|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001917291
007 cr mnu|||uuuuu
008 071119s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002026
The rise of Clostridium difficile in Florida
h [electronic resource] /
by Owen Bendixsen.
[Tampa, Fla.] :
b University of South Florida,
ABSTRACT: Clostridium difficile, a bacterium that causes diarrhea in hospitalized patients, is on the rise in the United States as well as in other countries. This study was done to determine the extent of the problem in Florida's acute care hospitals. The Agency for Healthcare Administration (AHCA) provided data for patients discharged from Florida's acute care hospitals for the years 1998 through 2004. This study will focus on changes in the prevalence of Clostridium difficile associated disease (CDAD) over time. The mortality and morbidity of patients that have CDAD will also be examined to show if the disease is increasing over time. Factors investigated in this study that may influence the prevalence of CDAD include gender, race, length of hospital stay, age, and the cost per patient discharged. In Florida the prevalence of CDAD has risen from 3.41 per 1,000 discharged patients in 1998 to 8.11 per 1,000 discharged patients in 2004. The mortality increased from 9.48% for CDAD positive patients in 1998 to 10.11% for CDAD positive patients in 2004. Age plays a role in both the prevalence and mortality of this disease. In 2004 the mortality of patients who were positive for Clostridium difficile was 4.1% for those individuals that were 30-40 years old compared to 0.54% mortality for those patients in the same age group that did not have CDAD. The corresponding mortality for the patients aged 70-80 for the year 2004 was 11.1% for persons who had CDAD and 3.58% mortality for patients discharged with no CDAD. The analysis showed that CDAD prevalence is increasing in Florida acute care hospitals. During 1998-2004 mortality rates for patients diagnosed with CDAD is also increasing. This analysis also indicates that age is a factor that increases the death rates for patients that are CDAD positive. A more concerted effort to implement hospital techniques that prevent the increasing prevalence of Clostridium difficile in Florida hospitals is recommended.
Thesis (M.S.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 89 pages.
Advisor: Aurora Sanchez-Aguiano, Ph.D.
x Public Health
t USF Electronic Theses and Dissertations.
The Rise of Clostridium difficile in Florida By Owen Bendixsen A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Public Health Department of Epidemiology/Biostatistics College of Public Health University of South Florida Major Professor: Aurora Sanchez-Anguiano, Ph.D. Donna Haiduven, Ph.D. Yougui Wu, Ph.D. Roger Sanderson, MA. Date of Approval: April 4, 2007 Keywords: Canada, CDAD, co-morbidity, mortality, AHCA Copyright 2007, Owen Bendixsen
i Table of Contents List of Figures vii List of Tables ix Abstract x Chapter One: Introduction 1 Chapter Two: Background on Clostridium difficile 4 HISTORY 4 CLINICAL FEATURES 5 Clostridium difficile associated disease 5 Pseudo-membranous colitis 6 Perforated colon and toxic megacolon 7 Septicemia 7 Death 7 EPIDEMIOLOGY 8 Age 8 Gender and racial distributions 8 Geographical distribution 9 Pre-existing medical conditions 9 MICROBIOLOGY 10 TRANSMISSION 11
ii DIAGNOSIS 12 TREATMENT 13 Antibiotics 13 Alternative methods 14 PREVENTION 14 Cleaner hospitals 14 Isolation/Precautions for CDAD 15 Hand hygiene 16 Drug therapy 17 Chapter Three: Literature review: The rising trend of Clostridium difficile 18 DISCHARGE DEFINITI ONS AND CODING FOR Clostridium difficile 18 REPORTING OF SUSPECTED DISCHARGES IN HOSPITAL RECORDS 18 CONDITIONS THAT ACCOMPANY Clostridium difficile 19 GENETIC TYPES OF Clostridium difficile 20 EPIDEMIOLOGY 21 Risk factors 21 Antibiotics 21 Age 21 Studies of Prevalence 23 Canada 23 United States 27 United Kingdom 29 Other countries 30
iii STUDIES TO PREVENT THE IN CREASING PREVALENCE OF CDAD 31 Chapter Four: Methods 34 STUDY DESIGN 34 STUDY POPULATION 34 INCLUSION CRITERIA 34 EXCLUSION CRITERIA 34 DATA SOURCES 35 DATA MANAGEMENT 35 DEFINITION OF VARIABLES 37 DATA ANALYSIS 37 Chapter Five: Results 39 STUDY POPULATION CHARACTERISTICS 39 Age 39 Gender 39 Racial distribution 40 Length of hospital stay 41 Cost per patient 42 Co-morbidities 42 CHANGES OF INTEREST IN POPULATION OVER TIME 43 Discharges 43 Gender for discharges in individual age groups 44 Racial distribution for the age groups 44
iv Length of hospital stay 44 Deaths 45 Chapter Six: Discussion 46 LIMITATIONS OF THE STUDY 46 STRENGTHS OF THE STUDY 46 INTERPRETATIONS OF RESULTS 48 Characteristics of the study population 48 Age 49 Age stratification 49 Length of hospital stay and cost per patient 50 Gender 51 Racial distribution 51 Death rate information 52 Co-morbidities and what they mean 53 Summary of bias, confounding, and chance 53 CONCLUSIONS 54 RECOMMENDATIONS 54 References 56 Appendices 62 Appendix A: Figure 1: Florida hospi tal patients, median age, discharged with and without CDAD, 1998-2004. 63 Appendix B: Figure 2: Florida hos pital patients, percent >60 years old, discharged with and without CDAD, 1998-2004. 64 Appendix C: Figure 3: Florida hospital patients, percent non-white, discharged with and without CDAD, 1998-2004. 65
v Appendix D: Figure 4: Average length of hospital stay for CDAD and non-CDAD discharged patients from Florida hospitals, 1998-2004. 66 Appendix E: Table 5: Florid a non-CDAD and CDAD patients average length of hospital stay, days, 1998-2004. 67 Appendix F: Figure 5: Florida averag e patient hospital co st at discharge, $/patient, for CDAD and non-CDAD patients, 1998-2004. 68 Appendix G: Figure 6: Florida hos pital discharges, toxic megacolon and perforated colon, for Clostridiu m difficile positive patients, 19982004. 69 Appendix H: Figure 7: Florida t oxic megacolon and perforated colon discharges per 1,000 positive disc harged CDAD patients, 19982004. 70 Appendix I: Figure 8: Florida Clostridium difficile positive patients discharged from hospitals, 1998-2004. 71 Appendix J: Figure 9: Florid a patients contract ing CDAD per 1,000 discharged patients, 1998-2004. 72 Appendix K: Figure 10: CDAD positive patients per 100,000 persons in Florida, 1998-2004. 73 Appendix L: Figure 11: Patients discharged from Florida hospitals, positive for Clostridium difficile by age group 1998-2004. 74 Appendix M: Figure 12: Fl orida CDAD positive patients per 1,000 patients discharged, by age group, 1998-2004. 75 Appendix N: Figure 13: CDAD pr evalence per 1,000 Florida discharged white male patients 1998-2004. 76 Appendix O: Figure 14: CDAD pr evalence per 1,000 Florida discharged white female patients, 1998-2004. 77 Appendix P: Figure 15: CDAD pr evalence per 1,000 Florida discharged black male patients, 1998-2004. 78 Appendix Q: Figure 16: CDAD pr evalence per 1,000 Florida discharged black female patients, 1998-2004. 79
vi Appendix R: Figure 17: CDAD pr evalence per 1,000 Florida discharged other race male patients, 1998-2004. 80 Appendix S: Figure 18: CDAD pr evalence per 1,000 Florida discharged other race female patients, 1998-2004. 81 Appendix T: Figure 19: Florida hospital patients, percent mortality, CDAD and non-CDAD, 1998-2004. 82 Appendix U: Figure 20: Florida hosp ital patients, percent mortality, nonCDAD by 11 age groups, 1998-2004. 83 Appendix V: Figure 21: Florida hos pital patients, percent mortality, CDAD positive patients by 11 age groups, 1998-2004. 84 Appendix W: Table 8: The mortality percent for CDAD discharges in Florida acute care hospitals, by age group, 1998-2004. 85 Appendix X: Table 9: The mortality of CDAD patients, in Florida, per 10,000 discharged patients, by age group, 1998-2004. 86 Appendix Y: Table 10: Discharges of Clostridium difficile in Floridas acute care hospitals with the co rresponding percent of discharges for each year, 1998-2004. 87 Appendix Z: Table 11: Total deaths for Florida non-CDAD discharged patients and the percent of deaths per year, 1998-2004. 88 Appendix AA: Table 12: Florida total death numbers for the CDAD patients and the percent of death for each year, 1998-2004. 89
vii List of Figures Figure 1: Florida hospital patients median age, discharged with and without CDAD, 1998-2004. 63 Figure 2: Florida hospital patients percent >60 years old, discharged with and without CDAD, 1998-2004. 64 Figure 3: Florida hospital patien ts, percent non-white, discharged with and without CDAD, 1998-2004. 65 Figure 4: Average length of hospital stay for CDAD and non-CDAD discharged patients from Florida hospitals, 1998-2004. 66 Figure 5: Florida average patient ho spital cost at discharge, $/patient, for CDAD and non-CDAD patients, 1998-2004. 68 Figure 6: Florida hospital discharg es, toxic megacolon and perforated colon, for Clostridium diffic ile positive patients, 1998-2004. 69 Figure 7: Florida toxic megacol on and perforated colon discharges per 1,000 positive discharged CDAD patients, 1998-2004. 70 Figure 8: Florida Clostridium difficile positive patients discharged from hospitals, 1998-2004. 71 Figure 9: Florida patients c ontracting CDAD per 1,000 discharged patients, 1998-2004. 72 Figure 10: CDAD positive patients per 100,000 persons in Florida, 1998-2004. 73 Figure 11: Patients discharged fr om Florida hospitals, positive for Clostridium difficile by age group 1998-2004. 74 Figure 12: Florida CDAD pos itive patients per 1,000 patients discharged, by age group, 1998-2004. 75 Figure 13: CDAD prevalence per 1, 000 Florida discharged white male patients 1998-2004. 76
viii Figure 14: CDAD prevalence pe r 1,000 Florida discharged white female patients, 1998-2004. 77 Figure 15: CDAD prevalence per 1, 000 Florida discharged black male patients, 1998-2004. 78 Figure 16: CDAD prevalence pe r 1,000 Florida discharged black female patients, 1998-2004. 79 Figure 17: CDAD prevalence per 1,000 Florida discharged other race male patients, 1998-2004. 80 Figure 18: CDAD prevalence per 1,000 Florida discharged other race female patients, 1998-2004. 81 Figure 19: Florida hospital pati ents, percent mortality, CDAD and non-CDAD, 1998-2004. 82 Figure 20: Florida hospital patien ts, percent mortality, non-CDAD by 11 age groups, 1998-2004. 83 Figure 21: Florida hospital patien ts, percent mortality, CDAD positive patients by 11 age groups, 1998-2004. 84
ix List of Tables Table 1 Population of Florida, total number of acute care discharges and the median ages of both the CDAD and non-CDAD groups of discharged patients for the years 1998-2004. 39 Table 2 The gender distribution of the CDAD and non-CDAD Florida acute care hospital patient groups for the years 1998-2004. 40 Table 3 The racial distribution of the CDAD and non-CDAD Florida acute care hospital patient groups for the years 1998-2004. 41 Table 4 The average length of stay in the hospital for the CDAD and nonCDAD Florida acute care hospita l patients for the years 1998-2004. 41 Table 5 The non-CDAD and CDAD patient populations average length of hospital stay in days for persons di scharged from Florida acute care hospitals for the years 1998 and 2004. 67 Table 6 The total cost per discha rge for CDAD and non-CDAD Florida acute care hospital patients for the years 1998-2004. 42 Table 7 The number of CDAD positive discharges and CDAD discharges per 1,000 Florida hospital acute care discharged patients for the years 1998-2004. 43 Table 8: The mortality percent for CDAD discharges in Florida acute care hospitals, by age group, 1998-2004. 85 Table 9: The mortality of CDAD patient s, in Florida, per 10,000 discharged patients, by age group, 1998-2004. 86 Table 10: Discharges of Clostridium difficile in Floridas acute care hospital with the corresponding percent of discharges for each year, 1998-2004. 87 Table 11: Total deaths for Florida non-CDAD discharged patients and the percent of deaths per year, 1998-2004. 88 Table 12: Florida total death numbers for the CDAD patients and the percent of death for each year, 1998-2004. 89
x The Rise of Clostridium difficile in Florida Owen Bendixsen ABSTRACT Clostridium difficile a bacterium that causes diarrhea in hospitalized patients, is on the rise in the United Stat es as well as in other countries. This study was done to determine the extent of the problem in Floridas acute care hospitals. The Agency for Healthcare Ad ministration (AHCA) provided data for patients discharged from Floridas acut e care hospitals for the years 1998 through 2004. This study will focus on changes in the prevalence of Clostridium difficile associated disease (CDAD) over time. The mortality and morbidity of patients that have CDAD will also be examined to show if the disease is increasing over time. Fact ors investigated in this study that may influence the prevalence of CDAD include gender, race, length of hospital stay, age, and the cost per patient discharged. In Florida the prevalence of CDAD has risen from 3.41 per 1,000 discharged patients in 1998 to 8.11 per 1,000 discharged patients in 2004. The mortality increased from 9.48% for CDAD positive patients in 1998 to 10.11% for CDAD positive patients in 2004. Age plays a role in both the prevalence and mortality of this disease. In 2004 the mo rtality of patients who were positive for
xi Clostridium difficile was 4.1% for those individuals that were 30-40 years old compared to 0.54% mortality for those pati ents in the same age group that did not have CDAD. The corresponding mortality fo r the patients aged 70-80 for the year 2004 was 11.1% for persons who had CDAD and 3.58% mortality for patients discharged with no CDAD. The analysis showed that CDAD prevalence is increasing in Florida acute care hospitals. During 1998-2004 mo rtality rates for patients diagnosed with CDAD is also increasing. This analysis also indicates that age is a factor that increases the death rates for patients th at are CDAD positive. A more concerted effort to implement hospital techniques that prevent the increasing prevalence of Clostridium difficile in Florida hospitals is recommended.
