Citation
Evaluation of Patient Risk Factors for Carbapenem-Resistant Enterobactericeae Infection

Material Information

Title:
Evaluation of Patient Risk Factors for Carbapenem-Resistant Enterobactericeae Infection
Creator:
Predic, Marko
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (65 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Epidemiology
Committee Chair:
PRINS,CINDY A
Committee Co-Chair:
COOK,ROBERT L
Committee Members:
IOVINE,NICOLE MARIE
Graduation Date:
12/17/2016

Subjects

Subjects / Keywords:
Antibiotics ( jstor )
Antimicrobials ( jstor )
Carbapenems ( jstor )
Comorbidity ( jstor )
Confidence limits ( jstor )
Hospital admissions ( jstor )
Hospitals ( jstor )
Infections ( jstor )
Point estimators ( jstor )
Predisposing factors ( jstor )
Epidemiology -- Dissertations, Academic -- UF
cre
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Epidemiology thesis, M.S.

Notes

Abstract:
Carbapenem-resistant Enterobacteriaceae (CRE) are a family of bacteria with high levels of resistance to many antimicrobials used in hospital settings today. We performed a retrospective chart review, from October 2014 to May 2016, in order to identify the risk factors contributing to CRE incidence among (i) cases versus controls, (ii) community acquired CRE infection and matched control infection, and (iii) hospital acquired CRE infection and matched control infection at UF Health Shands. The study consisted of 50 cases with positive CRE status and 100 controls with infections collected from similar anatomical sites as cases. Risk factors evaluated for significance were reasons for admission (planned versus emergency visit), antimicrobials, total length of stay, admitting location, discharge location, indwelling medical devices, and invasive procedures performed. Risk factors that were the most significant for cases versus controls were emergency admissions (OR 3.67; 95% CI 1.03-13.08), being administered inpatient medication (OR 5.88; 95% CI 2.27-15.24), admission from a long term care facility (Or 5.17; 95% CI 1.93-13.84), fluoroquinolone (OR 3.75; 95% CI 1.35-10.38), cephalosporin (OR 2.37; 95% CI 1.17-4.86), and invasive procedures with a scope (OR 4.57; 95% CI 1.31-16.02). Additional risk factors were significant for both community-acquired cases and hospital-acquired cases. Significant risk factors were inserted into logistic regression to develop a model to identify patients best suited for CRE screening upon admission. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2016.
Local:
Adviser: PRINS,CINDY A.
Local:
Co-adviser: COOK,ROBERT L.
Statement of Responsibility:
by Marko Predic.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Predic, Marko. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Classification:
LD1780 2016 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

EVALUATION OF PATIENT RISK FACTORS FOR CARBAPENEM RESISTANT ENTEROBACTERICEAE INFECTION By MARKO PREDIC A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2016

PAGE 2

2016 Marko Predic

PAGE 3

To my mom and dad for the unending encouragement and countless hours of support through my university career

PAGE 4

4 ACKNOWLEDGM ENTS I thank Dr. Cindy Prins for her encouragement, support, and aid through not only my thesis process but during my time in the Master of Science in Epidemiology program. I thank John Delano for the hours of assistance, data retrieval, and for introducin g and informing me about the field of infection control from the day I arrived at the department. I thank Elizabeth Tremblay for technological assistance, data retrieval, and willingness to help throughout my thesis process. I thank all the member of the I nfection Control department at UF Health Shands for all the encouragement I received and all the knowledge they shared with me during my time in the department. I thank my com m it tee members Dr. Nicole Iovine and Dr. Robert Cook, for aiding me in my thesis and providing the time to meet and assist me during this process. Finally, I want to thank all my friends who supported and encouraged me during this process.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 11 CHAP TER 1 GENERAL INFORMATION ................................ ................................ ..................... 13 Introduction ................................ ................................ ................................ ............. 13 Epidemiology ................................ ................................ ................................ .......... 13 Health Care Settings ................................ ................................ ............................... 16 Antimicrobial Stewardship ................................ ................................ ....................... 17 Comorbidity ................................ ................................ ................................ ............. 18 2 METHODS ................................ ................................ ................................ .............. 22 Study Population ................................ ................................ ................................ ..... 22 Factors ................................ ................................ ................................ .................... 23 Age ................................ ................................ ................................ ................... 23 Gender ................................ ................................ ................................ ............. 23 Infection and Colonization ................................ ................................ ................ 23 Admitting Locati on ................................ ................................ ............................ 24 Days until Positive Culture ................................ ................................ ................ 25 Length of Stay ................................ ................................ ................................ .. 25 Admitting Di agnosis ................................ ................................ .......................... 26 Total Hospital Admissions within Past Year ................................ ..................... 26 Indwelling Devices ................................ ................................ ............................ 27 Medication ................................ ................................ ................................ ........ 27 Comorbidity and Comorbidity Scores ................................ ............................... 28 Discharge Location ................................ ................................ ........................... 29 Invasive Surgeries and Invasive Surgeries with Scopes ................................ .. 30 Analysis ................................ ................................ ................................ ............ 30 3 RESULTS ................................ ................................ ................................ ............... 32 Demographics ................................ ................................ ................................ ......... 32 Cases vs. Controls ................................ ................................ ................................ .. 33 Community Acquired CRE ................................ ................................ ...................... 35

PAGE 6

6 Hospital Acquired CRE ................................ ................................ ........................... 37 4 DISCUSSION AND CONCLUSION ................................ ................................ ........ 49 Discussion ................................ ................................ ................................ .............. 49 Plan versus Emergency Admissions ................................ ................................ ....... 49 Ambulatory and/or Inpatient Antibiotics ................................ ................................ ... 49 Admitting Locatio n ................................ ................................ ................................ .. 50 Discharge location ................................ ................................ ................................ .. 51 Antibiotics ................................ ................................ ................................ ............... 51 Indwelling Devices ................................ ................................ ................................ .. 53 Invasive Procedures/ with Scope ................................ ................................ ............ 54 Limitations ................................ ................................ ................................ ............... 55 Applications of Find ings ................................ ................................ .......................... 56 LIST OF REFERENCES ................................ ................................ ............................... 60 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 65

PAGE 7

7 LIST OF TABLES Ta ble page 3 1 Demographics of Total Population ................................ ................................ ...... 39 3 2 Infection and Colonization by Community or Hospital Acquired Case s .............. 39 3 3 Infection and Colonization by Category for Cases ................................ .............. 39 3 4 Odds Ratio Emergency versus planned admissions ................................ .......... 40 3 5 Odds ratios Inpatient and Ambulatory Antibiotics ................................ ............... 40 3 6 Odds ratios admission from LTCF versus ACF ................................ .................. 40 3 7 Odds ratios individual admitting location ................................ ............................ 41 3 8 Odds ratios discharge to LTCF versus ACF ................................ ....................... 41 3 9 Odds ratios individual discharge location ................................ ........................... 42 3 10 Odds Ratio Antibiotics and Antibiotics Binary ................................ ..................... 42 3 11 Odds Ratio Indwelling Devices ................................ ................................ ........... 43 3 12 Odds Ratio Invasive Procedure & Invasive Procedures with Scope ................... 43 3 13 Odds Ratio Age ................................ ................................ ................................ .. 43 3 14 Odds Ratio Inpatient and Ambulatory Antibiotics (Community Acquired) ........... 43 3 15 Odds Ratio Binary Admission Location (Community Acquired) .......................... 44 3 16 Odds Ratio Individual Admitting Location (Community Acquired) ....................... 44 3 17 Odds Ration Antibiotics taken in Prior Month (Community Acquired) ................. 44 3 18 Odds Ratios Indwelling Devices (Community Acquired) ................................ ..... 45 3 19 Odds Ratio Age (Community Acquired) ................................ .............................. 45 3 20 Odds Ratio Total Length of Stay (Hospital Acquired) ................................ ......... 45 3 21 Odds Ratio Total Length of Stay Binary (Hospital Acquired) .............................. 45 3 22 Odds Ratio Admission Location (Hospital Acquired) ................................ .......... 45 3 23 Odds Ratio Admission Location Binary (Hospital Acquired) ............................... 46

PAGE 8

8 3 24 Odds Ratio Discharge Locations (Hospital Acquired) ................................ ......... 46 3 25 Odds Ratio Discharge Location Binary (Hospital Acquired) ............................... 46 3 26 Odds Ratio Antibiotics Taken in Prior Month (Hospital Acquired) ....................... 47 3 27 Odds Ratio Invasive Procedure & Invasive Procedures with Scope (Hospital Acquired) ................................ ................................ ................................ ............ 47 3 28 Odds Ratio Indwelling Devices (Hospital Acquired) ................................ ............ 48 3 29 Odds Ratio Age (Hospital A cquired) ................................ ................................ ... 48

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Patients with NDM CRE reported to the CDC as of April 2016 3 ......................... 20 1 2 Patients with OX 48 CRE reported to the CDC as of April 2016 3 ....................... 20 1 3 Patients with VIM CRE reported to the CDC as of April 2016 3 ........................... 21 4 1 Algorithm for Patients Arriving with Little Known Medical History ....................... 59 4 2 Algorithm for Patients Arriving with Known Medical History ............................... 59

