Icaac 2013

Top papers ICAAC 2013 September 13th, 2013

Fraser & Voss 1

Video from ICPIC 2013 (available on YouTube)

InfecEon control talks are generally rated as “therapeuEc” for HCWs with sleeping disorders.

Disclaimer As a non-‐naEve (Dutch/German) English speaker some of the things I say may sound “harsher” than meant to …

hUp://www.slideshare.net/iPrevent/voss-‐icaac-‐online

hUp://actu.epfl

.ch/ne

ws/slo

wing-‐the-‐aging-‐process-‐on

ly-‐w

ith-‐anE

bioE

cs/

Propionibacterium

Eur Spine J 2013 Apr 22(4): 689 + 690 + 697


Fraser & Voss 2

Conclusions The addiEon of anEbioEcs to therapeuEc regimens for uncomplicated severe acute malnutriEon was associated with a significant improvement in recovery and mortality rates.

Trehan et al. N Engl J Med 2013;368:5 Kluytmans et al. CID 2013;56:478

¤  One hundred forty-‐five ESBL-‐EC isolates from retail chicken meat, human rectal carriers, and blood cultures were analyzed using mulElocus sequence typing, phylotyping, ESBL genes, plasmid replicons, virulence genes, amplified fragment length polymorphism (AFLP), and pulsed-‐field gel electrophoresis (PFGE).

RESULTS: ¤  Three source groups overlapped substanEally when their geneEc

composiEon was compared. ¤  A predicEon model based on the combined data classified 40% of the

human isolates as chicken meat isolates. CONCLUSIONS: ¤  We found significant geneEc similariEes among ESBL-‐EC isolates from

chicken meat and humans … Chicken meat is a likely contributor to the recent emergence of ESBL-‐EC in human infecEons in the study region.

Kluytmans et al. CID 2013;56:478 Kluytmans et al. CID 2013;56:478

Kluytmans et al. CID 2013;56:478 Kluytmans et al. CID 2013;56:478

¤  Carbepeneames in the food-‐chain?


Fraser & Voss 3

Collignon P et al. EID Augustus 2013

Es9ma9on based on dutch data!

¤ EsEmate for NL ²  21 addiEonal death, ²  908 hospital bed-‐days

¤ EsEmate for Europe ²  1,518 addiEonal death, ²  67,236 hospital bed-‐days

¤ The ongoing use of 3GC in mass therapy and prophylaxis should be urgently examined and stopped, parEcularly in poultry, not only in Europe, but worldwide!

Collignon P et al. EID Augustus 2013

Ammerlaan et al. CID 2012;54:1342

¤ Increased nosocomial BSI rates due to ARB occur in addiEon to infecEons caused by ASB, increasing the total burden of disease.

Ammerlaan et al. CID 2012;54:1342

“… for to everyone who has, more shall be given, and he will have an abundance’


¤ Whole genome mapping creates high-‐resoluEon, ordered whole genome restricEon maps

¤  Access WGM for (LA-‐)MRSA

Bosch et al PLOSone 8(6): e66493


Fraser & Voss 4

¤ Whole genome mapping produced highly reproducible results

¤ Provided a much higher discriminatory power than spa-‐typing, PFGE, or MLVA

¤ Whole genome mapping can provide a comparison with other maps

Bosch et al PLOSone 8(6): e66493

¤  Samples from 71 ambulances from 34 different Chicago-‐area municipaliEes

¤  At least one S. aureus sample was found in 69% of ambulances tested à 12% MRSA.

James V. Rago et al. Am J Infect Control, April 20122

Na9onal MRSA Rates Run Along with Fair Play of Na9onal Football Teams: A Cross-‐naEonal Data Analysis of the European Football Championship, 2008

E. Meyer et al. InfecEon 2012 epublished August 5

r = 0.628 p = 0.038

cards / 100 m

in

MRSA % E. Meyer et al. InfecEon 2012 epublished August 5

Copper a day - Keeps MRSA away

Noyce et al. J Hosp Infect 2006;63:289-‐297

¤  Repeat of study in The Netherlands woud not be possible …


Fraser & Voss 5

Edmond & Wenzel New Engl J Med May 2013

Editorial: Huang et al. Targeted versus universal decolonizaEon to prevent ICU infecEon. N Engl J Med 2013. DOI: 10.1056/NEJMoa1207290.

¤ Ver9cal infec9on-‐preven9on strategy. ² VerEcal intervenEons are designed to reduce colonizaEon or infecEon due to a specific pathogen by detecEon and isolaEon , they typically have high resource uElizaEon, and costs

² The philosophical underpinning is one of excepEonalism: some pathogens are more important than others and merit special control measures.

¤ Horizontal strategy ² is populaEon-‐based, is applied universally, and uses intervenEons effecEve in controlling all pathogens transmiUed by means of the same mechanism.

² Includes hand hygiene, chlorhexidine bathing, and care bundles, and they oqen require modificaEon of the behavior of HCWs

Edmond & Wenzel New Engl J Med May 2013

O’Brien AM et al. (2012) PLoS ONE 7(1):e30092

Largest sampling of raw meat products for MRSA contaminaEon

to date in the U.S.

¤ 395 pork samples were collected from a total of 36 stores in Iowa, Minnesota, and New Jersey.

¤ S. aureus was isolated from 256 samples (64.8%)

S.aureus MRSA

Conven9onal

67.3%

95% CI 61.7%–72.6%

6.3%

95% CI 3.9%—9.7%

An9bio9c-‐free

56.8%

95% CI 46.3%–67.0%

7.4%

95% CI 3.0%–14.6%

convenEonal

alternaEve

convenEonal

O’Brien AM et al. (2012) PLoS ONE 7(1):e30092

van Rijn et al. PLoS ONE 8(6): e65594

¤ Regular consumpEon of poultry (OR 2.40; 95% CI 1.08–5.33)

¤ CaUle density per municipality (OR 1.30; 95% CI 1.00–1.70)

¤ Sharing of scuba diving equipment (OR 2.93 5% CI 1.19–7.21)

¤ CA-‐MRSA carriage was not related to being of foreign origin.

van Rijn et al. PLoS ONE 8(6): e65594


Fraser & Voss 6

Top Infec9on Preven9on Papers 2012 – 2013

Victoria J. Fraser, MD Adolphus Busch Professor and

Chairman of Medicine

Washington University School of Medicine St. Louis, Missouri

Disclosures

¤ Consultant: BaUelle ¤ Research Funding:

