Risk factors for healthcare-associated urinary tract infection and their ...

Journal of Hospital Infection 82 (2012) 219e226Available online at www.sciencedirect.comJournal of Hospital Infectionjournal homepage: www.elsevierhealth.com/journals/jhinReviewRisk factors for healthcare-associated urinary tractinfection and their applications in surveillance usinghospital administrative data: a systematic reviewC. King a, *, L. Garcia Alvarez a , A. Holmes a, b , L. Moore c , T. Galletly b , P. Aylin a, da The National Centre for Infection Prevention and Management, Imperial College, London, UKb Infection Prevention and Control, Imperial College Healthcare NHS Trust, London, UKc Department of Infectious Diseases, Imperial College Healthcare NHS Trust, London, UKd Dr Foster Unit at Imperial, School of Public Health, Imperial College London, London, UKA R T I C L EI N F OS U M M A R YArticle history:Received 12 January 2012Accepted 4 May 2012Available online 15 June 2012Keywords:Urinary tract infectionAdministrative dataRisk factorsHealthcare associatedSurveillanceUrinary catheterizationBackground: Healthcare-associated urinary tract infections (HCA UTI) account for a largeproportion of hospital infections, with recently launched surveillance in the UK focusing onreducing catheter-associated urinary tract infections. However, a wealth of administrativeinformation already collected routinely by hospitals is currently not used to its maximumpotential for surveillance.Aim: To quantify the evidence base of HCA UTI risk factors and to determine theirpotential for shaping and informing innovative surveillance tools using local hospital data.Methods: A systematic literature review was undertaken to find established risks for HCAUTI. Population-attributable risk percentages (PAR%) were calculated for these riskfactors, generating a hierarchy of risks. Administrative hospital data were subsequentlyinterrogated for these quantified risks.Findings: Over 30% of the risk factors identified from the systematic literature review wereindependent predictors of infection. The highest PAR% was associated with urinary catheterization,with the calculation that 79.3% of UTI would be prevented if catheterization wasnot performed. PAR% calculations were performed for 60% of the independent predictors forHCA UTI. Sixty-five percent of the identified independent risk factors were found to becoded within the administrative hospital dataset, including urinary catheterization.Conclusion: This work has quantified established HCA UTI risks and demonstrates thatthere is potential for more effective use of administrative hospital data for risk monitoringand surveillance of HCA UTI.Ó 2012 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.Introduction* Corresponding author. Address: Imperial College, Charing CrossCampus, 3rd Floor, The Reynolds Building, St Dunstan’s Road, LondonW6 8RP, UK. Tel.: þ44 (0) 2075940779; fax: þ44 (0) 207594 0854.E-mail address: c.king09@imperial.ac.uk (C. King).Healthcare-associated urinary tract infections (HCA UTI)contribute a significant burden to healthcare services globally. 1In the UK, it is estimated that the National Health Service (NHS)incurs 66,160 cases of HCA UTI per year, costing £68 million inexcess bed-days. 2 In the 2006 UK and Republic of Irelandprevalence survey of healthcare-associated infections (HCAI),0195-6701/$ e see front matter Ó 2012 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.jhin.2012.05.004

220UTI accounted for 19.9% of HCAI, second in frequency only togastrointestinal infections (20.6%). 3 The high rate of UTI withinhospitals is commonly attributed to the use of urinary catheters,with 31.8% of the NHS patient population having undergoneurinary catheterization or bladder instrumentation on theday of the survey or in the preceding seven days. 4Traditional surveillance has long been established as anessential component to any infection prevention and controlprogramme, 5 despite the fact that it can be costly to implementand time consuming to conduct. 6 In the UK, there aremandatory surveillance systems for several HCAI, such asClostridium difficile (since January 2004) and meticillinresistantStaphylococcus aureus (MRSA) bloodstream infections(since early 2001). 7,8 Surveillance of catheter-associatedurinary tract infections (CA UTI) is now performed as part of aninitiative launched by the UK Department of Health in January2011 in order to reduce patient harm. 9 However, despite thehigh burden that HCA UTI represents to the NHS, this surveillancescheme is voluntary, focuses exclusively on CA UTI, and isstill in its infancy. The potential utility of innovative HCA UTIsurveillance tools is apparent, with current work focusing onthe development of automated systems to identify UTI usingtraditional clinical definitions. 