• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Animal & Veterinary

  • Print
  • Share
  • E-mail

Section Contents Menu

Safety & Health

Presentation: Risk-Based Initiatives at the Center for Veterinary Medicine

by Dr. Barry Hooberman

DR. HOOBERMAN: Thank you for coming.


I would like to start off a little bit -- for context -- about some other activities that we are doing at the Center on risk-based initiatives.


So -- and some of you have been at several of our Public Meetings, have these slides before. Why a risk-based approach? And I am going to go kind of fast, so please slow me down if you think I am going too fast or you can raise your hand and ask questions.

The risk-based -- a risk assessment approach -- is just a logical way to organize data, and I don’t think it is anything more than that. It is a way to put things together so that you can come to a decision. You can use it as kind of a forecasting tool, and also, if you do it right and you emphasize the transparency behind all your decisions in the format of the risk assessment, it can serve as a very effective tool to ease the communication between decision makers, stakeholders and the public.


So what is -- we are doing a risk ranking, using risk assessment, and as you see here, the risks come in all shapes and sizes, and the whole purpose here is to try and rank them. That is what we are trying to do, worst risks at the top. You can do this in a scientific manner, but of course it doesn’t necessarily mean that those values prepared by the scientists are the same as done -- or those held by the risk managers, they may have other factors that will change the actions that the risk managers take, such as Congressional mandates or budgetary considerations and things like that.

So, actions based on the risk assessment may not always translate directly into management actions.


Okay, so -- covering quickly, what is risk assessment? There are four simple questions. What can go wrong -- that is what we refer to as hazard identification.

What are the consequences of things going wrong? That is the consequence assessment. In our case, it is the health consequences.

How can it happen? is the exposure assessment.

Finally, you put the consequence assessment together with the exposure assessment, and you answer the question: What is a likelihood that something is going to go wrong? And that is what we call the risk estimation.

I will point out, and you heard it a little bit earlier, is that in our feed safety approach, we are not actually estimating risks from hazards. It is all a relative ranking system. So we are not going to tell you that Dioxin at this level is associated with this health risk to humans. That is not the end point that we are after.


Okay, so the Center has several compliance programs. These are inspectional programs where we send out inspectors to follow up and look for violations and make sure everybody is compliant with our regulations and policies and procedures and things like that.

Here is a list of the ones we have in the Center -- BSE, drug residues, medicated feeds, drug process (cGMPs), animal feed contaminants, and then there are two more, BIMO, which is Biological Monitoring and pre-approval for new animal drugs. I am not going to talk too much about those last two because they -- for other reasons, they are not part of our risk-based approach at this time. There is a risk-based approach for the last one, the pre-approval process that follows along the one that is being implemented by our Center for Human Drugs, and that is a whole different topic.


Okay, so why are we doing these risk-based approaches for these programs, for BSE? We are trying to figure out where we should send our inspectors. How many do we need to send out? How many inspections do we need to plan? And where should they go? So we are trying to come up with a list of ranking, where we think the biggest risks are.

The two criteria for this particular BSE program is site hazards, and we will talk a little bit more about that, and violation history. So you come up with a -- what we call a “go/no go” decision made annually that we communicate to our Office of Regulatory Affairs and deliver what is on the list and say, “Here is where we need to send our inspectors.”


Okay, so a little bit more detail about the scoring method for the BSE program. Factors are like last inspection date, type of firm -- you can read what is on the slide. Do they handle prohibited material? Issues like that. And Shannon Jordre in the back can answer further questions -- right, Shannon? -- of you are wondering about the BSE program.


Additional questions that we use as criteria for establishing our ranking system are up on the screen, and in the sake of time, I am not going to read them for you. You can read them or see them in your hand-outs.


Okay, another program is drug residues, in other words, what we get from our sister agency, FSIS -- Food Safety Inspection Service. They do sampling of meats and milks, and the meat data comes back for -- tells us if there are any violations of drug residues, in other words, residues higher than the allowed tolerance. And then we have to make a decision about which one of those we are going to follow up.

Again, same type of approach -- we are trying to prioritize where we need to spend our resources.


Some of the criteria here are violation history, the same type things, and I am going to go more into the risk criteria in the coming slides here.


So, for drug residue, the things we are going to look at is the toxicity of the drug -- is it a highly toxic drug? If there is a violation, we may want to prioritize on that? If they are well above the tolerance, that may be a key criteria to saying that is something we need to follow up on.


This is a very busy slide, but this is basically our scoring criteria for drug residues.

Three factors. The hazard, which is the toxicity of the drug residue.

What we call the likelihood of exposure, which is represented by violation frequency as their past history. Does the firm, or the farm, have current violations?

Then the final one is also part of exposures: How high a violation is it? How many times the tolerance?

Somebody that is -- now you can see 250 times the tolerance, they are way over what they are allowed to do, which may indicate intentional violations or very poor controls. That gets a score of 8. And then you can see in red here are the scores. And then these are combined to -- mathematically to prioritize and come up with a list, and that is our targets. And we have a threshold cutoff. If the score is above a certain level, those are the ones we try to get to because we have to plan for the following year. And we do revisit it as new data comes out weekly from FSIS, and we revise the schedule.


Okay, then are medicated feeds, another compliance program. These are a little bit different, and Jo Gulley, who is here, can fill you in more information about that, but we can’t really pin it down to an individual drug. A facility may be licensed for one category of drug, which means they can do a whole range of drugs in that category, and therefore we are limited in what we can use for criteria.

So we have identified facility risks -- again, looking at inspection history. When is the last time we were there, or inspecting there? And also look at the overall category in terms of a product risk. And those are the criteria we use for coming up with our priority list for medicated feeds.


Then here is a little bit about our drug cGMP inspections. This gets a little bit more complicated.

We have three main areas -- a facility risk, product process risk, and what we call a “complacency risk,” which is basically: How long since the last inspection?

For some of these programs, we are mandated by Congress to go a certain number of times to investigate. We don’t always make that schedule, but it is a consideration in coming up with our priority list.


Okay, so briefly that is the kind of things we are doing. We have implemented most of these programs for this past fiscal year and we are revising and refining them for the coming fiscal year.

One of challenges is: How do we rank risks across these programs? Is the BSE program, the risk that we are trying to address there, comparable to what is a medicated feed risk or a drug residue risk?

Remember, a lot of these considerations are important for our budget. You know, we do have limited resources, so we have to try and prioritize where we want to put our money.

There are a number of ways to do this. We haven’t come up with a good way yet. We are always looking for ideas, and we are working on this. And just to give you a glance at the kind of factors that we are thinking about is one of the issues is that these programs are all for different reasons. Some of them are surveillance programs, some of them are for cause which means that we know there is a violation that we are trying to follow up on. So there are different needs for these programs and they are trying to address different things, and that is part of our challenge is trying to -- how do you compare those?


Again, another way you could characterize a program, some of them are known hazards. BSE, we know what the hazard is, and the range is either BSE is there or it is not there, and that is kind of where we are at, so that is -- for that program it is very different than for the feed contaminants where we know most of the hazards but the range is huge, from very toxic things to very nontoxic things and very high exposures to very low exposures.

So there are just -- the point is that there are different characteristics to all these compliance programs and it is proving to be a quite a daunting task to try and compare across these programs, compare risk across programs. And this is an important theme you will hear again towards the end, because one of the things we are trying to do with the Animal Feed Safety System is compare chemical risks versus microbial risks, and that is a huge challenge that a lot of people in government are facing, and nongovernment, too, and how to do that, and we can talk about that later.


So this is just repeating what kind of what I have said. The challenges are comparing programs with different objectives. There is a wide range of potential hazards and risks across these programs, and you don’t always have a specific hazard. You do in some programs. In others, you are talking about facility risks -- so how does a facility risk compare to an individual chemical risk, a contaminant risk? We have to try and come up with some sort of metrics to measure risks that will be able to be applied to all these programs. As it says there, finding appropriate metrics is a big challenge.

So, we are always looking for solutions. If anybody has got any, or think that they might have any, we are always willing to listen.

That is all I wanted to say, so that is moving along fairly quick.

Any questions or anything?

(No response)

DR. HOOBERMAN: No? Was that too fast? Everybody kind of get a feel for what we are trying to do? Okay.


So, the one I -- the one program I left out in describing was this one, the animal feed contaminants. It is another compliance program. It is the program that has supplied most of the data that we have, even as limited as it might be, for our Animal Feed Safety System risk model.


So you have seen this slide before. This is what the Animal Feed Safety System is. Again, we are trying to make it risk-based, and we are talking about risk to both animals and human health.


It is important to remember how we are going to use these ranked risks that we are going to come up with.

I will point to the bottom first. It is primarily at this point for internal use only. We hope to get it to a state one day where we could share it.

We have a -- we would have to do a lot of documentation because the paucity of data means that we are making a lot of assumptions. So in order to make this transparent to whoever may use it, we have to document all these assumptions and where they are, and hopefully, once it gets out there a little bit, some data may come in or some opinions come in, expert opinions, that will help us refine those assumptions and make them present a more accurate picture of the scenarios. So that is something we want to shoot for eventually, because we are making a lot of assumptions at this point in time.

So we are going to use this ranked risk to help us focus on where we need to focus our -- where we need to spend our resources. Do we need to -- are there contaminants we need to be looking for? And are there limits that we need to establish, analytical methods we need to develop? Can we provide any guidance to the manufacturers about process changes they may want to make that could reduce risk? And then, do we need to establish further surveillance sampling programs or refine the ones that we have to more appropriately address the bigger risks?


Okay. So here are the goals as we developed this risk model which is a fundamental part of the AFSS, to try and rank relative risks. Again, I will emphasize we are not telling you -- estimating true risk for each contaminant. We are saying we are more worried about this contaminant than this contaminant, so the goal is to come up with a ranking, but you can’t infer that because it is a high risk in the ranking, this is a really high risk to human health, because we are not saying that -- it is among the chemicals we are looking at, it is a higher risk.


Okay, just a quick reminder, and you are going to hear this a lot in Phares’s talk coming up. Risk is a function of hazard and exposure. By hazard, what we are calling is represented by health consequences. We have kind of divided health consequences into two factors, a severity factor, which is the severity of the health impact, and the potency, like how much do you need to cause the illness?

On the exposure side, we are looking at, what are the routes of exposure? And we will be covering more about that. And what is the likelihood that that exposure is going to occur?


Okay, so the AFSS model has this framework. We are starting off with hazards that are in feed ingredients, and we are going to follow it into individual ingredients, how those ingredients are processed during manufacturing because those processing methods may affect the levels of the hazards, either increasing the amount of hazard by a concentration effect or possibly diluting it out or removing it. We have what we are calling “modifying factors” that will address that. And then getting down to what is in the final feed. So we have hazards from each of the -- hazard ingredient pairs, and then when we put the feed together, we are totaling up the hazards that may be in from the different ingredients and estimating exposure to a given hazard in the feed.

Then we are going to follow that up to look at humans, if it happens to be a food-producing animal, what residues may exist to -- that would eventually pose a risk to humans.


Just an overview of where these contaminants might be coming from. Obviously, there are a lot of sources from different starting materials, and you can see from there that the main categories of hazards we are trying to look for are Mycotoxins, agricultural chemicals such as PCB or Dioxins, microbial pathogens, metals -- cadmium, mercury, lead -- or possibly drug residues and TSEs.

The model that we have now at this point in time is not really addressing drug residues or TSEs currently. We hope to expand it to that at one point in time, but we are mainly focusing on the first four categories of hazards.


Okay, so we have a feed contaminants program. It is where I started out at. We do do some sampling and have some results on these chemicals, or elements -- certain pesticides, certain heavy metals, Mycotoxins, microbial pathogens, and the Dioxins in PCBs.

So this is -- the feed contaminants program is our main source of data. We have been scouring and asking for additional data, but not received much at all. So we are always -- I will put out another pitch -- we are always glad to talk to somebody who might be able to provide us with some additional data. I know our colleagues in Canada have mentioned to us that they might have some relevant data, and we would love to get that, so we -- I have to continue those conversations.


Okay, so we are going to have data on some ingredient contaminant levels from our feed compliance program. We will also have some data on complete feeds, and this is going to provide an interesting check on our methods of estimating exposure. In other words -- and you will see this more for the chemicals than you will for the microbials at this point in time. We are estimating -- trying to follow levels of a hazard, a chemical hazard, that starts in the ingredients and go through into the final feed. And if we do a good job of estimating that, it should be fairly close to what we are finding in our mixed feed results. It means that we did a good job. And we have had pretty good luck with that, but that will be -- we will talk about that further under the chemicals talks.


Okay, so what you are about to hear is you are going to hear Phares Okelo talk about the risk-ranking model that we are developing. Up to now, before Phares has put in some pretty intensive effort, we had everything in spreadsheets.

Phares is working on putting a nice front end and refining things. He has made a lot of advancements, mostly on the microbial side, and we are hoping to pour a lot of the chemical portions of the model into the same sheet, into the same approach. And we thought -- think that it will be a good refinement for what we have done on the chemical contaminants.

Then, Jerry Rushin is going to talk about using that risk-ranking model and its application to microbial pathogens.

Then I am going to come back and talk about a risk-ranking model because we haven’t implemented it in the model that Phares and Jerry are going to be talking about, but we are able to use our existing model to come up with a ranking of chemical contaminants.

The last slide will be our ranking at this point in time, a little thing to look forward to, and hopefully I won’t be laughed off the stage when you see that list. And that is what is coming up. So, Phares is up next.

Are there any questions while I convert over?

(No response)