1 Chapter One: Introduction Clostridium difficile associated diseas e (CDAD) is the most common cause of nosocomial associated diarrhea (Mylonakis, Ryan, & Calderwood, 2001). In recent years, a noted increase in prevalence of CDAD has been documented in developed countries (Valiquette, Low, Pepin, & McGeer, 2004). This disease increase generates a concern for the well being of persons that are admitted to hospitals. The two most notable increases of prevalence for this disease have been documented in Canada and the United Stat es (US) (Valiquette et al, 2004). Encompassing the years 2002 to 2004 a region of Quebec, Canada experienced an increase from 2.1 discharges per 1000 admissi ons to a projected 18 discharges per 1000 by the end of 2004 (Valiquette et al, 2004). Th e increasing number of patient discharges positive for CDAD is not the only concern. Anot her concern is that the reports also show increasing death rates from this disease. In a region of Quebec, Canada the death rate from CDAD within 30 days after diagnosis increased from 4.7% in 1991/1992 to 13.8% in 2003 (Pepin, Valiquette, & Cossette, 2005c). Recognizing that Clostridium difficile infections are increasing has prompted researchers from the US to evaluate the ex tent of increasing CDAD in this country. McDonald and his colleagues noticed that since 1996 the prevalence of CDAD has risen from 31 per 100,000 persons in 1996 to 61 per 100,000 in 2003 (McDonald, Owings, & Jernigan, 2006). McDonald et al also noted that age as a factor that increases the
2 prevalence of CDAD. For persons > 64 the prevalence of CDAD increased from 150 per 100,000 population in 1996 to over 325 discharg es per 100,000 in 2003 (McDonald et al, 2006). The CDAD increasing prevalence in the US has caused some researchers to investigate the extent of this problem at th e state level. Florida is projected by 2010 to have over 3.5 million people that are 65 y ears and older (McCharen, 2000). Increasing CDAD prevalence in hospitals makes this dis ease a concern for doctors that are treating geriatric patients. McDonald et al noted that the region of the US that includes Florida has shown an increase in prevalence of CDAD from 20 per 100,000 persons in 1996 to over 40 per 100,000 persons in 2003 (McDonald et al, 2006). This study will use a cro ss sectional study type format to determine the prevalence of Clostridium difficile for patients discharged from acute care hospitals in Florida. The cross sectional study format is the best format for providing this information on the prevalence of CDAD in Fl orida and will provide ideas for further research that is needed to reduce this di sease in Florida. The AHCA data contains primary and secondary diagnose codes. These codes are presented in the data using the International Classification of Disease, 9th revision (ICD-9) format. The ICD-9 code for CDAD is 00845 and a discharged patient is considered positi ve for this disease if the code is present in either the primary or secondary dia gnosis code (Hart & Hopkins, 2004). The questions that this res earch will answer include: Question 1. Is the CDAD prevalen ce increasing in Florida? Question 2. Are the mortality rates among C DAD patients increasi ng in Florida?
3 Question 3. Is patient age a contributing factor for increasing CDAD prevalence in Florida? Question 4. Are patients with CDAD showing an increase in morbidity from the disease through the years of this study? Question 5. Can an assumption be made as to when the more virulent strain of Clostridium difficile entered Florida? Other factors discussed are th e change in length of hospital stay for patients that are infected with CDAD, and if gender and r acial ethnicity play a role in being more susceptible to Clostridium difficile These factors will be addressed in an effort to explain the prevalence of CDAD in Florida. Although both antibiotics and infection cont rol practices are pertinent topics for CDAD, unfortunately the AHCA data does not allow for a way to measure how these two factors are affecting hospitals in Florida. Information on these two subjects will be provided to give a more complete background of the disease and th e direction some of the research is headed. Information for this study was extracted from AHCA data. This study uses the hospital discharge data portion of the AHCA data. The data set includes all of the patients discharged from Floridas acute care hospitals, not includi ng federal hospitals. This study has been designed to research the effect of CDAD prevalence on patients discharged from Florid as acute care hospitals.
4 Chapter Two: Background on Clostidium difficile HISTORY Clostridium difficile is an anaerobic spore forming gram-positive rod that is the most common cause of nosocomial associated diarrhea. Originally named Bacillus difficilis Clostridium difficile was first described in the mid-1930s (Voth & Ballard, 2005). This bacterium typically has been considered a nuisance causing minor amounts of disease and delaying patients discharge from the hospital by a few days. In 1978, Clostridium difficile was identified as the primary cause of psuedomembranous colitis and was proven to be a primary isolate from the feces of humans undergoing clindamycin therapy (Voth & Ballard, 2005). Typically, Clostridium difficile has been difficult to remove from a hospital setting because of its spore production that allows the bacterium to survive longer in this environment. Te sting for the bacteria has been difficult and labor intensive in the past. Many new tests have been invented to make recognition of the disease quicker and easier. A report by the Canadian Medical Asso ciation discusses th at the old style hospitals in Canada were not equipped to care for Clostridium difficile patients (Eggertson, 2005). Eggertsons paper contains q uotes from other arti cles that hospitals built 100 years ago were not capable of cu rtailing a hospital induced problem like Clostridium difficile (Eggertson, 2005). Eggertsons ar ticle also discusses that some Canadian medical professionals are dow nplaying the increased prevalence of Clostridium
5 difficile (Eggertson, 2005). At the beginning of the 21st century the prevalence of CDAD in hospitals seemed to change and the first i nklings of these changes started to appear in peer reviewed journals. Egge rtson reports that at this time, especially in Montreal, a significant amount of increased patient discharg es with CDAD starte d to occur together with an increase in deaths due to Clostridium difficle (Eggertson, 2005). Some authors have even labeled the problem as an epidem ic (Eggertson, 2005). This bacterium is now starting to garner more press worldwide and has started to raise concerns because of increased prevalence of CDAD in more and more countries. These facts are one of the major reasons this resear ch project is being done. CLINICAL FEATURES Clostridium difficile associated disease Clostridium difficile manifests itself most readily as Clostridium difficile associated disease. Onset of symptoms usually occurs in 5-10 days after antibiotic use, but ranges from 1 day to up to 10 weeks afte r antibiotics are stopped (Oldfield, 2004). The extremes of the ranges (1 day and 10 weeks) for onset of symptoms are rare. Clinical symptoms include watery diarrhea, fe ver, loss of appetite, nausea, and abdominal pain/tenderness (CDC Clostridium difficile : Information for healthcare providers, 2005). This disease is the only nosocomial diarrhea from the Clostridium family that should readily be screened for in hos pital settings (Heimesaat, Gransow, Leidinger, & Liesenfeld 2005). Once the bacterium is in the intestinal tr act it causes disease by secreting toxins. These toxins cause the problems associated with infections from Clostridium difficile and also provide the means to more readily detect the disease (Voth & Ballard, 2005).
6 These toxins lead to more problems and contribut e to a higher virulence of the bacterium. The other complications of th e disease are asymptomatic co lonization, diarrhea, pseudomembranous colitis, sepsis, toxic megacol on, colonic perforati on, and death (CDC Clostridium difficile : Information for healthcare providers, 2005). Other conditions associated with the bacteria that cause mo re severe problems are white blood cell count of >20,000 or less than 1500 cells/mm cubed, a bdominal pain associated with a preexisting condition or post-operatively, or the presence of bowel wall thickening and ascites (determined by CT or abdominal ra diography) (McEllistrem, Carman, Gerding, & Zheng, 2005). Asymptomatic colonization is believed to be as high as 5% in normal adults. Some reports from long term care facilities set the disease carrier rate higher, but the numbers vary depending on the facilities poole d (Simor et al, 2002). Other reports show that this rate in adults may be as high as 20% of the people cultured (Oldfeild, 2004). This rate seems high compared to other publ ished reports, but it is mentioned here to show the diversity of CDAD carri er rates. A person may carry the bacteria in their colon and not exhibit any symptoms and then in 2 months spontaneously defecate the remaining bacteria and not be a carrier. Pseudo-membranous colitis One of the symptomatic signs of CD AD is the presence of psuedo-membranous colitis. The detection of psuedo-membranes is a highly specific indicator of CDAD and almost pathognomonic of the disease (Bouza, Munoz, & Alonso, 2005). Pseudomembranous colitis appears in less than 25% of the discharges of CDAD, and is characterized by white-yellowish plaques pres ent anywhere in the colon (Bouza et al,
7 2005). Pseudomembranous colitis is a good sign th at the disease is present, but as stated does not always occur and would require a very invasive test to prove that it is present. Perforated Colon and Toxic Megacolon Some manifestations of Clostridium difficile in advanced stages are toxic megacolon and perforated colon. Toxic meg acolon is a dilated colon with abdominal distention that is usually ma nifested in Crohns disease a nd ulcerative colitis patients (Nayer, Vetrivel, McElroy, Pai, & Koerner, 2005 ). The usual symptoms consistent with toxic megacolon are hemorrhagic and necrot ic colitis from caecum to lower sigmoid colon sparring of the rectum. Toxic megacol on can become very dangerous and can lead to bowel surgery to repair the damage. In CDAD patients toxic megacolon is rare, but is a recognized complication (Nay ar, Vetrivel, McElroy, Pai & Ko erner, 2005). Perforated colon is another advanced stage of CDAD causi ng perforations in th e wall of the colon. Septicemia Septicemia occurs when bacteria enters the bloodstream and causes an infection. This condition is very serious and if not treated can cause death. The treatment for septicemia is antibiotics. This condition, when discussing CDAD, can be caused by the disease or the treatment for septicemia can lead to an infection with Clostridium difficile Septicemia is a problem inherent when usi ng hospital discharge data, because the time for onset of septicemia is not recorded in the records. Death The increase in deaths of patients that have been diagnosed with CDAD is a growing concern in Canada, United States United Kingdom and other countries. McDonald and his colleagues reviewed the mortality from this disease for the year 2000
8 and 2003. They found that although the number of deaths had risen from 8000 in the year 2000 to 15000 in the year 2003, the percent mortality did not rise significantly in the years 2000 to 2003 (McDonald et al, 2006). Res earchers have begun to notice that death is becoming a more common occurrence among patients that have Clostridium difficile Death occurs in CDAD patients when the ba cteria become too much for the body to recover from the disease or the treatment is inef fective. It is feasible that patients with Clostridium difficile could have other illnesses that predispose them for death. EPIDEMIOLOGY Age A majority of the literature examined s uggests that age is an important factor causing CDAD. An older patient is more lik ely to contract the di sease (McDonald et al, 2006). The age factor is compounded by the fact that older people are more likely to be hospitalized. Increasing a persons stay-tim e in a hospital increases the likelihood that the person will be introduced to the bacteria. The other extreme, neonate and pediatri c patients, have a low prevalence for CDAD (Tang, Roscoe, & Richardson, 2005). The interesting fact about neonates is that they tend to have higher colonization of the b acteria, as high as 64%, in their bowels, but the prevalence of CDAD in neona tes is low (Tang et al, 2005). Tang and his colleagues recommended that routine testing of infants for Clostridium difficile is not necessary because of the low occurrence of diseas e among this age group (Tang et al, 2005). Gender and Racial Distributions Gender and racial distributions do not a ppear to predispose anyone for CDAD. No literature was found that suggested a factor associated with Clostridium difficile for
9 any gender or racial group to have a higher frequency of this disease than any other group. This makes sense because Clostridium difficile is caused by a bacterium that affects hospitalized patients, i ndependent of gender or race. Geographical distribution An article by the Canadian medical a ssociation journal reported that too many hospitals in Montreal ar e battling outbreaks of Clostridium difficile (Eggertson & Sibbald, 2004). This report has been follo wed by other reports from Canada attempting to define the CDAD problem and why it is occurring. Another re searcher from the Netherlands discusses outbreak s occurring in his c ountry and refers to similar problems that the United Kingdom (UK), Canada, and the United States (US) are experiencing (Brierley, 2005). The CDAD problem is also reported by a researcher in Sweden indicating that this country is also experiencing an increasing prevalence of CDAD (Noren, 2005). These reports suggest that the prevalence of this disease is increasing and according to published reports occurs more freque ntly in developed countries. Therefore, the findings of this research paper that focu ses on data generated in Florida may provide important information about CDAD prevalence. Pre-existing medical conditions Another factor considered for the epid emiology of CDAD is pre-existing medical conditions. Similar to the age factor, pre-ex isting complications from other diseases appear to create conditions for contracting CDAD. Any medical c ondition that requires a person to be placed on antibiotics and to ha ve a prolonged stay in the hospital will increase a persons risk for an infection with Clostridium difficile The risk of developing CDAD as a hospital outpatient is about 7.7 discharges pe r 100,000 patients, while the
10 inpatient risk rate can be as high as 25 to 60 discharges per 100,000, depending on the antibiotics used (Mylon akis et al, 2001). One pre-existing condition th at is believed to cause CDAD is the use of proton pump inhibitors (PPI). PPI are drugs that attempt to redu ce the amount of acidity in the stomach. One such drug is Nexium. Some authors believe that th is class of drugs has contributed to the increasing prevalence of CDAD (Dial, Alrasadi, Manoukian, Huang, & Menzies, 2004; Kazakova, et al, 2006). Other researchers have noted PPI use, as a cause of increase in CDAD discharg es, is heavily confounded by lengt h of hospital stay and age of the patient (Pepin et al, 2005b). Ot her researchers found that PPI use did not significantly cause increases in CDAD (Loo et al, 2005). In an effort to look into community acquired CDAD, researchers are c ontinuing to look at gastric acid and suppressive agents (such as PP I) as reasons why more persons that have no history of being in a hospital might be contracting C DAD (Dial, Delaney, Barkun, & Suissa, 2005). More research into this topi c needs to be performed to make an accurate assessment of the affect of PPI on CDAD prevalence. MICROBIOLOGY The primary effect of Clostridium difficile on the human body are the production of two toxins produced during bacterial secretion. These toxins secreted by Clostridium difficile are labeled toxin A and toxi n B. Toxin A is considered an enterotoxin and cytotoxin and toxin B is a cytotoxin (McD onald et al, 2006; Gerding, 2005). These toxins are considered the main virulence fact ors for the disease. The enterotoxin (toxin A) causes disease by increasing the release of wa ter, which will lead to an increase in the release of electrolytes. This increase in the release of both electrolytes and water leads to
11 the diarrhea portion of the disease. The cytotoxin (both A and B) acts by killing neighboring cells that will in turn cause the inflammation associated with the disease. Some evidence is provided, however, that pa tients having no toxin A and having toxin B still have the disease (V oth & Ballard, 2005). The toxins of Clostridium difficile are increasingly b ecoming important in identifying the extent of the disease. Voth and Ballard quoted that a relationship has been proven to exist between toxi n levels and development of psuedomembranous colitis (Voth & Ballard, 2005). These toxins have also been important in the genetic mapping of the organism. Mapping of the organism, in the future, may lead to new mechanisms to limit the growth of the organism. Clostridium difficile will grow in an anaerobic en vironment on normal blood agar, but this process is time consuming and other tests have been developed that shorten the time required for detecting CDAD toxin production (Voth & Balla rd, 2005). Spore production readily occurs in a hospital environmen t. These spores carry the bacteria in a capsule that can withstand environmental pr essures for many weeks and even months in some charges (Gerding, Johnson, Peterson, Mull igan & Silva, 1995). These spores are highly resistant to disinfectants and cleaning procedures. TRANSMISSION The most likely transmission of Clostridium difficile occurs from the hands of hospital personnel, fomites, or the envi ronment surrounding hospital or nursing home buildings (Simor et al, 2002). Therefore, locating the sources of this bacterium in hospitals is very important. A patient who has Clostridium difficile and shares a room with a patient who uses antibiotics increases the CDAD prevalence for the antibiotic user
12 (McConnell, 2002). Also, a person that shares bathroom facilities with an infected CDAD patient has an increased prevalence fo r CDAD. If the hosp ital facilitie s are not properly cleaned or the infected patients are not detected and isolated appropriately, then patients who share rooms with CDAD patients wi ll have a significantly higher chance of contracting CDAD. In Canada, Pepin and his colleagues have observed that the infrastructure of some of the Canadian hospitals contributed to the problem they are having with Clostridium difficile (Pepin et al, 2005b). Pepin and his colle agues refer to conditions such as old plumbing, walls that are decaying, and 4 to 6 patients per room as some of the infrastructure problems that persist in the hospitals of Quebec, Canada (Pepin et al, 2005b). Pepin also notes that the populati on of Quebec is aging, and along with an increased use of alcohol hand r ubs, may have led to the increase in discharges noted in that country (Pepin et al, 2005b). Hospital objects that te nd to carry the disease are commodes, bathing tubs, and electronic th ermometers. These objects need to be monitored closely because of th eir propensity to carry spores of the bacteria (Gerding et al, 1995). Besides antibiotics, chemotherapeutic agents also seem to place patients at risk for the disease (Gerding, 2005). DIAGNOSIS In the past, diagnosis of Clostridium difficile relied solely on culture for detection. Patients with leukocytosis and hypoalbuminemia associated with their diarrhea should be considered candidates to test for the presence of Clostridium difficile toxins in their stool (Oldfield, 2004). The current hospital laborato ry detection method is direct testing by enzyme immunoassay for toxin A and/or B wh ich has a sensitivity and specificity of 88-
13 93% and 99-100% respectively (Tang et al, 2005; Russmann, Panthel, Bader, Schmitt, & Schaumann, 2007). This test provides a much more rapid detection of the bacteria. Patients in a hospital that are on antimicrobial therapy and discharge 3 or more watery or unformed stools in a 24 hour pe riod should be tested for Clostridium difficile Cell cytotoxicity testing with a se nsitivity of 80-90% and specifici ty of 99-100% is considered the gold standard for detection of CDAD toxi ns, but can take up to 24 hours for detection so is not routinely used for rapid detec tion in a hospital settin g (Price et al, 2006). TREATMENT Antibiotics The first step in treatment of Clostridium difficile is to stop using the offending antibiotics (Oldfield, 2004). Some studies have shown that in 15-23% of CDAD discharges discontinuing use of the offending antibiotic can relieve symptoms (Aslam et al, 2005). In some instances changing trea tment to a more eff ective antibiotic is necessary to not only make the patient bett er but also to prevent the occurrence of CDAD. Metronidazole is the drug of choice when treating an infection with Clostridium difficile (Aslam et al, 2005). Oral Vancomycin is used in place of Metronidazole when patients fail therapy with Metronidazole, a more severe disease is encountered, or a patient is pregnant or less than 10 years old (Modena, Gollamudi, & Friedenberg, 2006). Metronidazole is considered the first c hoice to prevent further colonization of Vancomycin resistant enterococci (VRE) (O ldfield, 2004). Bacitracin and Teicoplanin have been used to treat CDAD, but are not cons idered the first line of antibiotics to treat the disease (Oldfield, 2004; Aslam et al, 2005).