PAGE 10

10 LIST OF ABBREVIATIONS ACF Acute care facilities are a category assigned to acute care hospit als, clinics, and homecare. ACH Acute care hospital is a hospital with patient population that stays on average less than 14 days per admission. BAL Bronchoalveolar lavage is a procedure where a bronchoscope is passed through the nose or mouth and a smal l amount of fluid is squirted into the lung for specimen collection. CRE Carbapenem resistant Enterobactericeae KPC LTACH Klebsiella pneumoniae carbapenemases Long term acute care hospital is a hospital with patient population that stays an average of 14 or greater days per admission. LTCF Long term care facilities are a category assigned to long term care hospitals, skilled nursing facilities, or hospice care. MRN Medical record numbers are personalized identifiers a patient receives upon admission to UF Health Shands. NDM New Delhi m etallo b eta l actamase OX 48 CRE producing OX48 like c arbapenemases VIM Verona integron encoded metallo beta lactamase

PAGE 11

11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EVALUATION OF PATIENT RISK FACTORS FOR CARBAPENEM RESISTANT ENTEROBACTERICEAE INFECTION By Marko Predic December 2016 Chair: Cindy Prins Major: Epidemiology Carbapenem resist ant Enterobacteriaceae (CRE) are a family of bacteria with high levels of resistance to many antimicrobials used in hospital settings today We performed a retrospective chart review, from Oct ober 2014 to May 2016, in order to identify the risk factors con tributing to CRE incidence among (i) cases versus controls, (ii) community acquired CRE infection and matched control infection, and (iii) hospital acquired CRE infection and matched control infection at UF Health Shands. The study consisted of 50 cases wi th positive CRE status and 100 controls with infections collected from similar anatomical site s as cases R isk factors evaluated for significance were reasons for admission (planned versus emergency visit), antimicrobials total length of stay, admitting l ocation, discharge location, indwelling medical devices, and invasive procedures performed. Risk factors that were the most significant for case s versus controls were emergency admissions (OR 3.67; 95% CI 1.03 13.0 8) being administered inpatient medicatio n (OR 5.8 8 ; 95% CI 2.2 7 15.24), admission from a long term care facility (Or 5.1 7 ; 95% CI 1.9 3 13.84), fluoroquinolone (OR 3.7 5 ; 95% CI 1.35 10.3 8 ), cephalosporin (OR 2.37; 95% CI 1.17 4.86), and invasive procedures with a scope (OR 4.57; 95% CI 1.3 1 16.0 2 ). Additional risk factors were significant for both

PAGE 12

12 community acquired cases and hospital acquired cases. Significant risk factors were inserted into logistic regression to develop a model to identify patients best suited for CRE screening upon admission.

PAGE 13

13 CHAPTER 1 GENERAL INFORMATION Introduction Carbapenem resistant Enterobacteriaceae (CRE) are a family of bacteria contributing to the $20 billion in direct excess healthcare costs and 23 000 deaths from antibiotic resistant organisms each year 26 The En terobacteriaceae family includes many bacteria found in the human intestines, such as Klebsiella species and Escherichia coli 1 CRE infections are rare in healthy individuals who do not have underlying health conditions, appearing mostly in individuals wh o reside within nursing homes, long term care facilities, hospitals, and other health care settings 1 CRE is mainly spread to new patients through contact transmission from an infected patient or healthcare worker 1 These bacteria are opportunistic, enteri ng the body through wounds or openings developed through the insertion of medical devices such intravenous catheters or urinary catheters 1 CRE are therefor e at the forefront of problems within our healthcare system yet no definitive criteria have been id entified for screening patients for CRE upon arrival at health care facilities. Most healthcare systems are less likely to view CRE as an emerging problem because it is difficult to evaluate trends due to the sporadic nature of cases 13 Multiple factors mu st be observed ranging from admitting location, procedures undergone at the healthcare facility, comorbidities, and antibiotic regiments to determine the prevalence of CRE infections along with a viable screening process for CRE within healthcare systems 16 Epidemiology CRE is derived from plasmid mediated resistance, which arises through the incorporation of a n extrachromosomal piece of DNA that carries antibiotic resistance 37

PAGE 14

14 These resistance plasmids tend to be conjugative, allowing for direct cell to cell DNA transfer, which is how the carbapenem resistance plasmid gets transferred to a formerly susceptible bacteri um 37 Wi thin the Unit ed States there are five carbapenem resistance plasmids that can potentially be incorporated within bacteria with th e most prevalent of all of the CRE being Klebsiella pneumoniae carbapenemases (KPC). KPC has become a problem in health care settings due to its ease of dissemination and limited antimicrobial treatment options 7 There has been an increase in the last deca de in the Carbapenem resistance reported in Klebsiella species with one percent having resistance in 2000, increasing to around eight percent of Klebsiella having resistance in 2007 7 KPC was first identified in the United States in North Carolina in 200 1 and first became endemic in New York and New Jersey, with the first outbreak outside the United States occurring in Israel in 2004 7 Through strain typing it was found that a single dominant strain (ST258) was responsible for the outbreaks in the United S tates, Israel, and later in India 7 38 Since the initial outbreak in the United States, a few fit lineages have managed to spread across the globe and KPC is now present in most of the developed w orld and developing countries. The next CRE is the New Del hi m etallo b eta l actamase (NDM) CRE which was first identified in a case in New Delhi, India in 2007 2 Since its discovery, NDM CRE has been reported on every continent except South America and Antarctica 2 One of the first cases reported in the United S tates occurred i n January 2012 when a woman who travelled to Cambodia and was hospitalized with spinal cord compression in Ho Chi Min City, Vietnam 2 Upon her return, she was immediately hospitalized with a diagnosis of lymphoma at her local hospital in Rhode Island and had a urine culture that was positive

PAGE 15

15 for NDM CRE During her stay on the hematology/oncology floor, it was found that one of the seven patients cohabitating the floor had also developed NDM CRE 2 The plasmid carrying NDM has been found to be highly transmissible to other bacteria and last s for 2 The recommendations to prevent spread that have been determined to be effective are high rates of hand hygiene, minimizing the us e of invasive medical devices, and preventing unnecessary antimicrobial exposure by using a robust antimicrobial stewardship program 2 Since identif ication of the first case in the United States in 2012 the NDM CRE has spread from the Northeastern US to a ll parts of the country with a total of 157 cases to date in the United States as of April 2016 2 (Figure 1 1). Another CRE that is tracked within the United States is CRE producing OX A 48 like Carbapenemases (OX A 48). The OX A 48 CRE was first identified i n Turkey in 2001 4 The first case of OX A identified until 2012 using retrospective review of cases 4 Of the 53 OX A 48 CRE identified by August 2015 it was shown that a majority (81%) were K. pneumoniae isolates 4 Among 35 patients who tested positive, those who provide d an age during infection were found to have a median age of 70 years (range 29 91) 4 Even with OX A 48 CRE occurring in clusters in the United States it is mostly believed to be contracted in patients who were admitted to health care facilities outside the US; 29 patients who tested positive provided travel histories and 19 (66%) had travelled internationally within the previous year of specimen collection 4 Since April 2016, 61 cases of OX A 48 CRE have been identified within the United States 3 (Figure 1 2).

PAGE 16

16 Another major CRE within the United States is Verona integron encoded metallo beta lactamase (VIM) CRE. The VIM CRE belongs to the Class B metallo lactamases group which also contains another of the five main CRE, the IMP CRE, which has very few identified cases in the United States and globally 6 The first reported case of VIM CRE within the United States was a women hospitalized after a Grecian cruise in July 2010 5 Upon f urther test ing of the strain obtained from this case it was identified to be non susceptible to all antimicrobials used to treat Klebsiella 5 Of 22 patients that cohabitated the same medical floor as the women no other patients tested positive for CRE 5 With only 17 cases VIM CRE currently has the lowest prevalence within the United States 3 (Figure 1 3) Health Care Settings CRE has become an urgent problem within the health care system. Between 2008 and 2014, the detection of CRE has increased five fold as reporte d by the CDC 8 Within the United States CRE has so far been isolated in 48 of the 50 states, excluding Idaho and Maine. CRE is of large concern because of the complications caused and the death rates associated with CRE infections. In 2013 the CDC reporte d that 3.9% of short stay acute care hospitals and 17.8% of long term acute care hospitals had reported at least one health care related CRE in the previous year 11 A study done in a Northeast Ohio healthcare system found that 75% of CRE patients were tran sferred in from a long term acute care hospital (LT A CH), with only one of 13 patients being discharged home after infection 12 In a meta analysis done on 356 articles published about CRE it was found that in seven studies that the death rate was 26% to 4 4%, and it was 3% to 4% in two studies done retrospectively 9 Death rates were two times

PAGE 17

17 higher in among patients who developed bacteremia with CRE as compared to those who had Carbapenem sensitive Enterobacteriaceae (CSE) 9 In Illinois, a study was perf ormed looking at rates of transmission between health care settings, acute care hospitals (ACH) and LTAC S to evaluate the rates of CRE as well as the frequency of transfer between hospitals that causes CRE within the receiving hospital. LT A CH were determi ned to play a central role in the distribution of CRE to other health care settings 10 A hospital that shared four or mor e patients with an LT A CH within a three month period was found to have an elevated CRE rate 10 Once the minimum four shared patients we re found within an ACH, the crude CRE rate was doubled 10 In a study conducted by Lee et al. measuring the prevalence of CRE with no organized control methods, unorganized control methods, and coordinated control methods, predicted nationwide prevalence of CRE would reach 11.1% with no organized control methods ( Fig. 1) 8 LTAC H would see the greatest impact after ten years with the prevalence reaching 28.9%, with the least prevalence being observed in ACH at 3.1% with no control methods 8 Antimicrobial S tewardship Anti microbial stewardship is another factor that is believed to play a role in CRE development and a helpful factor to identify patients for testing. Patients antimicrobial histories have to be evaluate d to see whether classifications of drug s, or combinations of drugs, are present in all patients testing positive for CRE. Limiting excess antimicrobial use along with en suring the completion of an antimicrobial regiment has been seen to affect CRE infection ra t es 14 In a case control study of C RE it was discovered that a history of fluoroquinolones and Carbapenem s were more common