10Hospitals routinely collect and store large amounts ofadministrative data, including information on admissions,microbiology, diagnoses and procedures. 11 Although thiswealth of data has not been widely exploited for HCAIsurveillance, integrative approaches towards the more effectiveuse of administrative data are being pioneered. 11e13Determining the utility of established risk factors for HCA UTIwithin administrative hospital data presents the opportunity todevelop and test innovative surveillance tools with thepotential to enhance traditional surveillance of UTI within thehealthcare environment.The aim of this study was to quantify the existing evidencebase of risk factors for HCA UTI, create a hierarchy of risk,identify these risks within administrative hospital data, andassess the potential of this data source for informing novelsyndromic surveillance tools.MethodsSystematic literature reviewA systematic literature review was conducted to establishan evidence base of risk factors for HCA UTI. EMBASE, PubMed/MEDLINE, Web of Science and the Cochrane Database weresearched using ‘UTI’, ‘urinary tract infection’ and ‘risk’ as thesearch terms. The Cochrane Database was included to assesswhether or not a comprehensive literature review on HCA UTIrisk factors had been undertaken previously. The searchstrategy and inclusion criteria were discussed and agreedbetween the authors, and the search was conducted by a singlereviewer in April 2011.The inclusion criteria were: (1) original research papers inEnglish; (2) clear and reasonable study design; (3) demonstratedstatistical analysis, and reported odds ratios (OR), relative risksor hazard ratios, and P-values; (4) a clear definition of HCA UTIused to diagnose ‘cases’, either based on published definitions(e.g. US Centers for Disease Control and Prevention) or positivemicrobiology, antibiotic prescription and clinical data; (5)C. King et al. / Journal of Hospital Infection 82 (2012) 219e226‘controls’ clearly stated as not diagnosed with an HCA UTI; and(6) all study participants had been exposed to ‘health care’,defined as either a hospital admission for 72 h or a readmissionwith prior exposure to health care. Those studiesinvestigating risk factors for acquiring an antimicrobial-resistantinfection vs an antimicrobial-susceptible infection wereexcluded as the controls did not meet the inclusion criteria.Critical Appraisal Skills Programme (CASP) tools, developed bythe British Public Health Resource Unit, for caseecontrol(available at: http://www.sph.nhs.uk/sph-files/casp-appraisaltools/Case%20Control%2011%20Questions.pdf/view)14 andcohort studies (available at: http://www.sph.nhs.uk/sph-files/casp-appraisal-tools/cohort%2012%20questions.pdf/view) 15were employed for critical appraisal of the papers in thisstudy as they were developed specifically for cohort andcaseecontrol studies, and have been widely used in reviewarticles. 16 For those papers describing prevalence and surveillancestudies, an ‘adjusted CASP cohort tool’ was applied,which included all the aspects considered by the CASP cohorttool 15 with the exception of the follow-up criteria. Themaximum and minimum possible CASP scores ranged from14 to 14 for cohort studies, and from 12 to 12 forcaseecontrol, prevalence and surveillance studies.Those risk factors that were associated with HCA UTI(P < 0.05) at the univariate and multivariate level wereselected for further analysis, and were defined as ‘significant’risks. An ‘independent’ risk factor was considered when itsassociated P-value was

C. King et al. / Journal of Hospital Infection 82 (2012) 219e226 221review in order to address the utility of administrative hospitaldata for the development of surveillance tools.Imperial College Healthcare NHS Trust (ICHNT) is one of thelargest NHS hospital groups in the UK, consisting of five hospitalswith 1540 beds and one central laboratory. Over 750 localdatabases have been identified within ICHNT, 11 comprisinga mixture of small departmental datasets, surveillance andlarger trust-wide systems. Data on patient demographics,diagnoses [International Classification of Diseases Version 10(ICD-10)], 19 operations and procedures (OPCS), and admissionand discharge data are stored in the large organization-widepatient administration system (PAS). Microbiology and laboratorydata are located in the laboratory information managementsystem (LIMS), and further information on antimicrobialprescribing and infection prevention and control is located inlocally performed surveillance.A comprehensive variable list and code book of the