14 There are antibiotics that are being studi ed as to their effectiveness against Clostridium difficile Researchers are trying to find antib iotics that are more effective at killing Clostridium difficile and preventing the bacterium to gain resistance. Some researchers have looked more into the eff ectiveness of Metronidazole (Musher et al, 2005). Other researchers have investigated gi ving oral antibiotics versus intravenous antibiotics preoperatively (Wren, Ahmed, Jamal, & Safadi, 2005). The research into new drugs has led investigators to look at two drugs called Nitazoxani de and fusidic acid (Wullt & Odenholt, 2004; Aslam et al, 2005). This information on antibiotic usage is presented because of the important role that anti biotics play in the prev alence of CDAD. Alternative methods Alternative methods have been shown to be somewhat effective in treatment of CDAD. The methods in use include, with lim ited effectiveness, probiotics (introducing helpful bacteria into the colon), immunoglobu lins, and steroids (Aslam et al, 2005). Some of the probiotics in use are yeasts, Bifid bacterium, Lactobacillus GG, L. rhamnosus, L casei, L. plantarum, and Entero coccus faecium (SF68) (Dendukuri, Costa, McGregor, & Brophy, 2005). Other methods in us e are the use of Cholestyramine, which binds the toxins of Clostridium difficile and inactivates those t oxins (Oldfield, 2004). Unfortunately, Cholestyramine inactivates Vancomycin when the two drugs are used together. PREVENTION Cleaner hospitals Ways to prevent the patient from co ntracting the disease rely on adequate cleaning of the hospital rooms, equipment, and environment. Clostridium difficile spore
15 production allows the bacteria to survive for up to 60 days in the hospital environment. This presents a problem for housekeeping in th e hospital. Appropria te agents that are sporicidal are needed when cleaning rooms that contained previous patients with Clostridium difficile (Wilcox et al, 2002). These agen ts can be tertiary ammonium, oxygen-based, hypochlorite disinfectants, and detergent based (Wilcox et al, 2002). Effective cleaning of hospital environments is also a topic that this study will be unable to access efficiently. The CDC reco mmendations in 2003 for appropriate cleaning agents to eradicate Clostridium difficile from the environment include meticulous cleaning of all surfaces followed by hypochlor ite based germicides as appropriate (Schulster & Chinn, 2003). Two other groups of researchers also investigated which cleaning agents are the most eff ective at reducing the amount of Clostridium difficile in the environment and also decreasing the num ber of CDAD discharges. These researchers noted varying degrees of effectiveness (F awley, Parnell, Verity, Freeman & Wilcox, 2005; Wilcox et al, 2003). Unfortunately, cu rrent data as to what cleaning agents are used routinely in Florida hospitals is not available. The CDC guidelines are the best recommendations to limit the amount of bact erium that persist in the environment. Isolation/Precautions for CDAD The CDC isolation precaution guidelines for Clostridium difficile are to place the patient in contact isolation for the dur ation of their illness (Garner, 2005). Contact precautions include placing the patient in a priv ate room with their own bathroom or if a private room is not available placing them in a room with another CDAD patient. When both of these conditions can not be met patient placement needs to be carefully considered because of the infective nature of Clostridium difficile. Gloves should be
16 worn when entering the room and the gloves should be removed before leaving the room. Gowns should be worn if the caregiver is exp ected to have substantial contact with the patient, environmental surfaces or items in th e patients room. Remove the gown before leaving the patients envi ronment (Garner, 2005). Equipment to be used on the patient should be limited only to the patient whenever possible. Sharing of any patient care equipment should be done only after thorough cleaning and disinfecting of the equipment before using on another patient (Garner, 2005). Avoidance of rectal thermometers and cleaning the rooms of infect ed patients should be done w ith a 1:10 bleach solution (Bartlett, 2006). Diarrhea in an adult that has a history of recent anti biotic use is also to be considered for contact isolation so as to prevent a potential spread of Clostridium difficile even though CDAD status has not made (Gardner, 2005). Hand hygiene It is worthy of note here that the CDC recommendations for hand hygiene to prevent a patient from contracting any diseases, including CDAD. No specific recommendations have been given for Clostridium difficile but in 2002 the appropriate recommendations for hand washing were gi ven by the CDC. The first and most important recommendation that is given is if hands are visibly soiled with blood or body fluids (including feces) that they must be washed with soap (antibacterial or nonantibacterial) and water (Boyce and Pittet, 2002). This recommendation is a category IA meaning it is strongly recommended and str ongly supported by studies on the topic. Other recommendations associated with C DAD are decontamination after contact with patients skin (category IB), decontaminati on of hands after cont act with blood or body fluids when not visibly soiled (category IA), decont amination of hands after contact with
17 inanimate objects in the immediate vicinity of the patient (category II), and the decontamination of hands after removing gloves (IB) (Boyce and Pittet, 2002). Drug therapy Monitoring of drug therapy is anot her effective means of preventing Clostridium difficile Over medication of patients is a way to increase the prevalence of CDAD in our hospitals. Clindamycin and the Cephalosporin family of drugs are antibiotics that appear to cause the most problems with infection rates of Clostridium difficile (Thomas, Stevenson, Willliamson, & Riley, 2002). Other dr ugs that have attributed to cause the disease are Fluoroquinolones, Macrolides, and intravenous beta-lactam/beta-lactam inhibitors (Pepin et al, 2005). Research into which routinel y used antibiotics are causing discharges of CDAD is well documented and ha s value when considering the best drug to give patients that require tr eatment with antibiotics (Sta rr, Martin, McCoubrey, Gibson, & Poxton, 2003; Gopal, Mahankali, and Star ke, 2003). Hospitals should responsibly monitor drug use and especially the drugs that are known to cau se discharges of CDAD.
18 Chapter Three: Literature Review: The rising trend of Clostridium difficile DISCHARGE DEFINITI ON AND CODING FOR CLOSTRIDIUM DIFFICILE The coding used in this study is the Inte rnational Classificati on of Diseases ninth revision (ICD-9), a coding method for hospital records. Clostridium difficile toxic megacolon and perforated colon have their own specific ICD-9 code, and these codes are specific for determining exact diagnoses. A patient having Clostridium difficile is defined as any person that has tested positive for the bacteria or toxins and has documented more than three loose stools in a ny given day with any of these symptoms: fever, loss of appetite, nausea, and abdominal pain/tende rness (Oldfield, 2004; Dubberke, Reske, McDonald, and Fraser, 2006). Established veri fiable psuedomembranous colitis is also a viable definition of the disease (CDC Clostridium difficile Information for Healthcare providers, 2005). REPORTING OF SUSPECTED DISC HARGES IN HOSPITAL RECORDS A person in a hospital has a medical chart filled out by a doctor. This chart upon discharge from the hospital is given to medical records coding personnel. These coders take the chart and use the ICD-9 code system to appropriately assign what the discharged patient was diagnosed with while in the hospita l. Additionally, the c oders are the persons who make it possible for the hospital to bill th e patient for services rendered while in the hospital. Hospital billing relies on the coder assigning the correct ICD-9 code as an acknowledgement of the procedures that were relied upon to treat the patient while in the
19 hospital. If the coder does not do their job correctly then the patient is over or under charged. The hospital relies on good coders to make the right decisions for accounting and auditing purposes. Researchers have used discharge informa tion that rely on ICD-9 codes for correct recognition of CDAD. These researchers have tested the validity of using this coding system to correctly identify prevalence of CDAD. Scheurer and his colleagues attempted to show that CDAD prevalence can be appr opriately determined by using discharged data. These researchers found that the ICD-9 codes closely approxima te the true amount of Clostridium difficile infections that were in their hospital (Scheurer, Hicks, Cook, & Schnipper, 2006). They did note that includ ing symptomatic patient s whose test results are readily available at time of discharge is helpful to make a more accurate assessment of disease status (Scheurer et al, 2006). Dubberke and his co lleagues also researched the of Clostridium difficile toxin assays with ICD-9 codes at tributed to discharged patients (Dubberke et al, 2006). They found that good correlation existed between the two sources and that large scale use of discharge diagnosis codes can be effective in demonstrated prevalence of th e disease (Dubberke et al, 2006). CONDITIONS THAT ACCOMPANY CLOSTRIDIUM DIFFICILE As discussed earlier there are conditions that Clostridium difficile can progress to. These conditions are perforated colon, toxic megacolon, sepsis, and death. Psuedomembranous colitis is not mentione d here because this condition cannot be differentiated from Clostridium difficile during coding and the disease can occur at the beginning of a Clostridium difficile infection. Sepsis is a nother condition that, using discharge data, is difficult to differentiate from whether or not a person had sepsis before
20 or after contracting Clostridium difficile The above factors make it difficult to determine which disease caused which condition. In cons ideration of these factors this study will only list perforated colon, toxic meg acolon and death as conditions that Clostridium difficile can progress to. GENETIC TYPES OF CLOSTRIDIUM DIFFICILE In the past few years a large amount of research has been completed to discover the gene structure of Clostridium difficile In addition to the topic of toxins, the scientific community has been searching for strains of Clostridium difficile that have been prevalent more often than other strains. The scientific community believes that the presence of at least one of the toxins, eith er A or B, is necessary for disease, with enhanced virulence when both toxins are pr esent (Martirosian, Szczesny, Cohen, & Silva, 2005). This is important when considering that the Center for Disease Control (CDC) and researchers from Canada have named the strain believed to be the cause of the majority of the new outbreaks as hyper virulent toxin type III ribotyp e 027 strain (Pepin et al, 2005c). This strain has also been referred to as North American pulso-type I (NAP1)/ribotype 027 (Louie, 2005). This strain has been implicated in a number of outbreak settings of CDAD (Cloud & Kelly, 2007) This strain has also been shown to produce about 15-20 times more toxin than a normal strain of Clostridium difficile (Louie, 2005). This strain, or a close cousin to it, has also been identified in the UK (Pepin et al, 2005c). This strain is belie ved to be causing the new problems that are present now: increased prevalence of disc harges, more co-morbidities, and higher mortality.
21 EPIDEMIOLOGY Risk factors Clostridium difficile associated diarrhea risk factors are important to understand and to be able to diagnose a nd treat hospitalized patients that may be susceptible to contracting the disease. Six c onditions that can lead to CDAD are: history of antibiotic use, anti-neoplastic agents, age > 60 years, gastrointestinal surgery, enemas or stool softeners, and enteri c feedings, especially post pyl oric (Oldfield, 2004). All of these conditions in some way demonstrate th e destruction of the normal commensal flora that exists in normal human large and small intestines. The two most prominent risk factors that have received the most atten tion are the use of antibiotics and age. Antibiotics Some antibiotics are more likely to cause Clostridium difficile than others, such as Clindamycin, but even Metronidazole and Vanc omycin (both of which are considered for treatment of the disease) have been proven to be a cause of the disease (Oldfield, 2004). A study conducted in Paris, France by Blot and his colleagues determined that chemotherapeutic agents could cause CDAD (Blot, Escande, Besson, Barbut,Granpeix & Asselain, 2003). Age The age of a patient is consistently a concern for persons infected with Clostridium difficile Patients >60 years old tend to have a hi gher prevalence rate, a worse prognosis, and have an increased chance for relapse. There are several reasons for these conclusions, most notably, older people ar e more susceptible to infections of all
22 kinds, have a higher chance of being in a hos pital, and a higher chance of being on antibiotics. Pepin and his colleagues performed a study on what types of people are considered the highest risk for having CDAD (Pepin et al, 2005a). They researched several characteristics that could conceivably cause C DAD. They separated the years 1991-2004 in two year periods with age groups of 17, 18-64, and 65. In the early part of this study the 18-64 age groups always seemed to have more discharges of Clostridium difficile but in 2003-2004 that changed drastically showing that the 65 age group has over twice as many discharges as the other 2 age groups combined (Pep in et al, 2005a). This is important because of the implications this disease has on geriatric patients. Pepin and his colleagues also no ticed that those persons 65 also have a significant chance of relapse when compared to the relapse rate for the 17 year old age group (Pepin et al, 2005a). The published results of this study are important because drug regimens and testing may need to be adjusted to treat th e increasing number of older patients that are contracting this infection. The opposite end of the age spec trum is of concern as well. With a rising trend of CDAD it is only logical to be concerned w ith the effect on the very young. A group of researchers from Ontario, Canada undertook a project to determine the effectiveness of treating infants who contracted Clostridium difficile (Tang et al, 2005). Their research showed that the patients improved regardless of disease status or whether metronidazole was given or not. This research demons trates, especially in the younger age groups, Clostridium difficile as a cause for diarrhea s hould not be routinely tested for, and that the diarrhea experienced by infants is routinely from a source other than Clostridium difficile (Tang et al, 2005).