PAGE 18

18 among patients testing positive for CRE 14 This study also showed that 85% of patients who tested positive for CRE were previously treated with cephalosporin 14 This study demonstrates the importance of considering antimicrobial treatment in the screening process for potential CRE patients. Another key factor in the development of CRE is antimicrobials that allow pathogens to colonize the intestinal tract 15 In a stu dy of CRE infection rates in mice, antibiotics leading to suppression of anaerobic microflora and with limited activity against CRE causing organisms were associated with the highest rates of CRE colonization 15 The reverse wa s seen in mice with polymyxin E and gentamicin which lead to suppression of CRE organism colonization 15 Comorbidity Lastly, comorbidities have been proven to be associated with an increased risk of CRE infections. The Charlson comorbidity index is used in studies to help determine u nderlying comorbidities of patients, determined by diagnostic codes found in administrative systems for each health care system 19 A study analyzing 481 patients who tested positive for CRE identified that 415 patients had a t least one comorbidity (91.4%) with a median Charlson comorbidity index of 2 20 The most common ly reported comorbidity was diabetes (201 patients [44.3%]) followed by neurological disorders (185 patients [40.7% ] ) 20 Among the 185 patients with neurological disorders 107 (57.8%) had a n indwelling urin ary catheter within two days of the positive culture, emphasizing the importance of this indwelling device in association with CRE 20 In a case control study of CRE risk factors performed using the Cumulative Illness Rating Scale (CIRS), a n alternative to the Charlson comorbidity index, higher prevalence of comorbidities was identified in CRE positive cases as oppose d to

PAGE 19

19 controls 21 Upon a naly sis of CIRS of 133 patients it was found that comorbidities increased the risk for CRE infection a mong elderly patients with immuno suppression or frail status 21 H igh CIRS severity was also the main risk factor for CRE colonization among their study population (Odds Ration 13.3; 95% CI 6.88 25.93) 21 These factors along with others need to be furt her evalua t ed in creating an effective surveillance system for the detection of CRE. F actors that influence development of CRE include c omorbidities, procedures undergone, presence of indwelling devices during, co infections with other resistant organisms, and length of stay Determining the risk factors associated with CRE infection and colonization will allow health care systems to develop screening systems for CRE that are applicable to patients coming from the surrounding community a nd limit factors tha t increase the risk of CRE within the health care setting.

PAGE 20

20 Figure 1 1. Patients with NDM CRE reported to the CDC as of April 2016 3 Figure 1 2. Patients with OX 48 CRE reported to the CDC as of April 2016 3

PAGE 21

21 Figure 1 3. Patients with VIM CRE re ported to the CDC as of April 2016 3

PAGE 22

22 CHAPTER 2 METHODS Study Population Risk factors of interest were identified using re trospective chart review Cases were patients who had a positive culture for CRE at a UF Health Shands facility between October, 201 4, and May, 2016 Controls were selected based on collection of any organism from a similar anatomical site within a date range of two weeks post or prior to the cases selection. Collection site and media of specimen w ere matched to that of the case infect ions. Cases and controls were identified using Theradoc clinical services and controls were matched using a 1:2 ratio, resulting in 50 cases and 100 controls. All pati ent charts were viewed using EPIC USERWEB. Participants were excluded from the study based on predetermined criteria. Cases were required to be eighteen years of age at time of sample collection. Participants with fields not having an admitting date, collection date, or discharged date based on medical record coding were excluded from part icipation. Cases who ha d previous CRE infections as identified by EPIC outside the study window were excluded due to inability to match controls for these cases using Theradoc Participants were evaluated on a case versus control basis and then further su b analysis was performed The sub analyses were community acquired CRE infections and matched control infections, and hospital acquired CRE infections and matched control infections. Community acquired infections were defined as having a positive sample co llected prior to day three of stay during the admission. Hospital acquired infections were considered any positive sample collected on day three or after upon admission.

PAGE 23

23 Factors Age Age of cases was determined using birth date recorded upon admission to U F Health Shands during the admission that lead to a positive CRE infection status. Age of the controls was identified in the same manner using age upon positive CRE infection admission as the recorded age for the participant. Age was then categorized as p articipants being under the age of 6 0 (Age <6 0) a nd equal to or over the age of 6 6 0). The age of 50 was selec ted as the dividing age due to 6 0 being the closest decade to the mean of the controls (mean= 59.21 ). Gender Gender for cases was recorded as either female or male based on gender identified on admission that produced th e positive CRE infection status. Gender for controls was recorded as either female or male based on gender identified on admission that produced the positive matched infection status. Infection and Colonization CRE status was categorized as infection or co lonization for cases Infection and colonization were distinguished by the site of the specimen collection. If the specimen was collected from within the body such as blood cultures or b ronchoalveolar lavage (BAL) it was considered to be an infection. S p ecimens collected from wounds, sputum, abscess drainage, or urine were considered to be colonization because they may not result in signs or symptoms. All specimens were categorized into four classifications; respiratory secretions, blood, urine, and wound /abscess drainage. Being in one of the four categories is not a determinant of being infection or colonization ; i n the respiratory

PAGE 24

24 secretion category, a BAL was categorized as infection while sputum was classified as colonization. Admitting Location Both c ases and controls were evaluated on the location they were most recently located in or admitted to prior to admission to UF Health Shands. All participants appeared from five possible locations. Participants who changed location within the recent past were still assigned an admission location based on most current location. An example is a participant who was discharged from the hospital to home on Jan. 1 and returned to the hospital on Jan. 3 ; this would still be counted as a home/residence admission due t o spending any amount of time at another location. Home/residence wa s assigned to any patient who wa s coming from their place of permanent residence. Traumatic events during every day activities, such as vehicle collisions or falls, are also under the home /residence admission label. Participants were also admitted from acute care hospitals that we re not affiliated with UF Health Shands. These are hospital systems that function similar to UF Health Shands and retain patients on average less than fourteen da ys. Participants were also admitted from long term care facilities (LTC F ) which are settings in which patients remain greater than fourteen days during a single admission. LTC F consist of skilled nursing facilities, which are locations that maintain and c are for an elderly population. LTCF also consist of hospice care facilities, which are locations that care for patients with terminal diagnosis and attempt control their pain symptoms for their remaining lifespan Finally, participants we re considered to b e LTC F participants when they arrived from LTACH which consist s of patients who stay an average of fourteen days or greater

PAGE 25

25 The final admitting location of participants wa s rehabilitation facilities Participants a t rehabilitation facilities have typica lly been admitted from hospitals after an admission or surgery in attempt to regain function of senses or limbs. Examples are rehabilitation for lower limb weakness or speech therapy after a stroke. Days until Positive Culture The number of days until p ositive culture was identified by using date of admission of either case/control as initia l event date (Day 1 ) and date specimen was collected as final event date. Dates were calculated using calendar days instead of 24 hour time periods. The first day of admission w as calculated as day zero regardless of what time of day the patient was admitted The same criteria were used when calculating specimen collection date Length of Stay Total length of stay at UF Health Shands was calculated using date of admiss ion of case/control to set event (Day 1 ) and release date as final event date. Similar to days until positive culture, calendar days were used instead of 24 hour time periods to calculated total days at UF Health Shands. Total length of stay was categorize d using the quartiles determined for the controls. The average of total length of stay was rounded to the nearest quartile and that quartile range was used as the reference group for comparison. Total length of stay at UF Health Shands was also categorized into binary categories by using the average amount of days all controls had during their relevant infections as the reference. This was a ten day average stay for controls.

PAGE 26

26 Admitting Diagnosis Admitting diagnosis was determined by the chief complaint part icipants reported upon admission. Participant complaints were reported both in the initial contact report along with final coding within H&P in the EPIC system. These two admissions reports were used to develop a comprehensive admission complaint. After t he admitting diagnosis was recorded participants were separated into two categories ; admissions for medical emergencies or trauma. Patients could be categorized as those with planned admittance which was admission for pre scheduled procedures Patients c ould also be categorized as those with emergenc y admittance which was those who arrive d for trauma or medical emergencies Participants who experience d complications after procedures and were readmitted were categorized as emergency admission unless a plan ned readmission was stated in the medical record before discharge. Total Hospital Admissions within Past Year Total hospital admissions were calculated by using date of admission on which positive CRE or matching infection were identified as final date, an d then observ ing previous admissions within the prior 365 days. Admissions were based on patient records as coded i n EPIC. Patients who were only at the emergency department but not coded for admission did not have this UF Health Shands encounter counted a s an admission. Patients who were transferred from an outside hospital to UF Health Shands had an admission added on to their total admissions in the past year even if the outside Additional admissions at outside hospitals were not counted, as they could not be identified.

PAGE 27

27 Indwelling Devices Participants were evaluated for presence and total number of indwelling devices in place prior to a positive CRE infection or matching infection collecti on date. Devices were considered relevant to CRE infection or matching infection for controls if the device was present on the day of positive culture collection or the device was removed less than two days prior to culture collection as per CDC criteria The indwelling devices were as follows: Urethral Catheter GI Tube CVC Double Lumen CVC Triple Lumen PICC Double Lumen PICC Triple Lumen Arterial Line Hemodialysis Catheter Peripheral IV Graft or Fistula Colostomy Midline CVC double lumen, CVC triple l umen, PICC double lumen, PICC Triple Lumen, and hemodialysis catheter were further group ed into central line devices. Medication Participants were evaluated for medication that was administered within the previous four weeks of positive CRE infection or m atching control infection. Only antimicrobials were recorded. These consist ed of all anti bacterial, anti viral, and anti fungal medications. Participants antimicrobials were averaged between determine d means of antimicrobials by participant group A ntivi ral medication or antifungal medication s were categorized under those titles. A nti bacterial medication s w ere categorized into pharmaceutical class es due to multiple medications of the same class

PAGE 28

28 being admitted over the duration of a participant s stay. Pa also recorded as extended if the participant was on any medication longer than fourteen consecutive days. All medication s that w ere recorded during the study w ere categorized as one of the following pharmaceutical classes: Carbap enem Cyclic Lipid Fl u oroquino lo nes Azithromycin Clarithromycin Imidazole Aminoglycoside Cephalosporins Glycopeptide Nitro i midazoles Antiviral Antifungal Pharmaceutical classes were turned into a binary variable for evaluation using the mean number of anti biotics taken by controls during admission associated with infection. The average value of antibiotics rounded to the nearest whole number was three pharmaceutical classes of antibiotics taken during the admission. Comorbidity and Comorbidity Scores Parti comorbidities were recorded based on the chart coding which was were identified as any secondary wa s not the primary diagnosis upon admiss ion (ex. Metastatic tumor, HIV/AIDS, Diabetes). Comorbidities that were Comorbidity Index in 1986 as well as the Charlson Comorbidity Index 2001 (CCI) 23 24 The 2011 CCI was update d from the Charlson Comorbidity Index of 1986, based on the release of the ICD 10 23 24. The recorded comorbidities were

PAGE 29

29 HIV/A IDS Cancer COP Dementia Diabetes Myocardial Infarction Metastatic Carcinoma Mild Liver Disease Moderate Liver Disease Parap legia/Hemiplegia Peptic Ulcer Disease Peripheral Vascular Disease Renal Disease Rheumatologic Disease Comorbidity score was determined by evaluating coded comorbidities based on both the CCI 1986 and the CCI 2011 23 24 25 The CCI 1986 score took into acco unt both comorbidities along with age value of +1 per decade after. Comorbidities for both the CCI 1986 and the CCI 2011 were assigned values based upon already existing scoring charts for both CCIs 2 4 25 The CCI is calculated during the end of a participant s stay at a health care setting. Participants arriving as trauma patients who were unable to be identified were unable to have a CCI associated due to lack of medical history. Discharge Location Discharge location is the physical setting a participant is released to after their discharge from UF Health Shands. The locations participants can be discharged to vary along with the level of care they receive at their discharge location vary. Participan ts discharged home are released to either self care or home care with a health care professional coming in and providing assistance within their home or residence setting Participants are also discharged to long term care facilities. These long term care facilities include locations such as LTAC skilled nursing facilities, and hospice settings.

PAGE 30

30 Participants discharged to these locations tend to be of a higher acuity and require ongoing care or devices such as ventilators. Participants discharge to ACH an d rehabilitation centers are transferred for lessening of their acuity score to one a local hospital can provide care for, and recovery after surgery. Participants tran sferred to rehab facilities have a lowered acuity and are discharged in order to regain function of limbs or learning to adapt to changes in limbs or body function. Participants declared deceased do not get discharge d but are still included in discharge l ocations for evaluation. Participants who died during their admission at UF Health Shands had their day of death considered the final day of the current admission. Invasive Surgeries and Invasive Surgeries with Scopes Invasive surgeries were obtained thro ugh data retrieval software using medical record codes. Each participant was evaluated individually a master list was created to ensure the invasive surgery occurred before date of positive culture collection. Invasive procedures with scopes were also cons idered. These scopes were not differentiated by method or location of insertion. A ll invasive procedures with scopes performed on participants who had community acquired infections were not counted towards a positive invasive procedure with scope. Analysis All data analysis was done using SAS 9.4 (SAS Institute, Cary, North Carolina ) All values were considered significant w ith a p or T value of less than or equal to 0.05 s are presented within a 95% confidence interval (95%

PAGE 31

31 CI) and were considered significant if the 95% CI does not contain the value of one. For any tables with cell counts les s than five significance values were evaluated using Cell counts of zero or one w ere excluded from the analysis

PAGE 32

32 CHAPTER 3 RESULTS C ase and control data w ere analyzed on three different levels. The first analysis was between the c ases and controls with no other factors used as exclusion criteria. The s econd analysis was looked at cases and controls who had community acquired infections ; these are infections that developed prior to hospitalization and that were identified within thr ee days of admission The third analysis included only cases and controls with hospital acquire d CRE infections. Demographics The study population was divided into 50 cases and 100 controls as based on our matching. The average age of cases was 58.9 years old and the average age of controls was 59.2 years old. The number of female cases was 22 (N=22) and the number of male cases was 28 (N=28) In the control group there were 45 females (N=45) and 55 males (N=55) (Table 3 1 ). Looking at specimen type to det ermine infection or colonization (Table 3 2) shows there is no significant difference. Of the specimens that we re conside red infection among cases, we observe d that there were 11 respiratory secretions (22% of all cases) and 7 blood specimens (14% of all c ases). Of the specimens considered colonization we observe d that there were 16 urine samples (32% of all cases) and 16 wound/abscess drainage samples among the cases (32%of all cases). The averages of the Charlson comorbidity index did not prove to be sig nificant, using both the criteria of the CCI 1987 and CCI 2011.

PAGE 33

33 Cases vs. Controls When we analyzed factors between cases and controls all odds ratios were calculated by comparing the means of risk factors for CRE positive participants to the means of the controls rounded to the closest category ( e.g. If the m ean of previously resistant organisms for controls wa s 0.100 this wa s rounded to zero previously resistant organisms). One significant factor that was observed between cases and controls was reasons for admission (Table 3 4). When observing the differences in participants admitted for emergency reasons versus those admitted for planned reasons cases we re 3.67 (95% CI 1.03 13.0 8 ) times more likely to have an unplanned, emergency admission than control s. There was a significant difference between cases and controls with respect to a mount of pharmaceuticals (counted by their subclass) (Table 3 5). Cases we re 5.8 6 (95% CI 2.26 15.24) times more likely to have had i n patient antibiotics in the month prior compared to controls. Taking ambulatory antibiotics with no i n patient antibiotics in the month prior wa s not significant between cases and controls. The cases were 23.5 (95% C I 3.8 4 143.8 7 ) times more likely to take both inpatient and ambulatory antibiotic s as compared to controls. Admitting location was also a factor that proved to be significant ly different between cases and controls. The binary variable (Table 3 6) of admission from a LTCF versus a n ACF indicate d that cases were 5.1 7 (95% CI 1.93 13.84) times more likely to be admitted from LTCF as opposed to controls. When admitting location wa s not binary (Table 3 7) the significant difference between cases and controls wa s admission from LT A CH with cases being 15 .43 (95% CI 3.09 76.98) times more like ly to be admitted

PAGE 34

34 from LTCH as compared to controls. Cases are also 2.9 4 (95% CI 1.18 7.31) times more likely to come from ACH hospitals as compared to controls. Discharge location was another factor observed on both the binary and individual level. On the binary level (Table 3 8); we s aw that cases we re 3.4 7 (95% C I 1.45 8.23) times more likely to be discharge to LTCF as opposed to controls. When observing the factors on an individual level (Table 3 9) cases we re 6.6 7 (95% CI 1.2 2 36.95) times more likely to be discharge d to hospice facilities as opposed to controls. Other significant discharge locations between cases and controls were LTCH ( OR = 6.25, 95% CI 1.8 2 21.4 7 ) as well as skilled nursing ( OR = 3.46, 95% CI 1.0 7 11.21). Next, differences were note d for antimicrobials taken and the number of pharmaceutical class of antibiotics Cases were compared to the average of the controls which was 1.34 and the thir d quartile of the controls was three (75% of controls) which falls under the 0 3 binary antibio tics category. Cases we re 4.2 4 (95% CI 1.75 10.24) times more likely to have taken more than three antibiotics in the past month as compared to the controls. When the antibiotics category wa s broken down in to individual antibiotics some antibiotics we re s ignificantly different between cases and controls. Cases we re 3.7 5 (95 CI 1.35 4 10.3 8 ) times more likely to have taken Fluoroquinolones, 2.37 (95 % CI 1.17 4.80) times more likely to have taken Cephalosporin, 5.44 (95 % CI 1.0 2 29.1 4 ) times more likely to ha ve taken Aminoglycoside s (Table 3 10), and 5.4 1 (95% CI 1.35 21.95) times more likely to have taken c arbapenem s as controls were Cases were at increased risk versus controls to have had indwelling devices (Table 3 11). Cases with were 4.4 1 (95% CI 1.05 18 .44) times as likely to have

PAGE 35

35 colostomy bags as controls. Cases were also 13.5 (95% CI 1.5 8 115. 50 ) times more likely to have a central line as compared to controls. The final factor that had observed significance wa s invasive procedures performed with sco pes (Table 3 12). C ases who had been scoped were 4.57 (95% CI 1.3 1 16.17) times more likely to have had a procedure performed with a scope as compared to controls. Invasive procedures performed without scopes were not found to be significant. Other fact ors that were collected and were found to be not significant we re age categorized (Table 3 13), Charlson Comorbidity Index both 1987 and 2011. Previous infection or colonization with resistant organisms also was not significant at the level of cases and co ntrols. Community Acquired CRE P articipants were then analyzed at the level of community acquired infections of CRE and the subsequent matched control infection. Community acquired infections we re defined as patients having a positive sample collected on the day of admission or within two days after. After excluding the hospital acquired cases there were 84 participants with community acquired infections. Of these participants 23 were cases There were 10 females and 13 males and an average age of 56.09 years. The controls consisted of 61 participants with 28 females and 33 males and an average age of 57.34 years. Community acquired cases differed from controls with respect to inpatient and ambulatory antibiotics taken (Table 3 1 4 ). Cases we re 10.15 (95% CI 2. 80 36.8 8 ) times more likely to have taken inpatient antibiotics compared to controls. Cases we re also

PAGE 36

36 27.5 (95% CI 3.9 8 190.04) times as likely to have both inpatient and ambulatory antibiotics For location of admission both binary and individual l ocations ha d significant difference s between community acquired cases and controls. Using binary admission locations (Table 3 1 5 ) cases we re 4.9 9 (95% CI 1.39 17.84) times more like ly to be admitted from a LTCF as opposed to controls. Investigating indivi dual admission locations (Table 3 1 6 ) cases were 18 (95% CI 1 81 178.8 1 ) times more likely to be admitted from a LTCH as opposed to controls. The remaining admission locations were not significant between cases and controls. A long with location of admiss ion, some antibiotics ha d significant differences between cases and controls among participants (Table 3 1 7 ). Cases we re 3.8 2 (95% CI 1.36 10.7 3 ) times more likely to have taken n itro i midazoles as compared to controls. The other significant antibiotics tak en we re Fluoroquinolones with cases being 26.6 9 (3.0 6 232.89) times more likely to have taken Fluoroquinolones in the month prior as compared to controls. The last significant factors within the community acquired CRE infections we re indwelling devices ( Table 3 1 8 ). Of the indwelling medical devices observed only urethral catheters proved to have a significant difference among community acquired CRE infection. Cases we re 4.495 (95% CI 1.57 15.65) times more likely to have a urethral catheter within two d ays prior up to date of positive culture collection as compared to controls. The factors that were not significant among the community acquired infections were CCI 1987 and 20 11, age categorized (Table3 19), and all indwelling devices.

PAGE 37

37 Hospital Acquire d C RE Within our population 66 participants had CRE infection and/or colonization attributed to hospital acquisition. Samples were considered hospital acquired if they were collected any time after the third day of admission The factors significant for th ese participants will differ from the community acquired infections due to different exposures and environment. The first significant factor wa s total length of stay between cases and controls (Table 3 19 ). The second quartile was established as the refere nce during analysis because this was the quartile closest to the mean of the hospital control infections total length of stay. Cases were 11.4 6 ( 95% CI 1.25 104.60) times as likely to stay between 15 and 34 days, and 19.83 (95% CI 1.2 9 171.83) times as lik ely to stay 35 days or greater as compared to controls. When total length of stay wa s made into a binary variable, (Table 3 2 1 ) 14 days was used as the dividing day total for the categories. C ases we re 18.0 8 (95% CI 2.22 147.10) times more likely to stay g reater than 14 days as compared to controls. Location where participants were admitted from wa s significant for participants admitted from LTCH (Table 3 2 2 ). Cases were 16.2 (95% CI 1.7 3 151.85) times more likely to be admitted from LTCHs as compared to co ntrols. When locations of admission we re made binary based on being LTCF, cases we re 6.4 8 (95% C I i 1.2 3 34.1 6 ) times more likely to be admitted from a LTCF compared to controls (Table 3 2 3 ). There was also a significant difference observed for discharge lo cation ( T able 3 2 4 ). Cases we re 14.4 ( 95% CI 1.3 8 155.24) times more likely to be discharged to a LTCH compared to controls. A similar trend wa s seen when discharge locations we re

PAGE 38

38 made binary ( T able 3 2 5 ). Cases we re 3.18 (95% CI 1.1 4 8.9 2 ) times more like ly to be discharge to a LTCF compared to controls. Antibiotics showed no significance among cases and control s (Table 3 2 6 ) When antibiotics we re categorized as greater or less than the control average of antibiotics during admission ( antibiotics ses were 3. 10 (95% CI 1.07 8.9 4 ) times more likely to have had greater than three antibiotics in the last month. The final significant factor that follow ed the same trend as the comparison at the case versus control level wa s that of invasive procedures wi th the use of a scope (Table 3 2 7 ). Invasive procedures performed without scopes we re not significant between the hospital acquired infections. Analyzing invasive surgeries performed with scopes cases we re 5.05 (95% CI 1. 20 21.2 9 ) times more likely to hav e an invasive procedure performed with a scope compared to the controls. Indwelling medical device presence was no t significant between cases and controls (Table 3 28 ) nor was age categorized (Table 3 29 )

PAGE 39

39 Table 3 1. Demographics of Total Population Controls Cases Total N 100 50 Age 59.21 58.92 Gender Male 55 28 Female 45 22 Charlson Comorbidity Index (CCI) CCI 1987 4.3 4.72 CCI 2011 2.72 3.14 Table 3 2. Infection and Colonization by Community or Hospital Acquired Cases Communi ty or Hospital Community Hospital Total Infection 11 19 30 Colonization 12 8 20 Total 46 54 100 Table 3 3. Infection and Colonization by Category for Cases Community or Hospital Community Hospital Total Respiratory Secretion 3 8 11 Blood 3 4 7 Urine 10 6 16 Wound/ Abscess Fluid 7 9 16 Total 23 27 50

PAGE 40

40 Table 3 4. Odds Ratio Emergency versus planned admissions Effect Point Estimate 95% Wald Confidence Limits Admission Reason: Emergency vs Planned 3.674 1.032 13.077 Table 3 5. Odds ratios Inpatient and Ambulatory Antibiotics Effect Point Estimate 95% Wald Confidence Limits Inpatient Antibiotics 5.875 2.265 15.241 Ambulatory Antibiotics 5.222 0.721 37.85 Both Inpatient and Ambulatory Antibiotics 23.5 3.839 143.867 Table 3 6 Odds ratios admission from LTCF versus ACF Effect Point Estimate 95% Confidence Limits Confidence Limits LTCF vs ACF 5.167 1.928 13.842

PAGE 41

41 Table 3 7. Odds ratios individual admitting location Effect Point Estimate 95% Wald Confidence Limits ACH vs. Home 2.939 1.181 7.312 Corrections vs Home >999.999 <0.001 >999.999 Hospice vs Home 3.429 0.206 57.183 LTCH vs Home 15.429 3.092 76.988 Skilled Nursing Home vs Home 3.429 0.789 14.892 Clinic vs Home 1.714 0.148 19.849 Rehab vs Home 3.429 0.2 06 57.183 Table 3 8. Odds ratios discharge to LTCF versus ACF Effect Point Estimate 95% Wald Confidence Limits LTCF versus ACF 3.468 1.452 8.283

PAGE 42

42 Table 3 9. Odds ratios individual discharge location Effect Point Estimate 95% Wald Confidence Limits ACH vs Home/Residence 2 0.321 12.463 Death vs Home/Residence 3.333 0.996 11.154 Homecare vs Home/Residence 1.579 0.464 5.376 Hospice vs Home/Residence 6.667 1.215 36.593 LTCH vs Home/Residence 6.25 1.819 21.469 Rehab vs Home/Residence 4 0.8 53 18.752 Skilled Nursing vs Home/Residence 3.461 1.069 11.21 Table 3 10. Odds Ratio Antibiotics and Antibiotics Binary Effect Point Estimate 95% Wald Confidence Limits Fluoroquinolone s 3.747 1.353 10.379 Cephalosporin s 2.373 1.172 4.804 Ni tro i midazole 1.714 0.787 3.734 Glycopeptide s 1.857 0.931 3.703 Aminoglycoside s 5.444 1.017 29.136 Cyclic Lipopeptid e 1 0.177 5.654 Carbapenem s 5.406 1.332 21.951 Erythromycin 4.125 0.365 46.625 4.235 1.751 10.243

PAGE 43

43 Table 3 11. Odds Ratio Indwelling Devices Effect Point Estimate 95% Wald Confidence Limits Colostomy 4.409 1.054 18.442 Central Lines 13.5 1.578 115.498 Arterial Line 0.762 0.309 1.876 Urethral Catheter 1.325 0.638 2.749 Hemodialysis Graft Fist 3.128 0.505 19.355 Table 3 12. Odds Ratio Invasive Procedure & Invasive Procedures with Scope Effect Point Estimate 95% Wald Confidence Limits Invasive Procedures 1.657 0.805 3.411 Inv asive Procedures w/ Scope 4.571 1.305 16.017 Table 3 13. Odds Ratio Age Effect Point Estimate 95% Wald Confidence Limits 0.887 0.449 1.751 Table 3 1 4 Odds Ratio Inpatient and Ambulatory Antibiotics (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits Inpatient Antibiotics 10.154 2.796 36.875 Ambulatory Antibiotics 7.333 0.933 57.611 Inpa tient and Ambulatory Antibiotics 27.5 3.979 190.04

PAGE 44

44 Table 3 15. Odds Ratio Binary Admission L ocation (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits LTCF vs. ACF 4.900 1.369 17.536 Table 3 16. Odds Ratio Individual Admit ting L ocation (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits ACH vs Home 3.75 0.953 14.764 Hospice vs Home 4.5 0.259 78.204 LTCH vs Home 18 1.812 178.808 Nursing Home vs Home 4.5 0.789 25.659 Rehab vs Home 4.5 0.259 78.204 Table 3 1 7 Odds Ration Antibiotics taken in Prior Month (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits Nitro i midazole 1.966 0.501 7.72 Cephalosporin s 3.819 1.36 10.727 Fluoroquinolone s 26. 250 3 .007 229.133 Glycopeptide 2.72 0.911 8.118 Cyclic Lipopeptide 2.725 0.163 45.473 Antibiotics 3 versus Antibiotics 3 8.998 0.885 91.461

PAGE 45

45 Table 3 18 Odds Ratios Indwelling Devices (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits Urethral Catheter 4.959 1.571 15.653 Colostomy 8.998 0.885 91.461 Arterial Lines 2.81 0.372 21.227 Table 3 19. Odds Ratio Age (Community Acquired) Effect Point Estimate 95% Wald Confidence Limits Age 60 1.258 0.479 3.303 Table 3 20. Odds Ratio Total Length of Stay (Hospital Acquired) Effect Point Estimate 95% Wald Confi dence Limits Total Length of Stay Quartile 1 <0.001 <0.001 >999.999 Total Length of Stay Quartile 3 11.455 1.254 104.601 Total Length of Stay Quartile 4 19.833 2.289 171.832 Table 3 21. Odds Ratio Total Length of Stay Binary (Hospital Acquired) Ef fect Point Estimate 95% Wald Confidence Limits Total Length of Stay Greater than 14 days 18.079 2.222 147.101 Table 3 22. Odds Ratio Admission Location (Hospital Acquired) Effect Point Estimate 95% Wald Confidence Limits ACH 2.7 0.798 9.139 LTCH 16.2 1.728 151.85 Skilled Nursing Facility 2.7 0.154 47.392

PAGE 46

46 Table 3 23. Odds Ratio Admission Location Binary (Hospital Acquired) Effect Point Estimate 95% Wald Confidence Limits Admission LTCF versus ACF 6.475 1.228 34.155 Table 3 24. Odds Ratio Discharge Locations (Hospital Acquired) Effect Point Estimate 95% Wald Confidence Limits ACH 2.25 0.111 45.722 Death 5.625 0.537 58.909 Homecare 3.6 0.257 50.33 Hospice 9 0.522 155.241 LTCH 14.4 1.375 150.807 Rehab 13.5 0.878 207.622 Skilled Nursing Facility 11.25 0.972 130.22 Table 3 25. Odds Ratio Discharge Location Binary (Hospital Acquired) Effect Point Estimate 95% Wald Confidence Limits Discharge LTCF versus ACF 3.181 1.135 8.918

PAGE 47

47 Table 3 26. Odds Ratio Antibiotics Taken in Prior Month (Hospital Acquired) Effect Point Estimate 95% Wald Confidence Limits Aminoglycosides 2.312 0.359 14.876 Glycopeptide 0.952 0.344 2.637 Nitro i midazole 1.228 0.448 3.366 Cephalosporin s 0.8 0.217 2.946 Fluoroquinolone s 0.957 0.24 2 3.774 Cyclic Lipopeptide 0.462 0.045 4.69 Carbapenem s 3.6 0.811 15.97 Erythromycin 3.04 0.262 35.332 3.095 1.072 8.938 Table 3 27. Odds Ratio Invasive Procedure & Invasive Procedures with Scope (Hospita l Acquired) Effect Point Estimate 95% Wald Confidence Limits Invasive Procedures 1.055 0.362 3.077 Invasive Procedures w/ Scope 5.051 1.199 21.289

PAGE 48

48 Table 3 28. Odds Ratio Indwelling Devices (Hospital Acquired) Effect Point Estimate 95% Wald C onfidence Limits Hemodialysis Graft/Fistula 2.312 0.359 14.876 Arterial Line 0.333 0.11 1.006 Urethral Catheter 0.361 0.128 1.02 Central Line 2.035 0.72 5.751 Colostomy 2.312 0.359 14.876 Table 3 29. Odds Ratio Age (Hospital Acquired) Effect Poi nt Estimate 95% Wald Confidence Limits 0.686 0.252 1.872

PAGE 49

49 CHAPTER 4 DISCUSSION AND CONCLUSION Discussion S ignificant risk factors were identified at all levels that were observed. More risk factors were observed at the most inclusive level of all cases and controls. Fewer risk factors were identified as significant in the sub analysis of community acquired cases and hospital acquired cases. Plan versus Em ergency Admissions Pati ents with CRE were 3.674 times more likely to present to UF Healt h Shands for emergency reasons, such as medical emergencies and traumas compared to patients without CRE infections. These findings have not been previously reported in literature this could result from the categories being very broad making them not as applicable within the screening process. Further research can be done into the types of planned admittance (e.g. transplants, and chemotherapy) and emergency admissions (e.g. feinting, and skull fracture). Ambulatory and/or Inpatient Antibiotics T aking an tibiotics in an ambulatory manner, out side of UF Health Shands, or taking antibiotics on an inpatient basis wa s a significant risk factor There wa s an increased risk for being on inpatient antibiotics for both cases versus controls and the community acqui red CRE cases CRE infected participants were 5.875 times more likely to have taken inpatient antibiotics on the case versus control level, and 10.154 times more likely to have taken inpatient medication on the community acquired level Taking only inpatie nt antibiotics among community acquired cases would result from antibiotics administered upon arrival or antibiotics given within the first two days of stay.

PAGE 50

50 Cases were also 23.5 times as likely to take inpatient and ambulatory medication on the case vers us control level, and case s were 27.5 times as likely to take inpatient and ambulatory medication on the community acquired level Taking both inpatient and ambulatory antibiotics results from use of antibiotics before admission and either continuance of a ntibiotics or being administered new antibiotics admission. Patients who present with these risk factors are potentially those most suited for CRE screening In a recent study done by Marchiam et al. antibiotic exposure in the thr ee months prior to colonization was observed, and they discovered that there is an increased risk of 11.4 (2 64.3) for cases of CRE (n=91) compared to those without CRE 14 Admitting Location Admission location was a significant risk factor for our particip ants at all three levels of evaluation. On the binary level of admission from a LTCF those participants who developed hospital acquired CRE ha d the highest risk from this factor. This could be due to colonization of CRE at these LTCFs and not being screen ed upon ad mission to UF Health Shands, creating increased incidence within hospital acquired CRE category which with screening would have been attributed to the outside community. When the admitting locations were separated out of their binary categories, we observe d only one admitting location to be significant among all three categories. All cases, both community and hospital acquired admission, were at increased risk of developing CRE infection if participants were admitted from a long term acute care h ospital. In a study by Marguez et al. similar findings we re found regarding LTACH and ACH rates, with their study showing incidence of 2.54 per 1000 patient days at LTACH as compared to 0.31 per 1000 patient days at ACH 35

PAGE 51

51 Discharge location Along with ad mitting locations results show there is increased risk of CRE infected patients being discharge to certain discharge locations. Looking at the binary variable of being discharge d to LTCF this wa s an increased risk for CRE infected participants on the cas e versus control level and for those with hospital acquired CRE infections. When observing all participants, cases we re 3.468 times more likely to be discharge d to a LTCF, and hospital acquired cases we re 3.181 times more likely to be discharge to the LTCF W hen discharge locations we re lo oked at by individual locations, cases we re at an increased risk of being discharge d to hospice facilities, skilled nursing facilities, and LTACH (OR=6.67, OR=3.401, and OR=6.25 respectively) Hospital acquired cases only ha d a significant increased risk of being discharge to LTACH (OR 14.4). A study by Lee et al. highlighted the importance of coordinating and controlling transfer and discharge locations of CRE infected patients The authors reported that with coordinated r egional control of the transferring of patients they averted 21.3% of CRE infection s compared to an uncoordinated effort 8 This demonstrates the need for discharge to locations equipped to accommodate CRE infected patients and the need for avoidance of he alth care settings not equipped to handle these CRE infections Antibiotics Antibiotics were the next risk factor found to be significant on two levels of analysis. Multiple individual antibiotics were found to be significant in at least one level of analy sis. Comparing cases to the controls, the participants w ith positive CRE infec tions were at greater risk for f luoro quinolone cephalosporin carbapenem, and aminoglycosides ( OR= 3.747 OR=2.373, and OR= 5.444 respectively). S ignificance of individual antibio tics was also observed on the community acquired case level C ases

PAGE 52

52 were at increased risk of having been administered cephalosporin and fluoroquinolone ( OR= 3.819 and OR= 26.687 respectively ). In a study conducted by Schechner et al. fluoroquinolone was det ermined to be a predictor for CRE renal carriage, as it was on our case versus control and community acquired levels, with an OR o f 4.27 (95% CI 1.10 16.6) 2 8 These positive samples were collected from a different anatomical site but do demonstrate fluoroq uinolone being a predictor for CRE infection. In another study by Gasink et al. s imilar results were found when observing fluoroquinolone and the development of KPC producing K. pneumoniae with an increased OR of 3.39 (95% CI 1.118 7.66) in a study with fifty six cases and 863 controls 29 The last significant risk factor among the case versus control level and community acquired level was cephalosporin ; cases were 2.373 times more likely to take cephalosporins as compared to controls on the case versus co ntrol level, and 3.819 times as likely on the community acquired level In the article by Gasink et al. which identified fluoroquinolone use as a predictor they also found with in the same population that cephalosporin had an OR of 2.55 (95% CI 1.18 5.52) 29 The antibiotics that were found to have significance only when all cases and controls were observed were carbapenem and aminoglycoside. In one study evaluating whether the proper antimicrobial course was followed for CRE positive patients eight of the 16 (50%) patients had received carbapenem s 30 Carbapenem administration, in th at study and among our participants, may be attributed to not knowing patient s CRE status at time of administration The final antibiotics that showed significance on the case control level w ere aminoglycosides. This is not unexpected within a CRE positive population because aminoglycosides are often administered to treat CRE infections.

