Part I Barry Cookson University College, London, United Kingdom Declarations: Gojo, Ecolab, Qiagen
Part II Christina Vandenbroucke-Grauls VU University medical center, Amsterdam, The Netherlands Declarations: Biomerieux
Part 1: Barry Cooksons Topics Improving evidence base in infection control Modelling Surveillance including Surgical, LTCF, Costings Typing/Tracking Organisms MRSA Interventions Screening/Suppression of MRSA Hand hygiene
Part 2: Christina Vandenbroucke-Grauls Topics
Antimicrobial resistance & antibiotic use ESBL and carbapenemases Clostridium difficile The hospital environment Gender The Flos
Sources for Presentations www.micro-blog.info Otter & Yezli http://i-prevent.blogspot.nl A.Voss www.haicontroversies.blogspot.nl Perencevich, Diekema, Edmond Acknowledge Colleagues help Sheldon Stone Olga Paniara Modelling and Big Data Ben Cooper, Sarah Deeny & Julie Robotham ICPIC (Eli Perencevich & Andreas Widmer) ICAAC (Andreas Voss) 2013 ECCMID (Benedetta Allegranzi & Robert Skov) presentations: complement, augment
Importance of ESCMID CPD library: many are resource slides as little time to discuss in-depth The Flos Special Award of the Year is Cardiff University Library, Cochrane Archive, University Hospital Llandough. 1) Improving the quality of the infection control evidence base: research inform guidelines/guidance for policies for safer patient care
Tools for improving the evidence base of scientific literature CONSORT for randomized controlled trials (1996) STROBE for observational studies (2007) ORION for intermittent time series & outbreaks (2007) Preferred Reporting Items for Systematic Reviews & Meta- Analyses (PRISMA) for systematic reviews & meta analyses (2009) EQUATOR Network (2006) http://www.equator-network.org/resource-centre/library- of health-research-reporting/ STROME-ID (2014) PLoS Med 10(8): e1001504. doi:10.1371/journal.pmed.1001504 PLOS Med a champion and published previously STROBE Evidence that CONSORT & PRISM improved the literature Published: August 27, 2013 Studies also show that the quality of reporting overall remains suboptimal as not all journals endorse or enforce the use of reporting guidelines
Comment: this is certainly true for infection control and related journals Reviewed outbreaks reported in So Paulo State, Brazil, & verify compliance with mandatory outbreak notification Explored potential for ORION to inform where need to to improve competencies Only 15/87 (17%) published outbreaks reported to authorities Poor/Varied compliance with ORION categories Background 32% , Objectives 75% , Participants 2%, Setting 46% , Infection-Related Outcomes13%, Interventions 52% & Culture-Typing 55% Issue: Missed opportunity to stratify by Journals requirements for ORION (expected for authors and evaluation by referees) Pires et al, AJIC 42 (2014) e47-e53 PLoS Med 11(2): e1001603. doi:10.1371/ journal.pmed.1001603 Differences in: 1) Publication format
2) Work processes
3) Author team Management
4) Statistical methods. Current Emerging Health Knowledge Ecosystems Modelling! PLoS Med 11(2): e1001603. doi:10.1371/ journal.pmed.1001603 2) Modelling A previous ECCMID Debate (2012) Audience split on utility! Everything should be made as simple as possible, but not simpler Albert Einstein We argue that usability & stability of a model is an outcome of the negotiation that occurs within the networks & discourses surrounding it.
PLoS ONE 8(10): e76277. doi:10.1371/journal.pone.0076277 PLoS Med 10(10): e1001540. doi:10.1371/journal.pmed.1001540 We have found evidence to suggest that identification of uncertainties, combined with their deproblematisation can act to stabilise the role of scientific modelling in decision-making Clin Microbiol Infect 2013; 19: 993998
Mathematical models play an important role in helping healthcare systems to respond to ongoing epidemics or plan the logistics of various theoretical scenarios Prediction & model-based management of epidemics in their early phase are quite unlikely to become the norm. far too complex to be predictable BC: so do we guess: sometimes nothing else is available? Flu: pnas.org/content/early/2011/10/24/1103002108.short Internet biosurveillance systems can detect an outbreak of an infection more rapidly than ever before. BC : this may be a bit premature?
