SPATIAL AND TEMPORAL ANALYSIS OF SALMONELLA ENTERITIDIS OUTBREAKS IN USA, 1990 - 2015 By Azam Ali Sher A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Epidemiology - Master of Science 201 8 A BSTRACT SPATIAL AND TEMPORAL ANALYSIS OF SALMONELLA ENTERITIDIS OUTBREAKS IN USA, 1990 - 2015 By Azam Ali Sher To address an endemicity of Salmonella Enteritidis (SE) infection, a food borne disease, in United States of America (USA), different food vehic les associated with SE outbreaks were examined in this study. Data of all SE outbreaks reported to Centers for Disease Control and Prevention ( CDC) from 1990 to 2015 was retrieved from CDC website. The d ata included following information about each outbrea k: year, month, state, implicated food vehicles, location, and number of SE cases. It was found that eggs - based dishes 273 (24%) were the high est reported followed by other implicated food items; meat 130 (11%), vegetables 96 (8%), chicken items 95 (8%), d airy products 55 (5%), and bakery items 8 (1%) in the country . Relative occurrence of food vehicles compared to eggs - based dishes implicated in SE outbreaks was examined by using negative binomial model which showed significant contribution of other food i tems in causing SE outbreaks in the country such as meat (exp ) =0. 51 , 95% CI 0.37, 0.69), chicken (exp ) =0. 42 , 95% CI 0.30, 0.58), vegetables (exp ) =0. 41 , 95% CI 0.29, 0.55), and dairy items (exp ) =0. 27 , 95% CI 0.18, 0.40). In addition, d ifferent trends of SE outbreaks were analyzed based on available dataset, a nd newly created categorical variables such as census region, HHS regions and seasons . T he study enhance d the existed knowledge of other implicated food items besides eggs in persistent occurrence of SE food borne disease in the USA. iii ACKNOWLEDGEMENTS I would like to express the deepest appreciations to my academic advisor and thesis committee chair, Dr. A. Mahdi Saeed, who has been a major source of motivation, knowledge acquisition and professional development for me throughout the degree program. He al ways supported me to explore and integrate horizons of different sciences to address complex and emerging research questions of fatal infectious diseases in the world. I am profoundly grateful to my committee members, Dr. Joseph Gardiner and Dr. Sue Grady for providing their valuable guidance and cooperation. Insightful questions and constructive comments from Dr. Gardiner always gave me a reason to learn different biostatical analyses thoroughly and to interpret statistical results prudently. I am very tha nkful to Dr. Grady for her immense collaboration and instructions to perform and complete spatial analyses. In brief, my research would have been impossible without the support and directions provided by my committee members. My sincere and heartfelt than ks to my beloved father, Ali Muhammad, and mother, Shahnaz Ali, and respected elders who have always been a source of inspiration, hardworking and dedication for me throughout my life to accomplish many personal and professional tasks. I extend sincere app reciations to my lovely siblings; Asim, Azmat, Asmat, Iqra and Ishrat for their unconditional love and fun what they always share with me. Next, I am thankful to my respected teachers, roommates, friends and colleagues for providing a great support to achi eve academic goals and making my stay in the USA memorable and fruitful. iv I would like to express gratitude to Fulbright Program Pakistan, Office of International Student (OISS) and the department of Epidemiology and Biostatistics at Michigan State Univers ity for their financial and administrative support throughout the program. I especially say thanks to the department chair, Dr. Claudia Holzman, graduate program director, Dr. Barondess, and academic program coordinator, Nancy Bieber, for their con sistent and generous support to finish this degree program successfully. I would like to acknowledge the sincerest effort and role played by my administrative advisors from International Institute of Education (IIE); Stephanie Sasz, Courtney Castillo, and from Uni ted State Educational Foundation in Pakistan (USEFP); Rooman Qayyum Khan, and from OISS at MSU; Nicole Namy to make my academic journey smooth and successful. Finally, I would like to thank everyone who helped and supported me throughout this study period . v TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ........................ vii LIST OF FIGURES ................................ ................................ ................................ ..................... viii KEY TO ABBREVIATIONS ................................ ................................ ................................ ........ ix CHAPTER 1: INTRODUCTION ................................ ................................ ................................ ... 1 CHAPTER 2: BACKGROUND ................................ ................................ ................................ ..... 3 2.1 Eggs ................................ ................................ ................................ ................................ ....... 4 2.2 Chicken ................................ ................................ ................................ ................................ . 8 2.3 Vegetables ................................ ................................ ................................ ............................. 9 2.4 Specific Ai ms ................................ ................................ ................................ ...................... 11 2.4.1 Aim 1: ................................ ................................ ................................ .......................... 11 2.4.2 Aim 2: ................................ ................................ ................................ .......................... 11 CHAPTER 3: DATA and METHODS ................................ ................................ ......................... 12 3.1 Study Population ................................ ................................ ................................ ................. 12 3.2 Food Vehicles ................................ ................................ ................................ ..................... 12 3.3 Season ................................ ................................ ................................ ................................ . 14 3.4 Location ................................ ................................ ................................ .............................. 14 3.5 Regions ................................ ................................ ................................ ............................... 14 3.6 Group of Years ................................ ................................ ................................ .................... 14 3.7 Analytic Plan ................................ ................................ ................................ ....................... 15 CHAPTER 4: RESULTS ................................ ................................ ................................ .............. 16 4.1 Food Categories Associated with SE Out breaks ................................ ................................ 16 4.2 Temporal Changes in Food Categories Associated with SE Outbreaks ............................. 18 4.3 Geographic distribution of SE outbreaks in t he USA ................................ ......................... 20 4.4 Seasonality of SE Outbreaks ................................ ................................ ............................... 24 4.5 Location of SE Outbreaks ................................ ................................ ................................ ... 25 4.6 Statistical Analyses ................................ ................................ ................................ ............. 26 CHAPTER 5: DISCUSSION ................................ ................................ ................................ ........ 28 5.1 Interpretations ................................ ................................ ................................ ..................... 28 5.2 Strengths ................................ ................................ ................................ ............................. 29 5.3 Limitations ................................ ................................ ................................ .......................... 30 CHAPTER 6: CONCLUSION ................................ ................................ ................................ ..... 32 APPENDIX ................................ ................................ ................................ ................................ ... 33 vi REFERENCES ................................ ................................ ................................ ............................. 42 vii LIST OF TABLES Table 1: Distribution of dif ferent food categories implicated in Salmonel la Enteritidis outbreaks in USA (1990 - 2015) ................................ ................................ ................................ ........ 17 Table 2: Food Vehicles Implicated in Salmonella Enteritidis Outbreaks in USA ........................ 26 Table 3: Relative Occurrence of Food Vehicles Compared to Eggs - based Salmonella Enteritidis Outbreaks ................................ ................................ ................................ ......................... 27 Table 4: Description of all possible combinati ons of different foods implicated in Salmonella Enteritidis outbreaks in USA (1990 - 2015). ................................ ................................ ..... 34 Table 5: Analysis of Maximum Likelihood Parameter Estimates ................................ ................ 35 viii LIST OF FIGURES Figure 1: Description of the trends of Salmonella Enteritidis outbreaks associated with different foods sources from 1990 - 2015 in the USA. ................................ ................................ .... 19 Figure 2: Distribution of Salmonella Enteritidis outbreaks based on main known food sources, excluding bakery items across the USA from 1990 - 2015. ................................ .............. 20 Figure 3 : Description of Salmonella Enteritidis outbreaks occurred in census regions of USA from 1990 - 2015. ................................ ................................ ................................ .............. 22 Figure 4: Distribution of number of Salmonella Enteritidis outbreaks based on different fo ods across four census regions of the country. ................................ ................................ ...... 23 Figure 5: Distribution of Salmonella Enteritidis outbreaks across ten HHS regions of the USA from 1990 - 2015. ................................ ................................ ................................ .............. 24 Figure 6: Distribution of different food categories implicated in Salmonella Enteritidis outbreaks in USA (1990 - 2015) ................................ ................................ ................................ ........ 36 Figure 7: Distribution of foods associat ed with Salmonella Enteritidis outbreaks across five distinct groups of years in USA (1990 - 2015) ................................ ................................ . 36 Figure 8: Description of Salmonella Enteritidis outbreaks in four census regions of USA: Nort heast, Midwest, South, and West. ................................ ................................ ............ 37 Figure 9: Distribution of Salmonella Enteritidis outbreaks across HHS regions which had over two outbreaks per year. ................................ ................................ ................................ ... 37 Figure 10: Description of Salmonella Enteritidis outbreaks occurring trend in different seasons: Winter, Spring, Summer and Fall across the US census regions from 1990 - 2015. ........ 38 Figure 11: The pattern of different seasons - based Salmonella Enteritidis outbreaks in the USA HHS regions from 1990 - 2015. ................................ ................................ ........................ 38 Figure 12: Number of outbreaks and case rate associated wi th Salmonella Enteritidis outbreaks in five distinct group of years in USA ................................ ................................ ............. 39 ix KEY TO ABBREVIATIONS SE Salmonella Enteritidis CDC Center s for Disease Control and Prevention HHS Human and Health Services CI Confidenc e Interval MOR Marginal Odds Ratio WHO World Health Organization USA U nited S tates of America USDA United States Department of Agriculture SEPP SE Pilot Project PEQAP Pennsylvania Egg Quality Assurance Program FDA Food and Drug Administration 1 CHAPTER 1 : INTRODUCTION Salmonella serotypes are the leading causes of diarrhea and gastroenteritis in the world. It has been estimated that 93.8 million of cases of nontyphoidal salmonellosis and 155,000 deaths occur every year in the world where 86 percent of all these illnesses are due to foodborne infections. N ontyphoidal S almonella serotypes are a major cause of foodborne infections and outbreaks in the world (1) . Increasing foodborne infections have become serious public health issue in both developing and developed countries. A study on estimating the burden of foodborne diseases in the USA was conducted by using active and passive surveillance data of 31 different food pathogens. It was estimated that every year, about 9.4 million foodborne infections occur in the USA. The study results showed that no ntyphoidal Salmonella caused over 1 million infections (28%) out of 3.6 million bacterial infections in the USA. It was reported that nontyphoidal salmonellae are a major cause of foodborne diseases in the USA (2) . Among nontyphoidal Salm onella serotypes, Salmonella Enteritidis (SE) is the most commonly reported serotype. Based on isolates of Salmonella Enterica collected from human laboratories of 37 countries from all over the world (2001 - 2007), Salmonella Enteritidis (SE) was ranked fir st with around 43.5% proportion of all Salmonella isolates. SE ranked first among Nontyphoidal Salmonella infections among humans in the USA (3) . In a recent study, it was reported that 38 % of reported nontyphoidal Salmonella outbreaks were caused by SE from 1973 to 2009 in the USA (4) . Similarly, it was estimated that SE isolates represent 13 - 19% of overall Salmonella isolates in the USA (4,5) . Another recent study also showed an increasing trend of SE isolates per 100,000 population over Salmonella Typhimurium serovar, a previous highly prevalent Salmonella serotype in the USA (6) . These findings clearly prove SE as a major Salmonella se rotype associated with outbreaks of foodborne Salmonella infections . More importantly, a 2 study on genome comparison of different Salmonella enterica serovars showed that SE has gained such diversity both genomic and metabolically that it can survive and infect multiple hosts (7) . Persistent occurrence of SE outbreaks across the country requires epidemiological studies to identify important risk factors and routes of disease transmission to address this public he alth challenge in the USA . In addition, Frenzen et al. showed that salmonellosis has a significant economic impact in the USA, it overall costs 0.5$ to 2.3$ billion yearly and SE is one of the major Salmonella serotypes , ranked one, in causing S almonellosi s in the country (8) . In the past, few studies were designed to address food sources associated with SE outbreaks in US. Those studies showed that raw and uncooked eggs products were the main source of SE outbreaks in the country. Implementations of differen handling, transportation and storage led to a tremendous decline in eggs associated outbreaks. However, there are still significant SE outbreaks occurring in the country yearly which can be due to consumption of other SE - contam inated foods that have not been addressed in previous studies. C onsidering the above factors, this study is designed to elucidate the role of different foods associated with SE outbreaks at various time points and places in the country. 3 CHAPTER 2: B ACKGROUND Numerous studies found that diarrhea is among the most common health issue s in developed countries including Australia, Canada and USA . Diarrhea was more likely reported in females and children under the age of five years (9) . It shows the importance of recognizing food borne diseases , which in most c ases, cause diarrhea in humans. T here are multiple bacterial, viral and parasitic causes of food borne illness , however, Salmonellosis is a major cause of diarrhea in the country . This was found from a study which was conducted using data from both active and passive surveillance of food borne infections caused by thirty - one pathogens from 2000 to 2006 in the US A (10) . It was found that non typhoidal Salmonella serotypes are the leading cause of hospitalization (35%) and death (28%) among all foodborne infections that occurred in the country during 2000 to 2006 (10) . A study ha s been conducted on Salmonella enterica serotypes - based outbreaks and associated food items reported to CDC from 1998 - 2008 in USA found that among single food implicated - outbreaks, SE caused more outbreaks of approximately 36% (144/403) followed by Typhimu rium (14%) Newport (10%) and Heidelberg (6%). It revealed that SE is causing more foodborne outbreaks than any other serotype of Salmonella enterica in the US A (11) . According to an annual report of food borne disease outbreaks in USA, 2015, it was found that Salmonella is second common cause of single - etiology confirmed outbreaks in the country, causing 149 of 920 (34%) outbreaks, and 3944 (39%) cases fol lowing Norovirus which caused 164 of 920 (37%) outbreaks, and 3893 (39%) cases. Among 164 Salmonella outbreaks, SE was the most common serotype and was reported in 51 outbreaks (35%) (12) . 4 2.1 Eggs SE infection has been the most prevalent type of salmonellosis among humans in the US for many years. It is a food borne infection and spreads through consumption of different SE - contaminated food vehicles. In the past, e ggs - containing products had been significantly associated with SE outbreaks in the USA. The data from National Surveillance Sources (1973 to 1984) showed that around 44% SE outbreaks were associated with egg - containing products. Those outbreaks were associated with eggs which were n either contamina t ed by farm soil n or were cracked. The first study conducted by St Louis et al. revealed that 27 (77%) of the 35 SE outbreaks, occurr ing during 1985 - 1987, were associated with Grade A shell eggs (13) . Later, shell eggs - associated SE outbreaks raised the possibility of transovarial transmission of SE bacteria in egg laying chickens (14) . In addition, u nder the National Poultry Improvement Plan (NPIP), SE was identified in diverse sampling from thirty - six environmental, six non - viable fertilized eggs, and fifteen birds from sanitation - monitored breeder flocks since 1990 [ A.R. Rhorer, unpublished data , (15) ] . Moreover , i t has been recorded that the prevalence of different PT of SE changed over time such as in the 1980s , the most prevalent PT of SE were PT13a and PT8 (16) while in early 1990, PT23 and PT8 were the most common causes of SE associated human outbreaks in the country (17) . After that, an emergence of P T4 occurred in the US when first PT 4 associated SE outbreak was reported in Texas state in 1993 followed by PT4 outbreaks in Utah and California. The s tudy showed that the emergence of PT4 in the west region caused more outbreaks than above - mentioned phag e types (PT8, PT13a, PT23) in the country (18) . For example, i t was estimated that 496 cases of SE were recorded in Los Angeles County, California in just four months (April to July) in 1994 (19) . All t h ese finding s projected a great concern for 5 public health department because the infected birds can spread th e disease in flock quickly and can cause the disease outbreaks among humans by contaminated eggs . After recognition of shell eggs as a potential source of SE outbreaks, traceback program was introduced by United States Department of Agriculture (USDA) to source and origin. Altekruse et al . conducted a study under the traceback program and found that SE phage types associated with human outbreaks were recovered from 100% and 88% of the environment al samples and internal organs of he ns respectively from fourteen different implicated poultry farms (17) . It shows that traceback program was helpful in tracing and confirming the source or origin of SE outbreaks among humans which were caused by contaminated eggs (17) . SE became the most frequently reported Salmo nella serotype in 1990 in the USA. It accounted for 21% of all salmonellosis in 1990. From 1985 - 91, there were 380 outbreaks (213 food vehicle - unconfirmed and 167 foods vehicle - confirmed). Of 167 food vehicle - confirmed SE outbreaks, 137 (82%) were egg - ass ociated. There were 59 SE outbreaks reported in h ospital s and nursing homes accounted for most of the deaths in this study period. Thirteen out of fifty nine also found eggs associated SE outbreaks (20) . Similarly, Patrick et al. studied SE outbreaks reported to CDC from 1985 - 1999 in the USA (21) . There were 841 outbreaks reported to CDC and only 371 outbreaks report s had information about food vehicle and they found that 298 (80%) were egg - associated outbreak s. These eggs - associated outbreaks were associated with very diverse egg - containing recipes such as raw egg dishes (egg - nog, tiramisu, C aesar salad dressing), lightly cooked egg - dishes (hollandaise, sauce, meringue) cooked egg dishes (lasagna, z iti, and s tuffing) and egg battered dishes (chiles rellenos, egg rolls etc.) It shows that egg - containing recipes cooked in several 6 styles were equally capable to cause SE infections in humans. It might suggest the occurrence of contamination while cooking, serving or any other raw food mixing with egg - containing dishes. Next, they found that of 371 SE outbreaks, 20 (27%) outbreaks were chicken or turkey, 8 (11%) beef, 6 (8%) shrimps, and 9 (13%) vegetables associated. It shows that SE outbreaks happened from 1985 - 19 99 were mostly egg associated which matches other studies and ours as well (21) . After identification of high numbers of eggs - based SE outbreaks i n 1990, vigorous preventive measures were advised and implemented in the country (15) . In February 1990, USDA implemented SE regulation after recognizing a highly increased number of SE outbreaks and cases. Mostly these outbreaks occurred in the north - east region of the country. Traceback program was introduced by USDA to find the origin of contaminated eggs associated with SE outbreaks in humans. After identifying implicated poultry farms, SE presence was tested in the environment and laying hens. In addition, a survey w as conducted by USDA in 1991 and 1995 to detect the SE prevalence in consumed hens at slaughterhouses and unpasteurized eggs at breaker - plants, and result of this study showed an increased prevalence of SE in all across the country (15,22) . Later, the USDA started a quality assurance program, SE Pilot Project (SEPP), in 1992 by a collaboration of research i nstitutions, University of Pennsylvania, Pennsylvania State University, and Pennsylvania Department of Agriculture, and Pennsylvania Poultry Federation. Under the SEPP program, samples from birds, rodents, and the environment from poultry farms were tested . If farms were SE positive , then eggs were tested, and if eggs were also SE positive then all eggs from that farm were sent to pasteurization plant before sending them to markets. Later, in 1994, the SEPP w as changed into Pennsylvania Egg Quality Assuranc e Program (PEQAP) (15) . In 1995, an environmental survey was conducted and found that layer farms participated in SEPP or PEQAP program had reduced i nfections and isolation of SE (23) . 7 Another study showed that the diversion of eggs fr om farm to pasteurization plant under the SEPP program resulted in a remarkable reduction of human SE cases and outbreaks in the north - east region (24) . It shows that overall PEQAP progra m played a key role in controlling of SE infections and outbreaks in the north - east region. Similarly, other major step taken by governmental agencies and eggs producing and storage companies was storage of eggs at efficient temperature to further reduce S E outbreaks . It was estimated in 1998 that the refrigeration of eggs after laying , packing , and shipping could reduce SE cases by 8% and 12% respectively (25) , and i n August 1998 , the USDA issued regulations of eggs storage and transportation at temperature no higher than 45°F (7.2°C) and labeling of containers with storage instructions (26) . Moreover, in 1999, Seventeen different states implemented a rule to print an - (27) . Next, Food and Drug Administration (FDA) also issue d a regulation in 2000 which requires proper refrigeration of eggs for sale at retail stores (28) . To control the SE illne Action Plan on December 10, 1999, stating the goal to reduce eggs associated SE infections by half by 2005 and eliminate them by 2010 (29) . In brief, multiple preventive measures against eggs associated SE outbreaks were taken since the early 1990s to reduce the disease outbreaks . In general , the preventive steps included consumer and producer awareness about the di sease, traceback policy, on - farm testing, PEQAP, disease education to vulnerable population such as children, elder ly , and pregnant women to avoid consuming raw and uncooked products of eggs. In addition, restaurant and catering places were informed to mak e sure clea nliness, proper storage, cooking and serving of eggs containing products . Although many successful interventions were implemented to reduce the eggs associated SE infections and outbreaks, there are still SE outbreaks occurring at a similar rate in the USA. 8 2.2 Chicken A case - control study conducted at FoodNet sites showed that chicken was also a major risk factor for SE infections in the USA. (30) USDA survey based on Hazard Analysis and Critical Control Points (HACCP) determined that SE was found in 2.4% of broilers at slaughter house (31) . Kim ura et al. found that people who ate chicken at dining places out of the home were more likely to have SE infection (MOR: 2.2, 95% CI, 1.4 3.4) than people those ate chicken at home (0.4; 95% CI, 0.3 0.6). It shows that ready to go or ready to eat products of chicken or any other food can be a potential source of SE outbreak occurrence (30) . In studies from the early 1990s, chicken was no t a common source of SE infection in the USA. However, it was the endemic and frequent cause of human illness spread through SE infected and contaminated broiler chickens across the world, especially European countries (32,33) . Later, specifically, after the emergence of SE PT4 in the USA , there is a shift in the route of SE infection to hu mans . Currently more outbreaks are occurring due to contaminated chickens instead of eggs , which were a common source in the early 1990s in the USA . Similarly , a study conducted in Thailand (1995) showed that PT4 was a frequent source of chicken contaminat ion in retail places. They identified PT4 among 17 (74%) out of 23 SE isolates collected from retail chicken samples. It is well known from previous and current studies that overall salmonella is more likely to infect chickens, contaminate chicken and meat carcasses and cause salmonellosis in humans as well. Similarly, SE can cause human illness by contaminating chicken and meat carcasses along with eggs (34,35) . Roberts et al. (36) reported that SE PT4 was isolated from frozen chicken samples collected under the United Kingdom retails survey in 1990. It means that PT4 contaminated chicken was a significant source of SE i nfections in Europe 9 while eggs were major risk factor of SE infections in the USA. It is also possible that PT4 contaminated chicken might have become now a potential source of chicken - based outbreaks in the USA , which has been increasing since 2009 accord ing to our study results. It suggests that phage typing of current circulating serotypes of SE and contaminated chicken might explain more about PT4 specific outbreaks and infections in the USA. However, From a study, the virulence of PT4 in chicks was fou nd significantly higher than other phage types (PT8, PT13a, PT14b) (37) . Similarly, it was reported in an annual report of foodborne outbreaks of 2015 that chicken was implicated in 22 outbreaks (11%). Of 22 chicken outbreaks, salmonella serotypes were implicated in 9 outbreaks which clearly shows that chicken is major source of Salmonellosis (12) . A recent source attribution study from Ahlstrom et al (38) showed that chicken was a promi nent source associated with salmonella human cases in Minnesota State. In addition, they found that among the salmonella serotypes, SE was a major serotype in causing human cases throughout the study period (2005 - 2014). It can be inferred from the study th at chicken and its products could be a main source of SE infection s among humans in Minnesota and the whole country as well. 2.3 Vegetables Vegetables ha ve been an integral part of many ready to eat products. In addition, fashion of eating raw vegetable s in the form of salads , burger s and sandwiches is growing a lot all over the world. Food poisoning, food borne illness and outbreaks has also been increas ing but vegetables as major food vehicle associated with SE outbreaks has not been addressed in US. A Mexican study has shown an isolation of SE and other Salmonella serotypes from samples of v egetables (39) . Another study was conducted in Malaysia to determine the prevalence of 10 different S almonella serotypes in vegetarian patties, burgers and salads. Authors et al. found the highest numbers of SE positive samples (35/175) among all Salmonella serotypes followed by Salmonella Typhimurium (25/175) . Mostly SE positive samples were lettuce, cab bage and leafy vegetables which might be due to high chance of Salmonella surface (40) Similarly, a study conducted an analysis on leafy vegetables associated with fo od borne outbreaks in USA from 1973 to 2012. They found out that Salmonella is the third main cause of leafy vegetables associated foodborne outbreaks in the US after Norovirus, and E. coli. Salmonella was confirmed or suspected in 10/162 of single leafy v egetables and in 22/444 of leafy vegetables - based salads implicated in food borne outbreaks from 1973 to 2012 (41) . According to annual report of food borne diseases outbreaks in USA, 2015. It was fo und that the most illnesses, 1504, were associated with overall vegetables - based outbreaks (12) . Of 1504 illnesses, 1311 infections were caused by Salmonella serotypes which shows that contaminated vegetables are also major source of salmonellosis (12) . Moreover, pathogens contaminated vegetables can spread food bo rne diseases easily and cause more illness than other food items. an important way to reduce the SE survival and growth on vegetables. A study show ed a decreasing trend in growth of SE on vegetables stored at 4 °C both in air and i n 4.9% CO 2 , 2.1% O 2 , 93% N 2 conditions (42) . However, it has been observed that Salmonella spp. can survive at wide range of temperatures such as from lowest tem perature (4 to - 8 °C) to highest temperature (44 °C) (43) . In addition, proper cooking of vegetables can play a key role in reducing vegetables - associated SE outbreaks and infections. 11 2.4 Specific Aims The main goal of this study is to investigate the numbers, spatial and temporal distributions of food s associated with SE outbreaks in the USA (1990 - 2015) and how they vary by locations, season and regions. To achieve this goal, two specific aims have been made for the study: 2.4.1 Aim 1: To characterize the foods associated with SE outbreaks in the USA (1990 - 2015). In study will explore other sources of foods such as meat products, fresh produce and dairy products etc. to uncover key factors relating to SE outbreaks t hat may be related to these unreported disease - causing sources. I hypothesize that fresh produce, vegetable containing dishes and meat products are playing a significant role in causing SE outbreaks and maintaining the disease endemicity in the country. 2.4.2 Aim 2: To study locations, seasons , and regions associated with SE outbreaks overall and for different food sources. I hypothesize that SE contaminated foods sources vary across 12 CHAPTER 3 : DATA and ME THODS 3.1 Study Population F oodborne outbreaks occurred in USA has been regularly reported to Centre for Disease Control and Prevention (CDC) since 1973 (44) . It is defined by CDC that occurrence of two or more cases of foodborne disease from same food vehicle will be termed a foodborne disease ou tbreak (44,45) . Reports of these foodborne outbreaks contain information on eti ological agents, cases, deaths, hospitalization, year, month, state, food and location. All this information has been posted on CDC website and is publicly access ible . In addition, the data available on CDC website is fully decoded and in summarized form a nd does not include such an information which can be used to identify and contact subjects. Current study includes all reported outbreaks of SE in USA from 1990 to 2015. Data was retrieved from CDC Food Tool, an online data source of reported foodborne out breaks in USA from 1998 onward . Data before 1998 was collected from papers - based line - listings of reported foodborne outbreaks available on CDC website. Data was verified from available online sources , organized, and transformed into SAS dataset by recodin g the following variables: location, season and food vehicle into categorical variables with various levels. 3.2 Food Vehicles A ll food vehicles involved in SE outbreaks are divided into the following 7 categories (Egg - based dishes, Chicken containing di shes, Meat/Seafood containing products, Vegetables, Dairy products, Bakery items, Unknown food). These categories are chosen and made based on the knowledge of primary ingredients of recipes and already known foods reported in previous SE outbreaks. It is already known that eggs and chickens are the main risk factors of SE outbreaks in US but in this study, we want to know about the critical role of other food vehicles, 13 vegetables, meat and dairy products in endemicity of SE disease in USA. For this purpos e, all the plant origin products, vegetables, sprouts and fresh produce are included in category of as bread, tortilla, chips etc. are valuable information about di fferent other risk factors of SE parallel to Eggs and Chickens containing dishes. There was much overlap between food categories where we found eggs with meat, or chicken etc. It is a challenge to divide the food vehicles into distinct categories. For exa mple, from initial classification of food vehicle, a table is generated below. In this table, it is attempted to consider all combinations of different food categories into the classification, but it would be hard to run and interpret statistical analysis by inputting above mentioned food vehicles into statistical model. Here, a strategy is designed to classify the food vehicles for the purpose and analysis of the study. Based on the known risk of food associated with SE outbreaks from previous literature, all food categories are ordered in the following way: eggs items > chicken items > meat items > vegetables > dairy items > bakery. In this way, if eggs are found as an ingredient of outbreak associated food, then the outbreak is considered eggs associated by ignoring the presence of other foods. Similarly, if eggs are absent and chicken is present with other foods then outbreak is assumed to be associated with chicken. Most important, meat and vegetables categories would not be including any eggs and chick en containing dishes. Even in the analysis, we might get less value estimate, but it can be reported and compared with 14 Thus, all food vehicles are divided int o seven categories (Egg - based dishes, Chicken containing dishes, Meat/Seafood containing products, Vegetables, Dairy products, Bakery items, and Unknown food). 3.3 Season Furthermore, based on month of outbreaks and meteorological classification of season , a categorical variable, season: summer, spring, autumn and winter, was created. 3.4 Location Based on previously reported studies on SE and other foodborne outbreaks, all locations of occurrence mentioned in the data set are categorized into 7 types (F ood party, Grocery stores, Nursing/Hospital, Restaurant, School/Prison, Private residence, Unknown place) by considering importantly locations of food preparation, distribution and serving methods. (46) 3.5 Regions It is also attemp ted to find any trend of SE outbreaks among different regions of US. Based on given information of state in the data we created a categorical variable, census region, with four different regions: Northeast, South, Midwest and West. Similarly, based on regional classification of Human and Health Services (HHS), we also created a categorical variable, HHS, with ten different regions of USA. These occurrence across the country. 3.6 Group of Years Knowing the year of outbreaks, we divided twenty - six years (1990 - 2015) into five groups: 1990 - 1994, 1995 - 1999, 2000 - 2004, 2005 - 2009, 2010 - 2015. From literature, it is seen that mostly interventions regard ing the control of eggs associated outbreaks were done in 1990 - 15 1995 and 1995 - 1999. This group classification will not only help to analyze temporal change in overall occurrence of SE outbreaks but also learn about trends of specifically eggs associated out breaks and impact of related interventions. 3.7 Analytic Plan First, d escriptive s tatistics (means, standard deviations, frequency and proportions) were calculated for all variables of interest . For statistical analysis, Poisson regression model and Neg ative binomial model w ere applied using SAS software (9.4 version) to measure an association between the number of reported outbreaks with different categorical variables (food vehicle, location, season, HHS and census regions). All geographic analysis an d mapping procedures were performed in ArcGIS software (10.5 version) . The following equations of Poisson regression were used to calculate estimates : Where are terms for categorical variables (covariates). A p - value of < 0.05 was used to de clare statistical signifi cance of covariates . 16 CHAPTER 4 : RESULTS 4.1 Food C ategories A ssociated with SE O utbreaks There were 1144 single - state SE outbreaks reported to CDC from 199 0 - 2015. Of 1144 SE outbreaks, 657 (57%) outbreaks were reported with known food items and 487 (43% ) were unknown source - based outbreaks. Of 657 known food implicated outbreaks, 273 (24%) are associated with eggs - based products and ranked as first cause of SE outbreaks cumulatively occurred from 1990 to 2015. M eat products are found as second key cause of SE outbreaks 130 (11%) followed by other implicated food items; vegetables 96 (8%), chicken items 95 (8%), dairy products 55 (5%), and bakery items 8 (1%) in the country. (Table 1) It was analyzed descriptively by making categories of outbreaks associat ed food vehicle i.e. simple (single food implicated) and complex (one main high - risk item along with low - risk multiple food items). Of 273 egg - based outbreaks, 93 (8%) were simple egg - based and 180 (16%) complex based outbreaks. Of 130 meat products, 75 (7 %) are simple and 55 (5%) were complex category outbreaks. Similarly, of 95 (8%) chicken, 53 (4%) simple and 42 (4%) complex . O f 96 (8%) vegetables outbreaks, 75 (5%) simple and 21 (2%) complex category - based outbreaks. Among 55 (5%) dairy products, 44 (4% ) simple and 11 (1%) complex category - based outbreaks. (Table 1, Figure 6 ) T hese categories were made based on the disease risk associated with food items reported in previous literature in the country and world. Accordingly , eggs - based dishes were highly associated with SE outbreaks and were given highest rank in the risk order list. 17 Table 1 : Distribution of different food categories implicated in Salmonella Enteritidis outbreaks in USA (1990 - 2015) Food Vehicles Sub - category Fre quency Percent Total Frequency Percent Egg - based dishes 273 23.86 Simple Egg - based dishes 93 8.13 Complex Egg - based dishes 180 1 5.73 Chicken Items 95 8.30 Simple Chicken Items 53 4 .63 Complex Chicken Items 42 3.67 Meat Products 130 11 .36 Simple Meat Products 75 6.56 Complex Meat Products 55 4.81 Vegetables 96 8 .39 Simple Vegetables 75 6.56 Complex Vegetables 21 1.84 Dairy Products 55 4.81 Simple Dairy Products 44 3.85 Complex Dairy Products 11 0.96 Bakery Items Simple Bakery Items Only 8 0.70 8 0.70 Unknown 487 42.57 487 4 2.57 *Foods are divided into two categories, simple and complex, where simple means that it was only major ingredient and complex means that there were some other ingredients along with this major ingredient. C hicken was ranked as a second major risk factor followed by meat, vegetables, dairy products and bakery items. There is a possibility that this classification may overestimate the high order food items association with SE out breaks such as eggs and chicken . For that reason, w e further analyzed all possible combinations of different food items to see which foods are more likely to be implicated together in SE outbreaks (Table 3 ) . In case of eggs - based complex 18 outbreaks (180), i nterestingly, the dairy items and eggs together were highly reported 21 % (37/180) followed by vegetables 13% (24/180), meat 10% (18/180) and bakery 8% (15/180). It gives an idea of how much other food items are involved and implicated along with eggs in ca using SE outbreaks reported to CDC from 1990 to 2015. There are chances of underestimation and underreporting of food items other than eggs implicated in outbreaks . This underestimation might be a reason of lack of regulations and preventive measures taken against these food items to reduce SE occurrence in the country. Similarly, in case of chicken, major food items implicated with chicken items in SE outbreaks are vegetables 55% (23/42) and meat 14% (6/42). Next, of 55 meat based complex outbreaks, vegeta bles are the main food item reported jointly with meat products in 27 (49%) outbreaks followed by dairy and meat items together in 15 (27%) outbreaks. Furthermore, of 21 vegetables based complex category outbreaks, both dairy and bakery items were reported separately with vegetables in 8 (38%) SE outbreaks (Table A) . Overall, vegetables are the most frequent food items which are found jointly with implicated food items such as eggs, chicken and meat items in SE outbreaks from 1990 to 2015. This shows that v egetables items and meat products are also playing key role in causing of SE outbreaks in the country which has not been addressed properly in past years . However, few regulations exist to reduce SE outbreaks caused by these food items . 4.2 Temporal C hang es in F ood C ategories A ssociated with SE O utbreaks Next, we explored the overall temporal changes among these food vehicles associated SE outbreaks in USA. A plot was drawn between the numbers of outbreaks per food vehicle versus years starting from 1990 t o 2015 (Figure 1) . 19 Figure 1 : D escri ption of the trends of Salmonella Enteritidis outbreaks associated with different foods sources from 1990 - 2015 in the USA. We only kept major known sources in this plot, removed the bakery an d unknown items (Figure 1) . Next, we also divided the study period (26 years) into groups of 5 years except the last group which is based on 6 years period (Figure 7 ) . In both figure s , eggs associated outbreaks were dominantly higher than other known food sources until 2005 and afterward, when other food sources i.e. meat, chicken, vegetables, caused more outbreaks or at similar rate as eggs. Importantly, the number of chicken associated outbreaks has been increasing from 2009 to 2015. Moreover, the occurre nce of SE outbreaks by food sources other than eggs remained a consistent source of outbreaks since 1990s. It shows that we need to pay attention to these other food sources instead of only focusing on eggs alone to control the diseases occurrence at singl e and multistate level outbreaks in the country. 20 4.3 Geographic distribution of SE outbreaks in the US A In the current study, we observed the distribution of SE outbreaks occurred in District of Columbia ( Washington D.C. ) and other 48 states of the US A. Figure 2 : D istribution of Salmonella Enteritidis outbreaks based on main known food sources, excluding bakery items across the USA from 1990 - 2015. In figure 2, the number of different food categories associated SE outbreaks w ere plotted across states in the country. It was found that overall, highest number of outbreaks were reported by CA (130) followed by NY (127), PA (98), OH (70), MD (69), MA (52), IL (49) and NJ (47) to CDC from 1990 - 2015. However, d uring th is study perio d, eggs - based dishes were the dominant source of associated SE outbreaks throughout the country . However, there was some variation in the second most attributed food source across states . M eat products w ere second 21 major cause in high ranked states i.e. CA, NY, CT, MD and PA while vegetables in (VA, NJ, WA) and chicken items in (MN, WI) were second major source of SE outbreaks after eggs (Figure 2) . It shows a diversity in food items associated with outbreaks across the country. E ach state may declare regula tions against specific food items which are more associated with the disease occurrence there. Next, we explored the distribution of outbreaks associated with different food sources across different c ensus regions. Of 1144 SE outbreaks, 402 (35%) outbreak s occurred in Northeast region followed by South 253 (22%), West 250 (22%), and Midwest regions 239 (21%). As this distribution is based on aggregate count of outbreaks from 1990 to 2015 so, we further analyzed the trend of outbreaks that occurred in the c ensus regions on a per year basis since 1990. We found (Figure 3) that the highest number of outbreaks occurred in Northeast d spread and become endemic through the country. It is clear from the figure that more outbreaks are currently being reported from South, West and Midwest regions than Northeast region. 22 Figure 3 : D escri ption of Salmonella Enteritidis outbreaks occurred in c ensus r egions of USA from 1990 - 2015. In a di fferent analysis, we plotted a graph between census region outbreaks and five distinct groups of years. It was observed that Northeast region reported higher number of outbreaks from 1990 - 1994 but interestingly in the next five years, from 1995 - 1999, West region reported higher number of outbreaks . This might be due to an emergence of PT4 in West region that caused too many outbreaks simultaneously . In addition, from 2010 to 2015, Midwest region showed an increasing trend of outbreaks followed by West, Sout h, Northeast region s . It shows a changing pattern of outbreaks across time and regions (Figure 8 ) . We were interested to see the changes in occurrence of SE outbreaks in census regions based on known and unknown food items. Based on summarized the data , Fi gure 4 shows that overall, N ortheast region reported high number s of outbreaks associated with both unknown and 23 major known food categories. South region reported meat products associated outbreaks as high as eggs - based. However, we saw difference s in food items associated at the state level , but in case of census regions, this effect was diminished, and all census regions showed similar trends of food vehicles associated with SE outbreaks. Figure 4 : D istribution of number of Sal monella Enteritidis outbreaks based on different foods across four census regions of the country. We further studied the distribution of different foods - based outbreaks in HHS regions of the country because most of the health policies are designed based on the HHS statistics. This regional distribution has much less variability than census regions. It was found that HHS region 3 reported the highest number of outbreaks 225 (20%), followed by HHS region 5, 210 (18%), HHS region 2, 174 (15%), HHS region 9, 155 (14%), HHS region 1, 130 (11%), and HHS region 4, 109 (10%). Figure 9 provides additional information regarding SE outbreaks occurred yearly 24 in HHS regions. Next, different patterns of HHS regional outbreaks based on food items can be observed in figu re 5, where almost all regions showed high numbers of outbreaks associated with eggs - based dishes. Particularly for eggs associated outbreaks, region 3 had higher number of outbreaks followed by HHS regions 5, 2 and 9. In case of meat associated outbreaks, region 3 has significant higher outbreaks followed by HHS regions: 9, 2, 3,4 and 5. For chicken, HHS regions 7 and 3 reported higher outbreaks than others. Similarly, HHS regions 2, 3, 1, 9, 5, and 4 reported more vegetables - based outbreaks (Figure 5) . Figure 5 : D istribution of Salmonella Enteritidis outbreaks across ten HHS regions of the USA from 1990 - 2015. 4.4 Seasonality of SE O utbreaks Current study also analyzed the seasonal effect on occurrence of SE outbreaks in the USA. It was found that the highest number of outbreaks 437 (38%) in Summer followed by Fall 25 282 (25%), Spring 260 (23%), and Winter 165 (14%). We created a plot between season and census regions to show the pattern of regional outbreaks across season s . We foun d that almost all four census regions showed a similar pattern of outbreaks over the four seasons (Figure 10 ) . Similarly, HHS regions showed a similar trend of season - based outbreaks like census regions where higher number of outbreaks occurred in Summer f ollowed by Fall, Spring and Winter. However, HHS region 2 reported high number of outbreaks in both seasons Summer and Fall at same rate (Figure 11 ) . Spatial visualization in Figure 12, clearly shows an endemicity of the disease in the country. It was see n that in distinct groups of years , there was a switching pattern of the disease outbreaks and associated case rate, calculated per state population, in the country. It might be due to delayed or less interest of state food authorities to develop new preve ntive measures or to implement policies swiftly against SE disease outbreaks. Initially, S E outbreaks were reported the trend of high outbreaks moved to West region, and now outbreaks are happening more i n South and Midwest. It seems that delay in response against SE infection or any other food borne disease can spread throughout the country after some time . 4.5 Location of SE O utbreaks Furthermore, we also explored the trend of foods - based outbreaks and the various locations where these outbreaks occurred. It was found that of 1144 outbreaks, there were 1038 outbreaks with known locations and 106 outbreaks were without unknown locations. Overall, the highest number of outbreaks 645 (56%) occurred at resta urant s or hotel s followed by private home residence s 190 (17%), institutions 84 (7%) and nursing facilities or hospital s 65 (6%). It was found that most of SE outbreaks occurred at restaurant and private residence in case of all 26 food categories . This is c onsistent with previous findings , for example , a recent study showed that restaurants with sit - down places were reported highly associated with food borne outbreaks 373 (48%) in 2015. It shows that restaurants are still the common places associated with fo od borne outbreaks in the country (12) . 4.6 Statistical Analyses N egative binomial and Poisson regression models were applied to assess the impendent and j oint association of food vehicles in SE outbreaks . We analyzed counts of SE outbreaks per year, classified by different food categories from 1990 to 2015 . In the final model, food categories were added as categorical variables and year as a continuous vari abl e to estimate the expected count of SE outbreaks associated with each food category. Table 2 : Food Vehicles Implicated in Salmonella Enteritidis Outbreaks in USA Food Vehicles Estimates Lower 95% CI Upper 95% CI P - value Bakery 1.77 0.81 3.85 0.1493 Chicken 6.02 4.42 8.20 <.0001 Dairy 3.85 2.72 5.44 <.0001 Meat 7.34 5.55 9.71 <.0001 Vegetable s 5.71 4.22 7.72 <.0001 14.40 11.38 18.23 <.0001 Estimate of expected count of SE of outbreaks by food vehicle s based on Negative binomial regression with food vehicle categories from 1990 to 2015 . Eggs - based dishes were considered a referent category in the model. The estimated count of SE outbreaks associated with eggs - based dishes per year is 14 .4 outbreaks in the USA w hen all vehicle categories are held constant. After eggs, the expected numbers of SE outbreaks associated with meat products per year in the USA will be around 7 .3 (Table 2) . Similarly, expected count of SE outbreaks associated with meat vegetables per yea r in the US will be around 5.7 and e xpected count of SE outbreaks associated with chicken - based products will be 6 .0 for a year in the US A . However, all food vehicles were found statistically 27 significant source to cause considerable number of SE outbreaks in the country instead of only bakery items ( Table 2 , Table 5 ) . In addition, we calculated r elative occurrence (rate ratio) of different food categories compared to eggs - based outbreaks based on the same above mentioned Negative binomial model. Table 3 : Relative Occurrence of Food Vehicles Compared to Eggs - based Salmonella Enteritidis Outbreaks Food Vehicles Estimates (Rate Ratio) Lower 95% CI Upper 95% CI P - value Bakery vs. Eggs 0.12 0.06 0.27 <.0001 Chicken vs. Eggs 0.42 0.30 0. 58 <.0001 Dairy vs. Eggs 0.27 0.18 0.39 <.0001 Meat vs. Eggs 0.51 0.38 0.69 <.0001 Vegetable vs. Eggs 0.40 0.29 0.55 <.0001 Th is table shows that overall, occurrence of meat products associated outbreaks is higher than chicken associated outbreaks com pared to eggs. Similarly, vegetables and chicken associated outbreaks showed almost equal estimates when both were compare individually with eggs - based outbreaks. It clea rly shows that vegetables and meat products are critical source of causing SE outbreak s in the country similar as eggs and chicken. In addition, dairy associated products showed significant estimate compared to eggs - based outbreaks. Overall, these finding prove our hypothesis that other food sources are also playing crucial role in occurren ce of SE disease along with contaminated eggs and chicken products in the country. 28 CHAPTER 5 : DISCUSSION 5.1 Interpretations In this study, all SE outbreaks , from 1990 to 2015, with available information on CDC website were included to analyze th e dis tribution and relative occurrence of different food vehicles associated with SE outbreaks in the USA. By considering different dominant food vehicles in this study, it gave us an idea that how could other foods be a potential source of spreading the diseas e in the country besides eggs - associated outbreaks. In the analysis, it was found that , overall, eggs - based dishes 273 (24%) were dominantly reported to be associated with SE outbreaks in the country. I t was seen that highest number of eggs - based outbreak s were reported in following years, 1990 - 1993 and 1997 (Figure 1) , and a fter 2005, a significant decline was observed in eggs - based outbreaks in the country. It clearly shows the success of different interventions introduced by USDA and FDA to control eggs - associated outbreaks in the country. These findings are consistent with previously published studies on eggs - based outbreaks in the USA (4,21) . However, it is an important to know that in the current study, these food categories were made based on the level of disease risk associated with fo od items reported previously in the country and the world. vehicles associated with SE outbreaks together in one study. To address this, we devised a strategy of making classification of different food items. Ac cording to that, eggs - based dishes are highly associated with SE outbreaks followed by chicken, meat, vegetables, dairy products and bakery items. It means eggs - based dishes, complex category, can contain any other food item along with eggs implicated in o utbreaks during the study period. Similarly, chicken complex category can contain rest of food items i.e. meat, vegetables, dairy and bakery, but not eggs products. Therefore , this classification may overestimate the risk and numbers of outbreaks 29 associate d with higher order food items i.e. eggs, chicken relative to meat, vegetables and dairy. I t was found that e ven after putting meat and vegetables on low er order than eggs and chicken, results show ed that meat and vegetable products are more implicated in SE outbreaks than chicken items reported to CDC from 1990 to 2015 (Table1) . Though, food classification added bias to eggs - based and chicken food categories, but it clearly showed the importance of other food items with less bias ; meat, vegetables and dair y products in causing SE outbreaks those were not well - addressed in past studies. Healthy people goal was made to reduce the Salmonella infections up to 11.4 until 2020 when baseline infection rate was 15.0 per 100,000 population in 2006 - 2008 (47) . It was found that average isolation rate of S almonella serotypes was 14 cases per 100,000 population f rom 2004 to 2011 which was higher than healthy people goal 2010 (6.8 per 100,000) (48) . Current study also shows increasing and consistent trend of SE outbreaks and cases in the country. About negative binomial model, it was found that addition of other categorical variables such as sea son, census regions, along with food vehicles in model g ave an outcome with underestimated coefficients for each food category associated SE outbreaks. For the current study, we calculated expected count of outbreaks, and relative occurrence of different f ood vehicles compared to eggs - based SE outbreaks by using negative binomial model. However, the variable, year, was added in our model as a continuous variable to keep the number of outbreaks per year constant throughout the study period. 5.2 Strengths U nlike the past studies, i vehicles associated with SE outbreaks in the US since 1990 to 2015. This study has addressed the importance of food vehicles which are playing a critical role in endemicit y and persistent 30 occurrence of SE outbreaks in the country. We observed changes in occurrence of SE outbreaks at state level, census and HHS regions, which showed a variation among these regional classifications and state level reported outbreaks. It sugg ests that polic ies should be designed individually based on the local state. Moreover, it adds a concern that how preventive policies can be differed based on chosen classification of regions in the country. 5.3 Limitations Dataset has following few limi tations; first, all outbreaks are not reported and investigated. O utbreaks data alone is not good marker to study endemicity and prevalence of the disease. However, outbreaks data can provide a considerable overview of the disease occurrence and severity i especially when there is not enough information available about prevalence of sporadic cases associated with SE disease. Currently, FoodNet collects an incidence data of different reportable diseases f rom 10 different regional sites of the country by active surveillance which gives a good estimate of current represent the actual incidence of the disease in other states and country. Second, almost half of outbreaks were reported with unknown food source. It might reduce estimates of other reported food vehicles. However, we believe that unknown food category would represent kind of same proportion of the reported foods. Third, this study analysis based on the date when we access ed the data from CDC website. There is possibility of modification or deletion of outbreaks and illnesses reported to CDC. In this way, findings may vary from already reported in previous studies. Additionally, all cases reported with outbreaks cannot give precise estimates of actual cases in the country. 31 In addition, we also excluded multistate outbreaks from all of study analyses thinking that it might influence overall estimates . Additionally, for the sake of consisten cy in all analys e s, multistate outb reaks were excluded otherwise, it was unable to see changes across states, census regions and HHS regions. However, recent report showed that 30 multistate food borne outbreaks were reported in 2015 and of 30 , 17 outbreaks were caused by Salmonella serotyp es, and specifically 2 of 17 outbreaks were caused by SE (10) . 32 CHAPTER 6 : CONCLUSION In this study, it was aimed to investigate different food vehicles associated with SE outbreaks beside eggs in the USA. By exploring the distribution and trend of these food associated outbrea ks across seasons, census regions, HHS regions, and different time points, we discovered significant details about foods implicated in SE outbreaks from 1990 to 2015 in the country. Study results showed that food vehicles besides eggs: meat, chicken, veget ables and dairy products share significant role in causing and spreading SE disease outbreaks all over the country. Even though, eggs and chicken were put on higher order risk food categories, there were still high number of SE outbreaks associated with me at and vegetables products almost more than chicken items implicated outbreaks. In addition, there was a variation noticed in an order of foods implicated in SE outbreaks at state level which suggests designing state specific preventive measures and regula tions. Similarly, study analyses show varied distribution of food vehicles among different census regions and HHS regions. For spatial analysis, we were not able to run geographical weighted regression because the number of outbreaks per state were less th an 20. However, it would be worth knowing to design this kind of study at county level in high reporting states to further specify the hotspots of the disease within a particular state. Analysis of different locations implicated in outbreaks showed that pr ivate residence was second common place to be associated with SE outbreaks after restaurant. It suggests educating people about the storage , cooking and consumption of foods associated with SE disease. This study findings suggest that there is dire need t o make policies for shipping, selling and serving of these food items particularly meat, vegetables and chicken items at restaurants to stop consistent occurrence of SE outbreaks in the country. 33 APPENDI X 34 Table 4 : Description of all possible combinations of different foods implicated in Salmonella Enteritidis outbreaks in USA (1990 - 2015) . Vehicle Frequency Percent Cumulative Frequency Cumulative Percent Bakery items 8 0.70 8 0.70 Chicken 58 5.07 66 5.77 Chicken Baker 2 0.17 68 5.94 Chicken Dairy 1 0.09 69 6.03 Chicken Meat 6 0.52 75 6.56 Chicken Meat Vege * 1 0.09 76 6.64 Chicken Mix ** 4 0.35 80 6.99 Chicken Vegetables 23 2.01 103 9.00 Dairy 44 3.85 147 12.85 Dairy Bakery 8 0.70 155 13.55 Dairy Bakery Meat 3 0.26 158 13. 81 Eggs 93 8.13 251 21.94 Eggs Bakery 15 1.31 266 23.25 Eggs Chicken Mix 1 0.09 267 23.34 Eggs Chicken 5 0.44 272 23.78 Eggs Dairy 37 3.23 309 27.01 Eggs Dairy Bakery 1 0.09 310 27.10 Eggs Meat 18 1.57 328 28.67 Eggs Meat Mix 1 0.09 329 28.76 Eggs Meat Vege 1 0.09 330 28.85 Eggs Mix 59 5.16 389 34.00 Eggs Vege Bakery 1 0.09 390 34.09 Eggs Vege Dairy 17 1.49 407 35.58 Eggs Vegetables 24 2.10 431 37.67 Meat 75 6.56 506 44.23 Meat Product 7 0.61 513 44.84 Meat Bakery 1 0.09 514 44.93 35 Table 4 ( Meat Dairy 15 1.31 529 46.24 Meat Mix 5 0.44 534 46.68 Meat Vegetables 27 2.36 561 49.04 Unknown 487 42.57 1048 91.61 Vegetables 75 6.56 1123 98.16 Vegetables Bakery 8 0.70 1131 98.86 Vegetables Dairy 8 0.70 1139 99.56 Vegetables Mix 1 0.09 1140 99.65 Vegetables Mix 4 0.35 1144 100.00 *Vege: Vegetables **Mix: it means that the implicated food in outbreaks had other ingredients along with mentioned ones. Table 5 : Analysis of Maximum Likelihood Parameter Estimates Parameter DF Estimate Standard Error Wald 95% Confidence Limits Wald Chi - Square Pr ob Intercept 1 2.67 0.12 2.43 2.90 500.60 <.0001 yearnew 1 - 0.03 0.01 - 0.04 - 0.02 23.74 <.0001 Vehicle Bakery Items 1 - 2.10 0.40 - 2.88 - 1.32 27.72 <.0001 Vehicle Chic ken Items 1 - 0.87 0.17 - 1.20 - 0.55 27.70 <.0001 Vehicle Dairy Products 1 - 1.32 0.19 - 1.69 - 0.95 48.25 <.0001 Vehicle Meat Products 1 - 0.67 0.15 - 0.98 - 0.37 19.06 <.0001 Vehicle Unknown 1 0.66 0.14 0.39 0.92 23.60 <.0001 Vehicle Vegetables 1 - 0.93 0.16 - 1.25 - 0.60 31.84 <.0001 Dispersion 1 0.16 0.03 0.10 0.24 36 Figure 6 : D istribution of different food categories implicated in Salmonella Enteritidis outbreaks in USA (1990 - 2015) Figure 7 : Distributio n of foods associated with Salmonella Enteritidis outbreaks across five distinct groups of years in USA (1990 - 2015) 37 Figure 8 : D escri ption of Salmonella Enteritidis outbreaks in four c ensus r egions of USA : Northeast, Midwest, Sout h, and West. Figure 9 : Di stribution of Salmonella Enteritidis outbreaks across HHS regions which ha d over two outbreaks per year . 38 Figure 10 : D escri ption of Salmonella Enteritidis outbreaks occurr ing tr end in different seasons: Winter, Spring, Summer and Fall across the US census regions from 1990 - 2015. Figure 11 : The pattern of different seasons - based Salmonella Enteritidis outbreaks in the U SA HHS regions from 1990 - 2015. 39 Figure 12 : Number of outbreaks and case rate associated with Salmonella Enteritidis outbreaks in five distinct group of years in USA 40 Figure 12 ( c ont ) 41 Figure 12 ( c ont ) 42 RE FERENCES 43 REFERENCES 1. Majowicz SE, Musto J, Scallan E, et al. The Global Burden of Nontyphoidal Salmonella Gastroenteritis. Clin. Infect. Dis. [electronic article]. 2010;50(6):882 889. (https: //academic.oup.com/cid/article - lookup/doi/10.1086/650733). (Accessed February 17, 2017) 2. Scallan E, Hoekstra RM, Angulo FJ, et al. Foodborne illness acquired in the United States - Major pathogens. Emerg. Infect. Dis. [electronic article]. 2011;17(1):7 15 . (http://www.ncbi.nlm.nih.gov/pubmed/21192848). (Accessed August 17, 2017) 3. Hendriksen RS, Vieira AR, Karlsmose S, et al. Global Monitoring of Salmonella Serovar Distribution from the World Health Organization Global Foodborne Infections Network Countr y Data Bank: Results of Quality Assured Laboratories from 2001 to 2007. Foodborne Pathog. Dis. [electronic article]. 2011;8(8):887 900. (http://www.liebertonline.com/doi/abs/10.1089/fpd.2010.0787). (Accessed August 17, 2017) 4. Wright AP, Richardson L, Ma hon BE, et al. The rise and decline in Salmonella enterica serovar Enteritidis outbreaks attributed to egg - containing foods in the United States, 1973 - 2009. Epidemiol. Infect. [electronic article]. 2016;144(4):810 819. (http://www.journals.cambridge.org/ab stract_S0950268815001867). (Accessed January 23, 2018) 5. Braden CR. Salmonella enterica Serotype Enteritidis and Eggs: A National Epidemic in the United States. Clin. Infect. Dis. [electronic article]. 2006;43(4):512 517. (http://www.ncbi.nlm.nih.gov/pub med/16838242). (Accessed August 17, 2017) 6. Boore AL, Hoekstra RM, Iwamoto M, et al. Salmonella enterica infections in the United States and assessment of coefficients of variation: A Novel approach to identify epidemiologic characteristics of individual serotypes, 1996 - 2011. PLoS One [electronic article]. 2015;10(12):e0145416. (http://dx.plos.org/10.1371/journal.pone.0145416). (Accessed April 28, 2017) 7. Langridge GC, Fookes M, Connor TR, et al. Patterns of genome evolution that have accompanied host a daptation in Salmonella . Proc. Natl. Acad. Sci. [electronic article]. 2015;112(3):863 868. (http://www.pnas.org/lookup/doi/10.1073/pnas.1416707112) 8. Frenzen PD, Riggs TL, Buzby JC, et al. Salmonella Cost Estimate Updated Using FoodNet Data. Clin. Infect . Dis. [electronic article]. 2004;Volume 382(2):10 15. (https://pdfs.semanticscholar.org/832a/6e678f4bbce474520917d542503e73119175.pdf) 9. Scallan E, Majowicz SE, Hall G, et al. Prevalence of diarrhoea in the community in Australia, Canada, Ireland, and t he United States. Int. J. Epidemiol. [electronic article]. 2005;34(2):454 460. (http://academic.oup.com/ije/article/34/2/454/747093/Prevalence - of - diarrhoea - in - the - community - in). (Accessed January 23, 2018) 10. Scallan E, Hoekstra RM, Angulo FJ, et al. Foo dborne illness acquired in the United States - Major pathogens. Emerg. Infect. Dis. [electronic article]. 2011;17(1):7 15. 44 (http://wwwnc.cdc.gov/eid/article/17/1/P1 - 1101_article.htm). (Accessed January 23, 2018) 11. Jackson BR, Griffin PM, Cole D, et al. Ou tbreak - associated salmonella enterica serotypes and food commodities, united states, 1998 - 2008. Emerg. Infect. Dis. [electronic article]. 2013;19(8):1239 1244. (http://www.ncbi.nlm.nih.gov/pubmed/23876503). (Accessed April 12, 2018) 12. Dewey - Mattia D, Ma nikonda K, Vieira A, et al. Surveillance for foodborne disease outbreaks United States, 2015: Annual Report. 2017;1 24. 13. Louis MES, Morse DL, Potter ME, et al. The Emergence of Grade A Eggs as a Major Source of Salmonella enteritidis Infections: New I mplications for the Control of Salmonellosis. JAMA J. Am. Med. Assoc. [electronic article]. 1988;259(14):2103 2107. (http://jama.jamanetwork.com/article.aspx?doi=10.1001/jama.1988.03720140023028). (Accessed January 27, 2018) 14. Louis ME St. The Emergence of Grade A Eggs as a Major Source of Salmonella enteritidis Infections. JAMA [electronic article]. 1988;259(14):2103. (http://jama.jamanetwork.com/article.aspx?doi=10.1001/jama.1988.03720140023028). (Accessed July 29, 2017) 15. Hogue A, White P, Guard - Pe tter J, et al. Epidemiology and control of egg - associated Salmonella Enteritidis in the United States of America. Rev. Sci. Tech. L Off. Int. Des Epizoot. [electronic article]. 1997;16(2):542 553. (https://www.oie.int/doc/ged/D9161.PDF). (Accessed January 23, 2018) 16. Patrick ME, Adcock PM, Gomez TM, et al. Salmonella Enteritidis Infections, United States, 1985 1999. Emerg. Infect. Dis. [electronic article]. 2004;10(1):1 7. (http://wwwnc.cdc.gov/eid/article/10/1/02 - 0572_article.htm). (Accessed January 23, 2018) 17. Altekruse S, Koehler J, Hickman - Brenner F, et al. A comparison of Salmonella enteritidis phage types from egg - associated outbreaks and implicated laying flocks. Epidemiol. Infect. [electronic article]. 1993;110(1):17 22. (https://www.ncbi.nlm.n ih.gov/pmc/articles/PMC2271964/pdf/epidinfect00037 - 0026.pdf). (Accessed January 29, 2018) 18. Boyce TG, Koo D, Swerdlow DL, et al. Recurrent outbreaks of Salmonella Enteritidis infections in a Texas restaurant: phage type 4 arrives in the United States. E pidemiol. Infect. [electronic article]. 1996;117(1):29 34. (http://www.ncbi.nlm.nih.gov/pubmed/8760947). (Accessed January 29, 2018) 19. Passaro DJ, Reporter R, Mascola L, et al. Epidemic Salmonella enteritidis infection in Los Angeles County, California. The predominance of phage type 4. West. J. Med. [electronic article]. 1996;165(3):126 30. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1303718/pdf/westjmed00348 - 0028.pdf). (Accessed January 27, 2018) 20. Ban Mishu, Jane Koehler, Lisa A. Lee, Daniel Rodr igue, France Hickman Brenner PB and RVT. Outbreaks of Salmonella enteritidis Infections in the United States, 1985 - 1991. JThe 45 J. Infect. Dis. [electronic article]. 1994;Vol. 169(3):547 552. (http://www.jstor.org/stable/30113775). (Accessed January 27, 2018 ) 21. Patrick ME, Adcock PM, Gomez TM, et al. Salmonella Enteritidis Infections, United States, 1985 - 1999. Emerg. Infect. Dis. 2004;10(1):1 7. 22. United States Department of Agriculture/Animal and Plant Health Inspection Service. Chickens affected by S almonella Enteritidis. Final Rule, 9 CFR Parts 71 and 82. Federal Register,. 1991. 23. Widte PL, Schlosser W, Benson CE, et al. Environmental Survey by Manure Drag Sampling for Salmonella enteritidis in Chicken Layer Houses. J. Food Prot. [electronic arti cle]. 1997;60(10):1189 1193. (http://jfoodprotection.org/doi/pdf/10.4315/0362 - 028X - 60.10.1189). (Accessed January 27, 2018) 24. Associated Salmonella Enteritidis Infections in Wisconsin: The Role of Eggs. J. Infect. Dis. [electronic article]. 1999;180(4):1214 1219. (https://academic.oup.com/jid/article - lookup/doi/10.1086/314984). (Accessed January 29, 2018) 25. United States Department of Agriculture Food Safety Inspection Ser vice. Salmonella Enteritidis risk assessment: shell eggs and egg products. Washingt. U.S. Gov. Print. Off. [electronic article]. 1998;(http://www.adiveter.com/ftp_public/articulo354.pdf). (Accessed January 29, 2018) 26. United States Department of Agricul ture Food Safety Inspection Service. Refrigeration and labeling requirements for shell eggs [Final rule]. Fed Register. 1998 63:45663 - 75 p. 27. U.S. General Accounting Office. Food safety: U.S. lacks a consistent farm - to - table approach to egg safety. Wash ington: 1999 Report no. GAO/RCED - 99 - 184. p. 28. Food labeling, safe handling statements, labeling of shell eggs; refrigeration of shell eggs held for retail distribution. Food and Drug Administration, HHS. Final rule. Fed. Regist. [electronic article]. 20 00;65(234):76092 76114. (http://www.ncbi.nlm.nih.gov/pubmed/11503723). (Accessed January 29, 2018) 29. action plan to eliminate Salmonella Enteritidis illnesses due to eggs. Washington: 1999. 30. Kimura AC, Reddy V, Marcus R, et al. Chicken Consumption Is a Newly Identified Risk Factor for Sporadic Salmonella enterica Se rotype Enteritidis Infections in the United Clin. Infect. Dis. [electronic article]. 2004;38(s3):S244 S252. (https://academic.oup.com/cid/article - lookup/doi/10.1086/381576). (Accessed January 29, 2018) 31. Un ited States Department of Agriculture FSIS. Salmonella serotypes isolated from raw meat and poultry January 26, 1998, to January 25, 1999. Washington: 1999. 32. Machado J, Bernardo F. Prevalence of Salmonella in chicken carcasses in Portugal. J. Appl. Bac teriol. [electronic article]. 1990;69(4):477 480. (http://www.ncbi.nlm.nih.gov/pubmed/2292513). (Accessed January 29, 2018) 46 33. Geilhausen B, Schütt - Gerowitt H, Aleksic S, et al. Campylobacter and Salmonella contaminating fresh chicken meat. Zentralbl. Ba kteriol. [electronic article]. 1996;284(2 3):241 5. (http://www.ncbi.nlm.nih.gov/pubmed/8837384). (Accessed January 29, 2018) 34. ALALI WQ, GAYDASHOV R, PETROVA E, et al. Prevalence of Salmonella on Retail Chicken Meat in Russian Federation. J. Food Prot. [electronic article]. 2012;75(8):1469 1473. (http://jfoodprotection.org/doi/pdf/10.4315/0362 - 028X.JFP - 12 - 080?code=fopr - site). (Accessed January 29, 2018) 35. YANG B, XI M, WANG X, et al. Prevalence of Salmonella on Raw Poultry at Retail Markets in China. J. Food Prot. [electronic article]. 2011;74(10):1724 1728. (http://jfoodprotection.org/doi/abs/10.4315/0362 - 028X.JFP - 11 - 215) 36. Roberts D. Salmonella in chilled and frozen chicken. Lancet (London, England) [electronic article]. 1991;337(8747):984 5. (ht tp://www.ncbi.nlm.nih.gov/pubmed/1678071). (Accessed January 29, 2018) 37. Gast RK, Benson ST. The Comparative Virulence for Chicks of Salmonella enteritidis Phage Type 4 Isolates and Isolates of Phage Types Commonly Found in Poultry in the United States. Avian Dis. [electronic article]. 1995;39(39):567 574. (http://www.jstor.org/stable/1591810%0Ahttp://www.jstor.org/stable/1591810%0Ahttp:// www.jstor.org/stable/1591810) 38. Ahlstrom C, Muellner P, Spencer SEF, et al. Inferring source attribution from a mu ltiyear multisource data set of Salmonella in Minnesota. Zoonoses Public Health [electronic article]. 2017;64(8):589 598. (http://doi.wiley.com/10.1111/zph.12351). (Accessed January 23, 2018) 39. Quiroz - Santiago C, Rodas - Suárez OR, Carlos R V, et al. Prev alence of Salmonella in vegetables from Mexico. J. Food Prot. [electronic article]. 2009;72(6):1279 82. (http://www.ncbi.nlm.nih.gov/pubmed/19610340). (Accessed April 10, 2018) 40. Nillian E, Ching CL, Fung PC, et al. Simultaneous Detection of Salmonella spp., Salmonella Enteritidis and Salmonella Typhimurium in Raw Salad Vegetables and Vegetarian Burger Patties. Food Nutr. Sci. [electronic article]. 2011;2:1077 1081. (http://www.scirp.org/journal/fns). (Accessed April 10, 2018) 41. HERMAN KM, HALL AJ, GO ULD LH. Outbreaks attributed to fresh leafy vegetables, United States, 1973 2012. Epidemiol. Infect. [electronic article]. 2015;143(14):3011 3021. (http://www.journals.cambridge.org/abstract_S0950268815000047) 42. Kakiomenou K, Tassou C, Nychas GJ. Surviv al of salmonella enteritidis and listeria monocytogenes on salad vegetables. World J. Microbiol. Biotechnol. [electronic article]. 1998;14(3):383 387. (https://link.springer.com/content/pdf/10.1023%2FA%3A1008817312330.pdf). (Accessed April 10, 2018) 43. B hunia AK. Foodborne microbial pathogens: Mechanisms and pathogenesis. New York, NY: Springer New York; 2008 (Accessed April 11, 2018) 1 - 276 p.(http://link.springer.com/10.1007/978 - 0 - 387 - 74537 - 4). (Accessed April 11, 2018) 47 44. Bean NH, Goulding JS, Lao C, et al. Surveillance for foodborne - disease outbreaks -- United States, 1988 - 1992. MMWR. CDC Surveill. Summ. [electronic article]. 1996;45(5):1 66. (http://www.ncbi.nlm.nih.gov/pubmed/8890258). (Accessed February 10, 2018) 45. CDC. Outbreaks Involving Salmone lla | CDC. CDC . 2016;(https://www.cdc.gov/salmonella/outbreaks.html). (Accessed October 27, 2017) 46. 2015;(https://www.statista.com/chart/3616/where - food - preparation - lead s - to - food - poisoning/). (Accessed October 18, 2017) 47. Office of Disease Prevention and Health Promotion UD of H and HS. Food Safety | Healthy People 2020. (https://www.healthypeople.gov/2020/topics - objectives/topic/food - safety/objectives). (Accessed May 2, 2018) 48. Boore AL, Hoekstra RM, Iwamoto M, et al. Salmonella enterica infections in the United States and assessment of coefficients of variation: A Novel approach to identify epidemiologic characteristics of individual serotypes, 1996 - 2011. PLoS One [electronic article]. 2015;10(12):e0145416. (http://dx.plos.org/10.1371/journal.pone.0145416). (Accessed January 23, 2018)