23 Studies of prevalence Canada Studies of the rising trend of Clostridium difficile have been ongoing because hospitals are concerned about the nosocomial in fection rates. Most reports on prevalence of Clostridium difficile prior to 2004 discuss a small prev alence for this disease with little threat for the disease becoming a problem. In 2004 the Canadian medical association, discussing infection rates after 2000 reported that a high rate of Clostridium difficile is occurring in hospitals in Montreal and 79 deaths had occurred so far (Eggertson and Sibbald, 2004). This preliminary report lacks baseline data or any data to show exactly how much the rates of infection have incr eased. The report does contain a number of quotes from people stating that the probl em will probably become worse with the possibility that Clostridium difficile infections could become a serious problem. This report does not show the extent of the problem but mentions that a problem may exist. The Eggertson and Sibbald report is significan t because after this report was issued, an influx of information on just how much the disease has increased has been reported and the seriousness of this problem persists in Canada based on the studies of other researchers. The increased prevalence of CDAD not only persists in Canada, but it occurs in other parts of North America as reported by a number of researchers. Valiquette and his colleagues note that not very good records ha ve been kept for this disease. They researched data accumulated duri ng 1- years for discharges of Clostridium difficile
24 reported for a hospital in Montreal, Canada They found that in January of 2002 the hospital had about 10 discharges of the dis ease, and in May of 2004 they had almost 40 discharges of the disease and with a high poi nt for the disease in February of 2004 of almost 70 discharges (Valiquette et al, 2004). The prevalence rates for this data is 2.1 discharges per 1000 admissions in 2002, 10 di scharges per 1000 in 2003, and a projected 18 discharges per 1000 admissions in 2004 (Valiq uette et al, 2004). This report shows why Canada is concerned and has led Jacques Pepin and some of his colleagues to begin a formal research study of the complexity of the Clostridium difficile problem for Quebec, Canada. Pepin and his colleagues researched the Canadian Clostridium difficile problem and issued four reports that attempt to define the various factors c ontributing to increased prevalence of Clostridium difficile The first report reviewed the prevalence of CDAD from 1991 to 2003 at the Centre Hospitalier Un iversetaire de Sherbrooke (CHUS). This study is a retrospective chart review inves tigating details such as age, gender, CDAD acquired place, immunosuppression, and antibiotic treatments (Pepin et al, 2004). This hospital provided a well defined population making it possible for the researchers to estimate population based preval ence. The prevalence of CDAD in this Canadian region increased from 35.6 per 100,000 people in 1991 to 156.3 per 100,000 people in 2003 with the 65 years old or higher age group increasing from 102.0 to 866.5 per 100,000 people (Pepin et al, 2004). This study indicates that Canada has a legitimate problem with this disease that needs addressing. The factors contributing to hi gher prevalence of CDAD as indicated by the resear ch report of Pepin et al include immunosuppression, tube feeding, and the place acquired (hos pital versus community).
25 Pepin and his researchers next investigat ed if death rates a nd length of hospital stay were significant factors for the increased prevalence of this strain of bacteria. The following report confined the research to dete rmine if co-morbidity contributed to the disease and also contributed to death in a patient that had CDAD (Pepin et al, 2005c). The database used for this study originated from records of the 687 bed tertiary hospital in Montreal, Canada. After 365 days the study sh owed that on average those patients that had been diagnosed with CDAD had a survival rate of about 65% as compared to the control subjects' survival rate for the same ti me period of about 80% (Pepin et al, 2005c). This indicates that a person who has CDAD has a higher probability of dying. Pepin and his colleagues also completed a 30 day co-mor bidity survival rate study and compared the values to the survival rates of patients after 365 days. They reported for the 30 day study that CDAD merely precipitated an event th at could have occurred in any discharge a few months later, but for the 365 day st udy, about one-sixth of the inpatients with CDAD died (Pepin et al, 2005c). This means that a person contracting CDAD has a higher chance of dying after one year than the group of people that are admitted to a hospital that have no record of CDAD. Pepin and his researchers attr ibuted most of their infections to the NAP1/027 strain (Pepin et al, 2005c). Therefore, the survival probabilities for this strain may not n ecessarily compare to other strains of Clostridium difficile Some issues that have also been report ed for Canada have been discussed earlier in this paper: a) the recurrence rate for me tronidazole treatment is increasing and b) the propensity of fluoroquinolones to cause more discharges of CDAD (Pepin et al, 2005a; Pepin et al, 2005c). Both of these reports provide adequate eviden ce that the prevalence
26 of CDAD is increasing in Cana dian hospitals. These report s provide significant factors contributing to the rising prevalence of CDAD and offer methods that the health community can use to reduce the rising prev alence of CDAD in our own communities. This report by Pepin and his colleagues s how that the recurrence of CDAD after metronidazole therapy increased from 15.2% in 1991-1992 to 47.2% in 2003-2004 (Pepin et al, 2005a). This paper contains specific data for only one hospita l and it could not be confirmed if the results of the survey a pply to any other Montreal hospitals. The November 2005 report by Pepin and his colleagues provides the most extensive research into antibio tics and their involvement with the disease. This report helps to visualize which drugs have the highest factors for contributing to the prevalence of CDAD in Montreal (Pepin et al, 2005b). The results repo rted may not apply to other hospitals reported by different researchers. The Pepin report also may not be useful because the patients antibiotics have been se parated into many groups that decreases the population of discharges reported for each group. One final study worthy of note that originat ed in Canada contai ns research data on the CDAD strain that emerged in Canada, discusses the toxin pr oduction for this new strain, and compares the prevalence of this st rain compared to previous strains for this bacterium. This new strain is labeled NAP 1/027 and is the strain that was discussed earlier in this paper. Thes e researchers found that this strain produ ces 16 times more toxin A and 23 times more toxin B than contro l strains (Warny et al, 2005). This greater toxin production led these rese archers to hypothesize that th e new deadlier strain would increase the prevalence of CDAD. Intuitivel y this makes sense beca use an increase in toxin production would increase the prevalence of CDAD. A pr oblem with this report is
27 that the report compares 15 NAP1/027 strains with 25 toxinot ype 0 strains (Warny et al, 2005). The report did not appear to have a high number of discharges and this lowers the significance of the report. Th e culture and growth curve an alysis results have been questioned by a group of researchers (F reeman, Fawley, Baines, & Wilcox, 2006). Freeman and her colleagues refute the methods to reproduce growth in vivo with in vitro analysis are not sound enough to give accurate results. The researchers did study the strains from the same geographical areas as the NAP1/027 strain s so that geography could not be a factor that confounded th e study (Freeman et al, 2006). The report by Warny and his colleagues does a good job of showing reasons why Clostridium difficile is starting to cause problems in other regions of the world, but the numbers for amount of toxin may be lower than reported by this research (Warny et al, 2005). United States Research results from Ca nada hospitals concerning Clostridium difficile prevalence may be important for US hospita ls because Canada is a close neighbor. Valiquette et al researched the Clostridium difficile rates for Canada, and other areas of the world to demonstrate that the disease is also increasing in othe r parts of the world (Valiquette et al, 2004). These researcher s used data from Or egon, Pittsburgh, and the US national nosocomial surveillance databa se. These three respective places all experienced some type of increase whether it was in CDAD prevalence or in mortality rates (Valiquette et al, 2004). More research into the dis ease in the United States needs to be done to determine the extent of the problem. A study was completed by Clifford McDonald and his colleagues. These researchers used the National hospital discha rge survey information records to conduct
28 their research. This study is one of the mo st comprehensive and thorough medical record studies performed on the topic of Clostridium difficile (McDonald et al, 2006). McDonalds report is significant because it defines that the prevalence rates for Clostridium difficile are increasing in the US. McDonald et al shows that the prevalence of Clostridium difficile increased from 31/100,000 people in 1996 to 61/100,000 people in 2003 (McDonald et al, 2006). This report not es that the increased prevalence in the 65 year age group is disproportionately highe r at 228/100,000 people as compared to the next younger age group, 45-64 year old, that had a prevalence of 40/100,000 people (McDonald et al, 2006). Anothe r item of note is that this article also observed higher prevalence rates across the entire US with the northeast region of the US having the highest prevalence rates. It is interesting to note the Sout h region, including Florida, has a prevalence rate of about 20 Clostridium difficile positive patient discharges per 100,000 people in 1996. This rate increased to just over 40 Clostridium difficile positive discharges per 100,000 people in 2003 (McDonald et al, 2006). This report demonstrates that the prevalence rates of CDAD are increas ing in the US and al so increasing in the south region of the US. A previous study done at a statewide le vel was performed in Oregon (Chandler, Hedberg, & Cieslak, 2007). The prevalence of CDAD in Oregon increased from 1.4 to 3.3 discharges per 1,000 discharged patients from 1995 to 2002 (Chandler et al, 2007). These researchers were unable to find any not able changes in hospita l practices that could have perpetuated the rise in prevalence. This reports use of a descriptive study is close in method type to the study that is being researched in this paper fo r the state of Florida.
29 United Kingdom The United Kingdom, similar to U.S. a nd Canada, has also been experiencing a rise in prevalence rates of Clostridium difficile accompanied with a rise in death rates (Brierley, 2005). Rob Brierley discusses sp ecifically the number of deaths was at 25 across 15 hospitals, but does not provide base line information for CDAD prevalence in UK. Brierley does refer to some preliminar y typing of some of the discharges that occurred at the Stoke Mandeville hospital, where the CDAD epidemic was believed to have begun. The preliminary typing indicates that the ribotype 027 strain has at least partial blame for the outbreak in the UK (B rierley, 2005). Valiquette and his colleagues report using information gleaned from the UK health protection agency communicable disease centre (Valiquette et al, 2004). This report shows that in 1986/87 the hospitals had <2000 discharges of Clostridium difficile as compared to >12,000 discharges in 2000/01 (Valiquette et al, 2004). This report de monstrates that the problem with this disease is not confined to Ca nada, but other parts of the wo rld are experiencing the same Clostridium difficile problem. Pepin and his colleagues also note that CDAD prevalence is increasing in the UK and that further research is needed to c ontrol these increases. They report that the prevalence of CDAD has doubled from 2001 to 2004 (Pepin et al, 2005a). The report contains preliminary information that ribotype 027 strain has spread to some areas of the U.K. and has contributed to the increas ed prevalence of CDAD for this country. The previous articles discuss ribotype 027 st rain typing to show th at this strain in England is similar to the one observed in North America. Research by Warren Fawley and his colleagues included iden tifying the important subtypes of the disease in Leeds,
30 United Kingdom (Fawley et al, 2005). Fawley s results showed on ward A that 95.2% of the Clostridium difficile infections were attributed to the PCR ribotype Ia, and in the other ward researched, ward B, the ribotyp e causing 97.5% of the discharges was type PCR ribotype Ia (Fawley et al, 2005). PCR ri botype Ia is similar to the NAP1/027 strain that has been identified in North America. Fa wley also discusses that this strain seems to be thriving and growing and is becoming the mo re predominant strain in other parts of the U.K. as well. Other countries Warren Fawleys typing of Clostridium difficile strains led him to report some instances of PCR type Ia, commonly referred to as ribotype 1, in Belgium and France (Fawley et al, 2005). An outbreak of CDAD was noted in Paris, France in 1996, but a specific strain of the bacteria was not identifi ed as the culprit (Blo t et al, 2003). This study attempted to define why Clostridium difficile is causing infections among chemotherapy patients. This report provides so me evidence that outbreaks of this disease are occurring outsi de of the U.S., U.K., and Canada. Norwegian researchers reported on Clostridium difficile infections observed in two university hospitals in Oslo, Norway. These researchers attempted to discover the extent of the infections caused by this b acterium by examining risk factors that may contribute to the increased pr evalence of CDAD (Berild et al, 2003). The report focused on reasons for the disease outbreaks in their facilities. Berild and colleagues discuss prevalence of CDAD in two hospitals. In 1993 one hospital reported a prevalence of about 4 discharges of CDAD per 100,000 people as compared to that same hospital
31 having a prevalence of about 17 discharges of CDAD per 100,000 people in 2001 (Berild et al, 2003). Sweden is another country reporting mortalit y rates caused by CDAD. Torbjorn Noren reports that the mortality rate for Clostridium difficile from a PCR ribotype 17 strain (serogroup C) was as high as 13% in the year 2000 (Noren, 2005) Norens report attempts to confirm that in Orebro county, Sweden the mortality rate is similar to rates reported by Canadian researchers. Norens report shows that high mortality rates are not confined to only North America but that ot her countries are expe riencing strains of deadly Clostridium difficile infections. STUDIES TO PREVENT THE INCREASING PREVALENCE OF CDAD The rising trend of CDAD is apparent and methods are needed to prevent the increased prevalence of this disease. Thom as J. Louie makes some suggestions based on various articles he has researched (Louie, 2005). His suggestions are reduction in patients in crowded wards, lower toilet to bed ratios, and increased infection control methods in hospitals to treat CDAD infec tions (Louie, 2005). These suggestions are given repeatedly by other re searchers and are not new. One problem with these suggestions is that they require hospital funds to make the necessary changes. Therefore, research that identifies le ss expensive methods for preven ting the prevalence of CDAD and growth of the bacteria are important. One of the factors noted by researchers that can reduce Clostridium difficile infection rates is modifying the antibiotics gi ven in hospital wards. In Western Australia Thomas and her colleagues accumulated data as they made efforts to curtail the prevalence of Clostridium difficile in their hospital (Thomas et al, 2002). These
32 researchers noted that the prevalence of Clostridium difficile has been increasing gradually at their hospital and methods were used to modify antibiotic usage in an attempt to curtail the increas ing infection rates (Thomas et al, 2002). This hospital almost completely abolished the use of thir d-generation cephalosporins by the year 2000 in an effort to prevent further growth of the bacteria. This method decreased the prevalence from almost 3 discharges of CDAD per 1000 patients in 1993 to below 1 discharge of CDAD per 1000 patients in 2000. Th e other statistic of note is that by 2000, the amount of CDAD per occupied bed days (20 patients per 100,000 bed days) was at the lowest rate in this hosp ital since 1983 (Thomas et al, 2002 ). This is convincing research that using antibio tics appropriately can aid in lowering infection rates for CDAD. This report fails to record the altern ative drugs used to treat patients. If the alternative drugs used in this hospital were Metronidazole and Vancomycin, then the data seems reasonable because Metronidazole and Va ncomycin are drugs that have been used successfully to treat CDAD. The report contai ns a note that they used Clindamycin on a very limited level, which means that Clinda mycin was not the drug of choice to replace the use of third-generation Ce phalosporins. The amount of people in the study was high providing confidence that the values give n are statistically significant. A report by Vesta and colleagues attemp ted to show that no specific drug is responsible for discharges of CDAD (Vesta, Wells, Gentr y, & Stipek, 2005). This research, although it had good intentions, di d not have enough data points to be statistically significant. Th e report contained data for only 144 patients and separating the data into 9 drug categories reduced the statistical significa nce of the report too low to be comfortable with the conclusions. As di scussed earlier, Pepin and his colleagues did
33 show that Floroquinolones, all Cephalospor ins, Macrolides, and Clindamycin were factors affecting higher prevalence of Clostridium difficile in Quebec hospitals (Pepin et al, 2005b). This report used a database cont aining 7421 discharges of CDAD. The Pepin et al report concludes that the drugs studied do increase prevalence of CDAD in hospitals (Pepin et al, 2005b). A previous paragraph noted that resear ch studies reported on the efficacy of Metronidazole and Vancomycin in tr eating the newly emerging strains of Clostridium difficile (Freeman, Stott, Baines, Fawley, & Wilcox, 2005b). At Leeds, U.K. the resistance of Clostridium difficile bacteria to Vancomycin and Metronidazole remains small according to their research. Another e ffort to find new and more efficient drugs was undertaken by Jane Freeman and her co lleagues. This report discusses the effectiveness of Ramoplanin compared to vancomycin in treating CDAD (Freeman, Baines, Jabes, & Wilcox, 2005a). This report uses hamsters and models treatment by randomizing injections into hamsters for CD AD treatment with either Ramoplanin or Vancomycin. This report showed that, while Vancomycin reduced the amount of bacteria in the intestines during treatment, Ramoplanin reduced both the amount of bacteria and the amount of spores, which Vancomycin failed to do (Freeman et al, 2005a). This suggests that ramoplanin coul d be a good alternative to Vancomycin. No reports on the effectiveness in humans we re found. The research on hamsters is convincing, but further studies into the effectiveness for humans needs to be shown.