The Flos Honorary Award of the Year goes to Prof Tim Berners-Lee Inventor of the World Wide Web 25 th Anniversary 3) Surveillance Lancet Infect Dis 2014;14: 16068 Dengue & Influenza reviewed
Complement not replace existing systems
Many issues described e.g. Lack of access to internet Populations vary in internet use and Health-seeking behaviour (BC: sample migration e.g. Facebook?) USA poor sensitivity and spatial resolution necessary to detect small, localised flu outbreaks Also see Big Data FT article Big data: are we making a big mistake? FT Magazine: March 28, 2014: Tim Harford Googles estimates of the spread of flu-like illnesses were overstated by almost a factor of two. Google cared about correlation rather than causation The End of Theory: with enough data, the numbers speak for themselves .. hopelessly naive where spurious patterns vastly outnumber genuine discoveries. When it comes to data, size isnt everything. Non random.. sampling error USA twitters are disproportionately young, urban/suburban & black (Sample migration; Super tweeters?)
Cape Town Sunday Times Lancet Infect Dis 2014;14: 16068 Journal of Hospital Infection 86 (2014) 34e41 Referral to:
Observed rates Supra Regional specialist Hospitals Created Same Collective Murchan et al, EMRSA-16 spread in England and Wales. J Clin Microbiol 2004; 57: 345-346.
HAI results, ECDC PPS 2011-2012
HAI prevalence: Overall: 13829/231459 patients with 1 HAI HAI prevalence: 6.0% Country range: 2.3%-10.8% 15000 HAIs; 1.1 HAI/patient HAI present at admission: 23% Same hospital: 55% 33% surgical site infection HAI during current hospitalisation: 76% of HAIs, prevalence: 4.5% Median time to onset of HAI: 12 days Microorganism reported: 45.9% Prevalence antimicrobial use for treatment of hospital infection: 6.4% 95% of patients with HAI received >=1 antimicrobial on day of survey
Source: ECDC PPS, 2011-2012 (1) incl. C. difficile infections 3.6% (2) incl. clinical sepsis 5.3% 23% 19% 20% 11% 8% 6% 4% 10% Pneumonia/LRTI Urinary tract Surgical site infection Bloodstream Gastrointestinal (1) Systemic (2) Skin/Soft tissue Other/unspecified Type of HAI 30 countries (29 EU/EEA countries + Croatia), 33 PPSs (networks)
2014:12;71. | Observed vs predicted HAI prevalence by country, ECDC PPS 2011-2012 0 2 4 6 8 10 12 14 16 18 20 Latvia Romania Lithuania Slovakia Bulgaria UK-Wales UK-N. Ireland Malta Hungary Czech Republic UK-Scotland France Germany Ireland Luxembourg Estonia Croatia UK-England Austria Italy Slovenia Poland Cyprus Belgium Sweden Netherlands Finland Norway Spain Greece Denmark* Iceland Portugal Patients with HAI (%) Observed HAI prevalence (%) (with 95% confidence interval) Predicted HAI prevalence (%) (based on case-mix) http://www.ecdc.europa.eu/en/publications/Publications/he althcare-associated-infections-antimicrobial-use-PPS.pdf Structure & process indicators: percentage of single room beds
Single room beds in participating hospitals (%): median = 11.1% Source: ECDC, 2012 (ECDC PPS data as of 23/11/2012) 0 20 40 60 80 100 N of single room beds*100 /Total beds UK-Wales UK-Scotland UK-Northern Ireland UK-England Sweden Spain Slovenia Slovakia Romania Portugal Poland Norway Netherlands Malta Luxembourg Lithuania Latvia Italy Ireland Iceland Hungary Greece Germany France Finland Estonia Denmark Czech Republic Cyprus Croatia Bulgaria Belgium Austria *Poor data representativeness Alcohol hand rub consumption in acute care hospitals, ECDC PPS 2011-2012 HAI PPS 4600 trained Susan Hopkins lead, HP Agency & HP Scotland HALT 2 PPS 1700 LTCFs staff trained. Fidelma Fitzpatrick & Tracey Dillane, HSE & HP Surveillance Centre 2013: http://www.ecdc.europa.eu/en/ publications/Publications/infect ion-control-core- competencies.pdf 2014-15: TRICE-Implementation Strategy includes: IC Course assessments IC/HH WIKI : ESCMID SGs to be involved Revisiting TRICE IC Resources Awaiting clearance for: ECDC-Funded SIGHT Project Should be in Lancet ID (Zing et al) Increasing Healthcare Delivery in the Community Increasingly aged populations Decreasing lengths of hospital stay: HAIs presenting in the community Increasing numbers of step-down from hospital and hybrid (Step-Down/Residential) facilities Issues with definitions of names of facilities Lack of surveillance, infection control standards, guidelines and audit . LOW PRIORITY IN MANY COUNTRIES See: Moro et al, ICHE 2010; 31 (suppl. 1); 559-62 (IPSE WP7) Cookson et al, J Hosp Infect 2013; 85: 45-53 (HALT 1)
Carl Suetens ECDC Championed HALT Antoon Gijsens DG SANCO Bea Jans HALT Lead LTCF a late addition Cookson et al, J Hosp Infect 2013; 85: 45-53. HALT-1 & 2 Reports
Katrien Latour, Abstract No ECCMI-4351 http://www.ecdc.europa.eu/en/Pages/home.aspx 5 th May 2014
Additional & increasing burden of AMR organisms Main increases are in AMR organisms MSSA continuing increases MRSA increases 7.6%/Y then decreases 4.8%/Y <> 2005
Surgical Site Infections (SSIs) Korol et al, PLoS ONE 8(12): e83743. doi:10.1371/journal.pone.0083743 1) Challenging: very rich literature many study designs, settings, categorizations & definitions
2) Despite this they consistently found associated SSI risk factors
3) Risk Factors relate to: reduced fitness, patient frailty, surgery duration, & complexity BC: Body warming, glucose control, triclosan stitches?
Risk score gauged independent influence of each risk factor New logistic regression approach of quantifying the influence of different operative types on SSIs Model used 181 894 operations for derivation, 181 146 for validation SSIRS captured 89.7% of the validation population Web based system available significantly better discrimination than NNIS Basic Risk SSI Index Also a simpler SSIRS index permits 30-day SSI risk bedside estimation without computational aids BC: Needs external validation. ECDC interested? van Walraven C, Musselman R. PLoS ONE 2013;8:e67167 Owens et al, JAMA. 2014;311:709-716 Retrospective analysis: 284, 098 ambulatory surgical procedures Databases of 1/3 US population in 8 dispersed USA states Needed postsurgical acute care visits General, orthopaedic, neuro, gynae, & urologic surgery Length of stay less than 2 days SSIs at 14days 3.09 /1000 ambulatory surgical procedures 30days 4.84 /1000 ambulatory surgical procedures 63.7% within 14 days of the surgery: limited risk factor analysis 93.2% needed inpatient treatment Low relative to all causes but significant adverse events The Post Discharge Surveillance An Elephant in the Room! http://en.wikipedia.org/wiki/Elephant Difficult to compare studies as differences in e.g. Lengths of stay Operative categories Definitions of infections Methods of data collection: Re-admissions only! Telephone or Patient or Healthcare worker reporting If staff/patients are trained Lack detail so work can be repeated
American Journal of Infection Control 41 (2013) 591-6 Journal of Hospital Infection 86 (2014) 127e132 American Journal of Infection Control 41 (2013) 549-53 Ann Intern Med. 2013;159:447-455.
Post Discharge Surveillance (PDS) J Hospital Infect 2013; 84: 267 (& refs to two previous letters) Disputes including English HPA (PHE) SSI system do not bench mark PDS data Brisbane group supported English HPA responding: Until a valid & cost-effective solution PDS is found do not bench mark. Hospitals are required to undertake PDS, so the data can be used for internal (i.e. local) quality improvement Should not be required to submit PDS data for public scrutiny or be penalized for not doing so.
