💊Methotrexate, 🔪Murder and 🤔Mysteries

Okay, here’s the honest truth.
I wrote the best intro that The Handover has ever seen. It had it all. The structure. The rhythm. The self-aware dry humour. But my dog literally just ate it 💔.
Dang Nabbit.

But… he didn’t fancy the medical news much. So here you go(just make sure read around the saliva.)

👋 Happy Friday. Here’s what we got:

If you want to read any previous editions of The Handover, you can on our website.

So you’ve prescribed methotrexate for your patient with rheumatoid arthritis.
Good. Sensible. It’s been the cornerstone of Rheumatoid Arthritis management for decades.

Cheap, effective, dependable.

Problem is, it’s got all these pesky side effects:
Mucositis, myelosuppression, pneumonitis and fibrosis keep popping up. It’s not exactly the friendliest of drugs.

Which is why we monitor. Relentlessly.

FBCs. LFTs. U&Es. The whole biochemical surveillance package. Unfortunately, these blood tests behave like toddlers. Leave them unsupervised, and you can be absolutely certain something expensive is being damaged somewhere.

In a study published in Arthritis Research & Therapy, researchers conducted a retrospective analysis to assess methotrexate's impact on kidney and liver-related adverse reactions in RA patients.

They looked at 10,319 adverse drug reaction reports where methotrexate was the suspected culprit. Outcomes were categorised as either:

So what did they find?
Out of those ten thousand cases, 1,082 were liver-related, 365 were kidney-related, and 67 involved both. On paper, liver toxicity was more common. 

But when it came to deaths, the kidneys were ahead. Among kidney-related side effects, fatalities occurred in 21.1% of cases compared to only 5.8% with liver toxicity. Suddenly, the liver looks like the least of your worries.

Here are the additional takeaways:

In their own words, the authors put it plainly:

"Because drug management in patients with RA using methotrexate is a complex matter, precise and standardised recommendations on when and how frequently renal function needs to be tested to detect early signs of renal impairment might be helpful to prevent fatal outcomes."

TLDR: Whilst LFTs are important for monitoring, maybe don’t let the kidneys feel left out.

My loyal servant of the National Health Service.

ChatGPT is good and all. But it’s just not for you.
By all means, feel free to ask it how to make a matcha, or how to survive yet another GMC appraisal 🙄

But DO NOT ask it a question about healthcare, for it could confidently lead you astray.
Citing American guidelines as its source. Or worse yet, confabulate the whole thing like a drunkard with Korsakoff's!

You needn’t restrict yourself to trawling through guidelines or asking your reg for help…

Instead, you can ask Medwise.ai

Imagine if ChatGPT entered a polyamorous relationship with NICE guidelines, Society Consensus Statement, and Royal College publication.

Then, 9 months later, popped out a beautiful AI baby trained on everything you need in your practice. 

You type a question and bam:
Instant Evidence-based answers
Accurate. Sharp. Clinically on point. 

It’s used by over 2,000 NHS organisations.
It’s completely free.
If you’re still not convinced, you get free CPD points just for asking questions too!

Start looking dangerously competent on the wards and use Medwise.ai today.
And if you use your NHS email address, you’ll get access completely free.

Try Medwise for free using the link below 👇

That, in essence, appears to be the message from the Department of Health and Social Care to GP practices that have yet to prescribe a single weight-loss injection.

Whilst everybody and their nan seems to want these drugs, and many a GP are happy to oblige them, there is a small group of GP practices who are yet to dish them out.

To sweeten the deal, the DHSC have made some changes to the Quality and Outcome Framework(one of the parameters that influence how much a GP practice gets paid), which incentivises prescription. 