hospitaldata available for this study was created and cross-checkedagainst the list of extracted risk factors for UTI. Keywordsearches of the ICD-10 codes were performed, based on the listof HCA UTI risk factors identified during the literature review,using the World Health Organization’s online search tool(http://apps.who.int/classifications/icd10). A Microsoft Excel(2007) reference spreadsheet was generated. ICD-10 codesidentified from any included studies were also recorded on thereference spreadsheet. The process was repeated with operationand procedure codes and pathology codes using referencedatabases provided by ICHNT.In order to use hospital data for research, ethical approvalwas granted by St. Mary’s National Research Ethics Committeefor the use of administrative hospital data from ICHNT in ananonymized format, i.e. one way anonymization and removalof all patient-identifiable fields prior to data analysis (Ref. No.:09/H0712/85).ResultsThe systematic literature search returned 4109 uniquepapers. One percent were excluded for not being in English,89% had titles that were not relevant to the study question, and6% were excluded as irrelevant after abstract review. Theresults retrieved from the Cochrane Database returned norelevant studies. In total, 154 studies underwent full-textreview and 23 were ultimately included in the study(Figure 1). Of the 23 papers included, three papers werecaseecontrol studies, 20e22 13 papers were cohort studies, 23e35six papers were surveillance and prevalence studies, 4,36e40 andone paper used data derived from a large randomizedcontrolled trial database (Table I). 41The majority of studies were performed in high-incomesettings, of which two were based in the UK. 4,22 Threestudies were undertaken in the middle-income countries ofBrazil, 24 India 23 and Egypt. 39 Two studies 23,34 involved paediatricpopulations, with the remaining 21 papers addressingadult populations, including patients from generalwards, 4,20,22,26,28,36,37,39,41 patients in intensive care units(ICUs), 29e31,33,38,40 surgical patients, 35 stroke patients, 32orthopaedic patients, 27 hip fracture admissions, 21 spinalinjury patients 25 and renal transplant patients. 24The CASP scores for the three caseecontrol studies rangedfrom nine 22 to six, 20 whilst the prevalence studies, using theadjusted CASP cohort study tool, scored between five 4,37 andTitle not relevant:3668Not English: 44Criteria not met: 13134 – lacking statistics60 – inappropriate cases/controls37 – not original researchInitial search: 4109Abstract review: 441Full-text review: 154Abstract notrelevant: 243Included:23Figure 1. Systematic literature review schematic. The numbersrepresent the number of studies at each stage (e.g. 441 studiesunderwent abstract review).11 36 out of 12. The cohort studies scored between three 38 and13 35 out of 14. Of the 23 papers included, almost 80% hada CASP value within the score’s upper quartile, being six andabove for the caseecontrol, prevalence and surveillancestudies, and seven and above for the cohort studies.Of the 143 risk factors assessed in all the studies, 96 (68%)were significant and 43 were independent risk factors for HCAUTI, with six studies not performing multivariate analyses. Ofthe independent risk factors, 29 were only assessed by onestudy. Among the remaining 14 risk factors that were assessedby multiple studies, half were found to be significant in morethan one of the studies in which they were included (femalesex, increasing age, diabetes mellitus, length of hospital stay,prior stroke, urinary catheter and duration of urinary catheter).Female sex was included in all the papers for analysis andwas found to be significant in 65% of the studies. One-quarter ofthe independent risk factors were associated with urinarycatheter use and management, including colonization of thedrainage bag, 41 having more than two catheters recorded 20 andintermittent catheterization. 22In order to assess the relative importance of the significantrisk factors for HCA UTI, PAR% values were calculated for thosefactors where sufficient data were reported in the correspondingstudies. Raw data (as described in the Methods) wereextracted directly from 14 studies, while ORs were used toobtain raw data in four studies. 25,33e35 Two studies did notpublish sufficient data to enable PAR% calculations, 27,40 andthree papers were not included as these were prevalencestudies. 4,36,37 Thus, 18 of the 23 papers contributed data to thePAR% calculations. PAR% values were calculated for 35 significantrisk factors for HCA UTI, and 26 were also independent riskfactors. The remaining 17 independent risk factors did not havesufficient data for PAR% calculations.