Raschka et al, AJIC 2013;41: 773-7 19% less selected HAIs over 4 years Cost avoidance of at least $9 million 80% in last two years! (so enormous potential: useful to quote internationally) Interesting methodology Vancouver regional Canadian Inf. control programme Standardized policies, procedures, and initiatives (including hand hygiene campaign) The Flos Extraordinary Anniversary of the Year
60 years since description of DNAs structure Watson, Crick, (Franklin and Wilkins) http://www.chemheritage.org/discover/online-resources/chemistry-in-history/themes/biomolecules/dna/watson-crick-wilkins-franklin.aspx 4) Typing & Tracking HAI Organisms
Medline database searches with terms
Infection & Molecular Epidemiology
No papers/100K/year
Lancet Infect Dis 2014; 14: 34152
20 items added to 22 item STROBE checklist
Should advance the quality & transparency of scientific reporting, with clear benefits for evidence reviews & health-policy decision making Lancet Infect Dis 2014; 14: 34152
Lancet Infect Dis 2014; 14: 34152
Educate users (STROME-ID) e.g. strengths and limitations Clear objectives to ensure optimises patient care/safety Interpretation needs detailed epidemiological data Golden fleece TIMELY automatic appropriate simplified discriminant data : (TATs not stated in many papers!) Dialogue with reference laboratories and others with expertise in this area to optimize its potential. Consider accreditation requirements for Reference labs National information considered National funding ref Public Health aspects Optimisation of NGS information for Infection Control Teams See also :Humphries & Coleman Letter: http://dx.doi.org/10.1016/j.jhin.2013.05.002 Ehrlich & Post JAMA Internal Medicine 2013;173: 1406-06 Struelens and Brisse. Euro Surveill. 2013;18(4):pii=20386. Editorial and several very useful papers in issue also look at: Price et al, CID 2014;58:60918 Invited paper: interesting analysis of potential MRSA Melo-Christino et al, Lancet 2013;382:205 (20 July 2013) vanA +MRSA strain in Europe from Portuguese patient USA MRSA type (ST105, SCCmec type II: common HAI strain in Portugal) No epidemiological links to USA vanA source may a wound VRE Treatment progressing and no spread thus far
Emerged in Brazil following multiple vancomycin courses No spread so far: died with GNR BSI Related to community USA 300 strain: has bsa CA MRSA operon PVL negative Perhaps element originated from a GRE in another patient in the same room: patient also positive (late) Transferable to other S aureus (WOW!!!!!) Worrying : USA 300 clades common globally Rossi et al, N Engl J Med 2014;370:1524-31 174 SA and MRSA of CC8 (including CA USA300) from 5 Continents 1957-2008 Phylogeny explored at 112 genetic housekeeping loci, AMR and diverse mobile genetic elements 9 clades: 8 independent SCCmec acquisitions : started in mid-1970s. 88% carried plasmidic rep gene sequences (5 rep genes and eight rep families) Increasing and stable AMR (9 classes) during the evolution of several lineages, including USA300. (Comment: check pedigree of different strains.) Diverse virulence determinants Perhaps WILL become a multiple resistant MRSA? (USA300 now a hospital MRSA)
Strommenger et al, J Antimicrob Chemother 2014; 69: 616622 4 new MRSA mecA/C negative from 3 ST types from Scottish patients Perhaps due to identified amino acid substitutions in their endogenous PBPs 1, 2 & 3? Not BORSA strains, as resistant to both oxacillin & cefoxitin Need for vigilance ref molecular MRSA assays & possible new drug target resistance MUST NOT ABANDON PHENOTYPIC TESTING Xiaoliang et al, J Antimicrob Chemother 2014; 69: 594597 A 55-pound or greater Hog produces at least 10 gallons of manure a day (!) Manure spread on surrounding Iowan fields MRSA can be aerosolized from this manure to human food or water sources. Carrel et al, Infect Control Hosp Epidemiol 2014;35(2):190-192 BetterPhoto.com McKinnell et al, ICHE 2013; 34(: 161170 Often ignored in studies & national screening programmes Systematic review 1966-2012: 23 papers: 39,497 patients Few studies looked at >1 such extra nasal site! ICU admission: detects ~1/3 more Hospital admission: >6% MRSA incidence: extra 37% <6% MRSA: extra 50% Sole sites (supports various guidelines) Oropharynx +21%: Rectum +20%: Wounds +17% : (Axilla +7%) Dont forget ~25% of new cases are first positive at manipulated sites e.g. wounds and catheters CDC EIP-ABCs sentinel laboratory-based MRSA case finding identified MRSA cultures in 9 US metropolitan areas from 2005 through 2011
