The BOAS tips below come from one of the interviews I did at Science Week last year. I'll be back at Science Week 2026 - this time not only podcasting, but also speaking. Come and catch me in the general practitioner stream for my ‘5 things I’ve learned through making the Vet Vault that have changed the way I practise’ talk, and ‘Is ChatGPT a better vet than you?’, where I’ll review findings from my AI as a clinical assistant tool research. If you’ve never been to Science Week, you are missing out. The mix of cutting edge full-nerd-mode content and very practical GP level stuff makes it one of my favourite educational events of the year.
If you like a bit more adventure with your CPD... I’ve just confirmed Prof Jill Maddison and Prof David Church as our speakers for our Africa ‘27 conference in April. Ngorogoro Crater, Serengeti, days starting with a game drive, followed by top notch CPD, all wrapped up with sunset African splendour and a stiff G&T. Keep an eye on vetsontour.com - ticket sales are opening up in the next day or two.
From Episode 220 Surgery Feed With Dr Ema Bowman
I’ve written before about owner-assisted recovery for brachycephalic surgery patients, where you have the owners present as their precious little freaks come around from their anaesthetic to stop them from spazzing out, as they do. But what we have not discussed is what you do with your BOAS surgery patient once it’s awake.
One of the very first episodes we ever did on the Vet Vault clinical series was a conversation about upper airway surgery in brachycephalics. This was back in 2020, and our specialist surgeon guest was horrified to hear that some GP surgeons were sending these patients home on the day of surgery. I’ve babysat a fair number of post-op BOAS patients in my time working in emergency, and I’ve certainly seen more than a few go from “I’m okay” to “I can’t breathe!” within minutes. I’d hate to be in the room if that happened at some doting owner’s home.
Which is why I found this concept from our conversation about anaesthetics in obese brachies with Dr Emma Bowman - who spends her days exclusively doing BOAS surgery - so fascinating:
“I’ve just ticked over a thousand BOAS surgeries, and I have sent the vast majority of those dogs home the same day, without complications.”
Here’s the thing: in those patients I cared for that fell to pieces after surgery, the vast majority went south because the patient was a stress head. They’d wake up, start freaking out - puffing and panting and making that horrendous bird-like noise we all know and love.
What you don’t want when major surgery in your upper airways is more irritation and inflammation. All that panting and carrying on puts you right back on what Dr Emma calls the ‘hamster wheel of inflammation and obstruction.’
The result is that we pretty much have to re-anaesthetise these patients to keep them quiet - something that is not without risk, and feels kind of pointless, because at some point you’re going to have to wake them up and get them home.
So why not just... wake them up and send them home?
“You can send these dogs home the same day and they do really well. They’re monitored so well at home by their owners, because they get treated more like children than dogs anyway.”
NB: This is not a universally accepted opinion. Many specialist hospitals still very much want highly skilled eyes and well-equipped hands on their post-op BOAS patients during that immediate post-op period. But it’s also not a cowboy move - several specialist centres are now adopting this approach, and there is some literature to back it up.
From Episode 222 ECC Feed With Dr Jake Wolf
You’re not going to see a great many patients with carboxyhaemoglobinaemia, but for the one or two that you do, you’ll want to know about these pearls from our conversation about haemoglobin abnormalities with criticalist Dr Jake Wolf:
You’ll really only ever see it in a smoke inhalation patient - i.e. carbon monoxide poisoning.
Yes, your vague memories from vet school are correct: they can have brick-red mucous membranes. BUT - they often won’t. Mucous membranes can be completely normal in colour, even in the face of clinically significant carboxyhaemoglobinaemia.
Here’s something that will trick you: if you put the pulse ox on a patient with carbon monoxide poisoning, you will get a normal reading of 99-100% saturation, even while your patient is trying to die from decreased oxygen-carrying capacity. This is because carboxyhaemoglobin has the same light absorption wavelength as oxyhaemoglobin, so your machine will be confused. You will be too.
So how will you know?
1. Your patient will be in a terrible state: shocky, collapsed, with the respiratory pattern of an animal that’s not getting oxygen down to tissue level.
2. If you have a fancy blood gas monitor with co-oximetry, you’ll be able to measure carboxyhaemoglobin levels directly.
And, bonus pro tip: If you do have that fancy blood gas machine, consider looking more carefully at the results in your confusing anaemia patients. There’s emerging data showing that immune-mediated haemolytic anaemia patients can have significantly elevated carboxyhaemoglobin levels compared to other anaemia patients. Not enough to completely hang your hat on - but it might be the extra piece of information you need to support your diagnosis in a less obvious IMHA case.
From Episode 225 Medicine Feed With Dr Linda Fleeman
But it can’t have Cushing’s, because it’s not showing …. Complete that sentence with your clinical sign or lab finding of choice. How often have you thought that about one of your “should I run a stim test?” patients?
Dr Linda Fleeman will help you finish that sentence with this about the subtle signs of Cushing’s disease that most of us are probably missing (or conveniently ignoring!). Some things to keep on your radar when the next chubby/alopecic/panting/PU/PD dog walks through your door:
Cushing’s cases hardly ever present for Cushing’s signs. They’ll come in for a vaccination or some other random consult - almost never for “my dog has a pot belly and is panting and drinking lots” - until it’s very far advanced. In other words: if you keep your eyes open, you’ll notice it before the owner does.
Speaking of clinical signs: did you know that PU/PD is a late-onset sign of HAC? AND - because it happens gradually, when you see that chubby dog with the naked belly and ask the owner “is she drinking more?” - they’ll likely say “no, no more than normal.” It’s the frog in boiling water.?
“No matter what clinical sign a dog or cat has for Cushing’s, everyone goes, ‘but it doesn’t have this one.’” But nobody said it has to have them all. The alopecic dog with a normal ALP an no pot belly? Probably needs a stim test. The dog with weird coat changes but no PU/PD? Stim test. You get the picture.
“I think we just need to stop trying to avoid the diagnosis. It’s a common condition that causes clinical signs and causes progressive decline of health, and treatment results in improved quality and length of life.”
2 Other Things
“Somebody's boring me. I think it's me.”
-Dylan Thomas
“A key prediction of the theory is ‘cognitive surrender’ - adopting AI outputs with minimal scrutiny, overriding intuition and deliberation.”
-Quote from this study
I want to double click on that term from our second quote - ‘cognitive surrender’ - as it applies to AI, and interrogate it with a veterinary lens.
It’s no secret that I’m bullish on AI in our profession. I’ve been thinking and talking about - and living - the challenges of vet life for close on 30 years now. Some days I feel like we just talk in circles about the same old problems - the pressure, the burnout, the money. True - many things have changed for the better. Many have not. And where there is progress, it’s often at glacial pace. But now - here is this tool that has the potential to have very real, very immediate impacts.
As a profession, and as individuals, we’re begging for it: high consequence, high cognitive load, limited resources, real time pressure, emotionally loaded environment. If AI can deliver on even half of what it promises, then, as they say: shut up and take my money.
But that term ‘cognitive surrender’ finally gives language to a niggle about AI that I’ve found hard to verbalise. Specifically when it comes to the core of our work as clinical vets: the actual diagnosing and treating of animals - vetting.
Hands up if you’re experimenting with an AI tool to help you with this part of the job.
🙋♂️
It’s good, right? How good? Well, I’m midway through my own little research project where I’m testing several LLM models in their ability to reason through and make recommendations on complicated medical cases, and then comparing their answers to what human GP vets would do. Not to get ahead of myself, but so far I’m pretty certain that for the most part, in 2026, AI is really good at vetting. Not perfect - not specialist level, but probably better than you. (Come listen to my talk at Science Week if you want to argue with me.)
I once interviewed someone about antibiotic stewardship. We were discussing why making informed, thoughtful decisions about antibiotics is so hard for most of us, and my guest said something that has stuck with me:
I think that right there goes a long way at explaining why AI outperforms us at reasoning. It’s not necessarily smarter, just less rushed, and less lazy. It’s never tired or distracted or upset because it missed lunch. And it has access to ALL the facts. (When was the last time you read a journal article?)
We should probably be using it. We’re still figuring out how exactly, but not incorporating AI into your clinical decision making is a bit like having a super smart, very up to date colleague in the building with you, then never asking them for any help.
If we using AI, as our quote above states, to think for us ‘with minimal scrutiny, bypassing both intuition and deliberate reasoning’, will we become dumber?
Here’s the catch-22 ahead of us: ignore the fact that we’re lazy and just focus on the ‘busy’. Busy, and also completely incapable of keeping up with what’s new. As my own knowledge increases through doing the clinical podcasts I am increasingly aware that my standard of care is often limited by the time constraints imposed on us by the veterinary business model. I know enough to know that there’s a lot more to know. There are all these things I want to do, and there’s so much to communicate with the client - who, by the way, has higher expectations than ever. Is it any wonder then that I’m rushing into the warm (cold?) embrace of cognitive surrender? AI could offer relief, efficiency… and potentially a silent downgrade in thinking. Because AI doesn’t just help us think better. It changes whether we think at all. (There’s already some early data on this outside of the vet space.)
In surrendering, will I atrophy those very skills and insights that are driving me into the arms of AI reasoning, and end up more dependent on AI as a result?
And in the process, I’m also removing one of the most satisfying bits of being a vet.
Can I (should I be allowed to) choose between the promise of better patient outcomes and a less exhausting workday on one hand, and my own development and personal satisfaction on the other?
Here’s an article about a German language professor at Cornell University who became so exasperated by watching her students hobble their own growth by using AI to do their assignments that she now forces them to complete coursework in the classroom - using only typewriters. What she’s trying to preservein her students is exactly what’s at stake for us: the act of engaging with the friction of thinking. (The students like it, by the way - their main complaint being that their feeble little touch-screen fingers weren’t strong enough to type on the stiff contraptions.)
Which has got me thinking: what are the equivalent of the professor’s typewriters in veterinary science? And how do we use our ‘typewriters’ in a way that stops us from getting dumber and yet supports our workflows to be better at our job?
I don’t have the answer - I really do want you to think about this. Here are some initial ideas to get you thinking. These are not anti-AI. They’re anti-surrender.
Think first, double check with AI. Do the hard work of mulling a problem over. Use AI to help you find information, but do your own reasoning. Then get AI to check your reasoning and find the gaps in your knowledge and the places where you were cognitively lazy. LEARN FROM IT, so that you can…
Get better than AI at a few things. There’s a very valid school of thought posting that AI will never replace genuine expertise. It’s baked into how AI models work: they give the most likely, most common answers. Experts are experts because they can pick the not-so-common edge cases. Become an expert, and let AI make you average-good at all the rest.
Keep learning. Active, ongoing learning, for the reason above, but also just for the sake of learning. Not that very long ago the idea of going for a run when you weren’t chasing or being chased would have been preposterous. And pointlessly lifting heavy weights or pretending to box in a gym… someone call the asylum! But now that we are whisked around in cars and do most of our work on computers, we’ve figured out that it’s very bad for us to never use our bodies. I think it will be much the same with our brains. Veterinary science makes a pretty fun brain gym. (There’s this very good veterinary clinical podcast I know about…)
The hard part isn't whether to use AI. It's deciding what you're not willing to hand over. I'd love to hear where you've drawn that line.
Much love,
Hugh
How to connect with us:
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.