The Visit He Chose to Leave

The room was exactly the kind of place where a therapy dog was expected to shine. Six residents sat in a loose circle while afternoon sunlight warmed the carpet, and the faint smell of coffee drifted in from the hallway. A woman in a lavender sweater was already smiling before the dog reached her. Another resident patted the empty space beside his wheelchair. The dog greeted them willingly, his body loose and his tail moving softly behind him. His handler watched with quiet satisfaction. Then, less than fifteen minutes into the visit, the dog walked away.

He did not bolt for the door or hide behind his handler. He simply stepped out of reach and stood several feet away. When someone called him back, he glanced toward the group but did not return. His handler could easily have shortened the leash and guided him forward. The dog knew how to behave. He probably would have complied. Instead, the handler loosened the leash and watched. After a moment, the dog turned toward the hallway. The handler thanked everyone, said they were finished for the afternoon, and followed him out.

That small decision reaches into one of the most important conversations taking place in canine welfare research: the difference between a dog who will do something and a dog who wants to do it. A 2025 pilot study specifically examined touch involving therapy dogs. When dogs were given the opportunity to move away from human interaction, they mostly chose to remain within reach. During forced-touch conditions, however, researchers observed more frequent ears-back behavior. The authors concluded that choice and agency are important considerations in therapy dog welfare. (ScienceDirect⁠)

Newer research makes the question even more interesting. A 2026 study examined 837 animal-assisted service sessions involving 63 dogs. Researchers found that canine responses varied according to the individual dog, handler, client, type of activity, experience and other circumstances. Another 2026 analysis of those sessions found that overt stress behaviors were generally within expert-defined acceptable ranges, yet positive behaviors such as play, voluntary lying down and broad tail wagging were often less frequent than expert-derived thresholds suggested they ideally should be. The finding does not mean those dogs were suffering. It raises a more subtle welfare question: Is the absence of obvious distress enough, or should good therapy work also provide opportunities for dogs to demonstrate positive engagement? (MDPI⁠)

That distinction matters because therapy dogs are often exceptionally cooperative animals. Their reliability can fool us. A dog may stand still while someone hugs him, remain beside a wheelchair when he would rather move away, or return when his handler quietly asks him to. Compliance can look remarkably similar to willingness unless the handler pays attention to what happened immediately before the command. Research published in 2025 observed avoidance behavior at least once in every recorded intervention among six dogs studied across seven sessions, even though most handlers rated their dogs’ stress as low. The sample was very small, so broad conclusions would be inappropriate, but the researchers identified handler education and advocacy as important welfare considerations. (PubMed⁠)

This does not mean every head turn, yawn, sniff or step away should end a visit. Canine behavior has context, and no single gesture tells an entire story. Nor should we assume therapy work is inherently stressful. A randomized prospective study of 14 therapy dog teams working in inpatient stroke rehabilitation found no evidence that the sessions induced overall stress, and some physiological measurements suggested the dogs were more relaxed afterward. (PubMed⁠) The lesson is not to become suspicious of every movement. It is to become curious about what our dogs are telling us.

In the hallway, the handler crouched beside his dog. The dog shook his coat, stretched his front legs across the cool tile and looked toward the exit. There would be another visit another day. No one had been cheated of anything. In fact, something important had been protected.

A therapy dog should never become a prisoner of his own goodness. The finest partnership is not built on how much a dog will tolerate because we ask him to. It is built on trust that runs in both directions. We ask our dogs to listen to us in crowded hallways, unfamiliar rooms and difficult moments. They deserve something equally sacred in return.

When they quietly tell us, “I’ve had enough,” we listen.