Safe Food_ Bacteria, Biotechnology, and Bioterrorism - Marion Nestle [27]
Counting Cases and Estimating Costs
If harmful microbes are widespread in food and if they make so many people sick, why isn’t everyone—the food industry, health officials, and the public—doing something to prevent them from getting into food? One reason is that most episodes of food poisoning are not very serious. Another is that it is difficult to collect accurate information about the number of cases and their severity. Attributing a bout of diarrhea to food rather than to other causes is no simple matter. Most of us eat several foods at a time, several times a day, in several different places. How could we possibly know which food might be responsible for our getting sick, especially if there is a delay in the onset of symptoms? I cannot imagine bothering to call a doctor about a brief stomach upset. Even if I did, the doctor might not suspect food as the source of my problem. Busy doctors rarely report such suspicions to health authorities. It usually takes an “outbreak”—the severe illness or death of more than one person eating the same food—before health officials learn about a foodborne illness and attempt to trace its origin.
For these reasons, counting cases is a formidable undertaking, and to this day there is no national system for doing so. The current surveillance system, such as it is, evolved piecemeal. In the 1920s, the Public Health Service started tracking diseases carried in milk. In 1961, the Centers for Disease Control (CDC), an agency of what is now the Department of Health and Human Services (DHHS), took over that task and began to issue annual counts of illnesses transmitted by food and drinking water. Five years later, the CDC initiated a voluntary program of state surveillance of outbreaks, meaning that states could choose whether or not to participate.
As early as 1970 the CDC realized that its counts were way too low. Nearly half the participating states were reporting no outbreaks or very few, suggesting considerable underreporting. In 1985, several federal and private agencies began to make more serious attempts to estimate annual cases of foodborne disease, based on two assumptions: (1) an episode of diarrhea counts as a foodborne illness, and (2) the proportion of reported cases to those that are not reported ranges from 1 out of 25 to 1 out of 100 or more. The agencies understood perfectly well that diarrheal diseases could be due to causes other than foodborne illness, and that foodborne illness also causes symptoms other than diarrhea. Nevertheless, they multiplied the number of cases of diarrhea by 25 to 100 to estimate the “real” number of cases. During the next few years, these confusing assumptions led to widely varying guesses about the number of annual cases (6.3 to 81 million) and deaths (500 to 9,000), depending on how the assumptions were interpreted.8
In 1996, the CDC initiated a new surveillance program, FoodNet—the Foodborne Diseases Active Surveillance Network—in just a few states and for just seven microbial pathogens. In its first year, FoodNet identified 8,576 laboratory-confirmed cases of foodborne illness, of which 15% resulted in hospitalization. In 1999, the CDC used this and other information from its surveillance networks to suggest that known pathogens caused 14 million illnesses, 60,000 hospitalizations, and 1,800 annual deaths. When they added these estimates to those for cases caused by unknown pathogens, they arrived at the annual totals mentioned earlier: 76 million illnesses, 325,000 hospitalizations, and 5,000 deaths.9 Because these estimates rely so heavily