9

Sometimes I have suggested that food that sat out a little longer than optimal may be safe to eat.

Granted, this is based on my experience, and I always qualify that it shouldn't be served to guests (and by extension, customers) and that one should determine their own tolerance to risk, but I want to know if there is information as to the methodology of the FDA guidelines. Specifically, what is the contamination rate at n hours, and what are the initial conditions of the food and environment? How much risk is there REALLY (in incidences per n-population and such) in a standard kitchen for food that has sat out 50%, 100%, 200%, etc. longer than recommended?

Cascabel
  • 58,065
  • 24
  • 178
  • 319
JSM
  • 719
  • 3
  • 8
  • If you have something you want to discuss to do with how people should answer questions, what should be closed, or why people vote the way they do, please take it to [meta]. Some general comments, before you post there... Users are free to vote how they like, and no site policy will or should stop them from downvoting answers they disagree with, but downvotes aren't the same thing as saying you're not allowed to post answers like that. And we do frequently close food safety questions as duplicates of one of a few canonical questions. – Cascabel Aug 07 '14 at 07:22
  • @Jefromi Fair enough. I just get a little frustrated when I answer a question no one bothered to, and get downvoted as a result. I always try to give the 'correct' answer, then my experience, and qualify with YMMV. I may take this up in Meta, but I really do want to know how the FDA/USDA determines their safety guidelines. As an engineer, knowing the methodology behind them would allow me to give better answers and qualify the risks, should I choose to answer those kinds of questions in the future. – JSM Aug 07 '14 at 15:44
  • Yes, I agree this is a great question. Happy to discuss the rest on meta or in [chat] if you'd like. – Cascabel Aug 07 '14 at 15:53
  • 2
    I have no idea how they measure, but I would assume they'd use something like the LRFD method used in civil engineering -- you try to minimize the risk of there being a problem based on statistics. (how much of a risk they design for, I have no clue ... 1%? 0.01%?) and of course, different foods are going to have slightly different risks that the 2hr guidelines don't account for. – Joe Aug 07 '14 at 17:03
  • That's what I figured, Joe. Not familiar with the LRFD method, but been engineering for 7-8 yrs now, so I am familiar with fault tolerances. I figure the number they give may also depend on cleanliness of prep area, ambient temp, processing of food prior to preperation, etc. I'm sure they take some sort of mean, but would like to see the raw data. Are there cases where 2 hours is too long, or where the food would be fine for half a day? Those are things I usually eyeball at this point, and would like a more definitive answer. – JSM Aug 07 '14 at 18:40
  • I don't think your question has an answer. Let's say that X bacteria cells of a given species will make the average person ill, and the FDA has calculated holding times such that a contaminated piece of food containing on average X*10^-7 bacterial cells will have at most X*10^-3 after being handled according to guidelines. Bacteria multiply at an exponential rate influenced by temperature, so it can happen that the same food will reach dangerous levels after 10 hours at 20 Celsius and after 1 hour at 25 Celsius, for example. And nobody knows the exact number of contaminating cells too. – rumtscho Aug 10 '14 at 11:56
  • 1
    @rumtscho I guess I am asking what the values of X are in your example. I realize that there are certain assumptions, and maybe the FDA doesn't make it public. If they do though, it would be nice to see it. That way the answer to I left x out y hours could be, "These are the guidelines. If you are in Minnesota in December, you may have n more hours. If you are in Alabama in August, maybe less. When in doubt, throw it out." I believe in nuanced answers, and a stock 'toss it' does not make me happy. – JSM Aug 11 '14 at 16:35
  • 1
    I agree that this is a great question, and one that I often wonder about as a cook and haven't been able to find info about online. It's a legitimate question about food preparation practices, not site guidelines, that shouldn't be relegated to meta. – Josh Aug 12 '14 at 04:22

1 Answers1

4

The recommendations (in the US at least) are based on a risk model, which takes into account a number of factors:

  • frequency of outbreaks and occurrence of illnesses
  • severity of illness, taking into account illness duration, hospitalization and mortality
  • likelihood of contamination
  • growth potential/shelf life
  • manufacturing process contamination probability/intervention
  • consumption
  • economic impact

Some of these factors are statistical (frequency, economic, for example), and some are measured (growth, aspects of severity). For the measured aspects, a large number of methods are employed, including animal testing, human histories, and toxicology reports (including LD measures).

Primarily, most home cooks should follow the FDA guidelines as the recommendations take into account the supply chain and common outcomes for most pathogens (for a given food). It's important to note that live pathogens are not the only risk in food safety, also toxins (as produced by the pathogens) are an important factor depending on the specific life form. These toxins are not mitigated by cooking (generally), and some pathogens are not completely mitigated by heat or cold. The risk is related to the combined risks, time, and supply chain.

The FDA site contains the regulations, methods, and sampling guidelines for various stages of food production and food safety.

This is not my speciality, but I did teach food safety many years ago (please refer to the current guidelines and methods for current recommendations).

References:

  1. Designating high risk foods
  2. Home food safety cheatsheet
  3. Example toxicity studies - I'm not certain these feed into these specific guidelines, but the methodology will be similar
Bruce Alderson
  • 2,765
  • 19
  • 25