Burden of Foodborne Illness: Questions and Answers

The paper on unspecified agents says that 38.4 million gastrointestinal illnesses are due to agents not yet described, unrecognized foodborne agents, and noninfectious agents. How does CDC know that these unspecified agents are foodborne? Can’t acute gastroenteritis be caused by conditions or vehicles other than foodborne transmission?

• Gastroenteritis can be caused by agents that are transmitted through food, water, direct contact with animals, or person-to-person contact.
• Based on the data for known pathogens, we estimated the proportion of acute gastroenteritis illnesses that was caused by foodborne transmission.
• We applied that proportion to the number of all gastrointestinal illnesses from unspecified agents to estimate the number of illnesses from unspecified agents that were foodborne.

It is important to note, however, that using this method assumes that the foodborne proportion is the same for illnesses from known agents and from unspecified agents. We believe that this is a reasonable assumption, given the limited information available, but it is certainly possible that the foodborne proportion differs for known and unspecified agents.

How can I get more information about the methods used to create the estimates?

Estimating US foodborne illnesses for 31 known foodborne pathogens

For each pathogen, we gathered data from surveillance systems and corrected for underreporting and under-diagnosis. We then multiplied the adjusted number by the proportion of illnesses that was acquired in the United States (that is, not during international travel) and the proportion transmitted by food to yield an estimated number of illnesses that are domestically acquired and foodborne. Then, we added the estimates for each of the pathogens to arrive at a total, and used an uncertainty model to generate a point estimate and 90% credible interval (upper and lower limits).

Estimating US foodborne illnesses for unspecified agents

Unspecified agents fall into four general categories:

• Agents with insufficient data to estimate agent-specific burden
• Known agents not yet recognized as causing foodborne illness
• Microbes, chemicals, or other substances known to be in food whose pathogenicity is unproven
• Agents not yet described

To estimate foodborne illnesses from unspecified agents, we used symptom-based data from surveys to estimate the total number of AGI and then subtracted the number of illnesses accounted for by known gastroenteritis pathogens. We then multiplied this number by the proportion of domestically acquired illnesses and of illnesses attributable to food, just as we did for the known agents. Finally, again as with the known-pathogens estimate, we used an uncertainty model to generate a point estimate and 90% credible interval (upper and lower limits).

Foodborne illnesses caused by chemicals that cause acute gastroenteritis are included in the estimate of illnesses due to unspecified agents. However, chemicals or unspecified agents that do not cause acute gastroenteritis are not included in the estimates.

Learn more about the methods used >

The paper says that comparing estimates from 1999 with estimates from 2011 would be meaningless because better data sources and methods were used for the new estimates. Were the 1999 estimates wrong? How can you be sure that the new estimates accurately reflect the burden of foodborne illness if they are so different from those in 1999?

The 1999 Mead et al. estimates were based on the best data available and used accepted methodologies of the time, so in 1999 they were the best possible estimates. Those estimates provided important information about the relative frequency of various pathogens and filled a major gap in knowledge that helped in decision making. Public health science is not static. Every year new findings emerge, new methods are developed for research and analysis, new data are added to surveillance systems, and new surveys are conducted. The 2011 estimates are the result of innovations in methodology and improvements in the data made over the past decade, which we believe have resulted in more accurate estimates than their predecessors. We hope future estimates will benefit from additional innovations and improvements.

Learn more about improvements made in the 2011 estimates >

Why is it important to estimate domestically acquired foodborne illnesses?

Because many persons and organizations are involved in food safety in the United States—farmers, product manufacturers, distributors, retailers, restaurant workers, and consumers. We all want the most accurate measure of how well we’re doing in making food safe. Every time someone gets sick from eating food in the United States, we know we have more work to do. Whether we are involved in developing new or improved practices or policies in industry, or in encouraging best consumer practices, specific estimates about domestically acquired foodborne illnesses tell us that improvements are still needed.

Can you tease apart the proportion of the big drop in the estimates that’s due to improved methods from an actual decrease in illness? Put another way, was there any actual decrease in illness? If the answer to the second question is “yes,” to what can the actual reduction in illness be attributed?

Estimates of burden are designed to estimate the overall burden of disease at a point in time. They are not designed or intended to gauge increases or decreases in illnesses. Surveillance systems, such as Foodborne Diseases Active Surveillance Network (FoodNet), which capture data in a standardized way at regular points in time, are more appropriate for looking at disease trends.

If the new estimates do not really reflect a drop in the number of foodborne illnesses, how can you tell whether or not the government’s policies and programs for preventing these illnesses are working or if changes are needed?

To determine the effects of policies and programs, you need trend data—a way to tell how the incidence of disease has changed over time. The Foodborne Diseases Active Surveillance Network (FoodNet)—provides the best data for watching trends in common foodborne diseases. These data offer insight into whether policies and programs are having the intended effects.

• FoodNet is a surveillance system that collects information from sites in 10 states about diseases that are caused by nine organisms transmitted commonly through food.
• It is designed to identify everyone in the 10 sites who went to a health care provider, had a sample tested in a laboratory, and was diagnosed (laboratory-confirmed) with one of these infections.
• FoodNet gives us important information about which foodborne diseases are becoming less common and which need more concentrated attention.

How does our burden of foodborne illness compare with those of other countries?

Studies estimating the burden of foodborne illness have been done in Canada, England and Wales, and Australia.

• Like our study, the Canadian and Australian studies found that norovirus is responsible for a large portion of foodborne illnesses (31% and 30%, respectively).
• The study in England and Wales found a much lower percentage of norovirus-related illnesses than the US estimates, but when study samples were reexamined with a different technique, the rates were higher.
• Salmonella and Campylobacter were found to be leading causes of foodborne illnesses in the three foreign studies, as in the US study.
• The three foreign studies attributed a large burden of foodborne illness to unspecified agents and estimated a similar proportion of gastrointestinal illness to be transmitted by food.

However, it is important to note that each study used different methods, and therefore their specific estimates are not directly comparable.

If norovirus is the biggest cause of gastroenteritis illness, why isn’t it reportable?

Each year, the Council of State and Territorial Epidemiologists, with input from CDC, recommends diseases to include on the list of nationally notifiable conditions (formerly called nationally notifiable diseases).

• Considerations include the frequency, severity, and transmissibility of the disease as well as other considerations, such as whether ill persons are given a specific diagnosis when they receive health care.
• Based on this recommendation, states mandate which diseases must be reported to the state health department.
• Reporting to CDC is voluntary.
• The list changes periodically as new pathogens emerge or incidence of an illness drops over time.
• Norovirus is currently not on the list of nationally notifiable diseases. One of the main reasons is that clinical laboratories do not routinely test for norovirus, so ill persons are usually not specifically diagnosed with norovirus infection.
• Usually, norovirus is specifically diagnosed only in an outbreak situation.
• Beginning in 2009, all norovirus outbreaks have been reportable through outbreak surveillance, including those not transmitted by food; previously only food- and waterborne outbreaks were reported to CDC.

CDC conducts surveillance for norovirus outbreaks through the National Outbreak Reporting System (NORS). Through NORS, states are able to report all outbreaks of gastroenteritis, including those caused by foodborne, waterborne, and person-to-person transmission of norovirus directly to CDC. When states send specimens for testing or sequencing, norovirus outbreaks may also be reported to CDC’s National Calicivirus Laboratory, within the Gastroenteritis and Respiratory Viruses Laboratory Branch. (Calicivirus is the family name for the family of viruses that includes norovirus.)

In addition, CDC recently has developed a national surveillance network for norovirus genetic sequences (like DNA “fingerprints”) called CaliciNet, which is modeled after PulseNet, a surveillance network for enteric bacteria. CaliciNet is a national network of public health laboratories that contribute to a database of genetic fingerprints from noroviruses identified in outbreaks. As more states participate, CaliciNet may find links to help identify multistate outbreaks, detect potential norovirus-contaminated food before preparation and serving, and identify the emergence of new norovirus strains.

Other countries have done surveillance for norovirus illnesses; what are the barriers to doing so for the US population?

The UK, the Netherlands, and Australia conducted large community-based studies to estimate the fraction of acute gastroenteritis attributable to norovirus and other pathogens. These studies are very expensive to implement, costing millions of dollars. It would be costly and complex, but could be done here if we devoted the time and money to do it. It is a matter of priorities. In the United States, public health authorities think it is more efficient to focus on improving surveillance, including norovirus in outbreak reporting. However, CDC is engaged in numerous efforts to better assess the burden of norovirus using a variety of methods. When the burden of foodborne disease is estimated again, we will likely use norovirus data from these special studies, outbreaks, or other surveillance gathered here in the United States.

Norovirus can be transmitted through contact with contaminated objects in the environment, so why is it being included in the estimate of foodborne illness?

Although norovirus is, indeed, easily transmitted by touching contaminated objects in the environment as well as by direct person-to-person contact with an infected person, it is also transmitted commonly through food.

• Most often, food is contaminated by infected food handlers.
• Food may also be contaminated if it is grown in or irrigated with water contaminated with norovirus, as has been shown with oysters, lettuce, and raspberries.
• As we did when we estimated the number of foodborne illnesses for each of the other known pathogens, we multiplied the total estimated cases of norovirus-related acute gastroenteritis by the foodborne fraction for the pathogen.
• The foodborne fraction was derived from outbreak surveillance data, risk factor studies, and a literature review.
• This gave us an estimate of the proportion of norovirus infections that were transmitted through food.

What is the government doing to lower the burden of norovirus-related illnesses?

Food handlers are often implicated in outbreaks of norovirus illnesses.

• Many local and state health departments require that food handlers and preparers with gastroenteritis not work until 2 or 3 days after they feel better.
• In addition, strictly enforced rules requiring hand washing after using the bathroom and before handling food items is important in preventing the spread of this virus, even when food handlers do not have symptoms of gastroenteritis.
• Food handlers who were recently sick can often be given different duties in the restaurant so that they do not have to handle food (for example, working the cash register or serving as a host or hostess).
• The presence of kitchen managers trained in appropriate food hygiene practices may also reduce contamination during food preparation.
• The US Environmental Protection Agency has designated several effective disinfectants for killing norovirus. These cleaners should be used in food service, child care, health care, and other settings where transmission of norovirus is likely.

CDC is engaged in additional research to evaluate the specific foods and points in the food distribution systems at which norovirus contamination occurs, as well as the factors that contribute to contamination. These efforts may help identify control points at which interventions can be made to reduce norovirus contamination of food and the associated burden of illness.

There is also ongoing research towards development of a norovirus vaccine. Preliminary evidence shows that this may indeed be an effective strategy, although a great deal of work remains before a vaccine could be available.

The paper on known agents says that more accurate assessment of the foodborne norovirus illness rate is largely responsible for the big drop in the overall number of foodborne illnesses. Can you explain that in plain language?

• In the 1999 estimates, which were based on the best available data which had substantial limitations, the proportion of norovirus-related illnesses that was foodborne was estimated to be 40%.
• In the 2011 estimates, which were based on data from recently reported outbreaks and studies from other developed countries, the proportion of norovirus illnesses that was foodborne was estimated to be 26%.
• We believe that the 2011 estimate (26%) is a more accurate number than 40%. The 2011 estimate would probably have been more accurate in 1999, too, but it was not available then.
• Because norovirus causes millions of illnesses, the drop in the proportion of norovirus-related foodborne illnesses resulted in a sizeable drop in the estimated proportion of foodborne illnesses from all known foodborne pathogens—from 36% in 1999 to 25% in the 2011 study.
• The new, smaller proportion was also used to estimate the foodborne proportion of illnesses from unspecified agents, thus yielding a different overall estimate of foodborne illnesses than in 1999.

The paper says that the estimates help direct food policy. Specifically, what decisions are made based on the burden estimates?

The new estimates identify and rank the most important known pathogens responsible for causing foodborne illness, hospitalizations, and deaths. They provide a foundation for priority setting for interventions, policy development, research and analyses, and advocacy. Regulatory agencies can use these data to conduct risk analyses required in the rulemaking process. Prevention efforts should focus on these pathogens, because these are the pathogens that are causing the biggest problem. Among the 31 known foodborne pathogens:

• Norovirus caused the most illnesses.
• Nontyphoidal Salmonella, norovirus, Campylobacter, and Toxoplasma caused the most hospitalizations.
• Nontyphoidal Salmonella, Toxoplasma, Listeria, and norovirus caused the most deaths.

From a broad public health perspective, what needs to happen to cause a major decrease in the burden of foodborne illness?

Many actions need to occur for major decreases to be seen:

• Research is needed to understand how foods become contaminated naturally in growing environments or during production, packaging, transporting, and/or preparing processes, so that all involved can take prevention measures at each step in the process.
• Policies need to be in place across the farm-to-fork continuum that support research, prevention, and early identification of and speedy response to problems.
• Everyone needs to adopt practices known to prevent foodborne illness— farms following best practices for safe production of produce or milk or meat, restaurants following practices outlined in the Food Code, and members of the public following the “clean, separate, cook, and chill” guidelines. We all have a role in preventing foodborne illnesses.

This paper gives us more accurate data on the pathogens causing foodborne illness in the United States. Next, we need to better understand what foods are causing these illnesses and where foods are being contaminated. CDC publishes data on the attribution of outbreak-associated illnesses to food commodities in the MMWR. We are currently enhancing this approach by using outbreak data to attribute the burden of illnesses caused by known agents and to food commodities using the new foodborne illness estimates.

Do you know if anyone is working on an estimate of the economic burden of foodborne illness using the figures in the new paper?

Yes, scientists have used the 2011 figures to update estimates of the cost and quality-adjusted and disability-adjusted life years of foodborne illness, including:

• Scallan E, Hoekstra RM, Mahon BE, Jones TF, Griffin PM. An assessment of the human health impact of seven leading foodborne pathogens in the United States using disability adjusted life years. Epidemiol Infect. 2015 Oct;143(13):2795-804. doi: 10.1017/S0950268814003185. Epub 2015 Jan 30.
• Minor T, Lasher A, Klontz K, Brown B, Nardinelli C, Zorn D. The per case and total annual costs of foodborne illness in the United States. Risk Anal. 2015 Jun;35(6):1125-39. doi: 10.1111/risa.12316. Epub 2015 Jan 2.
• Hoffmann S, Batz MB, Morris JG Jr. Annual cost of illness and quality-adjusted life year losses in the United States due to 14 foodborne pathogens. J Food Prot. 2012 Jul;75(7):1292-302. doi: 10.4315/0362-028X.JFP-11-417.
• Scharff R. Economic burden from health losses due to foodborne illness in the United States. J Food Prot. 2012 Jan;75(1):123-31. doi: 10.4315/0362-028X.JFP-11-058.

When will the next estimate of foodborne illness be done? Do you anticipate that further changes in technology and methods will cause the next estimate to change as dramatically as the 2011 estimate?

The burden of foodborne illness will likely be estimated again in about 10 years. New innovations and discoveries are occurring all the time in public health, so it is reasonable to think that methodologies and data sources will have changed by the time we do the next estimates. It is difficult to anticipate how those changes might affect the estimates, but if surveillance systems are maintained or expanded we expect that the accuracy will increase with each new improvement in data or methodology.

Where can I find more information about the seven pathogens that cause 90% of illnesses, hospitalizations, and deaths due to known pathogens?

CDC has information on these pathogens on its website.

• Salmonella
• Toxoplasma
• Listeria
• Norovirus
• Campylobacter
• E. coli O157
• Clostridium perfringens