34 Chapter Four: Methods STUDY DESIGN This thesis is a descriptive cross-se ctional study for the years 1998 thru 2004 to determine the prevalence of Clostridium difficile in Floridas acute care hospitals. Hospital discharge data was analyzed to comp are the prevalence, length of hospital stay, age, race, gender, and mortality over time for CDAD positive patients to determine if the values are changes from year to year. The da ta analysis also compared the mortality rates for CDAD positive patients to those patients that have no record of the disease. STUDY POPULATION The study population is persons discharged from Floridas ho spitals not including persons admitted to military hospitals. This da ta set contains all hospital discharges from Floridas non-federal hospitals regardless of where the patients pr imary residence is. INCLUSION CRITERIA A discharged patient is cons idered to have CDAD when the primary or secondary diagnosis ICD-9 code is 00845. The disc harge information from 1998 through 2004 is included in this study. The year that a person is discharged from the hospital is the year that that person is included in the data base. EXCLUSION CRITERIA This data set only contains information from Floridas non federal hospitals. Data from other states or countries, nursing home s in Florida, and federal hospitals (military
35 and Veterans Association) in Florida are excluded. Patients w ith CDAD and are not admitted to a hospital are also not included in this data set. Hospital discharge data before 1998 and after 2004 are not included in this study. DATA SOURCES This study analyses data from the Agency for Healthcare Administration (AHCA). The AHCA discharge data used in this study contains hos pital discharges for all non-federal hospitals in the state of Florid a. This data, with permission of the Florida Department of Health, is being used in this study to examine the effects of Clostridium difficile on people being admitted to Floridas acute care hospitals. DATA MANAGEMENT The large amount of information provided by this data required some preparation to make it usable for the purposes of this study. The biostatistical soft ware in use for this study is Statistical Analysis Software (S AS) version 8.1 or higher. The data was provided by year in separate SAS files. Th ere are at least 31 different categories of information. This information has been provided with no personal identifiers so no confidentiality has been breached. The disc harge information was provided in one-year increments. There are some minor differen ces in the names for categories during some years, and these differences have been ch anged so that the information could be combined into one main dataset that contai ns all information from 1998-2004. Important categories used in this study are year of disc harge, discharge status length of hospital stay, gross charges, gender, principal diagnoses code, sec ondary diagnoses code, patient age at admission, and patient race.
36 The information has been compiled on a yearly basis to determine the change over time for CDAD. The prevalence of CDAD has been determined by examining both the principal diagnosis codes and secondary di agnosis codes to esta blish if the patient was given the ICD-9 code of 00845. The di scharges of CDAD have been extracted from the ACHA data so that details of patient s who have been diagnosed with the disease can be analyzed. After the group of Clostridium difficile patients was extracted from the main data, the group was then compared to the remaining non-CDAD patients group to determine prevalence of CDAD. The comparison of Clostridium difficile positive patient population to the non-diseased hospital patient population was done for each year (1998 through 2004) to determine if there are any differences. Along with determining CDAD prevalence, th e next major variable of interest is discharge status. Discharge status is impor tant because this is the primary method for determining if a patient expired while in th e hospital. The information provided in the data that explains discharge status was ch anged from the 10-11 di scharge categories to two discharge categories, whether or not the discharged patient was alive or dead when they left the hospital. These changes allowed this study to determine the death prevalence rates for Clostridium difficile positive patients. Ch anges to the discharge status were performed for both the Clostridium difficile positive population and for the nonClostridium difficile positive population. Death as an outcome is not only important to establish for all patients, but also for all patients that are positive for Clostridium difficile
37 DEFINITION OF VARIABLES Variables are explained in these para graphs providing an understanding for interpreting the results of the data analysis. The length of hospital stay is determined by subtracting the date a person was discharged from the hospital from the time at which they entered the hospital. If a person is di scharged from the hospital and then readmitted to the hospital at some later time after they have been discharged this event would be considered a new admission and the length of hospital stay would st art over with the new admission regardless of what reason the new admission is for. Cost per patient is provided in the data and repr esents the total dollars char ged to the patient. A patients gender is given as male, fema le or other in the ACHA data. The race group is divided into 3 categories: white, bl ack and other. The other category includes persons from races that include American Indi an/Eskimo/Aleut, Asian or Pacific Islander, White Hispanic, Black Hispanic, and other. As discussed previously, Clostridium difficile is given an ICD-9 code of 00845. Other conditions of note for this study are to xic megacolon and perforated colon. These conditions are coded as 556.9 and 569.83 respec tively (Hart and Hopkins, 2004). These two conditions are discussed as th ey pertain to persons who have Clostridium difficile. DATA ANAYLSIS This data has been analyzed on a yearby-year basis to more fully understand the effect that Clostridium difficile has had on the Florida hospital population. Starting with the year 1998, the data has been manipulated to estimate means and sequencing operations was performed to defi ne if the variables important to this study are present. Values computed for this study include pe rcent female, percent non-white, mortality,
38 total patients per year that were diagnosed with Clostridium difficile mean age of population per year, average length of hospital stay per year, and total gross charges per patient per year. These values were then compared to the same values estimated for the Clostridium difficile positive group. The values we re used to compare the Clostridium difficile positive population with the full discharged patient popul ation for each year. The values for mortality among the CDAD patie nts were compared to the mortality for patients that did not have CDAD. To expand the discussion about ages th e data was divided into 11 age groups. The data was divided into 11 age groups to pr ovide a better understanding concerning the role age plays in the disease process. The 11 age groups are 1, > 1-10, >10-20, >20-30, >30-40, >40-50, >50-60, >60-70, >70-80, >80-90, and >90. These age groups were then compared on a year by year basis to consider if changes in average length of hospital stay, prevalence, mortality rate s, race distribution and gende r percentages occurred over the years of the study.
39 Chapter Five: Results STUDY POPULATION CHARACTERISTICS Age An analysis of the AHCA data yielded th e following results: The median age for the non-CDAD patients discharged from Florid a hospitals decreased from 56 in 1998 to 55 in 2004. The Clostridium difficile positive patients increased slightly from a median age of 74 in 1998 to a median age of 75 in 2004 (See appendix A). The percentage of persons with CDAD that are >60 year s old is 74.7% in 1998 and 76.8% in 2004 (see appendix B). These numbers are much higher than the percentage of persons without CDAD that are > 60 years old [45.8% in 1998 and 43.6 % in 2004] (see appendix B). Table 1: Population of Florid a, total number of acute care hospital discharges and the median ages of both the CDAD and non-CDAD groups of discharged patients for the years 1998-2004 Variables Year 1998 1999 2000 2001 2002 2003 2004 Population of Florida (millions) 15.310 15.680 16. 075 16.412 16.772 17.164 17.613 Total discharges 2.098 2.155 2.242 2.341 2.382 2.444 2.491 Ages (in years) Median for nonCDAD 56 56 56 56 55 55 55 Median CDAD 74 74 74 75 75 75 75 Gender The percentage of females in 1998 for the non-CDAD patient discharge population is 56.97% compared to 56.74% in 2004 (see table 2). The percentage of
40 females for the CDAD population group differs slightly from the non-CDAD group with 1998 having 57.98 % female as compared to 2004 CDAD patients having 58.67% female. The gender distribution will be discusse d, in reference to age group stratification, later in this chapter. Table 2: The gender distribution of the CDAD and non-CDAD Florida acute care hospital patient groups for the years 1998-2004. Variables Year 1998 1999 2000 2001 2002 2003 2004 Gender Non-CDAD male, (millions) .900 .926 .957 1.001 1.017 1.044 1.069 Non-CDAD female, (millions) 1.191 1.221 1.276 1.331 1.354 1.383 1.402 Non-CDAD percent Female 56.96 56.86 57.13 57.08 57.11 57.00 56.75 CDAD male 3,004 3,152 3,330 4,561 6,184 7,250 8,349 CDAD female 4,145 4,424 4,521 6,419 8,436 9,903 11,852 CDAD percent female 57.98 58.39 57.58 58.46 57.70 57.73 58.67 Racial distribution Race distribution was analyzed using the factor percent non-white. The data was separated into 8 racial classe s including no response to th e question and a category named other. The percent non-white grouping was us ed to analyze the data for patients who enter Florida hospitals. The non-CDAD patie nt population of Flor ida has a racial distribution of 29.33 percent non-white in 1998 and 33.79 percent non-white in 2004 (see Table 3). For the Clostridium difficile positive patients the per cent non-white is 18.17 in 1998 and 21.03 in 2004 (see Table 3).
41 Table 3: The racial distribution of the CDAD and non-CDAD Florid a acute care hospital patient groups for the years 1998-2004. Variables Year 1998 1999 2000 2001 2002 2003 2004 Race Non-CDAD white, (millions) 1.477 1.495 1.552 1.602 1.608 1.627 1.636 Non-CDAD black, (millions) .300 .304 .326 .342 .350 .366 .386 Non-CDAD other, (millions) .313 .348 .356 .387 .409 .433 .448 Percent non-white 29.33 30.27 30.52 31.23 31.92 32.94 33.79 CDAD white 5,850 5,976 6,189 8,795 11,770 13,870 15,953 CDAD black 617 643 774 995 1,388 1,635 2,154 CDAD other 682 957 888 1,190 1,462 1,648 2,094 Percent non-white 18.17 21.12 21.17 19.90 19.49 19.13 21.03 Length of hospital stay Length of hospital stay for Clostridium difficile positive patients is one of the largest differences between CDAD and nonCDAD patient populations. In 1998 the nonCDAD patient population averag e length of hospital stay was 4.94 days compared to 13.47 days for CDAD patients. In 2004 the average hospital stay for non-CDAD patients was 4.88 days and the average hospital stay for CDAD patients was 13.07 days (see table 4). Table 4: The average length of stay in the hospital for CDAD and non-CDAD Florida acute care hospital patients for the years 1998-2004. Variables Year 1998 1999 2000 2001 2002 2003 2004 Length of hospital stay (days) Mean non-CDAD 4.94 4.93 4.85 4.82 4.85 4.86 4.88 Mean CDAD 13.47 13.35 13.86 13.99 13.77 13.28 13.07 In addition to average hospital stay for the total CDAD and non-CDAD patient populations, this study determined the average length of hospital stay for 11 age groups. The age groups showing the greatest diffe rence between the non-CDAD group and those patients with CDAD are the age groups 1, >10-20, and >50-60 (see Appendix E).
42 Cost per patient The non-CDAD average patient cost increased from $14,401 total charges per patient in 1998 to $25,561 per patient in 2004. These values are much smaller than the patients who contracted CDAD in Flor idas hospitals [$40,773 in 1998 and $63,003 in 2004, respectively] (see table 6 and appendix F) If you multiply the number of persons that had CDAD in 2004 times the additional days in the hospital, the total charge is over 750 million dollars greater than for the non-CDAD patients. Table 6: The total cost per discharge for CDAD and non-CDAD Florida acute care hospital patients for the years 1998-2004. Variables Year 1998 1999 2000 2001 2002 2003 2004 Total charges per discharge Mean non-CDAD 14,401 15,459 16,756 18,472 21,204 23,549 25,561 Mean CDAD positive 40,773 42,983 48,179 51 ,287 57,603 61,047 63,003 Co-morbidities Perforated colon affected 27 Clostridium difficile positive patients in 1998 and 56 in 2004 (see appendix G). The rate of perforated colon per 1,000 CDAD positive discharged patients is 3.78 in 1998 and 2.77 in 2004 (see appendix H). Toxic megacolon occurred in a higher number of CDAD patie nts, 44 in 1998 and 92 in 2004 (see appendix G). The rate of toxic megacolon per 1,000 CDAD positive patients is 6.15 in 1998 and 4.55 in 2004 (see appendix H).
43 CHANGES OF INTEREST IN POPULATION OVER TIME Discharges The total number of Clostridium difficile positive patients from 1998 to 2004 was 7,149 to 20,201 (see Appendix I). The num ber of discharges of CDAD per 1,000 discharged patients is 3.40 in 1998 and 8.11 in 2004. Figure 10 shows that the prevalence of CDAD per 1,000 hospital patients continues to increase during the seven years of this study (see Appendix J). The tota l number of discharges in 2001 rose sharply to 10,980 as compared to 7,851 discharges in 200 0. This rise in CDAD is also reflected by the rate per 1,000 patients which rose from 3.50 in 2000 to 4.69 in 2001. This number is important because CDAD prevalence was relatively constant before 2001. The research presented in this study also s hows that in 2003 Florida had a rate of 99.9 discharges per 100,000 population of Florida (see appendix K). Table 7: The number of CDAD positive discharges and CDAD discharges per 1,000 Florida hospital acute care discha rged patients for the years 1998-2004. Variables Year 1998 19992000200120022003 2004 Population of Florida, (millions) 15.310 15. 68016.07516.41216.77217.164 17.613 Total discharges, millions 2.098 2.1542.2422.3412.3822.444 2.491 CDAD positive discharges 7,149 7,5767,85110,98014,62017,153 20,201 Total CDAD discharges per 1000 patients 3.40 3.513.504.696.137.02 8.11 This study also divides the patients in 11 age groups to determine the effect of CDAD on the groups. Figures 11 and 12 dem onstrate that not onl y is the amount of CDAD discharges increasing in Florida, but the prevalence is also increasing (see appendices L and M). One example is th e >80-90 age group. In 1998 the number of
44 CDAD discharges was 1735 and in 2004 the nu mber of CDAD discharges increased to 5483 (see Appendix L). The CDAD prevalence pe r 1,000 discharged patients in the >8090 age group is 6.95 in 1998 and 18.40 in 2004 (see Appendix M). Gender for discharges in the individual age groups The prevalence of CDAD discharges for gender and racial groups for 1998 through 2004 is shown in Appendices N thru S. Floridas discharge data was separated into male and female groups with three ra cial groups, white, black and other races. Generally the female gender in the three raci al groups demonstrates a greater prevalence difference when comparing the >1-10 age gr oup to the >10-20 age group than the male gender of the same race. White females have the highest prevalence of all the gender groups (see appendix O). The other male ge nder has the lowest prevalence in the older age groups (see appendix R). Racial distribution for the age groups Race was also investigated in this stu dy to determine if race is an important factor. Prevalence in the white race group had the highest numbers for both male and female genders at 18.1 discharges per 1,000 wh ite male >80-90 year old patients for 2004 and 21.2 discharges per 1,000 white female >80-90 year old patients in 2004 (see appendix N and O). The lowest prevalence for the >80-90 year old group was in the other racial groups with 12.1 discharges per 1,000 other male patients in 2004 and 12.6 discharges per 1,000 other female patients in 2004 for that age group, gender, and race. Length of hospital stay The biggest change in length of hospital stay occurred in the 1 age group with a 1998 average length of stay of 12.31 days fo r a patient with CDAD and in 2004 that
45 value increased to 17.85 days (see table 2). Also note in table 2 that the largest difference in length of days stayed in the hospi tal for non-CDAD patients compared to CDAD patients is for the 1 age group. Note in the >90 ag e group that the CDAD patients in 1998 had an average length of hospital stay of 11.60 days and in 2004 the value is 9.67 days. Deaths The number of deaths for persons with CDAD was 678 in 1998, 735 in 1999, 761 in 2000, 1,144 in 2001, 1,575 in 2002, 1,831 in 2003 and 2,043 in 2004. The percent mortality for the CDAD positive and CDAD nega tive patients is shown in figure 17 (see appendix R). The percent mortality for the CDAD positive patients versus CDAD negative patients increases by over a point when comparing 2004 to 1998 (4.64 and 3.57 respectively). This study investigated not only the mort ality rate for each individual year, but also mortality rates for each age group per year. Figure 19, appendix T, shows the percent mortality for the 11 age groups. The mortality percent and the mortality per 10,000 discharged hospitalized patients is s hown in Tables 8 and 9 (appendices W and X), respectively. The mortality of CDAD pa tients per 10,000 discharg ed patients for the >80-90 age group increased from 8.41 CDAD deaths in 1998 to 23.9 CDAD deaths in 2004. The >90 age group also had an increased death rate per 10,000 discharged patients of 10.4 CDAD deaths in 1998 to 32.6 CDAD deat hs in 2004 (see Table 9, appendix X). Total numbers of discharges and deaths for the individual age groups and years is listed in tables 10, 11, and 12 (see appendi ces Y, Z, and AA respectively).
