efficient ways to assess patient risk an improve patient safety in hospitals. (Am J Infect Control 2008;36:e1-e12.) Health care acquire infections (HAIs) have been associate with significant morbiity an attributable mortality, 1-8 as well as greatly increase health care costs. 4,5,8-10 Stuies conucte in the Unite States 30 years ago 11 an recently valiate in Germany 12 have shown that an integrate infection control program, with HAI surveillance as its cornerstone, can reuce the incience of HAIs by 30%, yieling economic benefits. From the Meical College of Buenos Aires an Bernal Meical Center, Buenos Aires, Argentina a ; University of Wisconsin School of Meicine an Public Health, Maison, WI b ; an The Centre for Healthcare Relate Infection Surveillance an Prevention an Queenslan University of Technology, Brisbane, Queenslan, Australia. c Aress corresponence to Victor D. Rosenthal, MD, International Nosocomial Infection Control Consortium (INICC), Lavalleja 305, Floor 9, Apt B (ZIP 1414), Buenos Aires, Argentina. victor_rosenthal@ inicc.org. Conflicts of interest: All authors report no conflicts of interest /$34.00 Copyright ª 2008 by the Association for Professionals in Infection Control an Epiemiology, Inc. oi: /j.ajic Most stuies on HAI have been conucte in hospitals in evelope countries Relatively little ata have been reporte from limite-resource countries. 1-8,16-24 We have recently shown that intensive care units (ICUs) in these countries have rates of eviceassociate HAI, incluing central line relate bloostream infection (CLAB), ventilator-associate pneumonia (VAP), an catheter-associate urinary tract infection (CAUTI), 3 to 5 times higher than rates reporte from US ICUs. 1,2 Most limite-resource countries o not have laws manating HAI control programs, an hospital accreitation is rarely require. Funs an resources for infection control are very limite, 25 nurse: patient staffing ratios are often far lower on average than in ICUs in evelope countries, an there are high proportions of inexperience nurses, all of which has been shown to have a powerful association with greatly increase risk of evice-associate infections. 26 Finally, the use of outate technology also may be a factor. For example, open intravenous infusion systems are use nearly universally in limiteresource countries instea of the close systems that are the stanar of care in evelope countries. 27 e1

2 e2 Vol. 36 No. 9 Rosenthal, Maki, an Graves It is clear that there is an urgent nee even a moral imperative to avance our unerstaning of the epiemiology an control of HAI to the many thousans of hospitals an billions of patients of the limite resources worl. HISTORY, GOALS, AND OBJECTIVES The INICC is an international nonprofit, open, multicenter, collaborative HAI control program with a surveillance system base on that of the US National Healthcare Safety Network 13 (NHSN), formerly the National Nosocomial Infection Surveillance system (NNIS). 28 Foune in Argentina in 1998 by Dr Victor D. Rosenthal, the INICC is the first multinational research network establishe to control HAIs in hospitals in limite-resource countries an also at hospitals in evelope countries without sufficient experience in HAI surveillance an control, through the analysis of ata collecte voluntarily by its member hospitals. 1,3-8,16-24 The INICC has the following goals: To create a ynamic global network of hospitals in limite-resource countries an in hospitals of evelope countries without sufficient experience in HAI surveillance an control, which conucts HAI surveillance through stanarize efinitions an establishe methoologies, promotes evience-base infection control practices, an conucts infection control research To provie training an surveillance tools to allow iniviual hospitals to conuct outcome an process HAI surveillance, measure their consequences, an assess the impact of infection control practices To improve health care safety an quality worlwie through systematize programs to reuce rates of HAI, associate mortality, excess length of stay an costs, an bacterial resistance To improve the use of anti-infectives in clinical practice, with the ultimate goal of controlling antimicrobial resistance To train infection control personnel in iniviual hospitals how to esign an carry out simple prospective research stuies to analyze the clinical impact an cost- effectiveness of infection control interventions To measure trens in HAIs an antimicrobial resistance in hospitals aroun the worl using riskajuste ata that allows meaningful intrahospital an interhospital comparisons for local, nationwie, an global quality improvement efforts To evelop new, simple, an inexpensive but effective strategies for HAI prevention of HAI at hospitals in limite-resource countries an at hospitals in evelope countries without sufficient experience in HAI surveillance an control. MECHANISMS OF MEMBERSHIP Most of the current participating hospitals joine the INICC after 2002, the majority on their own volition after hearing the INICC Chairman, Dr Rosenthal, speak in their country or in International Scientific Meetings, reaing a publishe INICC paper, watching a scientific poster at a scientific meeting, learning about the INICC from its website, or hearing from a colleague alreay participating in the INICC (Fig 1). Patient confientiality is protecte by coing the recore information, with patient ientities known only to the iniviual hospital s infection control team. Membership in the consortium requires that each intereste hospital meet the following criteria: Hospital interest in reucing rates of HAIs through membership in the INICC One or more ICUs A potential infection control team; with a at least 1 eicate infection control practitioner (ICP) or any HCW willing to receive training in that role; an infectious isease specialist, epiemiologist, critical care specialist, pathologist, or any other physician willing to unergo brief training an accept responsibilities as principle investigator (PI) hospital epiemiologist; an a microbiology laboratory using techniques an criteria for bacteriologic isolation an ientification an oing stanarize antimicrobial susceptibility testing Willingness to engage in formal infection control training, conuct surveillance, an use process feeback to improve infection control practices in the hospital. Willingness to publish the results of the infection control interventions in peer- reviewe journals an/or present them at scientific meetings. Access of the hospital infection control team to Internet electronic communication is highly esirable. INICC requisites from the participating hospitals are as follows: Institutional review an approval of the INICC protocol, incluing by an internal Institutional Review Boar/Research Committee Aministrative approval for participation Institutional signature of an INICC commitment letter by a PI, an ICP an, if esire, a hospital aministrator ICP conucting outcome an process surveillance using stanarize CDC NHSN/NNIS efinitions for HAI an surveillance methoologies, eicating at least 60 minutes per ay for ata collection for each 15 ICU bes

3 Rosenthal, Maki, an Graves November 2008 e3 Hospital submits inquiry to INICC HQ about membership If hospital fulfills membership criteria, INICC HQ sens INICC protocol an commitment letter to PI/ICN PI submits INICC protocol to hospital IRB for approval PI/ICN an IRB or Hospital Aministrator sign commitment letter an return to INICC INICC HQ sens hospital an PI/ICN INICC surveillance manual an surveillance forms, PI/ICN traine on-site by INICC Chair or by training manual with /phone support PI/ICN commences outcome an process surveillance, may submit queries to INICC HQ 24/7 PI/ICN submit complete surveillance forms to INICC HQ monthly INICC HQ ajuicate ata, if necessary with queries to PI/ICN, uploa valiate ata into INICC electronic atabase INICC HQ generates institutional monthly report an transmits electronically to PI/ICN, with most recent publishe annual INICC an CDCNHSN/NNIS benchmark reports PI/ICN promulgate monthly report for performance feeback INICC HQ generates quarterly feeback on hospital trens an suggests stuies or infection control interventions INICC Infection Control manual upate every 6 months Perioic visits to hospital by INICC Chair if possible Member hospitals an INICC leaership generate research reports for publication Fig 1. INICC methoology for gaining membership an operational proceures. HQ, Heaquarters; PI, principal investigator; ICN, infection control nurse; IRB, institutional review boar; CDC, Centers for Diseases Control an Prevention; NHSN, National Healthcare Safety Network; NNIS, National Nosocomial Infection Surveillance System. Monthly submission of ata to INICC heaquarters in Buenos Aires an monthly an quarterly posting of INICC institutional reports for process feeback Prompt response to INICC queries an surveys Vigorous promotion of evience-base infection control practices locally that can be applie for the prevention of HAIs Willingness to carry out infection control research stuies. Active membership of participant hospitals provie following major benefits in terms of hospital safety an improve care: Training of hospital PIs an ICPs in basic hospital epiemiology, surveillance methos, an basic ata analysis Immeiate access to current scientific knowlege relevant to the iagnosis, surveillance, prevention, an control of HAIs Training an tools to be able to conuct outcome an process surveillance that can permit timely recognition of patient safety problems an intervention with appropriate control measures, to assess the clinical an economic impacts of HAIs in their hospital, an to assess the impact of specific infection control practices

4 e4 Vol. 36 No. 9 Rosenthal, Maki, an Graves Training to etect relevant trens in HAIs an make intrahospital an interhospital comparisons with risk-ajuste ata that can be use for local, regional, an nationwie quality improvement activities Improve safety an quality of health care through implementation of systematize programs to reuce HAI rates, associate mortality, excess lengths of stay, excess costs, an bacterial resistance Training of PIs an ICPs to esign an unertake simple hypothesis-riven applie research Ongoing support an avice on surveillance activities an control programs I Improve use of anti-infectives for prophylaxis an therapeutic use, with the goal of helping to control antimicrobial resistance Avice regaring clinical cost-effectiveness of new technologies relevant to infection control Ongoing central analysis of surveillance ata at INICC heaquarters, with monthly feeback institutional surveillance reports Reuce HAI rates, length of stay (LOS), extra costs ue to HAI, an antimicrobial resistance Opportunity to coauthor research reports for publication in peer-reviewe journals yearly. BASIC STRUCTURE OF THE INICC The INICC structure inclues the following: researchers in each member hospital, the INICC Country Coorinators, the INICC Heaquarters Team, an the INICC Avisory Boar. The researchers are the ICPs who collect surveillance ata an submit complete ata forms to Buenos Aires heaquarters an implement infection control activities in their hospital, an the PI hospital epiemiologist who irects the institutional program. The Country Coorinators are regional representatives of INICC an boar members of local infection control societies who ai in recruiting an avising local hospitals. The INICC Heaquarters Team trains the researchers in each member hospital an avises them on an ongoing basis, answers queries an supports researchers, ajuicates the surveillance ata on HAIs, uploas the ata using specially esigne software, analyzes the ata, generates a monthly report for each hospital an forwars it electronically, an eit abstracts an manuscripts to be submitte to scientific meetings an peer-reviewe journals. The INICC Avisory Boar, international leaers in infection control an public health, serve in an avisory capacity. Two INICC Country Coorinators an INICC Avisory Boar meetings are hel annually. The ientity of all INICC hospitals, cities, an countries is confiential, in accorance with the INICC charter. From its inception, INICC has use the surveillance methos an efinitions for HAI evelope by the CDC s long-staning NNIS/NHSN program in US hospitals, 13,28 an has vigorously promote the consistent implementation of simple, inexpensive, high-priority evience-base measures for prevention of HAI. 7,16-18,20,27,29 The lack of knowlege regaring HAI worlwie, especially in the limite-resource countries, an at hospitals in evelope countries without enough experience in HAI surveillance an control, the nee for more precise measurements of HAI risks an outcomes in specific patient groups, an the basic importance of surveillance to a hospital program that can successfully reuce the risk of HAI le to the conceptualization, evelopment, an implementation of the 2 INICC surveillance components: outcome surveillance an process surveillance. INICC hospitals can esign their own surveillance programs by selecting the surveillance components or moules to use for the perio of time that they esire. This report escribes the current methos use by the INICC to collect an analyze surveillance ata an facilitate its use in feeback to iniviual hospitals to reuce the incience of HAI in their ICUs. CHARACTERISTICS OF PARTICIPATING HOSPITALS The hospitals participating in the INICC provie general meical-surgical inpatient services to ault, chilren, an newborns requiring acute care. They may be of any size an ownership, affiliate or unaffiliate with a meical school, an locate anywhere worlwie. Although participation is voluntary, hospitals must apply for membership in the INICC an have aequate personnel an support for infection control, as well as approval from hospital aministration to participate in the INICC (Fig 1). A total of 98 ICUs from 18 countries in Latin America, Asia, Africa, an Europe currently participate in the INICC. 2 The goal of the INICC is to achieve a membership with at least 5 countries per continent, 5 cities per country, an 1 hospital per city, which woul constitute a representative sample of the limite-resource countries an hospitals of the worl. Infection control practices Han hygiene compliance varies wiely in the ifferent countries an ICUs, ranging from 20% to 70%. 7,16-18,20 A recent stuy in participating INICC ICUs foun an overall 50% rate of han hygiene compliance, 30 similar to recent stuies in the Unite States an Europe. 31 Use of sterile ressings on central venous catheter (CVC) insertion sites also ranges wiely. 7,17 Open infusion systems (rigi or semirigi containers that must amit air to empty) rather than close infusion systems (flexible collapsing containers that o not amit air) or

5 Rosenthal, Maki, an Graves November 2008 e5 combinations of open an close systems are still use for aministration of intravenous fluis an meications in all of the member hospitals. 27 Laboratory techniques Ventilator-associate pneumonia. A eep tracheal aspirate from the enotracheal tube is obtaine for gram-stain, an aerobic culture or a bronchoscopic specimen is obtaine. CVC-associate bloostream infection. CVCs are remove aseptically, an the istal 5 cm of the catheter is amputate an culture, using the stanarize semiquantitative metho. 32 Concomitant bloo cultures are rawn percutaneously. Catheter-associate urinary tract infection. A urine sample is aseptically aspirate from the sampling port of the urinary catheter an culture quantitatively. As note earlier, in all hospitals, stanar laboratory methos are use to ientify microorganisms, an stanarize susceptibility testing is performe. 33 TRAINING OF THE PI AND ICP In Argentina, Brazil, China, Colombia, Costa Rica, Croatia, Inia, Malaysia, Mexico, Peru, an Turkey, the INICC Founer an Chairman, Dr Rosenthal, personally traine the PI an ICP in each member hospital. In Cuba, El Salvaor, Kosovo, Lebanon, the former Yugoslav Republic of Maceonia, Morocco, Nigeria, Pakistan, Philippines, an Uruguay, the institutional investigators were self-traine by a manual specifying how to conuct surveillance an complete surveillance forms. Some (eg, Chile) ha previous long-term training in HAI control an hospital epiemiology at the time their hospital joine the INICC. Institutional investigators have continuous telephone or access to a support team at the INICC Central Office in Buenos Aires, which respons to all inquiries within 24 hours. Queries an responses are reviewe by the Chairman (Fig 1). Definitions Clinically efine pneumonia. A patient with a chest raiograph that shows new or progressive infiltrates, consoliation, or cavitations, an at least 1 of the following: fever, leukocytosis, leucopenia, or altere mental status with no other cause in $ 70 years ol; an at least 1 of the following: new onset of purulent sputum, change in character of sputum, new onset or worsening cough, yspnea, tachypnea, rales, bronchial breath soun, or worsening gas exchange. 34 Laboratory-confirme bloostream infection. Criterion 1: Patient has a recognize pathogen culture from 1 or more bloo cultures, an organism culture from bloo is not relate to an infection at another site. Criterion 2: Patient has at least 1 of the following: fever, chills, or hypotension; positive laboratory results are not relate to an infection at another site; an common skin contaminant is culture from 2 or more bloo cultures rawn on separate occasions. 34 Clinical sepsis. A patient with a central line with at least 1 of the following: fever, hypotension, or oliguria; bloo cultures were either not obtaine or no organisms were recovere from bloo cultures; there is no apparent infection at another site; an the physician institutes antimicrobial therapy. 34 Symptomatic urinary tract infection. Criterion 1: Patient has at least 1 of the following with no other recognize cause: fever, urgency, frequency, ysuria, or suprapubic tenerness; an a positive urine culture, that is, $10 5 microorganisms per ml of urine with no more than 2 species. Criterion 2: Patient has at least 2 of the following with no other recognize cause: fever, urgency, frequency, ysuria, or suprapubic tenerness; an at least 1 of the following: positive ipstick for leukocyte esterase an/or nitrate; pyuria; organisms seen on Gram s stain of unspun urine; at least 2 urine cultures with repeate isolation of the same uropathogenwith $10 2 colonies/ml in nonvoie specimens; #10 5 colonies/ml of a single uropathogen in a patient being treate with an effective antimicrobial agent for a urinary tract infection; physician iagnosis of a urinary tract infection; physician institutes appropriate therapy for a urinary tract infection. 34 Crue excess mortality. The ifference between the crue overall case fatality of patients with a eviceassociate infection an the crue case fatality of patients hospitalize in the ICU uring that perio who i not acquire a evice-associate infection. INICC SURVEILLANCE COMPONENTS INICC s first efforts were focuse on surveillance an control of evice-associate infection in the ICU because it aresses the health care setting with the most vulnerable patients, who have the heaviest exposure to invasive evices an highest HAI rates. 35 Data are collecte from the following types of ICU: burn, surgical cariothoracic, meical, meical-surgical, peiatric, neurosurgical, surgical, trauma, an high-risk nursery. Two types of surveillance ata are collecte, outcome surveillance ata an process surveillance ata. The outcome surveillance component inclues the following moules: CLAB, VAP, an CAUTI rates per 1000 evice-ays in ault ICUs, 19,21-24,36,37 peiatric ICUs an newborn ICUs, microorganism profile, bacterial resistance, 19,21-24,36,37 antibiotic use, extra length of stay, extra costs, 4,5,8 extra mortality, 3,4,7,24,36 an risk factor analysis. The process surveillance component inclues compliance rates of the following four moules: han

6 e6 Vol. 36 No. 9 Rosenthal, Maki, an Graves hygiene, 16,18 vascular catheter care, 7,17 urinary catheter care, 29 an measures for prevention of VAP. 20 Iniviual surveillance components or moules may be use singly or simultaneously, but once selecte, must be carrie out for a minimum of 1 calenar month. Each hospital ecies which surveillance components or moule to use an for how long. All infections reporte to the INICC must have occurre in patients who were amitte to a hospital ICU an remaine in the unit for at least 24 hours. Infections are categorize by HAI sites using stanar CDC efinitions that inclue clinical an laboratory criteria. 34 Data are collecte on both infecte an uninfecte patients. Hospitals also have the option to collect aitional ata of special interest on infecte patients for their own use. Denominator ata inclue the number of patients (an total patient-ays) in the unit an number of ays of exposure to a central line, urinary catheter, an mechanical ventilator. Data reporte to the INICC must conform to the protocol of the selecte surveillance component or moule before they are entere into the central INICC international atabase. Outcome surveillance forms The INICC s surveillance forms are esigne to gather ata from all patients in the ICU, both those with an those without HAIs to continuously prompt the ICP an PI to suspect HAIs, because the form provies a continuous picture of every patient s course in the ICU: aily ata regaring the patient s maximum temperature, lowest bloo pressure, exposure to invasive evices, cultures one, imaging stuies, an antibiotic use. Outcome surveillance forms shoul be requeste by to Dr Rosenthal to victor_rosenthal@inicc.org Severity of illness scores, APACHE II, an Average Severity Illness Score (ASIS) 28 are recore for each patient at ICU amission. The ASIS is recore using the CDC NNIS/NHSN criteria. Points are totale, with 1 point for surgical patients requiring routine postoperative observation only, 2 points for physiologically stable nonsurgical patients requiring overnight observation, 3 points for patients neeing continuous nursing an monitoring, 4 points for physiologically unstable patients requiring intensive nursing an meical care, with the nee for frequent reassessment an ajustment of therapy, an 5 points for physiologically unstable patients in coma or shock who require cariopulmonary resuscitation or intensive meical an nursing care with frequent reassessment. Central ajuication of each reporte HAI an reporting Each HAI reporte by a hospital is ajuicate (ie, scrutinize to be certain that criteria are fulfille to justify its recoring as a HAI); the ajuication process also inclues the scrutiny of ata reporte for putatively uninfecte patients to permit etection of unreporte but true HAIs. When iscrepancies are encountere, the INICC hospital team is contacte by by the INICC heaquarters team to resolve the ifference; the jugment of the PI an ICP of the participant hospital is final. Ajuication is a unique feature of INICC outcome surveillance component an is consiere essential for maximizing the accuracy of surveillance ata. Also essential is to assess on an ongoing basis the capacity of the ICP an PI at each hospital to accurately ientify HAIs by comparing the iscrepancies between the HAIs reporte by the hospital team an those ientifie by the INICC heaquarters team after reviewing the surveillance work sheets. The INICC heaquarters team prepares an sens to each participating hospital a final monthly report on their institutional rates of evice-associate infection, microorganism profile, bacterial resistance, LOS, an mortality in their ICUs, an rates of compliance with han hygiene, CVC an urinary catheter care, an measures to prevent pneumonia. OUTCOME SURVEILLANCE COMPONENT Outcome surveillance moule: HAIs per 1000 evice-ays in ault an peiatric ICUs The outcome surveillance component focuses on patients hospitalize in ault an peiatric ICUs. Data are prospectively gathere uring the stuy perio from all patients whose ICU LOS excees 24 hours. A HAI is an infection that was not present or incubating at the time of the patient s amission to the ICU but became apparent uring the ICU stay or within 48 hours after transfer from the ICU. 34 Patients are followe for 48 hours after ischarge from the ICUs to etect infections acquire in the ICU but manifesting only after transfer to a non-icu patient care unit. The ICP at each INICC hospital is responsible for extracting patientsõ ata prospectively from meical recor, charts, patient inspection, an laboratory results, incluing raiographs an all cultures one. Data collection sheets are checke by the PI to confirm each HAI iagnosis. Hospitals with more than 1 ICU may carry out surveillance in any or all ICUs, but in the selecte units, every patient is monitore for CLAB, CAUTI, an VAP. The enominator ata collecte are the total number of patients in the ICU uring the month, total number of patient-ays, urinary catheter-ays, central line-ays, an ventilator-ays. Calculation of site-specific infection rates is base on the appropriate enominator (eg,