² CDC Epicenters Program ² NIH K24 Mid Career Award ² NIH CTSA Research & EducaEon Director ² NIH KM1 CER Career Development Program ² AHRQ R24 CER Infrastructure Grant ² BJH FoundaEon

¤ Husband VP @ Express Scripts, 3 kids


Fraser & Voss 7

Surveillance and Epidemiology are S9ll Key

NNIS à NHSN & CLABSI Surveillance: BACKGROUND & METHODS

¤ # ICUs reporEng á from 144 (1990) to 794 (2010) ¤  ICU days á 236,000 (1990) to 11.4m (2010) ¤ ProporEon of Large teaching hospitals â from 57% (1990) to

24% (2010) ¤ 34% (CI 31.3-‐36.6%) – 55% (CI 53.4-‐57%) fewer CLABSI in

2009 vs 2001 ¤ CriEcal care days obtained, CLABSI rates NNIS/NHSN, applied

adjusted CLABSI rates to criEcal care days ¤ 3 scenarios: 1) no adjustment, 2) NNIS CLABSI rates

2004-‐2006 to 1990-‐2004 (surveillance arEfact), 3) â NNIS rates by 1/2 of scenario 2: Monte Carlo simulaEons, adult pts.

¤ Account for Δ definiEons, hospital type, Δ to NHSN Wise ME et al. Infect Control Hosp Epidemiol. 2013;34:547-‐554

Na9onal Es9mates of CLABSIs in Cri9cal Care Pa9ents

Wise ME et al. Infect Control Hosp Epidemiol. 2013;34:547-‐554

Figure 3. Hospital-‐onset CLABSI rates (cases per 1,000 ICU pt days ) adjusted for CLABSI definiEon change, surveillance parEcipaEon changes, and system transiEon, excluding neonates, U.S., 1990–2010

Wise ME et al. Infect Control Hosp Epidemiol. 2013;34:547-‐554


Fraser & Voss 8

Incidence Trends in Pathogen-‐Specific CLABSI in U.S. ICUs, 1990–2010

Fagan RP et al. Infect Control Hosp Epidemiol. 2013;34:893-‐899

¤ Methods – AcEve ICU surveillance, CDC NNIS 1990-‐2004 & NHSN 2006-‐2010

¤ Results ² 60% â in ICU CLABSIs over the past decade ²  Incidence of HO-‐MRSA BSI â 11%/year 2005-‐2008 ² EsEmated 18,000 CLABSI/year since 2006 ² Since 2006

²  S. aureus -‐18.3% (CI, -‐20.8 to -‐15.8%) annual â ²  GNR -‐16.4% (CI, -‐18 to -‐14.7%) annual â ²  Enterococci -‐17.8% (CI, -‐19 to -‐16.1%) annual â ²  Candida -‐13.5% (CI, -‐15.4 to -‐11.5%) annual â

² No Δ in pediatric ICU for S. aureus

Risk of Acquiring ESBL Klebsiella & E. coli From Prior Room Occupants in the ICU

Ajao AO et al. Infect Control Hosp Epidemiol. 2013;34:453-‐458

¤ Methods ² MICU – SICU pts of U of MD 9/2001 – 6/30/2009, ² Perianal cultures (LOTS)

¤ Results ² 267/7651 (3%) pts acquired ESBL GNR in ICU ² 32/267 (12%) in room ¯ˉ𝑐  prior ESBL⊕ pt ² Prior room not significantly associated

² AOR 1.39 (CI 0.94-‐2.08) ² 6/32 (18%) had similar PFGE strain to prior occupant

Ajao AO et al. Infect Control Hosp Epidemiol. 2013;34:453-‐458

Discon9nua9on of CP for MRSA: A RCT Comparing Passive & Ac9ve Screening With Culture & PCR

Shenoy ES et al. Clin Infect Dis. 2013;57:176-‐184

¤ RCT @ MGH, MRSA prevalence 8% (Hx on admission) ¤ No rouEne CHG bathing or decolonizaEon protocol ¤ Pts MRSA Hx > 90 days; 12/2010 – 9/2011, could enroll

on anEbioEcs (but not D/C CP); admission alert ID pts ¤ NonintervenEon = usual care; HO orders 3 MRSA nasal

Cx, 24 hrs apart; OFF anEbioEcs, 1° team not noEfied of enrollment

¤ IntervenEon = Cx and PCR x3, 24 hrs apart (Cepheid)


Fraser & Voss 9

Discon9nua9on of CP for MRSA: A RCT Comparing Passive & Ac9ve Screening With Cx & PCR


¤ 634 eligible; 457 included (198 control, 259 intervenEon)

¤ 62/198 (31%) controls screened (1/2 on ICUs acEve screening); 259/259 (100%) intervenEon screened

¤ 19/198 (9.6%) controls; & 191/259 (73.7%) intervenEon completed screening

¤ SensiEvity= 90.9%, 95.5%, 100% (1st, 2nd, 3rd swab) ¤ CP stopped 4x more in intervenEon (CI, 2.3-‐7.1) ¤ ~ ½ off anEbioEcs at screen; ~ both arms NS

Discon9nua9on of CP for MRSA: A RCT Comparing Passive and Ac9ve Screening With Cx and PCR


¤ First PCR vs 3 Cx = sensiEvity 93.9% (95% CI, 85.4-‐97.6%), specificity 92% (95% CI, 85.9-‐95.6%), PPV 86.1% (95% CI, 75.9-‐93.1%) and NPV 96.6% (95% CI, 91.6-‐99.1%)

¤ Passive Cx, AcEve Cx, PCR, CP D/C rates (6.6, 26.6 and 63.8%) and â CP days 104, 418 and 1841 ¤  (55% â CP days)

¤ Annualized savings $86,950, $349,472, $1,539,180 ¤ No difference if pts on anEbioEcs

So Much Pain Over So Many Regula9ons: Has it Made a Meaningful Difference? Payment & Repor9ng Policies & HAI Impact

¤  10/2008 CMS eliminates payment for HAC (CVC BSI one example) ¤  1/2011 CMS requires all hospitals parEcipaEng in IPPS to report CVC BSI to

CDC NHSN ¤  32 states & Washington DC mandate CVC BSI reporEng ¤  Oregon ICU study; external validaEons & adjudicaEon á publicly reported

rates 27% ¤  “ReacEve measure” – measure that modifies phenomenon under study &

changes thing being measured ¤  “Shame & financial penalty” incents â sensiEvity ¤  2 studies of CMS nonpayment policy found NO measurable impact on CVC

BSI or other HAI rates & no difference in CVC BSIs in hospitals in voluntary or mandatory reporEng states

Krein SL et al. JGIM. 2012;27(7):773-‐779. Lee GM et al. NEJM. 2012;367(15):1428-‐1437.