46 Chapter Six: Discussion LIMITATIONS OF THE STUDY The following are limitations when using large amounts of patient discharged data. The AHCA data does not identify someone that has multiple admissions to a hospital in the same year. This is a c oncern because a patient may have multiple admissions with Clostridium difficile infections in the same y ear. The recurrence rate of CDAD can be as low as 5% or as high as 20-25% (Pepin et al, 2005a; Schroeder, 2005). Authors opinions vary on this topic and r ecurrence rates for Flor ida are unknown. Other limitations include non-specific use of diagnos e codes, transcripti on errors of those persons entering the data, and inaccurate diagnoses of patients disease. The discharge data does not indicate th e time at which a person contracted CDAD. Therefore, the length of hospital stay includes both the time a person was in the hospital before they contracted CDAD and th e time they were in the hospital during recovery from CDAD. In summary, although th e above five limitations do exist, these limitations are minor when compared to the in formation gleaned from the data and do not detract from the conclusions of this study. STRENGTHS OF THE STUDY Other than the pharmacy data, cleaning solvent, and background knowledge of certain patients, the AHCA data is complete and is the best available information for this large population. This data, a lthough there could be missing information, is capable of
47 providing adequate information to perform the analysis and determine prevalence rates of Clostridium difficile in Florida hospitals. A report by Peter Layde and his colleagues (Layde, et al, 2004) discusses the use of data sets similar to the AHCA data with the intent of showing the usefulness of these data sets. Layde et al, 2004 performed an an alysis studying medical injury using hospital discharge data. Their results us ing validation studies shows rela tively high specificity and sensitivity when comparing hospital discharge datasets to medical r ecord reviews. No mention was made of the sensitivity and speci ficity actual values, but ICD-9 codes were used to determine medical injury. This met hod of verifying medical injury is similar to the method that is used to determine Clostridium difficile infection rates in this study. The Layde report (Layde et al, 2004) supports th e conclusion that using AHCA data is an effective tool for obtaining prevalence for Clostridium difficile A pilot study using hospital discharge data to access the morbidity of Rotavirus was completed and this report also demonstrat es the usefulness of statewide discharge data (Parashar, Chung, Holman, Ryder, Hadler and Glass, 1999). The Parashar et al report, using the Connecticut statewide disc harge hospital database shows that 10.4% of all diarrhea associated hospitalizations were at tributed to Rotavirus. This value nearly matches the nationally reported values for the Northeast region of the U.S. (Parashar et al, 1999200). Parashar and his colleagues s how that using statewide discharge data provides a useful tool for a re search project and can be eff ective in measuring trends for disease in some states. The above two studi es demonstrate that us ing information from hospital discharge data, alt hough there may be a few limitatio ns, provides a useful tool for tracking and monitoring of di seases at a statewide level.
48 A study that discusses Clostridium difficile more specifically discusses the effectiveness of using ICD-9 codes to correct ly determine the amount of CDAD that is present in hospitals. Dubberke and his co lleagues show that good correlation existed between Clostridium difficile toxin assay and ICD-9 code determination of the disease (Dubberke et al, 2006). This is important wh en considering the information provided in the AHCA hospital discharge data. The informa tion provided in this study relies solely on the correct assignment of the ICD-9 code to CDAD to make an assessment of the persons that have the disease. The repor t by Dubberke (Dubberke et al, 2006) provides good evidence that using ICD-9 codes to dete rmine disease status, especially in the discharge of Clostridium difficile can be accurate and helpful to determine disease prevalence. INTERPRETATIONS OF RESULTS Characteristics of the study population This study shows patients discharged from Florida hospitals have an increasing trend for contracting Clostridium difficile The results of this study show that the number of discharged patients contracting Clostridium difficile has increased by more than 10,000 discharges when comparing 2004 to 1998. Th e disease rate per 1,000 patients showed almost a 2.5 times higher rate in 2004 than in 1998. The prevalence of CDAD change from 2000 to 2001 supports the theory that a new more highly virule nt strain of the bacteria began to infect Florida patients in that year (see Appendix J). C. McDonald and his colleagues established a rate of 45 di scharges per 100,000 U.S. population in the southern region of the United States for the year 2003 and the national average was established by McDonald as 61 per 100,000 U. S. population (McDonald et al, 2006).
49 These numbers are in contrast to the rate presented here which is 99.9 discharges per 100,000 population in Florida for the year 2003. These numbers show that CDAD is a significant patient problem in Flor idas acute care hospitals. Age Age is one of the most important fact ors with influence on the percentage of patients that become infected with Clostridium difficile Figure 2 demonstrates the importance of this factor by showing that over 70% of the CDAD positive patients are over 60 years old compared to the non-C DAD population that is under 50% (Appendix B). Our analysis demonstrates that older people have a higher prevalence of CDAD. Dividing the population into 11 age groups s hows the effects of CDAD among different age groups in the population. Age stratification McDonald and his colleagues used 4 di fferent age group categories to show how CDAD affects different aged individual s (McDonald et al, 2006). This study demonstrates, for CDAD infection rates, th at as a person becomes older their chances increase dramatically for contracting CDAD. The other conclusion from this study is that during the years of the study CDAD prevalence has increas ed in all age groups (see Appendix M). This is esp ecially evident among the olde r age groups. Another significant finding in this study occurs in th e >1-10 age group. The CDAD prevalence per 1,000 patients shows a fairly constant in crease among the age groups as they become older, but the >1-10 age group is almost twic e as high as the next age group, >10-20. The >1-10 age group partially follows the pattern of the >30-40 age group which is 2 more
50 age groups ahead of it. In attempt to explain this we did gender and r acial distributions to discuss this interesting phenomenon. Length of hospital stay and cost per patient This study shows that a new more virulent Clostridium difficile bacteria strain may be a good explanation for increasing preval ence of this disease. Length of hospital stay and total charges at patie nt discharge are also factors that may support the conclusion of this study. The increasing patient tota l costs may be enough incentive to motivate researchers to provide more time and resource s to patients that have been diagnosed with CDAD in Florida hospitals. This study demons trates that discharged patients who have contracted Clostridium difficile are in hospitals an average of 9 days longer than patients with no evidence of CDAD at a total increas ed cost of over $35,000 per patient. The additional dollars equates to over 750 milli on spent to treat this disease in 2004 as compared to what would have been spent if no patients contracted Clostridium difficile The 1 age group is puzzling especially wh en you view the change in average length of hospital stay for the CDAD patie nts for 1998 and 2004. The over 5 more days that 2004 patients spend in the hospital for th is age group is not easy to explain. The non-CDAD group does not seem to exhibit much of any change for the age groups from the year 1998-2004. The other major difference seen in Table 5 is the >90 year old age group CDAD population average le ngth of hospital stay change from the year 1998 to 2004. This number decreased by almost 2 days. It is not known if hos pitals are able to diagnose these problems quicker, antibiotics are given in a better dosage so as to work faster, or these patients are being discharged to a geriatric facility were they can be treated and monitored more closely.
51 Gender Although no previous published reports on Clostridium difficile have mentioned gender or race as a cause for this disease th is study shows the dist ribution of gender and race for each age group as shown in Appendices N through S. Notice in almost all of these appendices a higher prevalence for CDAD females in all three of the racial groups. In 2003 the prevalence of CDAD female discharg es for the >1-10 age group is over four times higher than the >10-20 age group for 2003 (see Appendix S). Two reasons for the higher female prevalence may be the lowe r prevalence for the > 10-20 age group and the other may be the higher prevalence for the female gender when compared to the male gender of the same race (see appendices N thru S) This higher prevalence is difficult to explain using the discharge data and more rese arch into these differences is needed. The older age groups show a consistent pattern fo r both gender and race groups with a higher prevalence for CDAD in the age groups over 50 years old. Racial distribution The race with the highest prevalence is white (see appendix N and O). Generally, the races tend to have a prevalence that is highe r as a person is older. The prevalence is relatively constant when comparing the different races to each other. Except for the spike in the >80-90 age group prevalence for the black female group (Appe ndix O), black male, other male, and other female have nearly th e same prevalence with the other racial group having a slightly lower preval ence than the black CDAD group (see appendix P thru S). Another difference is the lower CDAD preval ence shown in the female black group and
52 other race group >10-20 age group. These groups tend to have one half to one third the prevalence when compared to the prevalence for the other gender racial groups (see Appendix Q and S). Race at least in Florida does not seem to be a contributor to the rise in discharges for the CDAD disease. Death rate information The mortality rate for Clostridium difficile positive patients when compared to the mortality rate for patients who did not have th e disease shows some interesting figures. Generally speaking the majority of the age groups in th is study showed decreasing mortality rates for the patients who did not cont ract CDAD. This is in contrast to the CDAD positive patients who generally had a re latively constant mortality with the age groups that are >70 having incr easing mortality. This info rmation is consistent with reported increasing death rates observed in Canada, the United States, and the United Kingdom. This study also shows that older patient s are more susceptible to contract the disease and die. The data an alysis shows that death rates for age groups that >70 years old have mortality rates over 10%. This percenta ge is in contrast to the mortality rate of lower than 7% among patients in the same age groups who did not have Clostridium difficile When considering the change in mortality rates, of CDAD positive patients, over the years of this study, the >90 age group do es show an increasing trend. This is in contrast to the decreasing mo rtality rate for patients in the same age group that do not contract CDAD. A theory advocated at the beginning of this study was that a more virulent strain of Clostridium difficile would affect all age groups, not just the >60 age group. This is
53 the discharge for the >30-40 age group where the mortality rate increases from 2.89% in 1998 to 4.10% in 2004 (see appendix W). For th e >50-60 age group this is not the case because the mortality rate decreases from 8.89% in 1998 to 6.41% in 2004 (see appendix W). The mortality rate is higher for Clostridium difficile positive individuals, but the rates dont appear to be in creasing for all age groups. Younger patients are dying with this disease, but do not appear, generally, to have a consistent ly increasing mortality rate over the years of this study. The increase in CDAD discharges leads to a higher number of deaths per 10,000 discharged patients by ag e group as shown in Table 9. The mortality percent in the >50-60 age group (Table 8) show s a slight decrease for the years of this study while the CDAD deaths per 10,000 discharged patients (Table 9) shows an increase indicating that the prevalence of CDAD discharges is increasing. The CDAD death rate per 10,000 discharged patients demonstrates that CDAD deaths are increasing in Floridas acute care hospitals. Co-morbidities and what they mean The data in this study was analyzed to determine if toxic megacolon and perforated colon prevalence ar e also increasing. The analys is results demonstrate that prevalence for these two co-morbidity conditi ons are not increasing. It is not known if patients are dying faster and not having these conditions, if patients are becoming better before these conditions are diagnosed, or if doctors are just not reporting these conditions. Summary of bias, confounding, and chance This study was limited to patients discha rged from Floridas acu te care hospitals. This is important because information from nursing care facilities and federal hospitals is
54 missing from the information that was provided. The results of this study should not be used as a predictor of CDAD prevalence for nur sing care facilities or federal hospitals in Florida. This study also lacks the ability to be used as a co mparison to other states in the U.S. The major confounder that could occur is age. The study intentionally grouped the data for each year to reduce the effects of age as a confounder on the analyzed results. Age affects mortality, but because we are es timating rates per year and comparing these rates on an annual basis, and the age of the patient population does not change very much over the years of this study, it is concluded th at age and its effects as a confounding factor has been controlled for during the analysis. This study further limits the affects of age by stratifying the population into 11 age groups and looking at di scharge and death trends among the age groups. Chance is not a factor in this study. All of the known patient discharge data from acute care hospitals in Florida is included in this study, therefore, p-values are not calculated because this study uses population data. CONCLUSIONS Prevalence of Clostridium difficile in Floridas acute care hospitals is increasing. The mortality of patients who have CDAD rema ins relatively constant except for patients that are >70 that show an in creasing mortality through the year s of the study. This is in contrast to a decreasing mortal ity for those patients that ha ve no CDAD. These increases make this a serious disease for persons hospitali zed in Floridas acute care facilities. This study demonstrates that age is a factor for higher CDAD prevalence. The variables length of hospital stay, race and gender show little affect on the increased prevalence
55 Clostridium difficile infections. The more virulent st rain of CDAD most likely entered Florida in the year 2001. The values and fi gures in this study provi de the most current available information for affects of this di sease on patients discharged from Floridas acute care hospitals. The trends noted in th is study show the disease prevalence to be increasing and the emphasis should be to develop and implement preventative methods that decrease the prevalence of CDAD. RECOMMENDATIONS Further explanation is needed to discu ss the prevalence of C DAD. Adjusting this study to further explain the prevalence at a county level would help to explain exactly what areas of Florida are affected the most. Using the results of this statewide descriptive study for an in depth analysis is recommended for individual hospitals to find ways to reduce the prevalence of CDAD. This c ould be accomplished by creating a group of hospitals to act as satellites for a CDAD anal ysis, make changes in the group of hospitals that would decrease preval ence of CDAD, and compare the changed prevalence of CDAD with hospitals that are not part of the satellite gr oup. This CDAD analysis would give opportunities for health care professiona ls to implement more effective methods of treating the disease and preventing the dis ease from spreading to other patients.