222C. King et al. / Journal of Hospital Infection 82 (2012) 219e226Table ISummary of the studies identified as a result of the systematic literature reviewStudy design Critical Appraisal Skills Patient population Country Data source (electronic) Reference (year)Programme Score (total)Caseecontrol 6 (12) General Saudi Arabia Patient records 20 (2004)Caseecontrol 8 (12) Hip Sweden Patient records 21 (1999)Caseecontrol 9 (12) General UK Study/surveillance 22 (1999)Cohort 5 (14) Paediatric India Study 23 (2011)Cohort 9 (14) Transplant Brazil Study 24 (2006)Cohort 11 (14) Spinal Spain Study 25 (2000)Cohort 12 (14) General Australia Study/patient records 26 (2007)Cohort 10 (14) Orthopaedic Spain Study 27 (2001)Cohort 10 (14) ICU Canada Linked records (e) 29 (2002)Cohort 8 (14) ICU Canada Linked records (e) 30 (2005)Cohort 9 (14) ICU France Study 31 (2003)Cohort 9 (14) Stroke USA CASPR 32 (2006)Cohort 8 (14) Paediatric USA Patient records 34 (2009)Cohort 13 (14) Surgery Spain Patient records 35 (2003)Cohort 6 (14) General USA Study 28 (1986)Cohort 12 (14) ICU France Study 33 (2001)Prevalence 5 (12) General UK Prevalence survey 4 (2008)Prevalence 11 (12) General Germany Prevalence survey 36 (1997)Prevalence 5 (12) General Israel Patient records 37 (2005)RCT 8 (14) General USA Study 41 (1986)Surveillance 3 (12) ICU USA Patient records 38 (1999)Surveillance 6 (12) General Egypt Study 39 (2010)Surveillance 12 (12) ICU France Surveillance 40 (2008)ICU, intensive care unit; CASPR, California Acute Stroke Pilot Registry; RCT, randomized controlled trial.For each study included in the analysis, the study design, the Critical Appraisal Skills Programme (CASP) score and the study setting aredescribed. CASP scores are presented together with the total possible score (in parentheses). The data source describes the origin of the dataused by the included study for their statistical analysis, with ‘study’ meaning either that the data was collected for the purpose of the researchpublished, or that there was no further clarification on the origin of the data. An ‘e’ in parentheses denotes that the data used were electronicor automated, instead of using paper patient records or manual searching.Of the 35 risk factors for which PAR% values were calculated,functional independence (measured as the level ofassistance required with daily living tasks) was found to havea non-significant PAR% (1.47, 95% CI 0.62 to 3.55), despitebeing identified as an independent predictor of HCA UTI. 25 Theremaining 34 significant risk factors were stratified into a hierarchy,with the risk factor with the highest PAR% (79.34%, 95%CI 78.18e80.50) being urinary catheterization (Table II). Fiverisk factors demonstrated protective properties from thecalculated PAR%: creatinine level

Table IISummary of the population-attributable risk percentage (PAR%) values for significant risk factors of healthcare-associated urinary tractinfections, and their availability within local administrative hospital dataRisk factor PAR% 95% ConfidenceIntervalC. King et al. / Journal of Hospital Infection 82 (2012) 219e226 223PopulationsizeNumberof studiesLevel ofsignificanceAvailable inlocal dataDatasourceUrinary catheter 79.34 78.18 to 80.50 4652 2 Independent Y ICD-10/OPCSTransurethral catheterization 52.41 48.21 to 56.60 544 1 Independent N ePositive meatal culture 36.88 33.42 to 40.34 747 1 Significant Y LIMSASA score >2 32.69 31.07 to 34.31 3229 2 Independent N eStroke 27.62 26.39 to 28.86 5028 2 Independent Y ICD-10Urine meter drainage 26.00 23.74 to 28.22 1474 1 Independent N eDisconnection of collection 25.63 23.40 to 27.86 1474 1 Significant N ejunctionBladder dysfunction 25.13 20.31 to 29.96 311 1 Independent Y ICD-10Length of stay >10 days 24.92 21.28 to 28.55 544 1 Independent Y PASUC duration >6 days 23.29 21.13 to 25.45 1474 1 Independent N eFemale sex 23.00 22.51 to 23.44 31,408 16 Independent Y PASHypertension 21.85 18.71 to 25.00 663 1 Significant Y ICD-10Dead kidney donor 19.32 13.26 to 25.38 163 1 Independent Y OPCSColonization of drainage bag 15.69 13.83 to 17.54 1474 1 Independent Y LIMSDrainage bag change 14.43 12.64 to 16.23 1474 1 Significant N eIntermittent catheterization 13.25 10.40 to 16.10 544 1 Independent Y AuditAtrial fibrillation 12.79 10.25 