challenges in data needs for assessment of food product risk and attribution of foodborne illnesses...

Challenges in data needs for Challenges in data needs for assessment of food product risk and assessment of food product risk and attribution of foodborne illnesses to attribution of foodborne illnesses to food products in the United Statesfood products in the United States

Chuanfa Guo, Carl Schroeder, and Janell KauseChuanfa Guo, Carl Schroeder, and Janell Kause

Office of Public Health ScienceOffice of Public Health ScienceFood Safety and Inspection ServiceFood Safety and Inspection Service

United States Department of AgricultureUnited States Department of Agriculture

Fourth International Conference on Agriculture StatisticsFourth International Conference on Agriculture Statistics

October 22-24, 2007October 22-24, 2007

Beijing, ChinaBeijing, China

Better information. Better decisions. Risk assessment.

22

Food Safety and Food Safety and AttributionAttribution

Food safety risk assessment requires a Food safety risk assessment requires a farm-to-table approachfarm-to-table approachFarm Farm Processing Processing Retail Retail Consumer Consumer

The same concept applies to attribution The same concept applies to attribution of foodborne illnesses to food productsof foodborne illnesses to food products Where does contamination occur?Where does contamination occur? How does contamination occur?How does contamination occur? How can it be prevented?How can it be prevented?


33

Attribution of Foodborne Attribution of Foodborne Illnesses to Food ProductsIllnesses to Food Products


44

FoodNet Attribution FoodNet Attribution ActivitiesActivities

Foodborne Diseases Active Surveillance Foodborne Diseases Active Surveillance Network (FoodNet) attribution working Network (FoodNet) attribution working group and modeling subgroupgroup and modeling subgroup

Centers for Disease Control and Centers for Disease Control and Prevention (CDC)Prevention (CDC)

Food Safety and Inspection Service Food Safety and Inspection Service (FSIS)(FSIS)

Food and Drug Administration (FDA)Food and Drug Administration (FDA) State health departmentsState health departments


55

Use of Expert Elicitation Use of Expert Elicitation for Food Safety and for Food Safety and

AttributionAttributionFSIS and RTI International conducted FSIS and RTI International conducted expert elicitations in 2005 and 2007expert elicitations in 2005 and 2007

Rank the public health risks posed by Rank the public health risks posed by bacterial hazards in processed meat and bacterial hazards in processed meat and poultry productspoultry products

Attribution of foodborne illnesses to Attribution of foodborne illnesses to specific pathogens as a result of specific pathogens as a result of consuming or handling processed meat consuming or handling processed meat and poultry productsand poultry products


66

SalmonellaSalmonella Attribution Attribution ModelModel

Hald Hald et al.et al. A Bayesian approach to quantify the A Bayesian approach to quantify the contribution of animal-food sources to human contribution of animal-food sources to human salmonellosis. salmonellosis. Risk Anal.Risk Anal. 2004. 24(1):255-69. 2004. 24(1):255-69.

Microbial subtyping provides link between public Microbial subtyping provides link between public health endpoint and source of infection.health endpoint and source of infection.

Bayesian framework uses Markov Chain Monte Bayesian framework uses Markov Chain Monte Carlo simulation to estimate number of human Carlo simulation to estimate number of human salmonellosis.salmonellosis.

The approach quantifies the contribution of each of The approach quantifies the contribution of each of the major animal-food sources to human the major animal-food sources to human salmonellosis. salmonellosis.


77

Adaptation of the Adaptation of the SalmonellaSalmonella Attribution Model to U.S. DataAttribution Model to U.S. Data

Joint Joint effort by FSIS, CDC, FDA, and state effort by FSIS, CDC, FDA, and state partners under the FoodNet Attribution partners under the FoodNet Attribution Working Group and Modeling Subgroup. Working Group and Modeling Subgroup. Objectives:Objectives:

Estimate the number of cases of human Estimate the number of cases of human salmonellosis attributable to various food salmonellosis attributable to various food sourcessources

Support risk managers and regulators when Support risk managers and regulators when deciding how to allocate resourcesdeciding how to allocate resources

Identify data needs and gaps for future Identify data needs and gaps for future attribution studiesattribution studies


88

SalmonellaSalmonella Attribution Model Attribution Model ParametersParameters

SalmonellaSalmonella prevalence prevalence by serotype in a food by serotype in a food source (p)source (p)

AAmount of particular food consumed (M)mount of particular food consumed (M) Food source dependent factor (a)Food source dependent factor (a) Serotype Serotype dependent factor (q)dependent factor (q)


Expected number of salmonellosis cases ()

99

Attribution Data SourcesAttribution Data Sources Human salmonellosis cases, by serotypeHuman salmonellosis cases, by serotype