Dixon-‐Woods & Perencevich ICHE. 2013;34(6):555-‐557.

Effect of Nonpayment for Preventable Infec9ons in U.S. Hospitals

¤ NHSN Data 2006 – 2011, CLABSI and CAUTI vs VAP ¤ Quasi-‐exp, interrupted Eme series, 398 hospitals ¤ â secular trends for CLABSI, CAUTI & VAP LONG BEFORE POLICY IMPLEMENTED

¤ No change in rates post vs pre CLABSI (IRR 1, p=.97), CAUTI (IRR 1.03, p=0.08), VAP (IRR 0.99, p = .52)

¤ No difference in mandatory reporEng states, or by volume, size, type ownership, teaching hospital

Lee GM et al. NEJM. 2012;367:1428-‐1437


Fraser & Voss 10

Effect of Nonpayment for Preventable Infec9ons in U.S. Hospitals

Lee GM et al. NEJM. 2012;367:1428-‐1437

Figure 1. Incidence Rates of Infec9ons Reported by Hospital Units between January 2006 and March 2011. The dashed line in all three panels indicates the Eming of implementaEon of the Centers for Medicare and Medicaid Services policy, in October 2008.

Effect of Nonpayment for Hospital-‐Acquired, Catheter-‐Associated Urinary Tract Infec9on

¤ Retro Before-‐ Aqer, HCUP inpt data from MI; 2007 & 2009

¤ Non-‐CAUTI ² 5.2 – 17.1% (mean 10%, CI 9.5 – 10.5%) 2007 ² 5 – 20% (mean 10.3%, CI 9.8 – 10.9%) 2009

¤ CAUTI ² 0 – 1.1% (mean 0.09%, CI 0.06 – 0.12%) 2007 ² 0 – 0.95% (mean 0.14%, CI 0.11 – 0.17%) 2009

¤ 2009, 2.6% (CI 1.6 – 3.6%) HA UTIs coded as CAUTI ¤ Nonpayment for CAUTI only â payment (0.003%)

HospitalizaEons Meddings JA et al. Ann Intern Med. 2012;157:305-‐312

Effect of Nonpayment for Hospital-‐Acquired, Catheter-‐Associated Urinary Tract Infec9on

Meddings JA et al. Ann Intern Med. 2012;157:305-‐312

â

Rates of hospital-‐acquired non-‐CAUTIs and CAUTIs in 2009 and change in rates from 2007 to 2009. A hospital's rate of diagnosis was calculated as the percentage of each hospital's discharges of adults with the indicated diagnosis. CAUTI = catheter-‐associated urinary tract infec9on.

Figure Legend:

ON MY HANDS



Fraser & Voss 11

Scheithauer et al. BMC InfecEous Diseases 2013, 13:367

¤ 378 paEent cases were evaluated with 5674 opportuniEes for hand rubs (HR) and 1664 HR performed.

¤ Compliance increased from 21% to 29%, and finally 45%

¤ IntervenEons were aimed at increasing compliance as well as reducing the number of HR needed by improving workflow prac9ces.


For individual paEent care, the number of HH moments significantly decreased from 22 to 13 for non-‐surgical and from 13 to 7 for surgical paEents (both p<0.001)


¤  EvaluaEng and improving the workflow is an important (and oqen forgoUen) intervenEon to improve HH compliance in our “overloaded” HCWs!

Kirkland et a. BMJ Qual Saf 2012;21:1019–1026

¤ IntervenEons ² (1) leadership/accountability; (2) measurement/feedback; (3) hand saniEser availability; (4) educaEon/training; (5) markeEng/communicaEon

¤ Results ² HH compliance increased significantly from 41% to 87% (p<0.01), and improved further to 91% (p<0.01) the following year.

² Nurses achieved higher HH compliance (93%) than physicians (78%).

² There was a significant, sustained decline in the HAI-‐rate from 4.8 to 3.3 (p<0.01) per 1000 inpaEent days.

Kirkland et a. BMJ Qual Saf 2012;21:1019–1026


Fraser & Voss 12

Nothing new,

but s9ll nice to

see that HH

works ..

Kirkland et a. BMJ Qual Saf 2012;21:1019–1026 Allegranzi et al. Lancet Infect Dis 2013, August 23rd

Works gloabbly, too..

¤ Overall compliance increased from 51·∙0% before the intervenEon to 67·∙2% aqer.

¤  Compliance was independently associated with gross naEonal income per head, with a greater effect of the intervenEon in low-‐income and middle-‐income countries (OR 4·∙67).

Stone et al. BMJ 2012;344:e3005

¤ InvesEgate the associaEon between infec9ons and procurement

¤ 187 acute trusts in England and Wales ¤ Combined procurement of soap and alcohol hand rub tripled from 21.8 to 59.8 mL per paEent bed day

¤ Rates fell for MRSA bacteraemia (1.88 to 0.91 cases per 10 000 bed days) and C. difficile infecEon (16.75 to 9.49 cases). MSSA bacteraemia rates did not fall.


Aqer roll-‐out increase in

soap > alcohol


¤ Increased procurement of soap was independently associated with reduced C. difficile infecEon

¤ Increased procurement of alcohol hand rub was independently associated with reduced MRSA bacteraemia (delayed effect)



Fraser & Voss 13

¤ Q: Why is the overall compliance (for both groups) decreasing over Eme?

¤ A: ² wearing off of the novelty of cleanyourhands as other quality iniEaEves were introduced

² performance of the ward co-‐ordinators had not been monitored and no retraining offered

hUp://www.aricjournal.com/supplements/2/S1

¤  1600 stock-‐photos were evaluated.

¤  Most common mistakes were with regard to HCWs white coats and uniforms

¤  Of the photos ²  displaying doctors 89% were incorrect

²  displaying nurses 31% were incorrect

Spierings et al. P141, hUp://www.aricjournal.com/supplements/2/S1 Spierings et al. P141, hUp://www.aricjournal.com/supplements/2/S1

Conclusion ²  The results seem to reflect the real world with only

40% of stock photos displaying correct behavior and doctors shown as being worse than nurses.