to 15.34 663 1 Significant N eAge >60 years 11.83 10.91 to 12.76 4674 2 Independent Y ICD-10Age >50 years 11.70 10.36 to 13.04 2221 2 Independent Y ICD-10Prior UTI in current10.37 8.82 to 11.93 1474 1 Independent Y ICD-10/LIMShospitalizationDiabetes mellitus 9.56 8.89 to 10.22 7522 6 Independent Y ICD-10ASA score >3 9.54 6.78 to 12.30 435 1 Independent N ePrior indwelling8.35 6.94 to 9.77 1470 1 Significant N ecatheterizationAge >65 years 6.42 5.12 to 7.72 1368 2 Independent Y ICD-10Bag-outlet tube error 5.72 4.54 to 6.91 1474 1 Significant N eOther malignancy 4.24 2.54 to 5.93 544 1 Significant Y ICD-10Bladder instrumentation 3.62 3.07 to 4.17 4365 1 Independent Y OPCSUnresolved spinal injury 3.54 2.99 to 4.09 4365 1 Significant Y ICD-10Central venous catheter 2.07 0.87 to 3.27 544 1 Significant Y OPCSCreatinine

224C. King et al. / Journal of Hospital Infection 82 (2012) 219e226Risk factors assessed: 143(47 risk factors were notstatistically significant)Significant risks: 96Independent risks: 43Located in localhospital data: 39PAR% calculated: 35(one of which wasinsignificant)Located in localhospital data: 28PAR% calculated: 26(one of which wasinsignificant)PAR% and located in localhospital data: 24PAR% and located in localhospital data: 18Figure 2. Schematic of the numbers of significant and independent risk factors at different stages in the methodology. All those riskfactors that are independent are also significant, whereas those risk factors that are significant but not independent may only have beenconsidered in univariate analysis or became insignificant when confounders were taken into account in multivariate analysis. Significantrisk factors, P-value of

C. King et al. / Journal of Hospital Infection 82 (2012) 219e226 225how to code conditions, and information on when they are tobe recorded are provided by the NHS (available at: http://www.connectingforhealth.nhs.uk/codingclinic). 46 The mostrecent guidelines specify a group of comorbidities to bemandatorily coded within administrative hospital data if thereis any mention of them in the patient’s medical notes, includingurinary retention, renal failure and diabetes. 46 However,despite clear guidance, poor-quality coding is still reported. 47Issues arise from variability in coders and their experience,and the detail and accuracy of the information contained withinthe patient’s medical notes. In order to create surveillancetools using administrative data, these quality issues need to betaken into consideration, with an understanding of the potentialfor the under or over-recording of codes.In the absence of a real-time medical record, the applicationof surveillance for HCA UTI based on predictors at anindividual level is not plausible. The delay often encounteredwith administrative data means that data on risks, such asdiabetes mellitus, are only available electronically once thepatient has been discharged. Centres such as Salt Lake City,where electronic patient records have been implemented for20 years, have shown the potential for this technology in aidingsurveillance of HCAI. 48 Certain hospital data, such as pathologyand radiology information, are recorded electronically as theyare processed, suggesting that real-time syndromic surveillancecould be developed based on these data. This has beenimplemented successfully in the USA for catheter-associatedbloodstream infections. 49 However, the use of predictors todevelop surveillance at the hospital population level usingadministrative data is possible with prospective applications.This includes implementing a risk stratified framework for HCAUTI surveillance, the creation of benchmarking tools for HCAUTI which account for local patient populations, and theopportunity to monitor interventions and control measures.The use of different administrative hospital data systems todevelop novel surveillance systems is still an emerging discipline,and with over 750 individual databases located thus farwithin ICHNT, further relevant information on risk factors islikely to be stored routinely within the hospital data. Electronicsurveillance represents a cost-effective, time-efficient androbust approach to surveillance, although it is subjective toclinical opinion and a changing policy landscape. 