A surgical solution first proven in dogs has now reached human patients. In the Netherlands, surgeons recently implanted a custom 3D-printed titanium hip socket in a young adult with hip dysplasia—an approach shaped directly by years of veterinary orthopedic experience.

Hip dysplasia affects both dogs and people, often leading to cartilage damage, chronic pain, and osteoarthritis. In veterinary medicine, traditional correction can require multiple invasive pelvic cuts and lengthy rehabilitation. Seeking a less traumatic option, veterinary and human surgeons collaborated on a patient-specific implant designed to restore hip stability by augmenting the socket rather than reconstructing the pelvis.

The results in dogs have been striking. More than 70 canine patients of varying sizes have undergone the procedure, frequently with both hips corrected in a single surgery. Many are mobile within hours and discharged the next day, dramatically reducing recovery time and postoperative discomfort.

That veterinary success provided the confidence—and data—needed to move into carefully monitored human safety trials. While recovery demands may differ between two- and four-legged patients, the leap underscores the growing role of veterinary medicine in shaping solutions for shared diseases.

This is One Health in practice: insights gained in animal patients directly informing safer, less invasive care for humans. This is also a strong reminder that veterinary patients aren’t just beneficiaries of innovation, they’re often the reason it exists. This work reinforces the clinical, scientific, and translational value of veterinary medicine in advancing care across species.

For thousands of years, dogs and cats have been our dominant companion animals. But science suggests the story of domestication may be far from over.

A recent Gizmodo Giz Asks feature explores whether wild animals like raccoons could be on a slow path toward domestication. Researchers point out that domestication doesn’t always start intentionally. Cats, for example, evolved alongside humans by exploiting human-made environments, gradually becoming less fearful and more sociable over generations.

Urban wildlife today may be undergoing a similar process. Studies show that raccoons living close to humans are developing physical and behavioral changes linked to reduced aggression, a possible early signal of “pre‑domestication.” Foxes and opossums may be following comparable paths.

That said, experts are skeptical we’ll see a new pet achieve the global success of dogs or cats anytime soon. Those species already dominate the human-animal niche, making competition difficult. Still, local, niche domestication, especially in cities, may be more likely than we think.

The takeaway? Domestication isn’t a fixed historical event. It’s an ongoing evolutionary conversation between humans and the environments we create.

As cities expand and ecosystems change, the animals living among us may be adapting faster than we realize.

What species do you think might be next? Shoot us a reply to let us know. We are fine with Dinosaurs as the answer…

I have a confession… I hate statistics.

But it is pretty clear that one of the highest leverage things you can do to help understand AI, is to understand statistics. Particularly, understanding the metrics which matter most for clinical AI. That is why this week’s Field Notes focuses on the most recent issue of my favourite medical AI newsletter - Machine Learning for MDs.

Since 1954 it has been common knowledge that we can lie with statistics. But in the age of AI, this has proven even more pertinent. In this recent article “Three Metrics for Healthcare AI Evaluation,” physician and AI evaluation expert Dr. Sarah Gebauer makes a compelling case that one of the biggest threats to trustworthy AI isn’t a lack of data, it’s the wrong use of statistics. Healthcare AI vendors often lead with headline numbers like accuracy or F1 score.

While these metrics sound impressive, they can be actively misleading. As Gebauer notes, recent reviews in The Lancet Digital Health found that many commonly used AI performance measures are “improper,” meaning they can reward flawed models and obscure real‑world risk. Instead, she calls for focusing on three distinct metrics, each answering a different and essential question:

AI transparency isn’t about showing more numbers, it’s about showing the right ones. Without understanding what each metric represents (and what it hides), clinicians, leaders, and regulators can’t meaningfully evaluate risk, benefit, or accountability. Without transparency and the right metrics, clinicians can’t truly get informed consent for the AI systems they are using.

Unfortunately for me, statistical literacy is a prerequisite for AI governance, trust, and safety.

-RBA

Here are some of the other stories that caught our eye and we're following this week from around the veterinary world and animal kingdom: