When news broke this past fall of two separate Salmonella outbreaks tied to Foster Farms chicken, one of the most noteworthy aspects of each outbreak was the number of patients sickened. To date, the outbreaks have been connected to 615 reported Salmonella infections altogether, making Foster Farms chicken the largest known source of foodborne Salmonella in 2013. But what isn’t as readily apparent from reading outbreak reports is the number of people estimated to be sickened beyond those confirmed in official case counts. According to methods used by epidemiologists to estimate the true impact of foodborne illness, the Foster Farms outbreaks may have sickened closer to 18,500 people. That’s because, for each case of Salmonella that gets identified through clinical laboratory analysis, another 29 illnesses are estimated to go unreported. While there are numerous factors that play into those estimates, one thing is certain: For every person officially counted as part of an outbreak, far more cases go unnoticed. The Big Picture When epidemiologists and food-safety professionals speak about the burden of foodborne illness, they often cite the statistic that 48 million Americans are sickened each year with a foodborne pathogen from domestic food. That number doesn’t come from the total number of people reported to health departments with foodborne infections, but instead is an estimate based on what epidemiologists know about how each pathogen functions and affects humans. The U.S. Centers for Disease Control and Prevention published the most recent estimation in 2011 in an attempt to measure the impact of foodborne disease on public health and the U.S. economy and put the agency’s rulemaking in a food-safety context. The estimation of 48 million illnesses, 128,000 hospitalizations and 3,000 deaths is based on what health officials know about 31 known foodborne pathogens, as well as unspecified agents such as microbes and chemicals. To make those estimates, epidemiologists considered a range of variables about each particular pathogen: How severe are most infections from it? How many doctors test for it? How sensitive are the latest testing methods? “It’s so complex that it’s somewhat easy to explain,” said Elaine Scallan, Ph.D., professor of epidemiology at the Colorado School of Public Health and lead author of the 2011 foodborne illness estimates compiled by CDC. One of the main challenges in estimated foodborne illness is that no pathogens can be judged equally. Some cause more severe illnesses, while some are less common and might not be diagnosed by as many healthcare providers. For example, while only one out of every 31 Campylobacter cases is estimated to be reported, health departments are believed to track one of every three cases of Listeria monocytogenes, which causes much more severe illnesses and therefore hospitalizes a greater number of those who are infected. Only those patients who seek medical care stand a chance of ending up on an official case count for an outbreak. But visiting a doctor is just one link in the long chain between exposure to a pathogen and an eventual reported foodborne illness. The Hurdles to Diagnosing and Reporting For a patient to become counted as an official case in a foodborne illness outbreak, a number of factors need to line up just right. It all depends on how the patient reacts to their symptoms and how well their healthcare provider diagnoses those symptoms. First, even if someone is exposed to an outbreak pathogen, they need to become sick enough to notice the effects. Many people exposed to a pathogen may not fall ill, or may not suffer severe enough symptoms to suspect something is amiss. If a patient does come down with symptoms, the symptoms also need to be severe enough for them to seek medical care. This fact alone eliminates a large percentage of cases from ever getting reported, said Carlota Medus, Ph.D., principal epidemiologist with the Foodborne, Waterborne and Zoonic Disease Unit of the Minnesota Department of Health. Consider how often people come down with diarrhea, cramping or vomiting without visiting a clinic. By the nature of disease reporting, health agencies never become aware of the people who get a foodborne bug but stay at home to tough it out. Because of that, epidemiologists anticipate that more severe pathogens – or more severe strains of pathogens – will hospitalize a higher percentage of those exposed. The type of food source causing the outbreak could also play a role in determining the percentage of patients who ultimately end up as reported cases. If the food product is aimed at children, for example, a higher percentage of cases could end up hospitalized simply because children’s immune systems are more susceptible to infection. But even if a patient makes it to a healthcare provider, their infection is still far from being guaranteed to be reported. They still need to have a stool sample taken, which may only happen if the patient agrees to provide one and the doctor decides it’s worth taking in the first place. The healthcare provider’s opinions and tendencies could very well be the deciding factor in whether or not a patient’s foodborne infection is identified, Medus said. Some doctors may believe it’s not worth the effort to test a diarrheal stool sample because it will likely be a brief illness, while others may decide it is worth testing such samples to contribute to outbreak surveillance efforts. Even if the patient does produce a stool sample, the pathogen might not be present in their system– perhaps due to having taken antibiotics or because too much time has passed. The doctor would also need to order the right tests to be looking for the right bugs. With each successive step, more and more illnesses are excluded from joining the official case count, which requires a clinical laboratory to positively confirm a test sample and report it to a health department. “Just a fraction of people seek care, and then just a fraction of those people submit a stool specimen for testing, and then a smaller fraction of those samples are tested for the right pathogen,” Scallan said. Then, even if a test result makes it through all the hurdles at the healthcare stage, the laboratory sometimes doesn’t report it, or they may not be required to report it, depending on state law. Under federal reporting laws, laboratories are asked to report cases of Salmonella, Listeria, Shiga toxin-producing E. coli, Shigella, Vibrio and botulism. Some states may also monitor for additional pathogens, such as Campylobacter and Toxoplasma, but they are not required to do so under federal law. “There are so many factors that go into whether your illness gets reported, including what state you live in,” Medus said. Even less likely to be reported is Norovirus, which sickens an estimated 5 million people per year but generally is not tested for because illnesses typically only last a day or two. In total, here’s the breakdown of how many cases of a given pathogen are estimated to go unreported for every diagnosed case: •     Salmonella: 29 cases unreported for every diagnosed case •     E. coli O157:H7: 26 cases unreported for every diagnosed case •     Listeria monocytogenes: Two cases unreported for every diagnosed case •     Campylobacter: 30 cases unreported for every diagnosed case •     Vibrio: 142 cases unreported for every diagnosed case Challenging Detective Work Scallan described the work of estimating illness numbers as incredibly challenging. The publication of the 2011 estimates were an update to estimates published in 1999, which projected foodborne illness to affect 78 million Americans per year. However, just because the estimates have dropped from 78 million to 48 million, health officials do not give all the credit to new prevention efforts to reduce foodborne illness. The methods for making the estimates have been refined since 1999, and so comparing two data sets would be disingenuous. Officials do, however, believe that the 2011 estimates are more precise. That’s due in part to efforts made through FoodNet, a foodborne illness surveillance network of state and local health departments in 10 states. Data collected by FoodNet are seen as creating the foundation for foodborne illness estimates and subsequent policy and prevention efforts based on them. And, in the end, decision-makers can only work with the data available, which is why epidemiologists can only accurately report the small fraction of cases that are confirmed in the lab. “As a general rule, we report what we know, because we get criticized if we start reporting on speculation,” Medus said. “But if someone asks me if I think there are more cases in an outbreak, I’ll say, ‘Yeah, it’s very likely there are more cases.’”