An additional £3,000 will be paid to practices that prescribe Mounjaro(tirzeptide) to the maximum number of eligible patients, and £1,000 for simply referring them to weight loss clinics. Easy money 💸

Whilst it sounds like a good start, many doctors remain sceptical:

“While the headlines promise much, in reality, there will be no change to NHS England's eligibility criteria for patients to access injectable weight‑loss medication on the NHS,” - Dr Bramal of the BMA

"GPs do not withhold treatment or prescribe based on financial incentives. Decisions are guided by clinical judgement and what is safest and most appropriate for individual patients.” - Prof Victoria Brown of the RCGP

I mean, you can almost hear the SJT stem being read out in a monotone voice:

You are a GP partner in a busy practice. The Government has introduced financial incentives to increase prescribing of GLP-1 weight-loss medications. Your practice has not yet prescribed any. What do you do?

A) Prescribe to all eligible patients immediately to maximise practice income.
B) Refuse to prescribe on principle.
C) Consider each patient individually, following clinical guidance and discussing risks and benefits.
D) Forget it all and move to Australia.

Pretty hard choice ngl.

The tech industry has never once exaggerated a product’s capabilities to bump a stock price. Never. Not even once.

So when OpenAI unveiled “ChatGPT Health” earlier this year – a consumer-facing AI tool that connects to your Apple Health data and tells you whether your fatigue is caused by doomscrolling or Stage 4 metastatic colon cancer, we were hardly surprised. 

OpenAI insists it’s not replacing doctors. If something serious pops up, it will absolutely, definitely, pinky-promise refer you to a medical professional. 💯👍

So it’s a good thing that the good researchers at Mount Sinai decided to legit check Sam Altman’s latest side quest. 

This paper, published in Nature Medicine, aimed to do two things:

They fed GPT Health with 60 clinician-authored case vignettes across 21 specialities, generating 960 total responses. Three doctors created the mark scheme for each case(gold-standard triage) using a four-level scale: A (home), B (routine GP), C (urgent 24-48h), D (ED now).

And the findings were concerning to say the least…

Whilst it did well with the routine cases (93% accuracy), it failed very poorly with the level D case, under-triaging 52% of cases. 

Plus, bias had a significant part to play.
When researchers introduced anchoring bias (e.g. adding that a family member had reassured the patient that “it’s probably nothing”), the model shifted toward less urgent care 11.7 times more often.

Encouragingly, race and gender had no statistical significance on triage.

All the same, I wouldn’t be so quick to trust this particular Chatbot with your healthcare decisions.

The Galleri test.
The liquid biopsy.
The final boss of blood tests. 
A single blood test that screens for over 50 different types of cancers.
Before a single symptom emerged. 

The test works by analysing cell-free DNA fragments in a blood sample – fragments released by cancer cells into the bloodstream. Then, with the help of machine learning and AI spits out a result to tell you whether you're likely to have one(or more) of 50 cancers or not. 

It’s giving Elizabeth Holmes, Theranos vibes. Except with actual research and fewer black turtlenecks. 

Prior trials showed it had a specificity of 99.6%(correctly ruled out cancer in 99.6% of people who did not have the disease), but a sensitivity of 62%(corrently identify cancer in 62% of people who actually had it). 

In order to see if their test is any good at scale, Grail collaborated with the NHS to recruit over 140,000 patients across 150 sites in the UK for the NHS-Galleri Randomised Control Trial.

The primary aim was to see if the test could create a statistically significant reduction in the combined rate of stage III–IV cancers in people offered annual Galleri testing plus usual NHS screening, compared with those receiving usual screening alone.

And… it wasn’t met 😕

The trial showed the Galleri test was unable to show a statistically significant reduction in later stage cancers diagnoses. Which is a genuine shame. For patients and for the company running the trial, who’ve lost $150 million in trial costs and seen their shares drop by 50%. Medtech is not for the weak!

Handover Over 🫡

If you liked it, tell your mates
If you hated it, tell your enemies.

We’re trying a new format today. Just an experiment, what do you think? Do you like it or never try ts again?

Before you depart(if you haven’t voted already), please let us know your role in the poll below - It means the world when you do 🫶