PAGE 53

53 Since the 1950s, aminoglycosides have been preferred to colistin, with both being 1 1 When the average number of antibiotics for CRE positive participants was compared to that of the controls, there was a significant difference. The average number of antibiotics taken by controls in the last month was three or fewer ; case s were 8.998 times more likely to take greater than the average number of antibiotics for controls. Physicians could be contributing to this risk factor by not following best antimicrobial stewardship recommendations or overprescribing medications O ur CRE positive participants may also have had underlying conditions that we re being treated along with the CRE infection, which would result in increased antibiotic usage However, this is not very likely due to similar comorbidity scores indicating the p opulations we re in similar health. Indwelling Devices Indwelling devices were found to be significant on the cases versus controls level as well as the community acquired level On the case versus control level, we observe d that cases we re 4.4 1 times as l ikely to have had a colostomy and 13.5 times as likely to have had a ce ntral line compared to controls The finding of increased risk of CRE with colostomy is interesting because there is not much research looking into the association. In a study conducted by Chang et al., they observed 51 participants (17 with carbapenem resistance and 34 without) and they identified that none of their carbapenem resistant group had colostomies compared to three with carbapenem sensitivity 31 Our results were in contrast o f these findings potentially due to a small sample size or difference in population demographic by Chang et al 31

PAGE 54

54 Cases were also 13.5 times more likely to have had a central line compared to controls on the case versus control level A rticles have been published on single cases of possible CRE infection acquired through central lines but little research has been conducted looking at large populations and their risk of CRE development 32 On the community acquired level having a urethral catheter wa s an i ncreased risk factor for CRE infection with cases being 4.9 6 times more likely to have a urethral catheter as compared to controls In a study by Huang et al. observing carbapenem resistant Acinetobacter baumannii bacteremia, Foley catheters were identifi ed as a significant risk factor and we re found to be present in 83.9% of carbapenem resistant Acinetobacter baumannii patients and 65.2% of carbapenem susceptible Acinetobacter baumannii patients 36 Invasive Procedures/ with Scope The last observed factor was invasive procedures performed on participants prior to collection of positive CRE bacteria sample. Among all levels of analysis, there was no significant difference when all invasive procedures were compared between cases and controls. When only invas ive procedures with scopes were observed we s aw significance on the case versus control level as well as the hospital acquired participants. Cases were 4.571 times as likely to have had an invasive procedure with a scope as compared to a control C ases wi th hospital acquired infections were 5.051 times as likely to have had invasive procedures with scopes as compared to their controls We were not able to adequately analyze community invasive procedures and those with scopes due to the low number of proced ures. In a 2013 outbreak in Chicago, a side viewing duondenoscope (ERCP) was found to be responsible for infection and

PAGE 55

55 colonization of 38 patients 33 These devices pose a risk due to the challenging nature of cleaning the devices and accessibility to all p arts of these devices 33 Limitations Even with significant findings among all cases and controls, community acquired CRE infections, and hospital acquired CRE infection s there were limitations to the study. Limitations included the size of the study popul ation, coding of the electronic medical records (EMR) in Epic, and the lack of knowledge of past medical records. The total amount of cases was limited by the timeline of TheraDoc the software used to identify cases, being available at UF Health Shands T his left us with a relatively small N (N=50). The limited population size lead to very large confidence intervals, which do not allow for an evaluation of the significance of the impact of the risk factor on the outcome of CRE development. With the limited population size some risk factors on the sub levels of analysis, community acquired and hospital acquire infections, did not allow for references for comparison. This occurred in cases of admitting location whe re there was only one participant admitted f rom a correction facility so no conclusions could be drawn for the significance of this admitting location. The coding of the medical records was the next limitation encountered during the study. All EMRs were coded by UF Health Shands employees after the participant had been discharged and was to be billed for the admission. There is potential for variance in coding between various employees A ll information recoded was verified to the best extent possible by the researcher but discrepancies arose due to coder oversight. Errors occurred mostly in admitting locations with locations such as home and ACH being cited as admitting location during the same admission with evidence for both locations being found throughout the EMR. There was not enough evidence t o disprove

PAGE 56

56 either location so as the location that was identified in the History and Physical note at time of admission was designated as the admitting location The last limitation of the study arose due to the lack of past medical history or EMR of patie nts being admitted from other health care facilities or having visited a n outside healthcare facility in the past calendar month. This limitation mostly a ffected antibiotic counts, indwelling devices, and past invasive procedures. Counts for any of the ris k factors could have been affected by the availability of these medical records but due to the inability to acquire the records they were omitted and significance was determined on the medical information contained within the c urrent EMR at UF Health Shan ds. Application s of Findings Identifying the significant risk f actors for CRE infection allows us to create possible model s to identify the best individuals to be screen ed upon admission Using the determined risk factors two models will be developed to s creen individuals ; one for those who are admitted from the community with little known past medical history and the second for patients arriving with a known past medical history. The first model would be applied to patients arriving with little known past medical history. This would include a quick verbal interview, visual cues upon arrival, and limited to no available EMR. To develop the algorithm significant risk factors would uld be used to determine best fit for screening. The significant risk factors are drawn from the community acquired infections because these patients arrived to UF Health Shands and have a positive sample of CRE bacteria or matching infection prior to day three of admission with minimal opportunity for hospital procedures or interventions. The risk

PAGE 57

57 factors used within the algorithm are antibiotics taken ambulatory (AMB) and/or inpatient (IP) arriving from a LTCF, having taken fluoroquinolones in the prior month, and having a urethral catheter (Figure 4 1). All variables within the model are binary ; in terms of urethral catheter if the patient was admitted with a urethral catheter then the patient would have a one (1) inserted in the algorithm. Stepwise lo gistic regression was used to determine the risk factors that are of significance within the logistic regression model. The second algorithm will be for patients admitted with a known past medical history. This would include verbal interview, visual cues, EMR, and all past medical records from UF Health Shands and other health care settings admissions. The significant factors for this algorithm are drawn from all participants at the case versus control level of analysis. The significant factors for the algo rithm are emergency versus planned admission, antibiotics taken ambulatory (AMB) and/or inpatient (IP), arriving from a LTCF, having taken fluoroquinolones in the prior month having taken cephalosporin in the prior month, having taken aminoglycoside in th e prior month, taking greater than three antibiotics, having a colostomy, and having an invasive procedure performed with a scope device (IPS) (Figure 4 2). Stepwise logistic regression was used to determine the risk factors that are of significance within the logistic regression model. Future research can build on the study presented with two goals The first goal would be to identify more risk factors for CRE infection and to expand the two models proposed above to be more specific. Risk factors such can be evaluated at a more precise level, invasive surgery with scopes risk factor the type of scope procedure can

PAGE 58

58 be observed (e.g. ER C P colonoscopy, endoscopy). Emergency and planned admission can be looked at more precisely, with planned admissions being separated into reason for planned admission and emergencies being separated by severity (e.g. fever >103 versus skull fracture). A second goal of future research would be to identify the validity of the models produced from this study. To conduct this rese arch all patients admitted during a set time period or until a certain participant population is reach ed which can be determined through a SAS 9.4 power test ( SAS Institute, Cary, North Carolina ) All samples collected would be tested for CRE infection or colonization, and separately from the CRE testing the models (Figure 4 1 and Figure 4 2) would be applied to patients to determine a score based on these models. Then all positive CRE samples would be matched with the patients and their model score woul d be average d to create a threshold score. This threshold score would serve as a baseline at which if a patient arriving has a higher model score they would be selected for CRE screening. In conclusion, multiple risk factors of CRE infection were recorde d through the course of this study. Not all risk factors proved to be significant with regard to CRE infection but a few factors did pose an increased risk to the development of a CRE infection or colonization. These factors lead to the development of mode ls which could be applied to identify patients most suited for CRE screening. There are many more risk factors that may contribute to the development of CRE infection, but this study revealed risk factors for this specific pa tient population which can be b uilt upon to create a more specific model in the future.

PAGE 59

59 Figure 4 1. Algorithm for Patients Arriving with Little Known Medical History Figure 4 2. Algorithm for Patients Arriving with Known Medical History

PAGE 60

60 LIST OF REFERENCES 1. CDC. (201 5, February 15). Carbapenem resistant Enterobacteriaceae in healthcare settings Retrieved August 3, 2016, from https://www.cdc.gov/hai/organisms/cre/ 2. CDC. (2012, June 22). Carbapenem Resistant Enter obacteriaceae containing New Delhi Metallo Beta Lactamase in Two patients Rhode Island, March 2012 Retrieved August 3, 2016, from CDC Home, https://www.cdc.gov/mmwr/preview/mmwrhtml/m m6124a3.htm 3. CDC. Tracking CRE Retrieved August 3, 2016, from https://www.cdc.gov/hai/organisms/cre/TrackingCRE.html#CREmapKPC 4. CDC. (2015b, December 4). Notes from t he field: Carbapenem resistant Enterobacteriaceae producing OXA 48 like Carbapenemases United States, 2010 2015 Retrieved August 3, 2016, from https://www.cdc.gov/mmwr/preview/mmwrhtm l/mm6447a3.htm 5. CDC. (2010, September 24). Update: Detection of a Verona Integron Encoded Metallo Beta Lactamase in Klebsiella pneumoniae --United States, 2010 Retrieved August 3, 2016, from https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5937a4.htm 6. Nordmann, P., Nass, T., & Poiler, L. (2011). Global Spread of Carbapenemaseproducing. Emerging Infectious Disease 17 (10), 1796 1798. 7. Arnold, R. S., Thom, K. A., Sharma, S., Phillips, M., Johnson, K. J., & Morgan, D. J. (2010). Emergence of Klebsiella pneumoniae Carbapenemase (KPC) producing bacteria. 104 (1), Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3 075864/ 8. Lee, B. Y., Bartsch, S. M., Wong, K. F., McKinnell, J. A., Slayton, R. B., Miller, L. Resistant Enterobacteriaceae, an emerging threat to health care facilities, and the impact of the centers for disease control and prevention Toolkit. 183 (5), Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4772438/ 9. Falagas, M. E., Tanslari, G., Karageorgopoulos1, D., & Vardakas, K. (2014). Deaths attributable to Carbapenem Resistant Enterobacteriaceae infections volume 20, number 7 July 2014 emerging infectious disease journal. Emerging Infectious Disease 20 (7), Retrieved from http://wwwnc.cdc.gov/eid/article/20/7/12 1004_article 10. Ray, M. J., Lin, M. Y., Weinstein, R. A., Trick, W. E., Rush, 2, Center, M., Chicago (2016). Spread of Carbapenem Resistant Enterobacteriaceae among