²  It seems that the stereotype image of a doctor does not agree with the current hand hygiene guidelines.

²  If we aim for higher compliance rates with IC measures, we need to change the social image of HCWs

John A. Bergh


Fraser & Voss 14

¤  TV/adverEsing is expressly directed at ge�ng us to do something that is in the best interests of the adverEser, but not necessarily our own.

¤ Elementary school children see an average of 15 TV food ads per day ² 98% of these ads promote products high in fat and sugar).

¤ Children exposed to food ads during a cartoon ate significantly more of the snack food in front of them (45% more!)

We need ads

with

“correctly

behaving”

HCWs

on TV

Makary JAMA, April 17, 2013—Vol 309, No. 15 1591

Same system being used to evaluate OR Eme-‐out, compliance with isolaEon measures, …

The Wallstreet Journal 12 December 2012 Gustafson et al. Mayo Clin Proc 2000;75:705-‐8

Sheldon Cooper, The Big Bang Theory


Fraser & Voss 15

BACKGROUND: ¤  Minimal research has been published evaluaEng the effecEveness of

hand hygiene delivery systems (ie, rubs, foams, or wipes) at removing viruses from hands.

METHODS: ¤  Hands of 30 volunteers were inoculated with H1N1 and randomized

to treatment with foam, gel, or hand wipe applied to half of each volunteer's finger pads.

RESULTS: ¤  Treatments with all products resulted in a significant reducEon in

viral Eters (>3 logs) at their respecEve exposure Emes that were staEsEcally comparable.

Larson et al. Am J Infect Control 2012;40:806

Cleanliness is Next to Godliness

Effect of Daily Chlorhexidine Bathing on Hospital-‐Acquired Infec9ons

Climo MW et al. NEJM. 2013;368:533-‐42

¤ MulEcenter, cluster-‐randomized nonblinded crossover trial

¤ Daily bathing CHG-‐impregnated washcloths ¤ 9 ICU’s & BMT’s in 6 hospitals (7,727 pts), no-‐rinse 2% CHG cloths vs non-‐anEmicrobial cloths x 6 mos

¤ IR of MDRO & HA-‐BSI compared Poisson regression ¤ MDRO acquisiEon 5.6/1000 pt days vs 6.6/1000 pt days (p = 0.03); 23% lower CHG

¤ HA-‐BSI 4.78/1000 pt days vs 6.60/1000 pt days (p = 0.007); 28% lower CHG

Effect of Daily CHG Bathing on HAIs Caveats

¤ Study interrupted by recall of CHG cloths due to Burkholderia cepacia contaminaEon

¤ AcEve surveillance for MRSA & VRE; isolates submiUed for CHG resistance

¤ No Δ in MRSA acquisiEon ¤ â rates fungal CA-‐BSI ¤ Emergence of high level CHG resistance not seen, low toxicity



Fraser & Voss 16

Effect of Daily CHG Bathing on HAIs


Figure 2. Rates of Primary Bloodstream Infec9ons According to the Type of Hospital Unit. Incidence rates of hospital-‐acquired primary bloodstream infecEons are shown among units using daily bathing with either chlorhexidine-‐impregnated washcloths or nonanEmicrobial washcloths (control). BMT denotes bone marrow transplantaEon unit, MICU medical intensive care unit, and SICU surgical intensive care unit.

Effect of Hospital-‐Wide Chlorhexidine Pa9ent Bathing on Healthcare-‐Associated Infec9ons

¤ Hibiclens 4% CHG ¤ Monitored CLABSI, CAUTI, VAP, VRE, MRSA, CDI ¤ “Horizontal” infecEon prevenEon ¤ Ease of use, broad spectrum, prolonged residual effect

Rupp ME et al. ICHE. 2012;33(11):1094-‐1100


¤ Quasi-‐experimental, staged, dose-‐escalaEon x19 mos ¯ˉ𝑐  4 mo washout, 3 cohorts (2008-‐2010)

¤ Academic center, NE, all pts except infants/neonates ¤ CHG basin baths 3x/week or daily ¤ Adherence ICU (90%) vs (57.7%) non-‐ICU p = < .001 ¤ C diff â all cohorts 0.71 (95% CI, 0.57-‐0.89; p = .003) 3x/wk & .041 (95% CI, 0.29-‐0.59; p = .001) daily CHG

¤ Washout 1.85 (95% CI, 1.38-‐2.53; p = < .001)

Rupp ME et al. ICHE. 2012;33(11):1094-‐1100


Rupp ME et al. ICHE. 2012;33(11):1094-‐1100

Figure 1. Effect of chlorhexidine gluconate (CHG) bathing on Clostridium difficile infecEon. Trends in incidence of C. difficile infecEon are shown for the 3 cohorts of paEents over the course of the study. The long-‐dashed line depicts the 3-‐days-‐per-‐week bathing period. The solid line starEng aqer the 3-‐days-‐per-‐week bathing period in each cohort depicts the every-‐day CHG bathing period. The short-‐dashed line indicates the washout period. pt d, paEent-‐days.


Fraser & Voss 17

The Efficacy of Daily Bathing with Chlorhexidine for Reducing HAI BSIs: A Meta-‐analysis

¤ Similar efficacy cloth or liquid ¤ Wipes: OR= 0.41 [95% CI, 0.25-‐0.65], ¤ All others: OR=0.47 [95% CI, 0.31-‐0.69]) ¤ Similar sensiEvity CLABSI (OR 0.40 [95% CI, 0.27-‐0.59]),

¤ All BSI (OR 0.46 [95% CI, 0.31-‐0.69]) ¤ Greatest evidence in MICUs, single SICU no benefit, no benefit GNR, Heterogeneous studies

O’Horo JC et al. ICHE. 2012;33(3):257-‐267 O’Horo JC et al. Infect Control Hosp Epidemiol. 2012;33(3):257-‐267

Figure 3. Risk of healthcare-‐associated bloodstream infecEon (BSI) with chlorhexidine (CHG) bathing and comparator, using paEent-‐days in the analysis. “Events” refers to the study end point of central line–associated BSI or BSI, as defined in Table 1. Studies using a CHG-‐impregnated cloth are listed in the lower subgroup (1.2.2); all other studies are listed on top (1.2.1). CI, confidence interval; M-‐H, Mantel-‐Haenszel.