50 In thecontext of HCA UTI, this could potentially provide a system tocomplement traditional surveillance schemes and infectionprevention and control practices across healthcare settings.Despite the methodology in this paper being based upon theadministrative data found routinely within the UK NHS, theauthors believe that the principle is applicable to any healthsystem with electronic administrative data.This study has demonstrated the potential of local administrativehospital data for the development of innovativesurveillance for HCA UTI by applying the results froma systematic literature review to local administrative datausing a novel methodology, despite the various limitationsdiscussed. This work has provided the basis for a framework tofurther explore the benefits of using hospital data in thedevelopment of surveillance tools for HCA UTI. The next stepsare to investigate the utility and quality of the codes identifiedby this review in local data, apply this to developing riskprofiles and proxy indicators, and test the value of syndromicalgorithms to enhance traditional surveillance and infectioncontrol approaches.Conflict of interest statementNone declared.Funding sourcesThe UK Clinical Research Collaboration funds the NationalCentre for Infection Prevention and Management at ImperialCollege London, and also supports the UK National Institutefor Health Research Biomedical Research Centre fundingscheme. The Dr Foster Unit at Imperial is largely funded bya research grant from Dr Foster Intelligence (an independenthealth service research organization) and is affiliatedwith the Centre for Patient Safety and Service Quality atImperial College Healthcare NHS Trust, funded by theNational Institute for Health Research.AcknowledgementsThe authors wish to thank the Infection Prevention andControl, Pathology and Microbiology departments and theInformation Technology teams within Imperial College NHSTrust for their collaboration.References1. World Health Organization. WHO guidelines on hand hygiene inhealth care: first global patient safety challenge e clean care issafer care. Geneva: WHO; 2009.2. Plowman R, Graves N, Esquivel J, Roberts JA. An economic modelto assess the cost and benefits of the routine use of silver alloycoatedurinary catheters to reduce the risk of urinary tractinfections in catheterized patients. J Hosp Infect 2001;48:33e42.3. Smyth ETM, McIlvenny G, Enstone JE, et al. Four country healthcareassociated infection prevalence survey 2006: overview of theresults. J Hosp Infect 2008;69:230e248.4. Humphreys H, Newcombe RG, Enstone J, et al. Four countryhealthcare associated infection prevalence survey 2006: riskfactor analysis. J Hosp Infect 2008;69:249e257.5. Haley RW, Culver DH, White JW, et al. The efficacy of infectionsurveillance and control programs in preventing nosocomial infectionsin United States hospitals. Am J Epidemiol 1985;121:182e205.6. Brown C, Richards M, Galletly T, et al. Use of anti-infective serialprevalence studies to identify and monitor hospital-acquiredinfection. J Hosp Infect 2009;73:34e40.7. Berrington A, Borriello P, Brazier J, et al. National Clostridiumdifficile Standards Group: report to the Department of Health.J Hosp Infect 2004;56(Suppl. 1):1e38.8. Department of Health. Hospitals to monitor hospital acquiredinfection. Press Release 2000/0584. London: Department ofHealth; 2000.9. Department of Health. Safety express: guide to programmedelivery. London: Department of Health; 2011.10. Choudhuri J, Pergamit R, Chan J, et al. An electronic catheterassociatedurinary tract infection surveillance tool. InfectControl Hosp Epidemiol 2011;32:757e762.11. García Álvarez L, Aylin P, Tian J, et al. Data linkage betweenexisting a healthcare databases to support hospital epidemiology.J Hosp Infect 2011;79:231e235.12. Teodoro D, Choquet R, Pasche E, et al. Biomedical datamanagement: a proposal framework. Stud Health Technol Informat2009;150:175e179.13. Jen MH, Holmes AH, Bottle A, Aylin P. Descriptive study ofselected healthcare-associated infections using national hospitalepisode statistics data 1996e2006 and comparison with mandatoryreporting systems. J Hosp Infect 2008;70:321e327.14. Solutions for Public Health. Critical appraisal skills programme ecase control studies. UK: NHS; 2006.

226C. King et al. / Journal of Hospital Infection 82 (2012) 219e22615. Solutions for Public Health. Critical appraisal skills programme ecohort studies. UK: NHS; 2004.16. Sanderson S, Tatt ID, Higgins JP. Tools for assessing quality andsusceptibility to bias in observational studies in epidemiology:a systematic review and annotated bibliography. Int J Epidemiol2007;36:666e676.17. Last J. A dictionary of epidemiology. 4th ed. Oxford: OxfordUniversity Press; 2001.18. Pietrantojn CD. Four-fold table cell frequencies imputation inmeta analysis. Stat Med 2006;25:2299e2322.19. World Health Organization. ICD-10 classifications of mental andbehavioural disorder: clinical descriptions and diagnostic guidelines.Geneva: WHO; 1992.20. Al-Helali NS, Al-Asmary SM, Abdel-Fattah MM, Al-Jabban TM, Al-Bamri AL. Epidemiologic study of nosocomial urinary tract infectionsin Saudi military hospitals. Infect Control Hosp Epidemiol2004;25:1004e1007.21. Hedstrom M, Grondal L, Ahl T. Urinary tract infection in patientswith hip fractures. Injury 1999;30:341e343.22. Nguyen-Van-Tam SE, Nguyen-Van-Tam JS, Myint S, Pearson JC. Riskfactors for hospital-acquired urinary tract infection in a large Englishteaching hospital: a caseecontrol study. Infection 1999;27:192e197.23. Brindha SM, Jayashree M, Singhi S, Taneja N. Study of nosocomialurinary tract infections in a pediatric intensive care unit. J TropPediatr 2011;57:357e362.24. Dantas SR, Kuboyama RH, Mazzali M, Moretti ML. Nosocomialinfections in renal transplant patients: risk factors and treatmentimplications associated with urinary tract and surgical site infections.J Hosp Infect 2006;63:117e123.25. De Ruz AE, Leoni EG, Cabrera RH. Epidemiology and risk factorsfor urinary tract infection in patients with spinal cord injury.J Urol 2000;164:1285e1289.26. Graves N, Tong E, Morton AP, et al. Factors associated with healthcare-acquired urinary tract infection. Am J Infect Control2007;35:387e392.27. Herruzo-Cabrera R, Lopez-Gimenez R, Cordero J, Munuera L.Urinary infection after orthopedic procedures. Int Orthopaed2001;25:55e59.28. Larsen RA, Burke JP. The epidemiology and risk factors for nosocomialcatheter-associated bacteriuria caused by coagulasenegativestaphylococci. Infect Control 1986;7:212e215.29. Laupland KB, Zygun DA, Davies HD, Church DL, Louie TJ, Doig CJ.Incidence and risk factors for acquiring nosocomial urinary tractinfection in the critically ill. J Crit Care 2002;17:50e57.30. Laupland KB, Bagshaw SM, Gregson DB, Kirkpatrick AW, Ross T,Church DL. Intensive care unit-acquired urinary tract infections ina regional critical care system. Crit Care 2005;9:R60eR65.31. Leone M, Albanese J, Garnier F, et al. Risk factors for nosocomialcatheter-associated urinary tract infection in a polyvalent intensivecare unit. Intens Care Med 2003;29:1077e1080.32. Ovbiagele B, Hills NK, Saver JL, Johnston SC. Frequency anddeterminants of pneumonia and urinary tract infection duringstroke hospitalization. J Stroke Cerebrovasc Dis 2006;15:209e213.33. Tissot E, Limat S, Cornette C, Capellier G. Risk factors forcatheter-associated bacteriuria