56 References Aslam, S., Hamill, R. J., & Musher, D. M. (2005). Treatment of Clostridium difficile associated disease: Old therapies and new strategies. The Lancet Infectious Diseases, 5 (9), 549-557. Bartlett, J. G., Narrative Review: The New Epidemic of Clostridium difficile Associated Enteric Disease, (2006), Ann Intern Med ., 145, 758-764. Berild, D., Smaabrekke, L., Halvorsen, D. S., Lelek, M., Stahlsberg, E. M., & Ringertz, S. H. (2003). Clostridium difficile infections related to antibiotic use and infection control facilities in two university hospitals. The Journal of Hospital Infection, 54 (3), 202-206. Blot, E., Escande, M. C., Besson, D., Barbut, F., Granpeix, C., & Asselain, B. et al. (2003). Outbreak of Clostridium difficile -related diarrhoea in an adult oncology unit: Risk factors and microbiological characteristics. The Journal of Hospital Infection, 53 (3), 187-192. Bouza, E., Munoz, P., & Alonso, R. (2005). Clinical manifestations, treatment and control of infections caused by Clostridium difficile Clinical Microbiology and Infection : The Official Publication of the European Society of Clinical Microbiology and Infectious Diseases, 11 Suppl 4 57-64. Boyce, J. M., and Pittet, D. (2002). Guidelin es for hand hygiene in health-care settings, recommendations of the healthcare infection control practice advisory committee and the HICPAC/SHEA/APIC/IDSA hand hygiene task force. Morbidity and Mortality Weekly Report, 51(RR16), 1-44. Brierley, R. (2005). Clostridium difficile --a new threat to public health? The Lancet Infectious Diseases, 5 (9), 535. CDC Clostridium difficile information for healthcare providers. (2005). Retrieved 07/22, 2005 from http://www.cdc.gov.ncidod/hip/gast ro/ClostridiumDifficileHCP.htm. Chandler, R. E., Hedberg, & Cieslak, P. R., (2007), Clostridium difficile Associated Disease in Oregon: Increasing Inciden ce and Hospital-Level Risk Factors, Infection Control and Hospital Epidemiology Vol. 28, No. 2. Cloud, J. & Kelly, C. P, (2007), Update on Clos tridium difficile associated disease, Curr Opin Gastroenterol, 23:4-9.
57 Dendukuri, N., Costa, V., McGregor, M., & Br ophy, J. M. (2005). Probiotic therapy for the prevention and treatment of Clostridium difficile -associated diarrhea: A systematic review. CMAJ : Canadian Medical Association Journal = Journal De l'Association Medicale Canadienne, 173 (2), 167-170. Dial, S., Alrasadi, K., Manoukian, C., Hua ng, A., & Menzies, D. (2004). Risk of Clostridium difficile diarrhea in hospital in-pat ients prescribed proton pump inhibitors: Cohort and case-control studies. CMAJ : Canadian Medical Association Journal = Journal De l'Associ ation Medicale Canadienne, 171 (1), 33-38. Dial, S., Delaney, J. A., Barkun, A. N., & Suissa, S. (2005). Use of gastric acidsuppressive agents and the risk of community-acquired Clostridium difficile associated disease. JAMA : The Journal of the American Medical Association, 294 (23), 2989-2995. Dubberke, E. R., Reske, K. A., McDonald, L. C., & Fraser, V. J. (2006). ICD-9 codes and surveillance for Clostridium difficile -associated disease. Emerging Infectious Diseases, 12(10), Eggertson, L. (2005). Clostridium difficile may have killed 2000 in Quebec: Study. CMAJ : Canadian Medical Association J ournal = Journal De l'Association Medicale Canadienne, 173 (9), 1020-1021. Eggertson, L., & Sibbald, B. (2004) Hospitals battling outbreaks of Clostridium difficile CMAJ : Canadian Medical Association J ournal = Journal De l'Association Medicale Canadienne, 171 (1), 19-21. Fawley, W. N., Parnell, P., Verity, P., Fr eeman, J., & Wilcox, M. H. (2005). Molecular epidemiology of endemic Clostridium difficile infection and the significance of subtypes of the United Kingdom epidemic strain (PCR ribotype 1). Journal of Clinical Microbiology, 43 (6), 2685-2696. Freeman, J., Baines, S. D., Jabes, D., & Wilcox, M. H. (2005a). Comparison of the efficacy of ramoplanin and vancomycin in both in vitro and in vivo models of clindamycin-induced Clostridium difficile infection. The Journal of Antimicrobial Chemotherapy, 56 (4), 717-725. Freeman, J., Stott, J., Baines, S. D., Fawley, W. N., & Wilcox, M. H. (2005b). Surveillance for resistance to metronid azole and vancomycin in genotypically distinct and UK epidemic Clostridium difficile isolates in a large teaching hospital. The Journal of Antimic robial Chemotherapy, 56 (5), 988-989. Freeman, J., Fawley, W., Baines, S., & Wilcox, M. (2006). Measurement of toxin production by clostridium difficile Lancet, 367 (9515), 982-3; author reply 983-4. Garner, J. S., CDC Guideline for Isolation Precautions in Hospitals, (2005), Department of Health and Human Services, Centers for Disease Control and Prevention last modified 4/1/2005.
58 Gerding, Dale N. MD. (2005). Pseudomembranous Colitis ( Clostridium difficile ). Association for Professi onals in Infection Control and Epidemiology, Inc ., 75(1), 02/01/2005. Retrieved 09/01/2005. Gerding, D. N., Johnson, S., Peterson, L. R., Mulligan, M. E., & Silva, J.,Jr. (1995). Clostridium difficile -associated diarrhea and colitis. Infection Control and Hospital Epidemiology : The Official Journal of the Society of Hospital Epidemiologists of America, 16 (8), 459-477. Gopal Rao, G., Mahankali Rao, C. S., & Starke, I. (2003). Clostridium difficile -associated diarrhoea in patients with community-ac quired lower respiratory infection being treated with levofloxacin compared with beta-lactam-based therapy. The Journal of Antimicrobial Chemotherapy, 51 (3), 697-701. Hart, Anita C. RHIA, & Hopkins, Catherine A. CPC (Eds.). (2004). ICD-9-CM professional for hospitals (Sixth edition ed.). Salt Lake City, Utah: Ingenix. Retrieved October 1, 2004, from www.ingenixonline.com Heimesaat, M. M., Granzow, K., Leidinger, H., & Liesenfeld, O. (2005). Prevalence of Clostridium difficile toxins A and B and clostridiu m perfringens enterotoxin A in stool samples of patients with antibiotic-associated diarrhea. Infection, 33 (5-6), 340-344. Kazakova, S. V., Ware, K., Baughman, B., B ilukha, O., Paradis, A., et al, (2006), A Hospital Outbreak of Diarrhea Due to an Emerging Epidemix Strain of Clostridium difficile Arch Intern Med Vol. 166, pp. 2518-2524. Layde, P. M., Meurer, L. N., Guse, C., Me urer, J. R., Yang, H., Laud, P., Kuhn, E. M., Brasel, K. J. & Hargarten, S. W. (2004). Medical Injury Identification Using Hospital Discharge Data. Information from the Department of Family and Community Medicine, Medical Colleg e of Wisconsin. Vol. 2, pp 119-131. Loo, V. G., Poirier, L., Miller, M. A., O ughton, M., Libman, M. D., & Michaud, S., et al. (2005). A predominantly clonal mu lti-institutional outbreak of clostridium difficile -associated diarrhea with hi gh morbidity and mortality. The New England Journal of Medicine, 353 (23), 2442-2449. Louie, T. J. (2005). How should we respond to the highly toxogeni c NAP1/ribotype 027 strain of Clostridium difficile ? CMAJ : Canadian Medical Association Journal = Journal De l'Associatio n Medicale Canadienne, 173 (9), 1049-1050. Martirosian, G., Szczesny, A., Cohen, S. H ., & Silva, J.,Jr. (2005). Analysis of Clostridium difficile -associated diarrhea among patient s hospitalized in tertiary care academic hospital. Diagnostic Microbiology and Infectious Disease, 52 (2), 153-155.
59 McCharen, Kathy, spokeswoman from the Florida legislature (2000). Florida Demographic Summary. www.Florida-business-data.c om/demographics-statesummary.htm. McConnell, E. A. (2002). Prevent the spread of Clostridium difficile Nursing, 32 (8), 24, 26. McDonald LC, Owings M, Jernigan DB. Clostridium difficile infection in patients discharged from US short-stay hospitals, 19962003. Emerging Infectious Diseases [serial on the Internet]. 2006 Mar [ date cited ]. Available from http://www.cdc.gov/ncidod/EID/vol12no03/05-1064.htm McEllistrem, M. C., Carman, R. J., Gerd ing, D. N., Genheimer, C. W., & Zheng, L. (2005). A hospital outbreak of Clostridium difficile disease associated with isolates carrying binary toxin genes. Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America, 40 (2), 265-272. Modena, S., Gollamudi, S., & Friedenberg, F. (2006). Continuation of antibiotics is associated with failure of metronidazole for clostridium difficile -associated diarrhea. Journal of Clinical Gastroenterology, 40 (1), 49-54. Musher, D. M., Aslam, S., Logan, N., Nallach eru, S., Bhaila, I., & Borchert, F. et al. (2005). Relatively poor outcome after treatment of Clostridium difficile colitis with metronidazole. Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America, 40 (11), 1586-1590. Mylonakis, E., Ryan, E. T., & Calderwood, S. B. (2001). Clostridium difficile --associated diarrhea: A review. Archives of Internal Medicine, 161 (4), 525-533. Nayar, D. M., Vetrivel, S., McElroy, J., Pai, P., & Koerner, R. J. (2005). Toxic megacolon complicating Esch erichia coli O157 infection. The Journal of Infection, 52(4), 103-6. Noren, T. (2005). Outbreak from a high-toxin intruder: Clostridium difficile Lancet, 366 (9491), 1053-1054. Oldfield, E. C.,3rd. (2004). Clostridium difficile -associated diarrhea: Risk factors, diagnostic methods, and treatment. Reviews in Gastroenterological Disorders, 4 (4), 186-195. Parashar, U. D., Chung, M. A., Ho lman, R. W., Hadler, J. L., & Glass, R. A. (1999). Use of state hospital discharge da ta to assess the morbidity from rotavirus diarrhea and to monitor the impact of rotavirus immunization program: a pilot study in Connecticut. Pediatrics, 104(3), 489-494. Pepin, J., Valiquette, L., Alary, M. E., Villemu re, P., Pelletier, A., & Forget, K. et al. (2004). Clostridium difficile -associated diarrhea in a region of Quebec from 1991 to 2003: A changing pattern of disease severity. CMAJ : Canadian Medical
60 Association Journal = Journal De l' Association Medicale Canadienne, 171 (5), 466-472. Pepin, J., Alary, M. E., Valique tte, L., Raiche, E., Ruel, J ., & Fulop, K. et al. (2005a). Increasing risk of relapse after treatment of Clostridium difficile colitis in Quebec, Canada. Clinical Infectious Diseases: An O fficial Publication of the Infectious Diseases Society of America, 40 (11), 1591-1597. Pepin, J., Saheb, N., Coulombe, M. A., Alary, M. E., Corriveau, M. P., & Authier, S. et al. (2005b). Emergence of fluoroquinolones as the predominant risk factor for Clostridium difficile -associated diarrhea: A cohort study during an epidemic in Quebec. Clinical Infectious Diseases: An O fficial Publication of the Infectious Diseases Society of America, 41 (9), 1254-1260. Pepin, J., Valiquette, L., & Cossette, B. (2005c ). Mortality attributable to nosocomial Clostridium difficile -associated disease during an epidemic caused by a hypervirulent strain in Quebec. CMAJ : Canadian Medical Association Journal = Journal De l'Associatio n Medicale Canadienne, 173 (9), 1037-1042. Price, M. F., Dao-Tran, T., Garey, D. W., Graham, G., Gentry, L. O., et al, (2006), Epidemiology and incidence of Clostrid ium difficile asso ciated diarrhoea diagnosed upon admission to a university hospital, Journal of Hospital Infection 65, 42-46. Russmann, H, Panthel, K., Bader, R.-C .,Schmitt, C., & Schaumann, R., (2007), Evaluation of three rapid assays for detection of Clostridium difficile toxin A and toxin B in stool specimens, Eur J Clin Microbiol Infect Dis DOI 10, 1007/s100096-006-0251-7. Scheurer, D. B., Hicks, L. S., Cook, E. F., & Schnipper, J. L., (2006), Accuracy of ICD-9 coding for Clostridium difficile infections: a restrospective cohort, Epidemiol. Infect Schroeder, M. S. (2005). Clostridium difficile --associated diarrhea. American Family Physician, 71 (5), 921-928. Schulster, Lynne, & Chinn, Raymond (2003). Guidelines for Environmental Infection Control in Health-Care Facilities. Recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee (HICPAC). Pp. 84-85, 117-138. Simor, A. E., Bradley, S. F., Strausbaugh, L. J., Crossley, K., Nicolle, L. E., & the SHEA Long-Term-Care Committee. (2002). Clostridium difficile in long-term-care facilities for the elderly. Infection Control and Hosp ital Epidemiology : The Official Journal of the Society of Hospital Epidemiologists of America, 23 (11), 696-703.
61 Starr, J. M., Martin, H., McCoubrey, J., Gi bson, G., & Poxton, I. R. (2003). Risk factors for Clostridium difficile colonization and toxin production. Age and Ageing, 32 (6), 657-660. Tang, P., Roscoe, M., & Richardson, S. E. (2005). Limited clin ical utility of Clostridium difficile toxin testing in infant s in a pediatric hospital. Diagnostic Microbiology and Infectious Disease, 52 (2), 91-94. Thomas, C., Stevenson, M., Williamson, D. J., & Riley, T. V. (2002). Clostridium difficile -associated diarrhea: Epidemiologi cal data from western Australia associated with a modified antibiotic policy. Clinical Infectious Diseases : An Official Publication of the Infectio us Diseases Society of America, 35 (12), 14571462. Valiquette, L., Low, D. E., Pepin, J., & McGeer, A. (2004). Clostridium difficile infection in hospitals: A brewing storm. CMAJ : Canadian Medical Association Journal = Journal De l'Associatio n Medicale Canadienne, 171 (1), 27-29. Vesta, K. S., Wells, P. G., Gentry, C. A., & Stipek, W. J. (2005). Specific risk factors for Clostridium difficile -associated diarrhea: A prospective, multi-center, case control evaluation. American Journal of Infection Control, 33 (8), 469-472. Voth, D. E., & Ballard, J. D. (2005). Clostridium difficile toxins: Mechanism of action and role in disease. Clinical Microbiology Reviews, 18 (2), 247-263. Warny, M., Pepin, J., Fang, A., Killgore, G., Thompson, A., & Brazier, J. et al. (2005). Toxin production by an emerging strain of Clostridium difficile associated with outbreaks of severe disease in North America and Europe. Lancet, 366 (9491), 1079-1084. Wilcox, M. H., Fawley, W. N., Wigglesworth, N., Parnell, P., Verity, P., & Freeman, J. (2003). Comparison of the effect of de tergent versus hypochlorite cleaning on environmental contamination and incidence of Clostridium difficile infection. The Journal of Hospital Infection, 54 (2), 109-114. Wren, S. M., Ahmed, N., Jamal, A., & Safadi, B. Y. (2005). Preoperative oral antibiotics in colorectal surgery increase the rate of Clostridium difficile colitis. Archives of Surgery (Chicago, Ill. : 1960), 140 (8), 752-756. Wullt, M., & Odenholt, I. (2004). A double-blin d randomized controlled trial of fusidic acid and metronidazole for treatm ent of an initial episode of Clostridium difficile associated diarrhoea. The Journal of Antimic robial Chemotherapy, 54 (1), 211216.