Scary Hospital Outbreaks

Hospital Outbreak of MERS Coronavirus

¤ WHO reported iniEal 2 cases 9/2012 MERS-‐CoV ¤ Saudi Arabia, Qatar, Jordan, UK, Germany, France, Tunisia and Italy

¤ Novel lineage C – MERS-‐CoV ¤ 4/1/2013 – 5/23/13 = 23 confirmed & 11 probable, single monophyleEc clade cases in the eastern province of Saudi Arabia

Assiri A et al. NEJM. 2013;369:407-‐16


¤ Median age 56, most male ¤ Signs/symptoms: fever 87%, cough 89%, vomiEng or diarrhea 35%

¤ Onset ² ICU: median 5 days (1 – 10 d) ² MV: median 7 days (3 – 11 d) ² Death: median 11 days (5 – 27 d)



Fraser & Voss 18


¤ Survival 3/4 (75%) acEve surveillance vs 3/19 (16%) clinically idenEfied (p = 0.04)

¤ IncubaEon 5.2 d (95% CI, 1.9 -‐ 14.7 d) ¤ Person-‐to-‐person transmission in HD units, ICUs, inpt units in 3 faciliEes, 21/23 cases, 5 family members, 2 HCWs

¤ CP & droplet precauEons, surveillance & IC criEcal


Hospital Outbreak of Middle East Respiratory Syndrome Coronavirus


Figure 1 Epidemiologic Plot of Confirmed and Probable Cases of MERS-‐CoV InfecEon in Saudi Arabia, April 1–May 23, 2013. All confirmed and probable cases are shown, according to the locaEon of the most probable transmission. One of the five family contacts (PaEent M) who is included as having been exposed in Hospital A was also exposed through caring for the paEent at home and may have acquired the infecEon either in the hospital or in the community.

Hospital Outbreak of Middle East Respiratory Syndrome Coronavirus


Figure 2 Transmission Map of Outbreak of MERS-‐CoV InfecEon. All confirmed cases and the two probable cases linked to transmission events are shown. PutaEve transmissions are indicated, as well as the date of onset of illness and the se�ngs. The leUers within the symbols are the paEent idenEfiers (see Fig. S2 in the Supplementary Appendix).

Innova9ve Interven9ons to Reduce HAIs

Beyond the bundle – journey of a ter9ary care MICU to zero CLABSIs

¤ ObservaEonal cohort, 25 bed MICU, 1/2008 – 12/2011 ¤ MulEdisciplinary team; bundle, inserEon checklist,

demonstraEon of competencies for line maintenance & access, daily CL necessity checklist, quality rounds, surveillance & feedback

¤ Molecular epi, environmental Cx & cleaning, â VRE contaminants

¤ CHG bathing; RCA of all CLABSI ¤  IntervenEons to â contaminants needed to get to zero

Exline MC et al. CriEcal Care. 2013;17:R41


Fraser & Voss 19

Beyond the bundle – journey of a ter9ary care MICU to zero CLABSI

Exline MC et al. CriEcal Care. 2013;17:R41

Figure 1. Central line-‐associated bloodstream infecEons, compliance with central line inserEon and dressing maintenance during the study period. NHSN, NaEonal Health Safety Network.

Targeted versus Universal Decoloniza9on to Prevent ICU Infec9on

¤ PragmaEc cluster RCT: 1.  MRSA screening & isolate MRSA + 2.  Targeted decolonizaEon (screening, isolaEon &

decolonizaEon of MRSA carriers) 3.  Universal decolonizaEon (no screening, decolonize all)

¤ 43 hospitals, 74 ICUs, 74,256 pts randomized ¤ (BIG, Just amazing to implement)

Huang SS et al. NEJM. 2013;368:2255-‐2265


¤ 12 mo baseline 1/1/09 – 12/31/09; phase-‐in 1/1/10 – 4/7/10, 18 mo intervenEon 4/8/10 – 9/30/11

¤ Primary outcomes: ICU aUributable MRSA ⊕ Cx, ¤ 2° outcome: ICU aUributable MRSA BSI & all BSI ¤ Designed 80% power to detect 40% â in MRSA BSI rate in grp 2, & 60% â grp 3; ITT



¤ Grp 1: < 1.0% got mupirocin or CHG ¤ Grp 2: 90.8% (56-‐100%) MRSA carriers got mupirocin & 88.8% (54-‐98.4%) got CHG

¤ Grp 3: 86.1% (41-‐99.1%) got mupirocin & 80.8% (53.1-‐98.6%) got CHG (highest baseline BSI rate, BMT, Tx)

¤ Grps similar @ baseline; 7 adverse rash events





Fraser & Voss 20


¤ Universal decolonizaEon most effecEve â MRSA clinical Cx 37% & all BSI 44%

¤ Strengths: sample size, diverse se�ngs, usual pracEce, real world

¤ Need to decolonize 181 pts to prevent 1 + MRSA Cx & decolonize 54 to prevent 1 BSI

¤ Why did it work this way? ¤ 1) started on Day 1 no delay, 2) â environmental burden, 3) â skin colonizaEon


hUp://www.slideshare.net/iPrevent/voss-‐icaac-‐online Leverstein-‐van Hall et al. Lancet Infect Dis 2011;10:830

Emerging InfecEous Diseases • www.cdc.gov/eid • Vol. 19, No. 8, August 2013 Emerging InfecEous Diseases • www.cdc.gov/eid • Vol. 19, No. 8, August 2013

Where do I need to go …


Fraser & Voss 21

slide from Sunita Paltansing

Central A & Carabians: 25%

South America: 6%

North Afria: 40%

Middle Afria: 30%

South Afria: 12%

Middle East: 13%

Central Asia: 30%

South-‐east Asia: 34%

East Asia (China): 67%

South-‐Asia (India): 72%

Overall: 31% ESBL+ aqer travel Emerging InfecEous Diseases • www.cdc.gov/eid • Vol. 19, No. 8, August 2013

¤  Although 26 parEcipants had posiEve results for ESBL-‐E 6 months aqer travel, they were not all posiEve for the same enterobacterial strain that was idenEfied immediately aqer travel. ²  15 of 26 (57%) different

¤ What does that mean with regard to isolaEon/(de-‐)flagging of paEents?