PAGE 61

61 Illinois h ealthcare facilities: The role of patient sharing. Clinical Infectious Diseases doi:10.1093/cid/ciw461. 11. P erez F., & V an D uin D. (2013). Carbapenem resistant Enterobacteriaceae: A menace to our most vulnerable patients. Cleve Clin J Med 80 (4), doi:10 .3949/ccjm.80a.12182 12. Bonomo, R. A. (2010). Carbapenem resistant Acinetobacter baumannii and Klebsiella pneumoniae across a hospital system: Impact of post acute care fac ilities on dissemination. 65 (8), doi:10.1093/jac/dkq191. 13. Thaden, J. T., Lewis, S. S., Hazen, K. C., Huslage, K., Fowler, V. G., Moehring, rates of Carbapenem Resistant E nterobacteriaceae in community hospitals: A mixed methods review of Epidemiology and microbiology practices in a network of community hospitals in the southeastern United States. 35 (8), Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4217156/#R9 14. Kaye, K. S. (2012). Recent exposure to Antimicrobials and Carbapenem Resistant Enterobacteriace ae: The role of Antimicrobial stewardship. Infect Control Hosp Epidemiology 33 (8), doi:10.1086/666642. 15. Perez, F., Pultz, M. J., Endimiani, A., Bonomo, R. A., & Donskey, C. J. (2011). Effect of antibiotic treatment on establishment and elimination of in testinal colonization by KPC Producing Klebsiella pneumoniae in Mice Antimicrob Agents Chemother. 55 (6), doi:10.1128/AAC.00891 10. 16. Hayden, M. K. (2012). Transfer from High A cuity long term care facilities is associated with carriage of Klebsiella pneumoniae Carbapenemase Producing Enterobacteriaceae : A Multihospital study. Infection Control & Hospital Epidemiology 33 (12), 1193 1199. doi:10.1086/668435. 17. van Duin, D., Kaye, K S., Neuner, E. A., & Bonomo, R. A. (2013). Carbapenem resistant Enterobacteriaceae: A review of treatment and outcomes. Diagnostic Microbiology and Infectious Disease 75 (2), 115 120. doi:10.1016/j.diagmicrobio.2012.11.009. 18. Amjad, A., Mirza, I., Abbasi, S., Farwa, U., Malik, N., & Zia, F. (2011). Modified Hodge test: A simple and effective test for detection of carbapenemase production. Iranian Journal of Microbiology 3 (4), Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3330182/

PAGE 62

62 19. University of Manitoba. (2010). Concept: Charlson Comorbidity Index Retrieved August 11, 2016, from Manitoba Center of Helath Policy, http://mchp appserv.cpe.umanitoba.ca/viewConcept.php?conceptID=1098 20. A. J. (2015). Epidemiology of Carbapenem Resistant Enterobacteriaceae, 2012 201 3. JAMA 314 (14), 1479 1487. doi:10.1001/jama.2015.12480. 21. Meschi, T. (2014). Comorbidities and disease severity as risk factors for Carbapenem Resistant Klebsiella pneumoniae c olonization: Report of an experience in an internal medicine unit. PLOS 9 (10), Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4198186/ 22. Society, A. C. (2016). What is hosp ice care? Retrieved September 5, 2016, from http://www.cancer.org/treatment/findingandpayingfortreatment/choosingyou rtreat mentteam/hospicecare/hospice care hospice care 23. (2011). Updating and validating the Charlson Comorbidity index and score for risk adjustment in hospital discharge abstract s using data from 6 countries. American Journal of Epidemiology 173 (6), 676 682. doi:10.1093/aje/kwq433. 24. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Deve lopment and validation. Journal of Chronic Diseases 40 (5), 373 383. doi:10.1016/0021 9681(87)90171 8 25. Project, p. Charlson Age Comorbidity index (CACI) Charlson et al. 1994 Charlson Age Comorbidity index (CACI) Retrieved September 14, 2016, from http://www.pmidcalc.org/?sid=7722560&newtest=Y 26. CDC. Untreatable: Report by CDC details today s drug resistant health threats. Retrieved October 18, 2016, from https://www.cdc.gov/media/releases/2013/p0916 untreatable.html 27. Marchaim, D., State, W., Chopra, T., Bhargava, A., Bogan, C., Dhar, S., Antimicrobials and Carbapenem Resistant Enterobacteriaceae: The role of Antimicrobial stewardship | Cambridge core. Infection Control & Hospital Epidemiology 33 (8), 817 830. doi:10.1086/666642

PAGE 63

63 28. Schechner, V., Kotlovsky, T., Tarabeia, J., Kazma, M., Sc hwartz, D., Navon Carbapenem Resistant Enterobacteriaceae (CRE) among Patients with Known Infection Control & Hospital Epidemiology 32(5), pp. 497 503. doi: 10.1086/659762 29. Gasink, L. B., Edelstein, P. H., Lautenbach, E., Synnestvedt, M., & Fishman, N. O. (n.d.). Risk factors and clinical impact of Klebsiella pneumoniae Carbapenemase Producing K. Pneumoniae. 30 (12), Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2893218/ 30. Falagas, M. E., Rafailidis, P. I., Kofteridis, D., Virtzili, S., Chelvatzoglou, F. C., Risk factors of carbapenem resistant Klebsiella pneumoniae infections: A matched case control study. Journal of Antimicrobial Chemotherapy 60 (5), 1124 1130. doi:10.1093/jac/dkm356 31. Chang, H. J., Hsu, P. C., Yang, C. C., Kuo, A. J., Chia, J. H., Wu, T. L ., & Lee, M. H. (2011). Risk factors and outcomes of carbapenem nonsusceptible bacteremia: A matched case control study. Journal of Microbiology, Immunology and Infection 44 (2), 125 130. doi:10.1016/j.jmii.2010.06.001 32. Palasubramaniam, S. (n.d.). Imipenem resistance in Klebsiella pneumoniae in Malaysia due to loss of OmpK36 outer membrane protein coupled with AmpC hyperproduction. International Journal of Infectious Diseases 11 (5), 472 474. doi:http://dx.doi.org/10.1016/j.ijid.2007.01.005 33. Muscarella, L. F. (2014). Risk of transmission of carbapenem resistant. World J Gastrointest Endosc 6 (10), 457 476. doi:10.4253/wjge.v6.i10.457 34. Ray, M. J., Lin, M. Y., Weinstein, R. A., Trick, W. E., Rush, 2, Center, M., Health, C. C. 3, System, H., & Chicago (2016b). Spread of Carbapenem Resistant Enterobacteriaceae among Illinois healthcare facilities: The role of patient sharing. Clinical Infectious Diseases 63 (7), 889 893. doi:10.1093/cid/ciw461 35. Marquez, P., Terashita, D., Dassey, D., & Mascola, L. (2013). Popula tion based incidence of carbapenem resistant Klebsiella pneumoniae along the continuum of care, Los Angeles County. Infection control and hospital epidemiology. 34 (2), 144 50. Retrieved from http s://www.ncbi.nlm.nih.gov/pubmed/23295560 36. Huang, S. T., Chiang, M. C., Kuo, S. C., Lee, Y. T., Chiang, T. H., Yang, S. P., P. (2012). Risk factors and clinical outcomes of patients with carbapenem resistant Acinetobacter baumannii bacteremia. Journal of Microbiology, Immunology and Infection 45 (5), 356 362.

PAGE 64

64 37. Bennett, P. M. (2008). Plasmid encoded antibiotic resistance: Acquisition and transfer of antibiotic resistance genes in bacteria. British Journal of Pharmacology 153 (S1), S347 S357. doi: 10.1038/sj.bjp.0707607 38. Leavitt, A., Navon Venezia, S., Chmelnitsky, I., Schwaber, M. J., & Carmeli, Y. (2007). Emergence of KPC 2 and KPC 3 in Carbapenem Resistant Klebsiella pneumoniae strains in an Israeli Hospital 51 (8), Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1932543/

PAGE 65

65 BIOGRAPHICAL SKETCH Marko Predic is a graduate student working towards his Master of Science in epidemiology from the University of Florida. He has previously received hi s undergraduate degree, Bachelor of Science in microbiology and cell sciences from t he University of Florida in 2014 During his time working on his master s he has aided and learned the profession of being an infection control practitioner in the D epartm ent of I nfection Prevention and C ontrol at UF Health Shands in Gainesville, Florida Marko Predic will be graduating in December 2016 to further pursue a career in infection control.