7 Rosenthal, Maki, an Graves November 2008 e7 number of CAUTIs ivie by the total number of inwelling urinary catheter-ays). 28 On amission, every patient in the unit is assesse by irect observation of surveillance personnel, who assign an APACHE II score an ASIS score, an iniviual patient scores are combine an average into a monthly severity of illness score for the unit. To provie feeback to ICU staff in the unit, the IN- ICC heaquarters team sen charts on a monthly basis to the ICP an PI, proviing a running recor of rates of evice-associate infections compile by the ICP, which are then reviewe at monthly staff meetings an poste in a prominent location in the ICU. Outcome surveillance moule: HAIs in highrisk nurseries The high-risk nursery (HRN) surveillance component provies the option of surveillance in level III nurseries, which are efine as nurseries that provie multisystem support or critical care for unstable neonates an are staffe by a full-time peiatrician with special qualifications in neonatal meicine an by nurses specially traine in perinatal care. In the HRN component, all neonates in the level III nursery are monitore for infection. An HRN-associate infection is one that was not present or incubating at the time of the neonate s amission into the HRN but that became apparent uring the HRN stay or within 48 hours after transfer from the HRN. Very early neonatal infections thought to have been acquire uring parturition are inclue as HAIs, whereas those consiere most likely to have been acquire in utero are not. 34,38 Denominator ata are stratifie for each of the following 5 birth weight categories: # 1000 g, 1001 to 1500 g, 1501 to 2500 g, an g, an inclue the total number of patients in the HRN uring the month, total number of patient-ays, umbilical catheter/central line-ays, an ventilator-ays). Site-specific infection rates are calculate per 1000 evice-ays by the INICC heaquarters team. Outcome surveillance moule: Microbiological profile of HAI an bacterial resistance The INICC inclues surveillance of the microbiological profile of HAI in the participating ICUs that are confirme microbiologically in CLABs, CAUTIs, an VAPs, an antimicrobial susceptibilities are recore on a esignate form. 18,20-23,35,36 Outcome surveillance moule: Extra LOS an evaluation of HAI costs The ICU LOS is recore for each infecte an uninfecte patient, an the timing of the onset of infection is recore. To ate, the effect of HAI on LOS has been estimate by matching patients in the same ICU uring the surveillance perio by age, sex, ASIS score, an other variables. Differences in LOS have been attribute to HAIs. 4,5,8 This metho is wiely use but has some weaknesses. 39 Many factors are associate with ICU/hospital LOS. Matching on more than 7 factors exclues infecte patients for whom no match can be foun, which will inuce a selection bias. Matching on 6 factors, or even fewer, is unlikely to control much of the variation among LOS outcomes, inucing another source of bias. 39 The INICC is currently eveloping statistical moels of LOS that mitigate these problems an provie better estimates. Three issues are being aresse: 1. Nonnormal istribution of LOS: LOS ata are not normally istribute with a small number of observations emonstrating very long ICU/hospital LOS. 2. Feeback effects: Infection is associate with LOS, yet increase LOS increases the risk of HAIs Timing of events is important; efining HAI as a time-epenent covariate is important for moels that preict LOS in hospital. 41,42 Vali estimates of the excess LOS ue to HAI are powerful ata. They can be use to show the number of be-ays that will be release by preventing HAIs. Preventing HAIs will not save much money, because most of the costs of running a hospital cannot be easily avoie in the short term (ie, they are fixe costs). 43 Instea, the cost savings from infection control are the value of the be-ays release to ecision makers. How much ecision makers are willing to pay to access more be-ays will vary by country. In health care systems where this is excess eman for hospital services, this value is likely to be positive. INICC is committe to using rigorous economic methos 43 to estimate the changes to costs from preventing HAIs. Outcome surveillance moule: Excess mortality In the INICC, hospital mortality is also recore for each patient. The crue excess mortality is efine as the ifference between the overall case-fatality of patients hospitalize in the ICU uring the surveillance perio with a HAI an the case-fatality of patients hospitalize in the ICU uring that perio who i not acquire a HAI. To ate, excess mortality has been estimate using a matching proceure. 3,4,7,24,36 Like cost outcomes, the process of estimating the inepenent effect of infection on mortality is complex an is aresse in ongoing INICC research activities. Decision making an infection control If infection control can be shown to reuce costs an improve health outcomes, then a failure to implement infection control is unethical. Excess costs are incurre

8 e8 Vol. 36 No. 9 Rosenthal, Maki, an Graves while simultaneously harming patients. Aopting infection control is an economic win win situation, as costs are save an patient welfare increases. Infection control still can be cost-effective if costs increase overall (ie, infection control oes not pay for itself), an health benefits increase. The ratio of these numbers (ie, the cost per unit of health outcome) must fall below the ecision makersõ threshol. Builing these arguments is complex, an only goo-quality ata an moeling methos shoul be use. 44 PROCESS SURVEILLANCE COMPONENT Process surveillance is an essential feature of the IN- ICC, esigne to monitor compliance with important, easily measurable control measures, such as han hygiene, vascular catheter care, urinary catheter care, an measures to prevent VAP. Process surveillance forms shoul be requeste from Dr Rosenthal by at victor_rosenthal@inicc.org Process surveillance moule: Han hygiene compliance Han hygiene compliance by HCWs is monitore by the ICP through ranomly selecte 1-hour observations 3 times a week, uring all working shifts an incluing all HCWs accoring to a specific sequence set forth in the INICC protocol. 16,18 The ICP is well-known to the ICU staff, an, although the HCWs are aware that han hygiene practices will be monitore, they are not informe when these observations are taking place. The ICP recors the opportunities for han hygiene an compliance before contact with each patient; han hygiene process surveillance ata are recore on a esignate form. The INICC recore more than 85,000 opportunities for han hygiene in 16 countries over a 10-year perio an foun 50% compliance. 2,30 Process surveillance moule: Vascular catheter care Vascular catheter care compliance also is monitore, an the following ata are recore on a stanarize form 5 ays a week: han hygiene before an after catheter insertion or reressing an intravascular catheter; presence of a sterile gauze or polyurethane ressing on insertion sites; the ate of catheter insertion an the last aministration set change; replacement of the gauze ressing every 48 hours an transparent semipermeable membrane ressings at least every 7 ays, with ate an time of ressing changes recore; replacement of peripheral intravenous catheters within 72 to 96 hours; aministration change set replacements at least every 72 hours; an others. 