Public Health Laboratory Information System Public Health Laboratory Information System (PHLIS), 1998-2003 (PHLIS), 1998-2003

Foods - Foods - SalmonellaSalmonella prevalence, by serotype prevalence, by serotype Beef, ground beef, chicken, turkey, pork, and Beef, ground beef, chicken, turkey, pork, and

processed egg products, FSIS in-plant samples, processed egg products, FSIS in-plant samples, 1998-20031998-2003

Shell eggs, Shell eggs, Pennsylvania SE Pilot Project, 1993-Pennsylvania SE Pilot Project, 1993-19951995

Food consumption dataFood consumption data USDA/Economic Research Service, 1998-2003USDA/Economic Research Service, 1998-2003

Outbreak and travel informationOutbreak and travel information Salmonellosis cases reported to FoodNet, 2004Salmonellosis cases reported to FoodNet, 2004


1010

Preliminary Model ResultsPreliminary Model ResultsEstimated Percentage Distributions of Estimated Percentage Distributions of

Human Human Salmonellosis Cases, 1998-2003Salmonellosis Cases, 1998-2003


* Shell egg data from the Pennsylvania Pilot Project, 1993-1995

1111

Data Gaps and Limitations of Data Gaps and Limitations of SalmonellaSalmonella Attribution Model Attribution Model

The stochastic model does not attribute all The stochastic model does not attribute all observed salmonellosis cases to food observed salmonellosis cases to food sources.sources.

Model does not address other foodborne Model does not address other foodborne sources (produce, dairy, etc) of sources (produce, dairy, etc) of SalmonellaSalmonella..

Model does not attribute any salmonellosis Model does not attribute any salmonellosis cases to non-food source, environmental cases to non-food source, environmental exposures, pets, farm animals, water, etc.exposures, pets, farm animals, water, etc.

Egg data are limited (1993-1995 data Egg data are limited (1993-1995 data available)available)


1212

Challenges in Data Needs For Challenges in Data Needs For Food AttributionFood Attribution

To obtain data concerning pathogen prevalence and To obtain data concerning pathogen prevalence and distribution in a wide variety of potential food distribution in a wide variety of potential food vehicles and other for other important sources of vehicles and other for other important sources of human exposure, such as indirect sources of human exposure, such as indirect sources of contamination and non-food sourcescontamination and non-food sources

To ensure that existing data sources continue to To ensure that existing data sources continue to adequately represent the burden of foodborne adequately represent the burden of foodborne illnesses in the U.S. population and the distribution illnesses in the U.S. population and the distribution of the associated pathogen in food vehicles and of the associated pathogen in food vehicles and exposure sources of interestexposure sources of interest

To refine existing data so that the comparisons To refine existing data so that the comparisons between data from various sources are based on between data from various sources are based on similar units of observation at the necessary levels of similar units of observation at the necessary levels of discrimination for defined points along the farm-to-discrimination for defined points along the farm-to-table continuum table continuum


1313

Results of 2007 Expert Results of 2007 Expert ElicitationElicitation

Rank the public health risks posed by bacterial Rank the public health risks posed by bacterial hazards in each of 25 categories of processed hazards in each of 25 categories of processed meat and poultry productsmeat and poultry products

Score of 1 to 10 for likelihood of illness from Score of 1 to 10 for likelihood of illness from consuming or handling meat and poultry consuming or handling meat and poultry products among healthy adults and vulnerable products among healthy adults and vulnerable consumersconsumers 1 – least likelihood1 – least likelihood 10 – greatest likelihood10 – greatest likelihood

Attribute foodborne illnesses of specific Attribute foodborne illnesses of specific pathogens to consuming or handling pathogens to consuming or handling processed meat and poultry productsprocessed meat and poultry products


1414

Likelihood of Illness Among Likelihood of Illness Among Healthy AdultsHealthy Adults

Finished Product TypeFinished Product Type

MediaMedian n

ScoreScore(1-10)(1-10)

Level of Level of ConfidenConfiden

cece(1-3)(1-3)

Raw ground or otherwise non-intact Raw ground or otherwise non-intact chickenchicken 1010 2.62.6

Raw ground or otherwise non-intact Raw ground or otherwise non-intact turkeyturkey 99 2.32.3

Raw ground or otherwise non-intact Raw ground or otherwise non-intact poultry – no chicken or turkeypoultry – no chicken or turkey 8.58.5 1.81.8

Raw intact chickenRaw intact chicken 88 2.62.6Raw intact turkeyRaw intact turkey 88 2.52.5Raw intact poultry – other than Raw intact poultry – other than chicken or turkeychicken or turkey 88 1.91.9

Raw ground or otherwise non-intact Raw ground or otherwise non-intact beefbeef 88 2.52.5

1515

Likelihood of Illness Among Vulnerable Likelihood of Illness Among Vulnerable ConsumersConsumers

Finished Product TypeFinished Product TypeMedian Median ScoreScore(1-10)(1-10)

Level of Level of ConfidenConfiden

cece(1-3)(1-3)

Raw ground or otherwise non-intact Raw ground or otherwise non-intact chickenchicken 1010 2.62.6

Raw ground or otherwise non-intact Raw ground or otherwise non-intact beefbeef 9.59.5 2.52.5

Raw ground or otherwise non-intact Raw ground or otherwise non-intact turkeyturkey 99 2.52.5

Raw ground or otherwise non-intact Raw ground or otherwise non-intact poultry – no chicken or turkeypoultry – no chicken or turkey 99 2.02.0

Raw intact chickenRaw intact chicken 8.58.5 2.62.6Raw intact turkeyRaw intact turkey 88 2.62.6Raw intact poultry – other than Raw intact poultry – other than chicken or turkeychicken or turkey 88 2.12.1

1616

Attribution of Foodborne Illness of Attribution of Foodborne Illness of Salmonella (Non-Typhi) to Meat Salmonella (Non-Typhi) to Meat

and Poultry Productsand Poultry Products


22%

14%

8%9%7%

6%

34%

Raw intact chicken

Raw intact turkey

Raw ground or otherwisenon-intact beef

Raw ground or otherwisenon-intact chicken

Raw ground or otherwisenon-intact turkey

Raw otherwise processedpoultry

Other 19 productcategories

1717

Attribution of Foodborne Illness of Attribution of Foodborne Illness of SalmonellaSalmonella (Multidrug Resistant) to (Multidrug Resistant) to

Meat and Poultry ProductsMeat and Poultry Products


19%

8%

20%

6%8%

7%

32%

Raw intact chicken

Raw intact turkey

Raw ground or otherwisenon-intact beef

Raw ground or otherwisenon-intact pork

Raw ground or otherwisenon-intact chicken

Raw ground or otherwisenon-intact turkey

Other 19 productcategories

1818

Closing RemarksClosing Remarks

Assessment of food product safety and Assessment of food product safety and attribution of foodborne illnesses require attribution of foodborne illnesses require extensive data originating from various sources extensive data originating from various sources

Available data sources often suffer from Available data sources often suffer from methodological limitations and the unavailability methodological limitations and the unavailability of certain types of data often result in critical of certain types of data often result in critical data gapsdata gaps

Expert elicitation is useful when epidemiologic Expert elicitation is useful when epidemiologic data are lacking, are sparse, or are highly data are lacking, are sparse, or are highly uncertain to fill the critical gaps in food safety uncertain to fill the critical gaps in food safety studiesstudies

Ensure and maximize the quality, objectivity, Ensure and maximize the quality, objectivity, utility, and integrity of the datautility, and integrity of the data


1919

AcknowledgementsAcknowledgements

FoodNet Attribution Modeling SubgroupFoodNet Attribution Modeling Subgroup CDC: Fred Angulo, Mike Hoekstra, Elaine Scallan, CDC: Fred Angulo, Mike Hoekstra, Elaine Scallan,

Xin Tong Xin Tong FSIS: Carl Schroeder, Chuanfa Guo, Liane Ong, FSIS: Carl Schroeder, Chuanfa Guo, Liane Ong,

Kristin Holt, Patty Bennett, Bonnie Kissler, Kristin Holt, Patty Bennett, Bonnie Kissler, Evelyne Mbandi, Reza Roodsari, Jane Harman, Evelyne Mbandi, Reza Roodsari, Jane Harman, Alecia Naugle, Bonnie RoseAlecia Naugle, Bonnie Rose

Oregon Health Division : Paul CieslakOregon Health Division : Paul Cieslak Georgia Division of Public Health: Dana ColeGeorgia Division of Public Health: Dana Cole Decisionalysis Risk Consultants, Inc.: Decisionalysis Risk Consultants, Inc.: Emma Emma

HartnettHartnett


Expert Elicitation ActivitiesExpert Elicitation Activities RTI International: Shawn Karns, Mary Muth, Michaela RTI International: Shawn Karns, Mary Muth, Michaela Coglaiti Coglaiti FSIS: Janell Kause, Chuanfa Guo, Matthew Michael, Cynthia FSIS: Janell Kause, Chuanfa Guo, Matthew Michael, Cynthia Williams, Don AndersonWilliams, Don Anderson

2020

THANK YOUTHANK YOU


challenges in data needs for assessment of food product risk and attribution of foodborne illnesses...

Documents