In the engineering world, we often do something called a Failure Mode Effects Analysis, or FMEA. We had some FMEA training recently at work, and this is what resulted from my mind wandering to farming topics during the presentation. Winking smile There are two flavors we use in our company: design, or dFMEA and process, or  pFMEA. The idea is to try, in advance, to think of all the ways a design or manufacturing process could fail, rate the possible failures for how bad they’ll be, and then tackle the worst ones to try to reduce or eliminate them before they become a problem.

It’s funny how many engineering principles relate well to farming, and this one is no exception. Farming is full of risk: financial risk, technical risk, process risk, weather risk. In recent months, I’ve been thinking about vaccinations, so here are some thoughts on an lFMEA, or a livestock FMEA!

A Simple Table

There are different methods for doing FMEAs, some very fancy that involve specialized software apps. But this is a simple version that works well for many applications. You start with a table, and in the first column, you list all the “parts” of the design or process. So for a sheep flock, we might list all the components that could have problems: the ewes, rams, lambs, breeding events, feed crop, etc- whatever we’re worried about.

The next column identifies all the ways that “part” can “fail” or have some kind of problem. The third column explains the result of the failure: what the customer would see, or what loss the business would incur. The fourth column identifies a potential cause of the failure mode. If the failure mode has more than one potential cause, each one gets its own row.


Next, you score the severity of the defect, 1 through 5, 1 being the least severe, 5 being the most severe. Then you assess the rate of occurrence, or how often you think this failure mode can or may happen. If you can base this on past data, all the better, but sometimes you don’t know, so you have to extrapolate. Again, use 1-5, one being a low rate of occurrence, 5 being high.

Last you evaluate your ability to detect the defect and do something about it before the customer sees it. In engineering, we call this “testing-in quality” where we may fail to design away a defect; but as a last resort, if we can catch and cull the problem in testing so it’s not seen by customers, sometimes we can live with that. Usually only if the rate isn’t too high. The ability to detect is also rated 1-5, with 1 being a reliable ability to detect, 5 being unable to reliably detect.


These numbers get multiplied together, and they create the RPN, or risk priority number. You can see that the numbers compound each other, high results are “bad,” low results are “not as bad.”

After assessing dozens or hundreds of failure modes, this reasonably objective method will help you prioritize which defects are the most concerning, that should be addressed first. Next you think of ways you can either eliminate or mitigate the potential cause, and then re-score the failure mode to see if you are happy with the new RPN, or if more work is needed. In the engineering world, FMEAs are always done by a team, or otherwise peer reviewed, because it’s hard for one person to think through all the possible risks. Being thorough and objective is the goal.

FMEA for Farms

Now, actually doing one of these for a small farm may be a little silly and overly formal. But maybe not. We probably all do this logic every day in our heads. It’s interesting, anyway, to walk through a snippet of an example, especially because I’ve been weighing this very decision in my head: vaccinating ewes. Here is the grid:


The concern is abortion, and I had a small rate last year. I had a 7% fetal loss, and another 7% loss of full-term lambs during or immediately post-birth, which could be caused by a lot of things, but abortifacients are one consideration. If I spell out the main suspected causes of abortion here on this farm, they would be vibrio, chlamydia, toxoplasmosis, a mineral supplement change, and rough handling or stress (caused by the border collies, needing to change pastures, me handling the sheep during maintenance activities, etc).

Clearly the severity score of abortion is high, since the outcome is a percentage loss of the lamb crop. I had to guess at the rate of occurrence for each cause. I based it partly on the symptoms of loss last year, and also general potential: for instance, vibrio can be spread by birds, and we’ve had an outbreak in our state. So even if I’ve never had it, I could get hit any time. The ability to detect gets a poor score of 5 for most abortion causes, because by the time you know about the problem, it’s too late, and it can sweep through the entire flock. Chlamydia is the exception, because if you do detect it early, you can treat with antibiotics and possibly stop the problem.

Simple Conclusion

So, you can see that vibrio is the highest priority risk factor, and post-mitigation, or after vaccinating ewes, the score gets down to a reasonable level. There is still “residual risk,” because vaccine doesn’t have 100% efficacy and detection ability is still poor. But this hints at vaccinating for vibrio to be a pretty good place to start.

There is still a cost factor to consider: if the vaccine cost threatens to outweigh the potential cost of loss of lambs, then it doesn’t make sense. But at a few bucks per ewe, plus some labor for me, if it saves a few extra lambs, then it’s worthwhile.

Circular References: Mitigations Become Risks

But it isn’t quite that simple, as you may have spotted from the last row of risk. You see, vaccinating comes with its own set of risks. Vaccine is toxic, and can cause adverse reactions in the body, which sometimes require medical attention. It can cause the patient to go into shock and die, if epinephrine isn’t administered promptly. Ever since its invention, vaccine has been implicated in all sorts of health concerns, from cancer to autism. Some allegations may be proven false someday, until then, there is a lot we still don’t know. But inarguably, vaccine does carry some level of risk, whatever it may be. Vaccinating livestock is higher risk that vaccinating pet dogs or children, because the cleanliness just isn’t there. (And in fact, usually in the wrestling match, the vaccinator ends up vaccinating him or herself with a contaminated, bloody needle, but that’s yet another risk topic… Open-mouthed smile).

So, I weighed that too, and this is what it comes down to: does the risk of vaccinating outweigh the benefits? In this case no: the objective scoring mechanism of an FMEA points out that the score for vibrio is much worse than the score for vaccine reactions. I still have to acknowledge that I could sicken or even kill a small percentage of ewes by vaccinating. But the potential percentage loss from abortions is far greater than the very small percentage loss anticipated from vaccinosis.

This does bring to light that vaccinating for mild diseases probably does not make good risk sense: if the severity of the disease only scores a 1, and the incidence is medium (3), detectability (treatability) a 2; then the overall RPN is a 6. Lower than the RPN of the vaccine itself (a 10), so not worth it!

So vibrio vaccine it is: the ewes all got their first dose last weekend!