63 Appendix A Figure 1: Florida hospital patien ts, median age, discharged with and without CDAD, 1998-2004.75 55 75 75 75 74 74 74 55 55 56 56 56 560 10 20 30 40 50 60 70 80 1998199920002001200220032004YearDischarged patient median age CDAD positive Non-CDAD
64 Appendix B Figure 2: Florida hospital patients, percent >60 years old, discharged with and without CDAD, 1998-2004.20 30 40 50 60 70 80 90 1998199920002001200220032004YearPercent patient popula tion > 60 years old CDAD positive Non-CDAD
65 Appendix C Figure 3: Florida hospita l patients, percent non-white, discharged with and without CDAD, 1998-2004.10 15 20 25 30 35 1998199920002001200220032004YearPatient non-white percent CDAD positive patients Non-CDAD patients
66 Appendix D Figure 4: Average length of hospital stay for CDAD and non-CDAD discharged patients from Florida hospitals, 1998-2004.0 2 4 6 8 10 12 14 16 1998199920002001200220032004YearHospital average length of stay, days CDAD positive patients Non-CDAD patients
67 Appendix E Table 5: Florida non-CDAD and CDAD patients average length of hospital stay, days, 1998-2004. Year and population Age group 1998 non-CDAD pop. 1998 CDAD pop. 2004 non-CDAD pop. 2004 CDAD pop. 1 3.57 12.31 3.80 17.85 >1-10 3.96 11.33 3.41 12.52 >10-20 4.42 14.97 3.87 14.65 >20-30 3.33 12.54 3.52 13.26 >30-40 4.26 14.28 4.08 12.95 >40-50 4.94 14.13 4.88 13.61 >50-60 5.18 15.04 5.18 15.90 >60-70 5.43 14.46 5.33 14.61 >70-80 5.70 13.64 5.52 13.21 >80-90 5.98 12.31 5.70 11.40 >90 5.93 11.60 5.59 9.67 Total mean Length of stay 4.94 13.47 4.88 13.07
68 Appendix F Figure 5: Florida average patient hospital cost at discharge, $/patient, for CDAD and non-CDAD patients, 1998-2004. 0 10000 20000 30000 40000 50000 60000 70000 1998199920002001200220032004YearHospital cost, $/patient CDAD patients Non-CDAD patients
69 Appendix G 2744 2151 2159 2870 3186 46100 5692 0 10 20 30 40 50 60 70 80 90 100 110Total number of cases 1998199920002001200220032004YearFigure 6: Florida hospital discharges, toxic megacolon and perforated colon, for Clostridium difficile positive patients, 1998-2004. Perforated colon Toxic megacolon
70 Appendix H 6.15 3.78 6.73 2.77 7.51 2.67 6.38 2.55 5.88 2.12 5.83 2.68 4.55 2.77 0 1 2 3 4 5 6 7 8Cases per 1,000 positive CDAD patients 1998199920002001200220032004Year Figure 7: Florida toxic megacolon and perforated colon discharges per 1,000 positive CDAD discharged patients, 1998-2004. Toxic megacolon Perforated colon
71 Appendix I Figure 8: Florida Clostridium difficile positive patients discharged from hospitals, 1998-2004.20201 17,153 14620 10980 7851 7576 71490 5000 10000 15000 20000 25000 1998199920002001200220032004YearTotal patients with Clostridium difficile
72 Appendix J 3.41 3.52 3.5 4.69 6.13 7.02 8.11 0 1 2 3 4 5 6 7 8 9 10CDAD patients per 1,000 discharged patients 1998199920002001200220032004Year Figure 9: Florida patients contracting CDAD per 1,000 discharged patients, 1998-2004.
73 Appendix K 46. 7 48. 3 48. 8 66. 9 87. 2 99. 9 114. 7 0 20 40 60 80 100 120 140C D A D patients per 100,000 persons i n Florida. 1998199920002001200120032004Year Figure 10: CDAD positive patients per 100,000 persons in Florida, 1998-2004.
74 Appendix L Figure 11: Patients discharged from Florida hospitals, positive for Clostridium difficile, by age group, 1998-2004.0 1000 2000 3000 4000 5000 6000 7000< = 1 > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age groupPatients with CDAD 1998 1999 2000 2001 2002 2003 2004
75 Appendix M 0 5 10 15 20 25CDAD patients per 1,000 discharged < = 1 > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age group Figure 12: Florida CDAD positive patients per 1,000 patients discharged, by age group, 1998-2004. 1998 1999 2000 2001 2002 2003 2004
76 Appendix N 0 5 10 15 20 25CDAD prevalence per 1,000 patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age Group Figure 13: CDAD prevalence per 1,000 Florida discharged white male patients 1998-2004. 1998 1999 2000 2001 2002 2003 2004
77 Appendix O 0 5 10 15 20 25CDAD prevalence per 1,000 patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age Group Figure 14: CDAD prevalence pe r 1,000 Florida discharged white female patients, 1998-2004 1998 1999 2000 2001 2002 2003 2004
78 Appendix P 0 5 10 15 20 25Prevalence per 1,000 discharged black male patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age group Figure 15: CDAD prevalence per 1,000 Florida discharged black male patients, 1998-2004 1998 1999 2000 2001 2002 2003 2004
79 Appendix Q 0 5 10 15 20 25CDAD prevalence per 1,000 patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age group Figure 16: CDAD prevalence per 1,000 Florida discharged black female patients, 1998-2004. 1998 1999 2000 2001 2002 2003 2004
80 Appendix R 0 5 10 15 20 25CDAD prevalence per 1,000 patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age group Figure 17: CDAD prevalence per 1,000 Florida discharged other race male patients, 1998-2004. 1998 1999 2000 2001 2002 2003 2004
81 Appendix S 0 5 10 15 20 25CDAD prevalence per 1,000 patients > 1 1 0 > 1 0 2 0 > 2 0 3 0 > 3 0 4 0 > 4 0 5 0 > 5 0 6 0 > 6 0 7 0 > 7 0 8 0 > 8 0 9 0 > 9 0Age group Figure 18: CDAD prevalence per 1,000 Florida discharged other race female patients, 1998-2004. 1998 1999 2000 2001 2002 2003 2004
Appendix T Figure 19: Florida hospital pa tients, percent mortality, CDAD and non-CDAD, 1998-20040 2 4 6 8 10 12 1998199920002001200220032004YearPercent mortality CDAD positive CDAD Negative
83 Appendix U Figure 20: Florida hospital patients, percent mortality, non-CDAD by 11 age groups, 1998-2004.0 1 2 3 4 5 6 7 8 9 10 1998199920002001200220032004YearPercent mortality <=1 >1-10 >10-20 >20-30 >30-40 >40-50 >50-60 >60-70 >70-80 >80-90 >90
84 Appendix V Figure 21: Florida hospital pa tients, percent mortality, CDAD positive patients by 11 age groups, 1998-2004.0 2 4 6 8 10 12 14 16 18 20 1998199920002001200220032004YearPercent mortality <=1 >1-10 >10-20 >20-30 >30-40 >40-50 >50-60 >60-70 >70-80 >80-90 >90
85 Appendix W Table 8: The mortality percent for CDAD di scharges in Florida acute care hospitals, by age group, 1998-2004. Age Group Year 1998 1999 2000 2001 2002 2003 2004 >1-10 3.53 0.93 1.85 0.82 0.76 0.69 0 >30-40 2.89 2.78 2.99 2.14 4.20 4.25 4.10 >50-60 8.89 7.19 7.44 6.97 6.75 7.10 6.41 >60-70 9.35 9.07 9.02 9.40 8.76 9.94 8.82 >70-80 10.2 10.2 10.4 10.4 12.3 10.7 11.1 >80-90 12.1 13.4 13.3 14.7 14.1 14.3 13.0 >90 15.1 15.8 13.2 18.5 18.4 18.0 18.6 Totals 9.48 9.70 9.69 10.4 10.8 10.7 10.1
86 Appendix X Table 9: Florida mortality of CDAD pa tients per 10,000 discharged patients, by age group, 1998-2004. Age group Year 1998 1999 2000 2001 2002 2003 2004 >1-10 0.74 0.24 0.47 0.22 0.22 0.21 0 >30-40 0.44 0.44 0.47 0.40 1.06 1.12 1.27 >50-60 2.66 2.07 2.42 2.81 3.59 4.29 4.40 >60-70 3.77 4.00 3.89 5.11 6.45 8.23 8.66 >70-80 4.52 5.82 5.82 7.88 12.9 13.3 15.7 >80-90 8.41 9.28 9.30 15.3 18.5 22.2 23.9 >90 10.4 11.7 10.6 20.3 24.4 28.2 32.6 Totals 3.23 3.41 3.40 4.88 6.60 7.49 8.20
87 Appendix Y Table 10: Discharges of Clostridium difficile in Floridas acute care hospital, per cent of discharges for each year, 1998-2004. Age Group Year Total Number of Discharges 1998 Discharges (%) 1999 Discharges (%) 2000 Discharges (%) 2001 Discharges (%) 2002 Discharges (%) 2003 Disc harges (%) 2004 Discharges (%) 1 943 84 (1.17) 120 (1.58) 132 (1.68) 144 (1.31) 163 (1.11) 150 (0.87) 150 (0.74) >1 10 844 85 (1.19) 108 (1.43) 108 (1.38) 122 (1.11) 131 (0.90) 144 (0.84) 152 (0.75) >10 20 875 68 (0.95) 110 (1.45) 102 (1.30) 103 (0.94) 142 (0.97) 162 (0.94) 188 (0.93) >20 30 1500 160 (2.24) 161 (2.13) 151 (1.92) 199 (1.81) 236 (1.61) 269 (1.57) 324 (1.60) >30 40 3279 311 (4.35) 324 (4.28) 334 (4.25) 420 (3.83) 571 (3.91) 612 (3.57) 707 (3.50) >40 50 5419 502 (7.02) 467 (6.16) 491 (6.25) 700 (6.38) 913 (6.24) 1073 (6.26) 1273 (6.30) >50 60 7754 585 (8.18) 598 (7.89) 726 (9.25) 976 (8.89) 1362 (9.32) 1620 (9.44) 1887 (9.34) >60 70 13284 1123 (15.7) 1246 (16.45) 1253 (16.0) 1638 (14.91) 2249 (15.4) 2596 (15.13) 3172 (15.7) >70 80 25164 2157 (30.2) 2256 (29.78) 2268 (28.9) 3185 (29.0) 4361 (29.8) 5107 (29.8) 5830 (28.9) >80 90 22039 1735 (24.3) 1806 (23.84) 1867 (23.8) 2894 (26.4) 3741 (25.6) 4513 (26.3) 5483 (27.14) >90 4429 331 (4.64) 380 (5.01) 425 (6.01) 600 (5.46) 751 (5.13) 907 (5.28) 1035 (5.11) Totals 85530 7149 7576 7851 10980 14620 17153 20201
88 Appendix Z Table 11: Total deaths for Florida non-C DAD discharged patients and the percent of deaths per year, 1998-2004. Age Group Year Total Number of Deaths 19982004 1998 Deaths (%) 1999 Deaths (%) 2000 Deaths (%) 2001 Deaths (%) 2002 Deaths (%) 2003 Deaths (%) 2004 Deaths (%) 1 7422 964 (1.74) 1035 (1.81) 988 (1.78) 1096 (1.92) 1090 (1.93) 1161 (2.10) 1088 (2.02) >1 10 1016 172 (0.31) 155 (0.27) 160 (0.29) 138 (0.24) 138 (0.24) 127 (0.23) 126 (0.23) >10 20 2128 266 (0.48) 295 (0.52) 329 (0.59) 307 (0.54) 291 (0.52) 304 (0.55) 336 (0.62) >20 30 4017 564 (1.02) 576 (1.01) 565 (1.02) 575 (1.01) 566 (1.00) 553 (1.00) 618 (1.15) >30 40 9208 1428 (2.58) 1346 (2.36) 1359 (2.44) 1340 (2.35) 1282 (2.27) 1221 (2.21) 1232 (2.28) >40 50 20247 2683 (4.84) 2656 (4.65) 2867 (5.16) 2934 (5.15) 3036 (5.38) 3020 (5.46) 3051 (5.66) >50 60 32241 4258 (7.73) 4358 (7.63) 4424 (7.96) 4680 (8.21) 4740 (8.40) 4913 (8.89) 4868 (9.03) >60 70 58684 8637 (15.6) 8708 (15.3) 8216 (14.8) 8342 (14.6) 8353 (14.8) 8184 (14.8) 8244 (15.3) >70 80 111793 16484 (29.7) 16845 (29.5) 16273 (29.3) 16573 (29.1) 15939 (28.2) 15198 (27.5) 14481 (26.9) >80 90 111357 15561 (28.1) 16323 (28.6) 15922 (28.6) 16209 (28.4) 16217 (28.7) 15795 (28.6) 15330 (28.4) >90 32576 4361 (7.87) 4792 (8.39) 4493 (8.08) 4780 (8.39) 4802 (8.51) 4796 (8.68) 4552 (8.44) Totals 390689 55416 57089 55596 56976 56454 55272 53926
89 Appendix AA Table 12: Florida total death numbers for th e CDAD patients and the percent of death for each year, 1998-2004. Age Group Year Total Number of Deaths 19982004 1998 deaths (%) 1999 deaths (%) 2000 deaths (%) 2001 deaths (%) 2002 deaths (%) 2003 deaths (%) 2004 deaths (%) 1 12 0 (0) 0 (0) 0 (0) 4 (0.35) 4 (0.25) 1 (0.05) 3 (0.15) >1 10 9 3 (0.44) 1 (0.14) 2 (0.26) 1 (0.09) 1 (0.06) 1 (0.05) 0 (0) >10 20 22 1 (0.15) 2 (0.27) 2 (0.26) 3 (0.26) 2 (0.13) 8 (0.44) 4 (0.20) >20 30 49 6 (0.88) 8 (1.09) 6 (0.79) 5 (0.44) 9 (0.57) 11 (0.60) 4 (0.20) >30 40 116 9 (1.33) 9 (1.22) 10 (1.31) 9 (0.79) 24 (1.52) 26 (1.42) 29 (1.42) >40 50 266 22 (3.24) 26 (3.54) 32 (4.20) 34 (2.97) 45 (2.86) 56 (3.06) 51 (2.50) >50 60 545 52 (7.67) 43 (5.85) 54 (7.10) 68 (5.94) 92 (5.84) 115 (6.28) 121 (5.92) >60 70 1220 105 (15.5) 113 (15.4) 113 (14.8) 154 (13.5) 197 (12.5) 258 (14.1) 280 (13.7) >70 80 2750 220 (32.4) 231 (31.4) 237 (31.1) 331 (28.9) 536 (34.0) 548 (29.9) 647 (31.7) >80 90 3007 210 (31.0) 242 (32.9) 249 (32.7) 424 (37.1) 527 (33.5) 644 (35.2) 711 (34.8) >90 771 50 (7.37) 60 (8.16) 56 (7.36) 111 (9.70) 138 (8.76) 163 (8.90) 193 (9.45) Totals 8767 678 735 761 1144 1575 1831 2043