7,17 Process surveillance moule: Urinary catheter care Urinary catheter care compliance also is monitore, an the following ata are recore on a stanarize form: silicone catheter, close catheter rainage, unobstructe catheter position above (not below) the leg, urine collecting bag below the level of the blaer, no contact of the collection bag with the floor are recore 5 ays a week, an others. 29 Process surveillance moule: Mechanical ventilator care Mechanical ventilator care compliance also is monitore 5 ays a week: noninvasive ventilation, if feasible, orotracheal rather than nasotracheal tube, suction port above the enotracheal cuff, elevation of hea of the patient s be 30 to 45 egrees (unless contrainicate meically), heat-moisture exchanger humiification, absence of poole flui within the ventilator tubing, absence of obstruction by mucous, presence of a water trap, nonturbi flui in the humiifier reservoir, close-system enotracheal suctioning, oral care at least aily, absence of poole pharyngeal secretions, an others. 20 Process surveillance monitoring performance feeback an outcome surveillance feeback The INICC heaquarters team prepares an sens to each INICC participant hospital monthly reports, showing bar charts with global rates per 100 patients an per 1000 be-ays, HAIs per 1000 evice-ays (CLABs per 1000 central line-ays, CAUTIs per 1000 catheterays, an VAPs per 1000 ventilator-ays), microbiological profile, bacterial resistance, extra mortality by type of HAI, extra LOS, han hygiene compliance, central line an urinary catheter care compliance, an measures to prevent pneumonia to post them in the ICU in a prominent location, to provie feeback to the ICU HCWs. The ata also are reviewe at monthly meetings of ICU staff (Fig 2). 16,18,30 STATISTICAL METHODS INICC uses EpiInfo version 6.04b an SAS software to analyze hospital ata. Device utilization rates are calculate by iviing the total number of evice-ays by the total number of ICU patient-ays. Rates of VAP, CLAB, an CAUTI per 1000 evice-ays are calculate by iviing the total number of HAIs by the total number of specific evice-ays an multiplying the result by Differences among treatment groups are analyze using the x 2 test or, when appropriate, Fisher s exact test for ichotomous variables, an Stuent s

9 Rosenthal, Maki, an Graves November 2008 e9 Fig 2. INICC outcome surveillance report to a member hospital on CLAB over a 16-month perio. t-test for continuous variables. Relative risk ratios, 95% confience intervals, an P values are etermine for primary an seconary outcomes. The INICC also is using survival analyses, competing-risks moels, an multistate moels. 41 DISCUSSION Surveillance of HAIs, especially in high-risk hospital settings such as the ICU, 13 has become an integral feature of infection control an quality assurance in evelope countries hospitals since the capacity of surveillance to reuce the risks of HAIs was emonstrate in the CDC s SENIC Stuy more than 30 years ago. 11 Stanars for institutional surveillance have been aopte in the Unite States, 13 the Unite Kingom, 45 Australia, 46 Germany, 12 an others. There is a vast boy of literature showing that HAIs are a major cause of patient morbiity an mortality in the evelope countries, 9 an evice-associate infections, particularly VAP, 47 CLAB, 48 an CAUTI, 10 pose the greatest threat to hospital safety in the ICU. 49 Surveillance of evice-associate HAIs has been stanarize by the CDC s NNIS/NHSN stuy by proviing simple, unambiguous efinitions. 34,38 Targete surveillance an calculation of evice-associate infection rates per 1000 evice-ays allows benchmarking with other similar hospitals an etection of unique institutional problems in nee of reress. As note earlier, most of the publishe stuies of ICU-acquire infection come from hospitals in the inustrialize Western countries, 15 an far less ata have been reporte from limite-resource countries, especially rates of ICU evice-associate infections using stanarize efinitions. 3-8,16-24 We reporte the initial finings of INICC surveillance from 2002 through 2006, pointing up very high rates of HAI in the ICUs of limite-resource countries. 1,2 Since the inception of organize programs for control of HAIs in the Unite States in the late 1960s, surveillance has been avocate for recognizing HAI problems an for targeting preventive measures. 50 Its efficacy in helping to reuce HAI has been reporte by multiple investigators, 51 an surveillance has been promote by the American Hospital Association, the Joint Commission, an the Health Care Financing Aministration. 52,53 Although the Joint Commission establishe infection surveillance as a responsibility of the meical staff in 1964, a stanarize approach to conucting surveillance generally was not available until 1969, when the CDC reporte its first surveillance stuy of HAIs. 54,55 The INICC s surveillance form is esigne to collect ata from all patients, both those with an those without HAI (Fig 2). The CDC s NNIS/NHSN program in US hospitals an surveillance systems use in other countries collect ata only from patients with infections acquire in the ICU. 12,13,28 In contrast, the form use by the INICC is specifically esigne to continuously prompt the surveillance officer to suspect HAI, because the form provies a panoramic view of what is happening each ay to every patient in the ICU in terms of vital signs, exposure to invasive evices, culture results, an antibiotic therapy. This approach is especially useful in cases in which no cultures have been one or the culture results are equivocal or negative, such as with pneumonia or sepsis, an that may not be otherwise recognize as a HAI. 1,19,21-24,36,37 Furthermore, by collecting ata on all patients in the ICU, it is possible to easily match patients with an without HAI for such features as age, sex, unerlying iseases, service, amission iagnosis, severity of illness score, time of year, exposure to specific invasive evices, an several others to calculate ae LOS an costs of hospitalization, attributable mortality, an risk factors for infection. 3-5,8,24,36 We believe that the INICC methoology further improves the accuracy of surveillance because

