When Martin Levy, an orthopedic surgeon and Border Collie enthusiast, began training his dogs to navigate agility courses nearly 20 years ago, he never expected to one day use the same techniques to train medical residents.
“Over time, I started to realize that we had better tools for training our dogs than our residents,” says Levy, the residency program director for orthopedic surgery at Montefiore Medical Center in the Bronx. Levy decided to use these tools to help new doctors learn the tricks of the trade.
When the American Board of Orthopedic Surgeons challenged residency directors a few years ago to come up with teaching tools to help new doctors quickly gain skill at basic surgical techniques—tying knots, positioning surgical instruments, and handling power tools—Levy teamed up with animal trainer Karen Pryor, one of the founders of the clicker training movement to come up with a teaching program he could implement with his residents. He broke down basic procedures, like drilling into bone, into simple steps: how to grip a drill, how to position it at the correct angle over a practice PVC pipe, and how to stabilize the drill tip. When a young doctor performed a movement correctly, the instructor noted the achievement with an event marker—a click, the flick of a flashlight, or simply the word “good” spoken in a neutral tone.
The event marker is a non-emotional stimulus, says Theresa McKeon, study author and founder of TAGTeach International, a company that has adapted clicker training techniques for human learning. “In that moment, the event marker lets the student know what they are doing is correct without having to process a lot of emotion at the same time,” she says. “That’s a key part of the learning experience.”
Clicker training has its roots in the reward-based operant conditioning techniques pioneered by behaviorist B.F. Skinner in the 1930s—the idea that new behaviors can be learned through positive reinforcement. For dogs, the click of a mechanical device marks a desired behavior, such as sitting on command. Once a dog hears the click, the owner rewards him or her with a treat. The treat elicits a pleasurable response in the dog. The owner then repeats the sequence until the dog begins to associate sitting on command with the positive outcome and no longer needs the treat. Over time, performing the behavior correctly itself becomes the reward.
With surgical residents, operant learning rarely involves a clicker (and never a dog biscuit), but the basic principle—reinforcing proper technique with reward—stays the same. Traditional surgical teaching methods, says Levy, have focused far more on deriding residents for making mistakes than acknowledging skills performed correctly. “There’s a saying in the field: without tension there’s no retention,” he says.
The findings are plausible, says psychologist and applied behavior analyst Susan Friedman, a professor emeritus at Utah State University, who now leads workshops and seminars on animal behavior and learning. Rewards act to maximize behaviors where punishments act to minimize them, she said. “When your motive is to escape an adverse stimulus, such as shouting or ridicule, you behave only as much as you need to, but when the motive is to get something you want, you do more than is needed,” says Friedman, who was not involved in the experiments.
Where dogs may need an additional reward, such as a treat, to reinforce a behavior, surgical residents understand the stakes from the onset, and need only achievement to reinforce proper technique. “For a highly motivated individual, having a teacher acknowledge that you hit your target is in itself a pretty huge reward,” says Levy.
When Levy analyzed two specific behaviors, tying and locking, sliding knot, and making a low-angle drill hole, he found that residents in an operant learning group and a traditional learning group could perform the two tasks at about the same speed. However the students in the operant group were more precise in their movements, and performed the tasks more proficiently. He published the findings in the Clinical Orthopaedics and Related Research last fall and says that many of the teaching doctors at Montefiore now favor the new techniques.
The next step, he says, is to measure the relative success of their operant teaching methods beyond the pilot study. “When we talk to the attending surgeons, they can say anecdotally that residents who learned with the new methods seem better prepared,” he said. Levy and colleagues are now working on studies to quantify and analyze how quickly residents become fluent at their new skill sets using these methods.
Levy isn’t the first to apply reward-based operant conditioning to human learning. Modified clicker training has been used for gymnasts, golfers, veterinarians, and children with autism. In addition to dogs, similar techniques have been used to train marine mammals, horses, birds, and other companion animals. Operant conditioning may be effective across different species, says Friedman, because “all animals are biologically prepared to behave for an effect. Changing what one does based on the consequences of that behavior is a fundamental phenomenon of the natural world.”