Detailed risk factor analyses for HA- & CA-MRSA
In 2011 invasive MRSA infections (80 461) 31% lower than in 2005
Dantes et al, JAMA Intern Med. 2013;173(21):1970-1978. Dantes et al, JAMA Intern Med. 2013;173(21):1970-1978. CA MRSA Stable: epidemiology incomplete e.g. home/LTCF Interventions Culture, Organisational & Behavioural Aspects Culture Excellent reviews: De Bono et al, J Hosp Infect. 2014; 86: 1-6 and Borg, J Hosp Infect 2014; 86:161-68 Essential www site of Geert Hofstede http://geerthofstede.com/dimensions-of-national-cultures Hand Hygiene Studies Utilizing Shared Accountability and Financial Incentives Talbot et al, ICHE 2013; 34: 1129-1136 Positive deviance study: Marra et al, AJIC 2013;41:984-8 Real-time assessment practice using a Theoretical Behavioural Domains Framework. Fuller et al, AJIC 2014;42:106-10 Hot Topics! Universal MRSA Screening Universal MRSA Decolonisation/Suppression Hammers to crack walnuts? See posts on: http://haicontroversies.blogspot.com & NEJM Letters English Universal MRSA Screening 1) NOW Audit: Fuller et, al PLoS ONE 8(9): e74219 Implementation of universal screening was poor Admission Screening performed on: Emergency admissions 61% (median 67.3%) Electives 81% (median 59.4%) Very low MRSA admission prevalence: Emergencies 1%: Electives 0.6% Inpatient MRSA prevalence 3.3% (6% for original model)
2) Modelling Consultation underway suggesting stopping universal screening Universal Decolonisation/Suppression A Horizontal Strategy Lee et al, 2013 doi 10.11.36/bmjopen-2013-003126
Derde et al, Lancet Infect Dis 2014; 14: 3139 Huang et al, N Engl J Med 2013. DOI: 10.1056/NEJMoa1207290 First RCT Universal Antiseptic Use Risks (Horizontal Strategies) At what rates of resistant organisms is it cost effective? Mupirocin needed: how effective is it in reducing infections? Increasing side effects? How and how often used, rotate them? What is best antiseptic to use e.g. octenidine? Inactive chlorhexidine/soap formulations? Increasing disinfectant/antiseptic resistance Increased quantities used Sumps of bacteria e.g. leaking abscesses, suppurating tracheostomies
Universal Antiseptic Use Risks (Horizontal Strategies) Surveillance issues Locally: short lengths of stay, not detect damaged organisms Nationally: no surveillance Resistance cut-off agreements review? Maillard et al, Microb. Drug Res., 2013 doi:10.1089/mdr.2013.0039 Morissey et al, PLoS One. 2014; 9: e86669. doi:10.1371/journal.pone.0086669
Universal versus Targeted MRSA Screening? Deeny et al, JHI 2013
More efficient use of resources
Less potential for resistance to antiseptics
J Hosp Infect 2013; 85: 33-44
Hand hygiene
Only 39 quasi-experimental & 4 cluster & 2 RCTs /8,148 studies over 12 years Insufficient studies to assess single interventions! How can we design bundles? No of bundle components was not associated with greater effect: Include one or two interventions tripled compliance Include >two produced doubling of compliance Two bundles effective and Three studies of each: Education, Reminders, Feedback: OR (pooled) : 1.45 (1.12, 1.94) If add Administrative Support, & alcoholic handrubs : OR (pooled): 1.82 (1.69-1.97)
Schweizer, et.al, Clinical Infectious Diseases 2014;58:24859 Rock et al, American Journal of Infection Control 41 (2013) 994-6 Conclusion: HH before donning nonsterile gloves does not decrease already low bacterial counts on gloves. HH before donning nonsterile gloves may be unnecessary? Comment: Saves time: will HCWs confuse with sterile glove usage especially when stressed?
Use of scents in training1/3 increase in compliance Comment: Effectiveness for: different staff? Sustained? In real workplace? Pavlovian approach ethical? Own life.