Victoria Blackwell watches her 4-year-old daughter’s continuous glucose monitor at night until she can no longer keep her eyes open. Then her husband takes over. The Tylertown, Mississippi, family hopes a trained Labrador named Emma can eventually add a working nose to that overnight watch.
Hudsyn “Huddy” Blackwell was diagnosed with diabetes at age 2. She wears a continuous glucose monitor that tracks her blood sugar around the clock, but her mother says control remains difficult. Huddy has also been hospitalized for diabetic ketoacidosis, a serious and potentially life-threatening condition that occurs when the body lacks enough insulin.
The family says Huddy experiences severe glucose fluctuations, particularly overnight, and that her monitor’s readings can sometimes lag or update late. Their concern is that an alert may arrive only after her blood sugar has already moved dangerously low or high.
It would be another set of eyes a nose, rather that would help us keep an eye on her, Victoria said of Emma. Because I swear, some nights it feels like I’m watching her monitor, and when I can’t hold my eyes open, I’m passing it off to my husband.
Emma is a 2-and-a-half-year-old trained Labrador selected to monitor blood sugar levels and alert the family when Huddy’s levels drop. The dog costs $30,000. At press time, the family’s fundraiser had collected about half that amount, and Emma had not yet been purchased or placed with Huddy.
What a diabetes alert dog can and cannot do
Diabetes alert dogs are trained to perform an obvious attention-seeking behavior when they detect an odor associated with glucose moving outside a target range. Depending on the program and dog, an alert can involve staring, pawing, licking, nudging or retrieving testing equipment.
The odor mechanism is still being studied. In a preliminary University of Cambridge investigation involving eight women with type 1 diabetes, researchers found that the naturally occurring breath chemical isoprene rose during hypoglycemia. They proposed that dogs may detect that chemical change, although the finding did not establish isoprene as the sole cue used by working dogs. The Cambridge research instead offers a plausible biological basis for scent-based alerts.
Training requires more than recognizing an odor. One established protocol paired hypoglycemic scent samples from a future handler with rewards, then reinforced specific alert behaviors. After placement, handlers were expected to check glucose following an alert and reward the dog only when the reading confirmed that it was appropriate. That distinction matters because reinforcement continues to shape performance throughout the dog’s working life.
The strongest available studies also argue against treating a dog as a replacement for monitoring technology. An owner-independent study of eight accredited diabetes alert dogs found that dogs alerted more frequently when glucose was outside the handler’s target range. However, sensitivity to hypoglycemic episodes varied from 33.3% to 91.7%, with a mean of 55.9%. The researchers concluded that these dogs can be an additional part of a diabetes plan but should complement technology rather than replace it.
Overnight performance deserves particular caution in Huddy’s situation. A separate study of 14 diabetes alert dogs compared recorded alerts with continuous-monitor data. Sensitivity during sleep was 22.2% for low-glucose events and 8.4% for high-glucose events, and only one of 11 dogs evaluated overnight performed statistically better than chance. Results varied substantially among individual dogs.
These findings do not determine how Emma would perform for Huddy. They do establish the appropriate frame: a trained dog is a living member of a working partnership, not an infallible alarm. Reliability can depend on the individual dog, the handler’s response, continued reinforcement and ongoing assessment. For the Blackwells, Emma is intended as another layer alongside Huddy’s continuous glucose monitor not a substitute for it or for her parents’ oversight.
The family had reached roughly half of its $30,000 target, helped by relatives, friends, neighbors and other supporters. That progress leaves the central tension unresolved: Emma has been chosen and trained, but the Blackwells had not yet secured her for Huddy.
What’s your dog’s job, official or self-appointed? Tell us in the comments.
By Tom Whitaker — Outdoor magazine editor who has reported on search-and-rescue, herding, field dogs, and working-dog handlers.