10 e10 Vol. 36 No. 9 Rosenthal, Maki, an Graves each reporte infection is ajuicate externally; however, the vast majority of ICU-acquire infections in both the CDC NNIS/NHSN system an the INICC are base on positive cultures, an we oubt whether the 2 surveillance systems iffer materially in their sensitivity for etecting evice-associate infections, except perhaps for VAP or clinical sepsis. The INICC s surveillance system has some limitations. First, we o not yet consier the ata to be aequate to represent any entire single country; however, with ata now being collecte in 98 ICUs in 18 limite-resource countries, we believe that our finings are becoming representative of the eveloping worl an most likely unerestimate the magnitue of the problem, because we believe that the INICC participants generally represent the best hospitals in their countries, hospitals with the most resources an commitment to patient safety in terms of controlling HAIs. Secon, we must rely on the member hospitalsõ laboratories to reliably ientify infecting pathogens an elineate bacterial resistance patterns. Different laboratories have varying levels of expertise an resource availability; however, similar concerns can be raise about any multi-institutional surveillance program or stuy. Finally, the frequency of culturing an the use of other iagnostic tests are beyon the control of infection control programs; in hospitals where culturing an other laboratory testing is infrequent an suspecte infections are treate empirically, the capacity of the surveillance program to etect most nosocomial infections is likely to be low. The limitations of INICC ata nee to be consiere in such eneavors. Nonetheless, 10 years of INICC surveillance have been valuable for conucting outcome an process surveillance, with the results use in feeback an eucation to prevent HAIs. Surveillance of HAIs efining the magnitue an nature of the problem is the first step towar reucing the risk of infection in vulnerable hospitalize patients. The next step is to implement more consistently essential infection control practices that have been shown to prevent HAIs We are confient that knowlege of the magnitue of the problem of evice-associate infections in the INICC ICUs will continue to provie a powerful impetus for instituting neee change, an we have alreay seen ample evience of prouctive change; process surveillance, targete performance feeback programs for han hygiene an central venous catheter, ventilator, an urinary catheter care have alreay translate to ocumentation of major reuctions in the incience of ICU-acquire infections in iniviual member hospitals, 7,16-18,20 an we have recently ocumente a highly significant (46%) reuction in CLABs (P,.001) in 73 INICC ICUs uring their first 8 months as members of the Consortium. 62 INICC ata are now being use by national health care planners in the member countries to evelop strategies an target resources for control of HAIs. The INICC s goals for the future inclue: 1. Strengthening infection control activities in the member hospitals, eliminating practices of ubious value an aing more key evience-base control measures Expaning outcome an process surveillance to inclue surgical site infections an key process measures that govern the efficacy of surgical antimicrobial prophylaxis Incorporating biohazarous exposures an their management into outcome an process surveillance 4. Aressing more restrictive an improve antibiotic utilization Unertaking multicenter comparative trials of promising control measures an novel inexpensive technologies for prevention of evice-associate HAIs 6. Developing an online atabase for uploaing an retrieving surveillance ata. Clearly, HAI is a huge an largely unrecognize threat to patient safety in the hospitals of the eveloping worl, a far greater threat than in the evelope countries, we believe, rivaling the huge buren of iarrhea of chilhoo, tuberculosis, an malaria. It is our hope that the initial successes of the INICC, combine with our ongoing efforts to more consistently implement simple, inexpensive measures for prevention, will lea to wier acceptance of infection control practices an continue reuctions in evice-associate infections, not only in the hospitals of the consortium, but also in their innumerable neighboring hospitals as well. The authors thank the many health care professionals at each member hospital who assiste with the conuct of surveillance in their respective facilities, incluing the surveillance nurses, hospital epiemiologists, clinical microbiology laboratory personnel, an the physicians an nurses proviing care for the patients uring the stuy, without whose cooperation an generous assistance the INICC woul not be possible. The authors also thank the INICC country coorinators (Altaf Ahme, Carlos A. Alvarez-Moreno, Luis E. Cuellar, Euaro A. Meeiros, Bijie Hu, Hakan Leblebicioglu, Ajita P. Mehta, Lul Raka, an Toshihiro Mitsua) an the INICC Avisory Boar (Carla J. Alvarao, Martin S. Favero, Gary L. French, Nicholas Graves, William R. Jarvis, Elaine Larson, Patricia Lynch, Dennis Maki, Russell N Olmste, Diier Pittet, an Wing Hong Seto), who have so generously supporte this unique international infection control network. Special thanks are ue to Patricia Lynch, who has inspire an supporte our vision espite obstacles. References 1. Rosenthal VD, Maki DG, Salomao R, Moreno CA, Mehta Y, Higuera F, et al. Device-associate nosocomial infections in 55 intensive care units of 8 eveloping countries. Ann Intern Me 2006;145: Rosenthal VD, Maki DG, Mehta A, Alvarez-Moreno C, Leblebicioglu H, Higuera F, et al. International Nosocomial Infection Control Consortium (INICC) report, ata summary for , issue January Am J Infect Control 2008; in press.