in a medical intensive care unit.Eur J Clin Microbiol 2001;20:260e262.34. Traxel E, DeFoor W, Reddy P, Sheldon C, Minevich E. Riskfactors for urinary tract infection after dextranomer/hyaluronic acid endoscopic injection. J Urol 2009;182(Suppl.):1708e1712.35. Vazquez-Aragon P, Lizan-Garcia M, Cascales-Sanchez P, Villar-Canovas MT, Garcia-Olmo D. Nosocomial infection and related riskfactors in a general surgery service: a prospective study. J Infect2003;46:17e22.36. Kampf G, Gastmeier P, Wischnewski N, et al. Analysis of riskfactors for nosocomial infections e results from the first nationalprevalence survey in Germany (NIDEP study 1). J Hosp Infect1997;37:103e112.37. Mnatzaganian G, Galai N, Sprung CL, et al. Increased risk ofbloodstream and urinary infections in intensive care unit (ICU)patients compared with patients fitting ICU admission criteriatreated in regular wards. J Hosp Infect 2005;59:331e342.38. Rosser CJ, Bare RL, Meredith JW. Urinary tract infections in thecritically ill patient with a urinary catheter. Am J Surg 1999;177:287e290.39. Talaat M, Hafez S, Saied T, Elfeky R, El-Shoubary W, Pimentel G.Surveillance of catheter-associated urinary tract infection in 4intensive care units at Alexandria university hospitals in Egypt. AmJ Infect Control 2010;38:222e228.40. Vanhems P, Baratin D, Voirin N, et al. Reduction of urinary tractinfections acquired in an intensive care unit during a 10-yearsurveillance program. Eur J Epidemiol 2008;23:641e645.41. Platt R, Polk BF, Murdock B, Rosner B. Risk-factors for nosocomialurinary-tract infection. Am J Epidemiol 1986;124:977e985.42. Dindo D, Muller MK, Weber M, Clavien P- A. Obesity in generalelective surgery. Lancet 2003;361:2032e2035.43. Jankovic S, Radosavljevic V. Risk factors for bladder cancer.Tumori 2007;93:4e12.44. Connor J, Norton R, Ameratunga S, et al. Driver sleepiness and riskof serious injury to car occupants: population based case controlstudy. BMJ 2002;324:1125.45. Madigan MP, Ziegler RG, Benichou J, Byrne C, Hoover RN.Proportion of breast cancer cases in the United States explainedby well-established risk factors. J Natl Cancer Inst 1995;87:1681e1685.46. National Health Service Classification Service. Coding clinic. 2011.47. Schweizer ML, Eber MR, Laxminarayan R, et al. Validity ofICD-9-CM coding for identifying incident methicillin-resistantStaphylococcus aureus (MRSA) infections: is MRSA infectioncoded as a chronic disease? Infect Control Hosp Epidemiol2011;32:148e154.48. Evans RS, Burke JP, Classen DC, et al. Computerized identificationof patients at high-risk for hospital-acquired infection. Am J InfectControl 1992;20:4e10.49. Woeltje K, McMullen K, Butler AM, Goris AJ, Doherty J. Electronicsurveillance for healthcare-associated central line-associatedbloodstream infections outside the intensive care unit. Infect ContHosp Epidemiol 2011;32:1086e1090.50. Lazarus R, Kleinman K, Dashevsky I, deMaria A, Platt R. Usingautomated medical records for rapid identification of illnesssyndromes (syndromic surveillance): the example of lower respiratoryinfection. BMC Public Health 2001;1:9.