Here are some basic facts we know about detection dogs, thanks in large part to all the federal grants that have gone to university laboratory studies of canine olfaction for the purpose of explosives detection. The facts give us a starting point, but they don’t paint a very clear picture about whether or not gluten detection is a good use of the dog’s incredible nose. For years I said, “No” to gluten detection because gluten really truly is everywhere, but after hearing story after story from Celiac patients across the country I realized that if it were me or my child going in and out of the hospital, dealing with chronic pain and failure to thrive (just to name a few of the many complications Celiac disease may bring) that I would train a dog to detect gluten, even though I know it’s not a perfect solution. It’s not as reliable as I would like it to be, it’s not an instinct-modified behavior the way traditional detection is (which I believe makes it is more prone to mistakes), and even with the best training and maintenance of the dog, false responses and misses will still occur. Even knowing all of that, if it saved me or one of my kids from a going through even half of the reactions of a typical year, I think I would do it. I guess those are pretty much the same reasons I started training allergen detection dogs in the first place. For those of you who have been wondering where we stand on gluten detection now or why we have changed our position, that’s why we have decided to pursue it. I still don’t think it’s a decision that is right for everyone with Celiac disease, nor is it a decision that should be made lightly by even those with severe cases. It is a lot of work, a big investment of time and money, and requires you to make some changes in daily life.
Alright, now back to the science…
1.) Dogs can detect odor at levels down to parts per billion, sometimes even parts per trillion!
Now, given, that is in a lab with NO distracting odors. Real life detection capability is virtually un-measurable. We can set out specific amounts of gluten for them to detect (measured in grams, milligrams, tenths of milligrams) but that isn’t the same as “parts per…”. To complicate things a bit further, the actual amount of the substance they are detecting is a separate thing from the amount of odor that is being given off from that substance. One milligram of old, hard gluten may give off much less odor than 0.1 mg of fresh gluten in a gooey form that has been spread over a larger area or 0.01 mg of powdered, airborne gluten.
Lots of factors including surface area, temperature, state of matter, air current and even what other substances the gluten is mixed with all play a part in how many odor molecules are being released from the gluten source and are reaching the dog’s nose. For the dog, it’s all about the number of odor molecules that make it up the schnoz and much less about the mass of the substance itself. For the person, the amount of gluten ingested is critical. So figuring out if the dog is going to be able to detect at a level the person really needs is a bit like comparing apples and oranges.
2.) Dogs detect all odors separately.
This means they smell the lettuce, the carrots, the cucumbers and the residue from some croutons someone accidentally put on your salad and then took off when they remembered you ordered a gluten free meal. That is true, but extraneous odors DO decrease the reliability of the detection- a host of scientific studies show this to be true. Extraneous odors are all the other odors besides the one the dog has been trained to detect. The more extraneous odors there are, the more distracting they are to the dog. Common sense and experience (not experiments) show that the more delicious those extraneous odors are, the more distracting they are to the dog as well. It is going to be a lot harder to get reliable results searching your steak dinner than your salad! Not that it is impossible, not that it is not trainable, just that it is harder to get reliable results. (Focus here is on reliability, not what is possible.)
3.) The places we want to use dogs the most are the areas of least reliability.
Most people really want to use dogs to check food in restaurants, but these restaurants are usually LOADED with gluten. Whether is pasta boiling, bread baking or just all the pastries in the coffee shop, the air is likely to be heavy with the odor of gluten products. Asking a dog to detect tiny amounts of cross-contamination in these environments is difficult at best. It is possible, something that we train for, but it is not as reliable as detection in other areas that are not “super saturated” with the trained target odor.
To put this task in human terms, think of it in visual terms: it’s easy to see your Indiglo light on your watch in the dark, when there’s not a lot of light around. Your accuracy at detecting when your watch is lit up is probably 100%. But, the more light that is in the surrounding environment, the harder it is going to be to tell if there is any light coming off your watch. When you are in a super bright area, your accuracy at detecting this tiny amount of light is going to fall. You may help overcome these difficulties by putting your eye right up to it, shading it with your hand to block out some of the surrounding light, or moving into a darker place. These would be compensatory measures, and we can figure out a few ways to do things like this for the dog too, but it is still difficult. This isn’t a perfect example because noses and eyes are different, but I think it gets the point across for us non-scent focused creatures.
There’s more science out there to be sure. There are more arguments out there both for and against using dogs for gluten detection, but the long and short of all of it is that it’s not perfect and it never really will be. But if it helps, I’m all for it.