AE Andersson, et al. AJIC 2012, Jan 28 epublished

¤  High levels of CFU correlated with total traffic flow per operaEon and the number of persons in the OR

¤  Traffic flow, number of persons present, & procedure duraEon explained 68% of the variance in total CFU


¤ 177 (33.5%) = necessary ²  40 = expert consultaEons ²  137 = supplies & equipment

¤ 184 (35.7%) = semi-‐necessary ²  76 = surgical team members entering or leaving ²  134 = breaks

¤ 168 (31.8%) = unnecessary ²  30 = logisEcs, like planning other operaEons / ²  45 = social ²  93 = no detectable reason



Fraser & Voss 22

¤ 77 (13.8%) = necessary ²  40 = expert consultaEons ²  37 = supplies & equipment

¤ 76 (13.6%) = semi-‐necessary ²  76 = surgical team members entering or leaving

¤ 402 (72.4%) = unnecessary ²  134 break , 100 supplies & equipment

²  30 = logisEcs, like planning other operaEons / ²  45 = social ²  93 = no detectable reason


¤ OperaEng-‐suit aqer 4-‐8h the worst ¤ No clothing – (no) shedding

Hill et al. Lancet , November 9, 1974

“Naked below the elbow” really works

Merollini et al. AJIC 2013 in press

Methods ¤ Baseline use of anEbioEc prophylaxis (AP) was compared with no anEbioEc prophylaxis (no AP), anEbioEc-‐impregnated cement (AP + ABC), and laminar air operaEng rooms (AP + LOR).

¤ A Markov model was used to simulate long-‐term health and cost outcomes of a hypotheEcal cohort of 30,000 total hip arthroplasty paEents.


Conclusion ¤ PrevenEng deep SSI with anEbioEc prophylaxis and anEbioEc-‐impregnated cement has shown to improve health outcomes among hospitalized paEents, save lives, and enhance resource allocaEon.

¤ Based on this evidence, the use of laminar air opera9ng rooms is not recommended.



Fraser & Voss 23

Gastmeier et al. . J Hosp Infect 2012;81:73-‐78

Conclusions: It would be a waste of resources to establish new operaEng rooms with LAF, and quesEonable as to whether LAF systems in exisEng operaEng rooms should be replaced by convenEonal venElaEon systems

Bischoff et al. J Infect Dis 2013 Jan 30

Bischoff et al. J Infect Dis 2013 Jan 30

¤  Subjects with influenza-‐like symptoms

¤  QuanEtaEve impact air samples

¤  43% of subjects emiUed influenza-‐virus ²  19% super-‐spreaders (32x more than others)

¤  Emission >50% of human infecEous dose at 1, 3, and 6 feet distance

C. Makison Booth et al. J Hosp Infect 2013

Methods

A dummy test head aUached to a breathing simulator was used to test the performance of surgical masks against a viral challenge. …


Findings

Live influenza virus was measurable from the air behind all surgical masks tested. A surgical mask will reduce exposure to aerosolised influenza virus; reducEons ranged from 1.1-‐ to 55-‐fold (average 6-‐fold), depending on the design of the mask.



Fraser & Voss 24

Conclusion

The results show limitaEons of surgical masks in this context, although they are to some extent protecEve.


Might there be a difference if the mask is molded or not?

Gedik et al. AnEmicrobial Resistance Infect Control 2013;2:22

Bacterial ContaminaEon of an Automated Pharmacy Robot Used for Intravenous MedicaEon PreparaEon

Cluck et al. ICHE 2012;33:517-‐520

no need for humans to cause outbreaks

3 isolates from the robot and 3/6 isolates from lidocaine dispensed by the robot had idenEcal B. cereus isolates.

Cluck et al. ICHE 2012;33:517-‐520


Fraser & Voss 25

TO ERR IS HUMAN, but to really foul things up you need a robot

TO ERR IS HUMAN, to blame it on

someone else shows management potenEal

Cluck et al. ICHE 2012;33:517-‐520

Policing One-‐track mind

Guidelines

Guidelines

Guidelines

Guidelines

More guidelines

IC needs to be to-‐the-‐point



Fraser & Voss 26

¤ Clinical MRSA isolates from 30/31 Orange Co., CA hospitals ² 10/08 – 4/10, ER excluded, 100/hospitals or 12 mo, + up to 20 blood isolates/mo

¤ spa type t008 = CA-‐MRSA, sample got MLST & PFGE ¤ 46% isolated spa t008 (CA-‐MRSA) (range 14%-‐81%), next most common t002 (15%), t242 (21%)

¤ spa t008 was USA300 by PFGE ¤ Of CA-‐MRSA, 66% wounds, 14% respiratory, 9% other, 8% blood, 3% urine, also 37% of HO-‐MRSA

Murphy CR et al. ICHE. 2013;34:581-‐587

Murphy CR et al. Infect Control Hosp Epidemiol. 2013;34:581-‐587

Figure 1. Percentage of all isolates that were community-‐associated methicillin-‐resistant Staphylococcus aureus (CA-‐MRSA; gray) and percentage of isolates that were CA-‐MRSA and associated with hospital-‐onset infecEon (black), by hospital. Hospitals that collected fewer than 50 isolates are marked with a star; fewer than 20 isolates were collected from hospitals 3, 4, 6, 7, and 30.

¤ 6 PA hospitals, clinical research data 2007-‐08 (MedMined/MediQual)

¤ ⊕ toxin > 48 hrs ¯ˉ𝑝  admit, > 8 weeks ¯ˉ𝑝  previous ⊕

¤ 1:3 matching ¯ˉ𝑐  non-‐cases, HO-‐CDI had higher mortality (11.8% vs 7.3%, p<.05), longer LOS (median interquarEle range 12 days (9-‐21) vs 11 days (11.059-‐38.429) p<0.01), higher cost (median interquarEle $20,804 ($11,059-‐$38,429) vs $16,634 ($9,413-‐$30,319) p<.01)

¤ AUributable effect HO-‐CDI 4.5% mortality (95% CI 0.2-‐8.7% p<.05), 2.3 days (95% CI 0.9-‐3.8 p<.01), $6,117 ($1,659-‐$10,574 p<.01)

Tabak TP et al. ICHE. 2013;34:588-‐596

¤ NaEonal burden esEmates based on 2009 HO-‐CDI data

¤ 216,000 acute care discharges ¯ˉ𝑐  CDI (HO-‐CDI ~ 65%) ² 140,000 HO-‐CDI discharges 2009 (Epicenters)

¤ 300,000 á hospital days, > $850m á costs, > $6,000 deaths/year

¤ Strengths: 90% matching propensity scores ¤ Limits: retrospecEve, generalizable, toxin tests



Fraser & Voss 27

Tabak TP et al. Infect Control Hosp Epidemiol. 2013;34:588-‐596

Figure 1. A, Mortality rate among Clostridium difficile infecEon (CDI) cases versus noncases before and aqer matching. B, Mortality or postdischarge care rate among CDI cases versus noncases before and aqer matching. C, Average length of stay among CDI cases versus noncases before and aqer matching. D, Cost per case among CDI cases versus noncases before and aqer matching.

¤  Kyne: Cost $3,669, LOS 3.6 days, no á mortality @ 3 or 12 mo (only infected pts, HO-‐CDI). ¤ CID 2002;34:346-‐353.

¤  Dubberke: Cost $2,454-‐$7,179/case (LR vs propensity-‐matched prs) at end of admit or 180 days, LOS 2.8 days, readmission 19.3%, death 5.7% (nonsurgical pts, all CDI). ¤  ICHE 2009;30:57-‐66, Emerg Infect Dis 2008;14:1031-‐1038, ¤ CID 2008;46:497-‐504.

¤  O’Brien: Cost $13,675, LOS 2.95 days (MA hosp admin data, 20 discharge dx only, not 1°. ¤  ICHE 2007;28:1219-‐1227.

¤  Miller: Endemic mortality 1-‐2%, epidemic mortality 7-‐17%. ¤ CID 2010;50:194-‐201.


Shepard J et al. JAMA Surg. doi:10.1001/jamasurg.2013.2246

¤ Methods – RetrospecEve, 4 JH hospitals, record review collect APR-‐DRG by ICP; 2007-‐2010

¤ Results ² SSI $7,493 vs $7,924 (p=.99) Cost ² 10.56 d vs 5.64 d (p<.001) LOS ² 51.94 vs 8.19/100 (p<.001) Readmissions

Shepard J et al. JAMA Surg. doi:10.1001/jamasurg.2013.2246 Shepard J et al. JAMA Surg. doi:10.1001/jamasurg.2013.2246


Fraser & Voss 28

Shepard J et al. JAMA Surg. doi:10.1001/jamasurg.2013.2246 Shepard J et al. JAMA Surg. doi:10.1001/jamasurg.2013.2246

Figure 1. Mean Daily Total Charges for a Pa9ent vs the Length of Stay for the Pa9ent

¤  JHH Pilot Study 20 pts VRE & other MDROs, cultured 5 pairs of “sterile supplies” from room @ discharge; & H2O2 vapor

¤  7/100 supplies VRE⊕, 4/20 rooms (20%), H2O2, none H2O2 (p=.014)

¤  9/100 supplies MDRO⊕ , 6/20 rooms (30%), H2O2, none H2O2 (p=.003)

¤  50% recovered organisms DID NOT match pt isolate ¤  ~ Direct annual cost discarded supplies $387,055 ¤  Can disinfect supplies in rooms undergoing H2O2 vapor

disinfecEon OUer JA et al. ICHE. 2013;34:472-‐478

OUer JA et al. ICHE. 2013;34:472-‐478

BeUer Methods to Clean & Disinfect the Environment


Fraser & Voss 29

¤ Metallic Cu⧺ intrinsic broad spectrum anEmicrobial acEvity

¤ In vitro Cu⧺ surfaces â bacterial concentraEons 7 logs in 2 hrs

¤ MUSC, MSKCC & Ralph Johnson VA ¤ 8 Cu⧺, 8 std rooms, 650 admissions 6/2010 – 7/2011 ¤ Cu ⧺ bed rails, overbed tables, IV poles, arm chairs, call buUon, mouse, computer palm rest, bezel touch screen monitor

¤ Weekly cultures Salgado CD et al. Infect Control Hosp Epidemiol. 2013;34:479-‐486

¤ Bivariate analysis: á APACHE II score assoc á HAI & colonizaEon (p=.011)

¤ Bivariate: infecEon on admission; HAI 16.6 vs 5.7 p=.047 non-‐Cu⧺ vs Cu⧺

¤ MV analysis: APACHE II score (p=.011) and Cu⧺ (p=.027)

¤ Significant associaEon env bioburden & HAI ¤ Bioburden 17% vs 50% (0.76 log â p<.0001) Cu⧺ ¤ Foot-‐board bioburden similar (2,786 vs 2,388 CFU/100cm2) No Cu⧺ Salgado CD et al. ICHE. 2013;34:479-‐486

¤ 53.4% of pts in Cu⧺ rooms had at least 1 object removed (non-‐study bed)

¤ 13.4% non Cu⧺ rooms exposed to Cu⧺ chair ¤ Caveats, cleaning, tarnishing, impact on different organisms, environments, cost benefit

Salgado CD et al. ICHE. 2013;34:479-‐486

Salgado CD et al. Infect Control Hosp Epidemiol. 2013;34:479-‐486 Salgado CD et al. Infect Control Hosp Epidemiol. 2013;34:479-‐486

Figure 2. QuarEle distribuEon of healthcare-‐acquired infecEons (HAIs) straEfied by microbial burden measured in the intensive care unit (ICU) room during the paEent’s stay. There was a significant associaEon between burden and HAI risk ( ), with 89% of HAIs occurring among paEents cared for in a room with a burden of more than 500 colony-‐forming units (CFUs)/100 cm2.


Fraser & Voss 30

Anderson DJ et al. ICHE Epidemiol. 2013;34:466-‐471

¤ Methods ² Env cultures for VRE, C. diff, Acinetobacter before and

aqer UV Rx

² 2 hospitals, 39 rooms of pts colonized ¯ˉ𝑐  VRE, C. diff, Acinetobacter (Tru-‐D SmartUVC; Lumalier)

¤ Results – UV-‐C ² Any organism (1.07 log10 â p<.0001) ² Target pathogen (1.35 log10 â p<.0001) ² VRE (1.68 log10 â p<.0001) ² C. diff (1.16 log10 â p<.0001) ² Acinetobacter (1.71 log10 â p=.25)

Anderson DJ et al. ICHE. 2013;34:466-‐471

Figure 1. Change in proporEon of posiEve plates for target organisms before and aqer use of an automated ultraviolet-‐C emiUer.

