First steps

Ellile is ready to go

I started my journey on a walk with our dog Ellie, while listening to a podcast. The podcast was Revisionist History, and the episode was “The Dog Will See You Now”. In it, Malcolm Gladwell explores the use of dogs to detect the presence of cancer and Covid-19 in humans. He imagines a world in which you needn’t submit to the awfulness that is the prostate exam or colonoscopy, a world in which Covid testing results are immediate. This world is possible because of dogs and their unimaginably precise sense of smell.

This was a case of great timing, when I was thinking about doing something dramatically different and was open to all possibilities. The concepts in the episode weren’t new to me, as I imagine they aren’t to you. I knew that academic studies had demonstrated that dogs can accurately detect health issues. I was aware of K9 units in law enforcement. And like most international travelers, I’ve been on the sniffing end of canine-assisted customs searches and discovered how much more pleasant an inspection is when it’s conducted by a dog walking around the baggage carousel than by humans rifling through luggage for non-existent contraband.

Although I had an awareness of dogs being used to assist us with their sense of smell, I had never been open to thinking about this as something I could contribute to. Suddenly, I could. Since that day, I’ve been reading and learning. Today, here are three of the most notable things I’ve learned.

All cancers share a unique odor (according to dogs)

Absorbing this idea requires a fundamental shift in how we think about diagnosis. So let this sink in. Dogs trained to identify by smell one specific cancer extrapolate that smell into the ‘cancer odor.’ Trained on one, they can identify multiple different types.

The Prostate Cancer Foundation has been early to invest in canine detection. One of the foundation’s blog posts tells the story of Claire Guest, a psychologist, registered pet behaviorist and co-founder and Chief Scientific Officer of Medical Detection Dogs, and her dog named Daisy. Trained to detect bladder cancer, Daisy quickly learned to detect the cancer odor, and was able to detect prostate cancer. She also detected breast cancer (‘very early, very deep’) in Guest herself, which may have saved her life. Daisy wasn’t trained on each different type of cancer she successfully detected. She figured out what cancer smells like.

Modern medicine is built on specializations: oncology itself is a specialty in the medical profession, and oncologists specialize to a single type of cancer. Screening and diagnostic tests are all specialized by type of cancer: they look for different markers, they use different technologies. Medical standards for health maintenance in women include regular mammograms, pap smears, colonoscopies, not to mention the monthly at-home breast checks for every woman of every age.

We are so used to thinking about cancers as different diseases requiring different diagnostics that it is truly stunning to consider that there is a cancer signature odor that dogs can detect. A breath or urine sample could be sniffed by a dog who could immediately detect the presence or absence of the cancer odor. Quite aside from the cost and time savings, the experience itself would be transformative. If the screening process were more humane, how many more people would be screened, and screened earlier? How many lives could be saved?

The dogs’ accuracy isn’t just hype, and how accurate are the medical screenings we depend on?

Assessing the accuracy of a screening method uses two measurements: sensitivity and specificity. Sensitivity is the accuracy rate of positive results (presence of the disease) and specificity is the accuracy rate of the negative results (absence of the disease). Study after study has demonstrated very high sensitivity and specificity rates in dogs’ detection of cancer. Here’s a sampling (links are courtesy of cancerdogs.ca):

How do these results compare to traditional screening methods?

All these years, I’ve been trusting that the health maintenance screening I’ve undergone has been spot-on accurate. Not only did I overestimate the accuracy of these tests, but also, I was completely unaware of a more accurate diagnostic tool, in the form of a trained dog, no stirrups or speculum needed.

It doesn’t cost an arm or a leg (but treats are always appreciated)

The labor and skill required to hone the skills of a detection dog are significant, and dogs who do this work can generally only work about three 30-minute shifts in a single day. This is intense cognitive work for the dogs. They must eliminate all the other scents that aren’t the target scent: that is, they must smell and dismiss all those odors that are specific to the human that gave the sample plus all other competing smells in the room.

The people who train the dogs must invest in extensive training, and in turn each dog must be trained over time and the training reinforced. The people who handle the dogs are also trained, and each team (dog and handler) must be certified and regularly re-certified.

At first, I assumed the delivery costs must be high, as concierge services are. In our economies, we’re used to paying more for a service that is personally delivered than for an automated technical solution. While the investment cost is substantial compared to dog obedience training, it’s a pittance compared to medical screening costs. In the Revisionist History episode, Michael Mina of the Harvard School of Public Health estimates the cost of canine detection of Covid is about $2.50 per screening, which covers all the concierge costs listed above. As a bonus, the results are immediate.

A sniff of a hand, versus a giant Q-Tip up your nose. Which would you prefer? If you haven’t already, do listen to the podcast, and click through to the links above. There are many people working on the promise of this idea. Listen to them. This isn’t a crazy idea; it’s a smart one.

One thought on “First steps

Leave a Reply

Your email address will not be published. Required fields are marked *