11 Rosenthal, Maki, an Graves November 2008 e11 3. Rosenthal VD, Guzman S, Orellano PW. Nosocomial infections in meical-surgical intensive care units in Argentina: attributable mortality an length of stay. Am J Infect Control 2003;31: Rosenthal VD, Guzman S, Migone O, Crnich CJ. The attributable cost, length of hospital stay, an mortality of central line-associate bloostream infection in intensive care epartments in Argentina: a prospective, matche analysis. Am J Infect Control 2003;31: Rosenthal VD, Guzman S, Migone O, Safar N. The attributable cost an length of hospital stay because of nosocomial pneumonia in intensive care units in 3 hospitals in Argentina: a prospective, matche analysis. Am J Infect Control 2005;33: Moreno CA, Rosenthal VD, Olarte N, Gomez WV, Sussmann O, Aguelo JG, et al. Device-associate infection rate an mortality in intensive care units of 9 Colombian hospitals: finings of the International Nosocomial Infection Control Consortium. Infect Control Hosp Epiemiol 2006;27: Higuera F, Rosenthal VD, Duarte P, Ruiz J, Franco G, Safar N. The effect of process control on the incience of central venous catheter associate bloostream infections an mortality in intensive care units in Mexico. Crit Care Me 2005;33: Higuera F, Rangel-Frausto MS, Rosenthal VD, Soto JM, Castanon J, Franco G, et al. Attributable cost an length of stay for patients with central venous catheter associate bloostream infection in Mexico City intensive care units: a prospective, matche analysis. Infect Control Hosp Epiemiol 2007;28: Jarvis WR. Selecte aspects of the socioeconomic impact of nosocomial infections: morbiity, mortality, cost, an prevention. Infect Control Hosp Epiemiol 1996;17: Tambyah PA, Knasinski V, Maki DG. The irect costs of nosocomial catheter-associate urinary tract infection in the era of manage care. Infect Control Hosp Epiemiol 2002;23: Haley RW, Quae D, Freeman HE, Bennett JV. The stuy on the efficacy of nosocomial infection control (SENIC) project: summary of stuy esign. Am J Epiemiol 1980;111: Gastmeier P, Geffers C, Brant C, Zuschnei I, Sohr D, Schwab F, et al. Effectiveness of a nationwie nosocomial infection surveillance system for reucing nosocomial infections. J Hosp Infect 2006;64: Ewars JR, Peterson KD, Anrus ML, Tolson JS, Gouling JS, Dueck MA, et al. National Healthcare Safety Network (NHSN) report, ata summary for 2006, issue June Am J Infect Control 2007;35: Vincent JL, Bihari DJ, Suter PM, Bruining HA, White J, Nicolas- Chanoin MH, et al. The prevalence of nosocomial infection in intensive care units in Europe. Results of the European Prevalence of Infection in Intensive Care (EPIC) Stuy. EPIC International Avisory Committee. JAMA 1995;274: Safar N, Crnich CJ, Maki DG. Nosocomial Infections in the intensive care unit associate with invasive meical evices. Curr Infect Dis Rep 2001;3: Rosenthal VD, McCormick RD, Guzman S, Villamayor C, Orellano PW. Effect of eucation an performance feeback on hanwashing: the benefit of aministrative support in Argentinean hospitals. Am J Infect Control 2003;31: Rosenthal VD, Guzman S, Pezzotto SM, Crnich CJ. Effect of an infection control program using eucation an performance feeback on rates of intravascular evice associate bloostream infections in intensive care units in Argentina. Am J Infect Control 2003;31: Rosenthal VD, Guzman S, Safar N. Reuction in nosocomial infection with improve han hygiene in intensive care units of a tertiary care hospital in Argentina. Am J Infect Control 2005;33: Ramirez Barba EJ, Rosenthal VD, Higuera F, Oropeza MS, Hernanez HT, Lopez MS, et al. Device-associate nosocomial infection rates in intensive care units in four Mexican public hospitals. Am J Infect Control 2006;34: Rosenthal VD, Guzman S, Crnich C. Impact of an infection control program on rates of ventilator-associate pneumonia in intensive care units in 2 Argentinean hospitals. Am J Infect Control 2006;34: Leblebicioglu H, Rosenthal VD, Arikan OA, Ozgultekin A, Yalcin AN, Koksal I, et al. Device-associate hospital-acquire infection rates in Turkish intensive care units: finings of the International Nosocomial Infection Control Consortium (INICC). J Hosp Infect 2007;65: Mehta A, Rosenthal VD, Mehta Y, Chakravarthy M, Toi SK, Sen N, et al. Device-associate nosocomial infection rates in intensive care units of seven Inian cities: finings of the International Nosocomial Infection Control Consortium (INICC). J Hosp Infect 2007;67: Salomao R, Rosenthal VD, Grinberg G, Nouer S, Blecher S, Buchner Ferreira SI, et al. Device-associate infections rates in critical patients of brazilian hospitals: International Nosocomial Infection Control Consortium (INICC) finings. Pan Am J Public Health 2008; in press. 24. Cuellar L, Fernánez Malonao E, Rosenthal VD, Castañea Sabogal A, Rosales R, Mayorga Espichan MJ, et al. Device-associate infections rates an mortality in intensive care units of Peruvian hospitals: International Nosocomial Infection Control Consortium (INICC) finings. Pan Am J Public Health 2008; in press. 25. Chanra PN, Milin K. Lapses in measures recommene for preventing hospital-acquire infection. J Hosp Infect 2001;47: Archibal LK, Manning ML, Bell LM, Banerjee S, Jarvis WR. Patient ensity, nurse-to-patient ratio an nosocomial infection risk in a peiatric cariac intensive care unit. Peiatr Infect Dis J 1997;16: Rosenthal VD, Maki DG. Prospective stuy of the impact of open an close infusion systems on rates of central venous catheter associate bacteremia. Am J Infect Control 2004;32: Emori TG, Culver DH, Horan TC, Jarvis WR, White JW, Olson DR, et al. National nosocomial infections surveillance system (NNIS): escription of surveillance methos. Am J Infect Control 1991;19: Rosenthal VD, Guzman S, Safar N. Effect of eucation an performance feeback on rates of catheter-associate urinary tract infection in intensive care units in Argentina. Infect Control Hosp Epiemiol 2004;25: Rosenthal VD, Salomao R, Leblebicioglu H, Akan O, Sobreyra- Oropeza M. Han hygiene compliance in Argentina, Brazil, Colombia, Inia, Mexico, Morocco, Peru an Turkey: finings of the International Nosocomial Infection Control Consortium (INICC). Proceeings an abstracts of the 33r Annual Scientific Meeting of the Association for Professionals in Infection Control an Epiemiology, June 11-15, p Wisniewski MF, Kim S, Trick WE, Welbel SF, Weinstein RA. Effect of eucation on han hygiene beliefs an practices: a 5-year program. Infect Control Hosp Epiemiol 2007;28: Maki DG, Weise CE, Sarafin HW. A semiquantitative culture metho for ientifying intravenous catheter relate infection. N Engl J Me 1977;296: Villanova P. Minimum inhibitory concentration interpretive stanars M7-A4. Wayne (PA): National Committee for Clinical Laboratory Stanars Horan TC, Anrus M, Dueck MA. CDC/NHSN Surveillance efinition of health care associate infection an criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36: Maki DG, Kluger DM, Crnich CJ. The risk of bloostream infection in aults with ifferent intravascular evices: a systematic review of 200 publishe prospective stuies. Mayo Clin Proc 2006;81: Moreno CA, Rosenthal VD, Olarte N, Gomez WV, Sussmann O, Aguelo JG, et al. Device-associate infection rate an mortality in intensive care units of 9 Colombian hospitals: finings of the International Nosocomial Infection Control Consortium. Infect Control Hosp Epiemiol 2006;27: