Problems to solve

For humans, smell is the third of our five senses (sight, touch, smell, taste, hearing). Smell alerts us to danger, such as smoke, gas leaks or even mold. It enhances our ability to taste and to experience the beauty of the natural world.

Although subconsciously we can detect the subtle presence of molecules like pheromones, our conscious use of smell is not especially discriminatory. Walking past a recently felled tree last week, my husband and I could smell the tangy, acrid smell of the cut wood. The tree had been enormous; we couldn’t miss the scent. Our dog Ellie? I have no doubt she not only smelled the tree but could also identify the specific squirrels who relieved themselves in the tree over time.

Although there’s no definitive ranking of all five senses for dogs, smell is the dog’s first sense. Dogs can smell volatile organic compounds (VOCs), identify them, and categorize them. They can recognize molecules in parts per trillion. This enables dogs to detect all kinds of things that we need to know are present, but cannot detect ourselves: explosives, drugs, missing people, and disease.

Law enforcement has built an impressive infrastructure for incorporating trained dogs into their efforts to detect the unlawful presence of explosives and drugs, and to assist in searches for missing people. Although scientific research has well documented the effectiveness of canine detection of cancers and Covid-19, and training facilities are actively working to develop these specific skills in dogs, we lack the necessary infrastructure to establish canine medical detection as a standard alternative to current technology-based screening procedures.

We need four essential components to make canine detection available on a large scale: standards, education, acceptance by the healthcare community, and – most importantly – market demand. Before diving into each of these, let’s pause to consider what scale might mean.Size of the needAccording to the World Health Organization (21 September 2021), the most common causes of cancer death worldwide in 2020 were:

  • Lung cancer (1.8 million deaths)
  • Colon and rectal cancer (935,000)
  • Liver (830,000)
  • Stomach (769,000)
  • Breast cancer (685,000)

In 2020, the US had 3,358,814 deaths total, 377,883 (11.3%) of the total were from Covid-19 (and 2021 deaths will surpass it). The World Health Organization estimates that the worldwide death rate from Covid-19 in 2020 exceeded 3 million, based on excess mortality, while 1,813,188 Covid-19 deaths were reported – a difference of 165%. Using this metric, Covid-19 is resulting in close to double the number of deaths as the leading cancer. Covid-19 testing In the 21 months since January 2020:

Data from

Testing alone does not bring down the death rate from Covid-19. As the data above show, the US has performed almost triple the number of tests per 1M people than New Zealand, yet its death rate is almost 420 times higher (more cases, more testing).

Achieving a reduction in contagion to eliminate viral spread requires comprehensive testing and contact tracing to identify the asymptomatic carriers of the virus, quarantine, and vaccination. Barriers such as masks and social distancing are required only because without comprehensive testing, it is unknowable who is carrying and shedding the virus and who is not. Areas of indoor public activities (schools, workplaces, entertainment, dining, shopping, travel) have become loci of hostility due to enforcement of medical masks and proof of vaccination.

If non-invasive testing with immediate results were established at sites of public congregation, these barriers could be removed and people could enjoy much more freedom than they do today. This is the size of the need that, over time, we must address. For Covid-19 detection alone, this works out to about 360 million tests/year in the US; 6 million in Canada. If we double that for cancer detection, that’s 60 million tests per month for the US, and 1 million tests/month in Canada.

That’s a lot of dog sniffing. We need to think big.


The National Institute of Standards & Technology’s (NIST) Dogs & Sensors Subcommittee is working to create standards. Their work is critical for establishing an authoritative set of standards such as these:

  • Best materials for screening accuracy: use of specimens or in-person interaction
  • How material is to be prepared and presented to the dog
  • Treatment of dogs
  • Training, certification, and ongoing re-certification
  • Documentation and transparency; publication of documents

These standards will enable consistency and reliability in medical use. It is essential for trust and accuracy.


Dog trainers working in medical detection are developing procedures and protocols based on scientific research and their own expertise. When standards have been created, it will be possible for trainers to work from the same script.

But how would it be possible to create a training infrastructure capable of meeting the immense needs for testing on a large scale?

In my professional career, I always found it more helpful to think about abundance rather than scarcity when trying to solve seemingly impossible problems. Thinking about what you don’t have only makes you feel powerless in the face of an obstacle. Instead, ask yourself: Where do we have rich resources that could be leveraged to address a problem?

First, we have a lot of dogs: the American Veterinary Medical Association estimates there are almost 77 million dogs in the US. We also have a lot of trainers who work with many of those dogs. About 15,000 people who are not self-employed identify their profession as Animal Trainers, according to the Bureau of Labor Statistics, and about 7,000 of them work with dogs. Since there are about 19,000 incorporated cities, towns, and villages in the US, counting only the 4,000 larger towns and cities with populations greater than 5,000, these 7,000 dog trainers would average fewer than 2 dog trainers per city. I assume there are more dog trainers than the official count, but whatever the actual number is, it’s not close to meeting the need.

You can call yourself a dog trainer without any formal education, certification, or license. However, associations and companies do offer training and certifications, which differ widely in the financial and training investments required. Their output is, as we can see in the employment numbers, not close to meeting the total need for medical detection.

Additionally, trainers who teach pet obedience skills are judged subjectively by their clients. For medical detection, more is needed.We also have an abundance of colleges and universities, which are sized to meet the market demand of the degrees they confer. (Those top-tier universities that reject more applicants than they accept do this not because they’re incapable of meeting applicant demand, but because they are perfectly sized to meet the demand of a relatively scarce commodity, the top-tier university degree.) Colleges could create new degree programs for canine medical detection, based on animal behavior and psychology, chemistry, forensics, and clinical science. These programs could be developed jointly with those working in dog training and canine detection already. Accreditation could standardize these programs, to ensure consistency and the robustness of the credentials awarded to graduates.

We could create armies of dog trainers, using methods and means that are well established. There is an abundance of possibility.

Enrolling the healthcare community

Healthcare infrastructure is heavily invested in the current technological methods of screening for cancers. It’s important to think of canine detection not as a replacement but as an augmentation of screening services.

When a dog detects the presence of a cancer, for instance, the information we get from the dog is limited: the dog takes an action (like sitting) to flag that the smell was detected. We don’t necessarily know what kind of cancer (if the dog has been trained on multiple cancer VOCs), or what stage the cancer is in, or where it is in the body. It’s just a starting point, although a critical one.

Any institution requires hard data to make a change. It needs to see for itself the cost-benefit results through the entire network – how it affects patients, healthcare providers, insurance companies and governmental agencies. Prototypes and pilots facilitate understanding, and they would be essential to making the case for healthcare to adopt this novel screening method.

Starting with under-served communities could be an effective strategy. The National Cancer Institute’s Community Oncology Research Program (NCORP) enables people living in under-served communities to participate in clinical trials for preventing, controlling, and treating cancer. The NCORP network serves 32 communities and 14 under-served communities. This program has created a contained and diverse healthcare environment that could be a terrific starting point to demonstrate the benefits of low-cost, reliable screening.

An informed public

Healthcare providers will only sign on to this alternative method if they’re convinced that canine detection services can be provided reliably. Trainers and handlers won’t invest deeply in canine detection services if they’re not assured of a market.

How do you build a market before the products or services are available? I think this starts with educating people over time, and I trust that an informed public will want this alternative for health screening, since it hits all the marks: simple, inexpensive, fast.

Building widespread understanding and acceptance will take more time, and people’s minds and behaviors will only be changed through narrative, not numbers. You can play a part in getting the story out.

Leave a Reply

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