Sitzlar B et al. ICHE. 2013;34:459-‐465

¤ Methods ² 3 sequenEal intervenEons

²  Fluorescent markers to monitor & feedback on cleaning ²  Automated UV adjuncEve disinfecEon ²  Enhanced disinfecEon, dedicated team, supervision & clearance

² Cleveland VA, 21 months


¤ Results ² Fluorescent marker improved cleaning thoroughness

(47-‐81%, p<.0001) ² ⊕ Cx â 14% (p=.024), 48% (p<.001), 89% (p=.006),

(intervenEons 1, 2 & 3) ² Baseline 67% CDI rooms had ⊕ Cx aqer disinfecEon vs

57%, 35% & 7% (intervenEons 1, 2 & 3)

² 35% of CDI rooms had ⊕ Cx ¯ˉ𝑝  UV treatment ² Daily disinfecEon, dedicated team (Clorox Germicidal

Wipes) & cleaning/supervision of cleaning ATP bioluminescence (CleanTrace; 3M)


Fraser & Voss 31

Sitzlar B et al. Infect Control Hosp Epidemiol. 2013;34:459-‐465

Figure 1. Effect of sequenEal environmental cleaning and disinfecEon intervenEons on thoroughness of cleaning (determined on the basis of fluorescent marker removal) and on disinfecEon of Clostridium difficile infecEon (CDI) rooms (determined on the basis of environmental cultures for C. difficile). IntervenEon 1 (January 1, 2011, through February 28, 2012; 14 months) involved educaEon in combinaEon with monitoring of fluorescent marker removal from high-‐touch surfaces with feedback to housekeepers; intervenEon 2 (March 1, 2012, through June 30, 2012; 4 months) included addiEon of an automated ultraviolet radiaEon device for disinfecEon of CDI rooms; intervenEon 3 (July 1, 2012, through September 30, 2012; 3 months) included enhanced standard cleaning through formaEon of a 3-‐person dedicated daily disinfecEon team for high-‐touch surfaces in CDI rooms and implementaEon of a process requiring that terminally cleaned CDI rooms be “cleared” for the next paEent by environmental services supervisors and/or infecEon control staff. Each intervenEon was divided into 3 Eme periods, which are indicated by separate bars.


Figure 2. Improvement in thoroughness of cleaning of high-‐touch surfaces with the fluorescent marker intervenEon.

Abbo LM et al. Clin Infect Dis. 2013;57:631-‐638

¤ Methods – Cross-‐secEonal mulEcenter electronic survey of 4th year med students’ knowledge, a�tudes & percepEons re: anEmicrobial use & resistance, quanEty & quality of educaEon on AB Rx; Miami, JH, U WA

¤ Results – 317/519 (60%) responded; 57% ♀, mean age 27, all had anEmicrobial stewardship programs


Medical Students’ Percep9ons and Knowledge About An9microbial Stewardship: How Are We Educa9ng Our

Future Prescribers?

¤ Differences in educaEonal resources used, perceived preparedness & knowledge

¤ 90% wanted more knowledge & educaEon, mean correct knowledge 51%, only 15% had done ID

¤ Those who did ID ranked quality of AB educaEon higher (3.93 vs 3.44, p=.0003), no difference in knowledge scores

¤ Only 1/3 reported “adequate” fundamental knowledge of anEmicrobial prescribing


“Everyone Else is Worse Than Me/Us”

¤ Students perceived anEbioEc overuse more naEonally than at their hospital

¤ Residents & Sr MDs agree “other doctors” overprescribe anEmicrobials compared to “themselves”; anEbioEcs overused “naEonally” compared to “their own pracEce”

Arch Intern Med 2002;162:2210-‐2216 Arch Intern Med 2004;164:1662-‐1668

ICHE 2011;32:714-‐718 ICHE 2006;27:1274-‐1277


Fraser & Voss 32

Abbo LM et al. Clin Infect Dis. 2013;57:631-‐638 Abbo LM et al. Clin Infect Dis. 2013;57:631-‐638




Fraser & Voss 33

Bacterial Infec9on as a Likely Cause of Adverse Reac9ons to Polyacrylamide Hydrogel Fillers in Cosme9c Surgery

Christensen L et al. Clin Infect Dis. 2013;56:1438-‐1444

¤ Widespread and á use of gel fillers for cosmeEcs ¤ á long-‐lasEng adverse reacEons (FDA reports tripled 2008-‐2011; 457 to 1309)

¤ Hyaluronic acid hydrogels, longer-‐lasEng collagen, RXN rates 4.5% -‐ 18.6%

¤ $22.5 billion filler market; incidence of gel injecEons not registered; # and brands of gel syringes used not public



¤ RXNS = lumps, nodules, swelling granulomas ¤ EEology unclear; Rx = anE-‐inflammatory agents, steroids, anEbioEcs

¤ MulEcenter case-‐control study; 59 paEents, 28 controls, biopsy & cytology, Cx, 16S rRNA, Gram stain, FISH



¤ Bacteria in 98% cases, (S epi, Propionibacterium) up to 5 years aqer injecEon, none in controls

¤ Long-‐term infecEons likely due to biofilms ¤ Sterile technique and adequate skin prep mandatory ? Explore anEbioEc prophylaxis ??? (Please don’t)

Grade 2 (A and B) and grade 3 (C and D) reac9ons in biopsies from lip and hand 9ssue seen ater 3 weeks and 1 year, respec9vely.

Christensen L et al. Clin Infect Dis. 2013;56:1438-‐1444 © The Author 2013. Published by Oxford University Press on behalf of the InfecEous Diseases Society of America. All rights reserved. For Permissions, please e-‐mail: [email protected].

Three-‐dimensional confocal laser scanning microscopy (CSLM) of a biopsy from a grade 3 reac9on following gel injec9on into the cheek 2 years previously.

Christensen L et al. Clin Infect Dis. 2013;56:1438-‐1444 © The Author 2013. Published by Oxford University Press on behalf of the InfecEous Diseases Society of America. All rights reserved. For Permissions, please e-‐mail: [email protected].

¤ Acknowledgements: ¤ Andreas Voss for pu�ng up with me ¤ Washington University ID Faculty & Fellows ¤ BJC & BJH InfecEon PrevenEon Specialists ¤ CDC, AHRQ, HHS, CMS, NIH and Industry for funding these great studies

¤ The InvesEgaEve Teams who did this research

Icaac 2013

Technology

clin infect

pcr shenoy

ann intern

ec isolates

ac9ve screening

2013 fraser

al plosone

pcr