It may not have been the tomatoes after all. After a fruitless (!) search for contaminated tomatoes, FDA is beginning to look elsewhere for the source of the salmonella "outbreak":
As sicknesses continue beyond the shelf life of tomatoes — it’s been almost three months since the first Salmonella case — officials are also starting to question whether another vegetable is to blame, according to this morning’s USA Today..
“We’re broadening the investigation to be sure it encompasses food items that are commonly consumed with tomatoes,” Robert Tauxe of the Centers for Disease Control and Prevention said.
I think the good doctor is just trying to reassure us. "Food items that are commonly consumed with tomatoes" sounds suspiciously like "the general food supply"; after all, what's not "commonly consumed with tomatoes"? This means that any one of us could be exposed to salmonella at any time. Why, you could be eating a big bowl of salmonella-tainted spaghetti right now.
I originally planned to write this up as a possible instance of observation bias:
- FDA detected a random cluster of reported salmonella cases and put out the alert. It didn't know whether the cluster was due to chance or to a genuine outbreak, but it gets in trouble for not warning, so it figured it was safer to over-warn.
- Doctors everywhere, alerted to a possible outbreak of salmonella poisoning, began looking for salmonella poisoning.
- The symptoms of salmonella poisoning are headache, diarrhea, nausea, fever, and abdominal cramps, symptoms that doctors ordinarily would have been content to diagnose as "stomach virus." But doctors instead screened these patients for salmonella poisoning. They asked them if they'd recently eaten tomatoes -- and many had, since many of them had recently eaten food. Rather than send them straight home with some Tylenol and Phenergan, the doctors sent them to the lab to be tested for salmonella.
- Presto. Lots of unremarkable and (epidemiologically) invisible cases of stomach virus became confirmed cases of salmonella poisoning.
As I said, that was my original take. But I now think this story is too charitable to FDA. It doesn't sound like there was an outbreak at all.
FDA has counted 810 cases since April 23, a period of two months. 810 cases every two months is equivalent to a rate of 4,860 cases per year. 4,860 cases per year is not an epidemic. It's not even the background rate:
It is estimated that from 2 to 4 million cases of salmonellosis occur in the U.S. annually. (Source: FDA Bad Bug Book) ... Every year, approximately 40,000 cases of salmonellosis are reported in the United States. Because many milder cases are not diagnosed or reported, the actual number of infections may be twenty or more times greater. (Source: excerpt from Salmonellosis (General): DBMD)
(I like that FDA gives its epidemiology publications cute, nursery-rhymey names.)
We are a nation of 300 million people who occasionally consume undercooked poultry and bad eggs and other food that tasted a little funny at the time. Some of us will get salmonellosis. "Some" of 300 million is "a lot" -- on average, 10 times as many confirmed cases as the Great Salmonella Poisoning Outbreak of 2008. And that's before subtracting the extra cases doctors have detected this year because they were specifically looking for them.
Thus, to save the general population from a risk of salmonella poisoning well below the background risk, FDA has given restaurants a $100 million butt-kicking and forced the rest of us to go for weeks without a decent salad or fajita.
They know they've screwed up:
“It’s bad, and I think everyone will be very apologetic” if it turns out tomatoes weren’t the source, said Tim Jones, Tennessee’s state epidemiologist, describing himself as “increasingly concerned” about whether tomatoes are to blame.
That's how an epidemiologist says "D'oh!"