  • Daniel_Cohen

    When disease reporting is used for policy decisions it would make more sense to describe the actual lab-confirmed cases first and then the varying multiplication factors and their rationale. Particularly when cross disease estimates are aggregated to draw conclusions about relative risks. Thus dead are compared to dead, hospitalized to hospitalized, ill and reported to ill and reported for different outbreaks, diseases and categories of transmission in the initial reporting. But when it comes to policy, as for FSMA implementation, the assumptions used for the purpose of estimation of overall disease burdens — useful in that context — are speculative for policy. What are the error statistics on the multiplication factors and the confidence limits? When estimates change from 78 million to 48 million the figures cannot be taken as written in stone. They are useful guestimates in context, better than “we know there are more cases”.

  • MeDoctor

    Our sickcare system is incredibly effective at discouraging doctor visits. Long wait times to schedule an appointment (if you can get through the typical answering system), then long delays at the doctors office sitting with a lot of other sick people, followed by a 3 minute evaluation and a prescription for lab tests and some general purpose antibiotics to make you feel it was worth the hassle. This lunacy is repeated over and over until you may seek another physician and start over. How many people go years with a condition unsatisfactorily resolved? Could it be the system is designed by Big Pharma, for Big Pharma, at our expense?

    • johnmarkcarter

      Well, it’s not just Big Pharma. And it’s more than a conspiracy. It’s a complex as powerful as our military industrial complex, comprising the pharmaceutical industry, health insurance, the American Medical Association, and ambulance chasing lawyers.