iiinova — where your digital self keeps evolving

Raza @transfusion.health · 16/09/2026

I think it starts with just approaching and involving them, building them up, and eventually preparing them to lead Have to go out of our way for this to happen!

000

Raza @transfusion.health · 04/08/2026

Link to Japanese study?

000

Raza @transfusion.health · 08/06/2026

In medical science, we do a terrible job of enabling non-physicians to do healthcare research There is no notion of protected research time, specific-grant funds, or recruiting their perspectives in the things we write Meanwhile non-physicians vastly outnumber physicians in healthcare provision

130

Reposted by Raza

Canadian Society for Transfusion Medicine @cstm-scmt.bsky.social · 04/06/2026

Quote of the morning from our AI workshop: if all you can do is prompt AI, then AI *will* take your job. @transfusion.health #cstm2026scmt #transusionmedicine #AIinmedicine

111

Raza @transfusion.health · 30/05/2026

Your standards for writing are rather low if you believe this

000

Raza @transfusion.health · 24/05/2026

Same shoes. Make an AI agent do it for you!

000

Raza @transfusion.health · 24/05/2026

I miss Mart Janssen. A brilliant data scientist and wonderful human. I still read his papers and miss him.

000

Raza @transfusion.health · 20/05/2026

Today I discovered Bradley-Terry, a very cool way to preserve information while analyzing data from multiple pairwise comparisons. I naively used to think multinomial regression would work here! #stats en.wikipedia.org/wiki/Bradley...

en.wikipedia.org

Bradley–Terry model - Wikipedia

000

Raza @transfusion.health · 08/04/2026

Survival analysis is among the most beautiful areas of biostats. As a clinician I am blown away by the ingenuity of concepts such as partial likelihoods, martingales, subdistributional hazards, transformation models. How many lives have been saved because we had the right tools for the job!? #stats

130

Raza @transfusion.health · 06/02/2026

Over-optimism for AI comes partly from the coincidence that the people developing it (coders) happen to perform the economic task it is the most adept at (coding). This creates a very strong self-reinforcing loop of hype and (imo) naive over-extrapolation of capabilities to other domains

010

Raza @transfusion.health · 04/02/2026

Although I disagree with this author that bias isn't error (it is absolutely a component of error), I agree with their argument that deciding when to address bias should guided by alignment on how the model is used, not only its generic contribution to raw accuracy

000

Raza @transfusion.health · 28/01/2026

Thank you for the resource. I wish biostatisticians would publish some sort of consensus guidelines/heuristics that us dumb clinicians can rely on to check that we are mostly doing the right thing (I am told statisticians will never agree enough to produce such a thing). An old paper I did like:

220

Raza @transfusion.health · 27/01/2026

I find 'alpha spending' to be an incredibly awkward concept. How many hypotheses can I test? At what point does it stop being kosher? How much do I need to correct for multiple testing? How many tests can I run and how many I should just never test again? Is pre-specification the only answer? #stats

210

Raza @transfusion.health · 02/01/2026

One of the best pieces of advice I got as a doctor in training was to share my uncertainty and transparently lay out my thought process with patients. It goes against conventional wisdom that one may appear incompetent, unconfident, or indecisive. I've yet to meet a patient who didn't like it #meded

010

Raza @transfusion.health · 02/01/2026

If you have a statistician, ML engineer, or some similar methods expert on your project, give them commensurately high author line credit. They usually contribute far more and consequentially to studies than most other authors, yet often end up low middle-author or sometimes left off entirely #stats

010

Raza @transfusion.health · 10/12/2025

I think it’s potentially valid to do this in some even-point scales (I can think of some safety related endpoints). Makes 0 sense in odd-point scales

110

Reposted by Raza

JAMA @jama.com · 09/12/2025

Among adults undergoing cardiac surgery at high risk of bleeding and platelet transfusion, cryopreserved platelets did not meet noninferiority for controlling postoperative bleeding and were less effective than liquid-stored platelets. #CCRdownunder ja.ma/4rEC06z

JAMA article titled 'Cryopreserved vs Liquid-Stored Platelets for the Treatment of Surgical Bleeding'. Key points: cryopreserved platelets may be less effective/safe than liquid-stored for cardiac surgery bleeding; didn't meet noninferiority criterion.

034

Raza @transfusion.health · 10/12/2025

Headline feels inaccurate

000

Raza @transfusion.health · 09/12/2025

The way the exponential form generalizes so many common, seemingly unrelated distributions (and all the properties you get for free, like sufficiency) has to be one of the most elegant things I’ve learned in #stats, and in general Good reader: brunomaga.github.io/Exponential-...

050

Raza @transfusion.health · 23/11/2025

I agree, I think these are mostly distinctions of intent, not modeling approaches. It’s like when statisticians say all ML is just redressed stats, which is a valid statement if all you are considering is the model architectures. The distinctions are still helpful for real world practice

030

Raza @transfusion.health · 23/11/2025

I also like generative vs analytical AI distinction. AI is a relatively flexible term, conceptually, that you could use it to describe a lot of ML applications, but most of that is not going to be generative AI

250

Raza @transfusion.health · 10/11/2025

Could Wasserman be considered intro?

000

Raza @transfusion.health · 10/11/2025

I honestly recommend powerpoint DAG package outputs just don't end up looking very publishable in my experience (and if your DAG is so complex that you need a package to update it, you start to question the value of visualizing it at all)

010

Raza @transfusion.health · 29/10/2025

In the era of evidence-based medicine, why do the similar patient face different treatment strategies at hospitals in the same healthcare system? We analyzed 800k+ admissions at 24 hospitals to examine transfusion practice at teaching and non-teaching hospitals link.springer.com/article/10.1...

link.springer.com

Red Cell Transfusion in General Medicine Patients at Teaching and Non-teaching Hospitals: A Retrospective Cohort Analysis - Journal of General Internal Medicine

Background Red blood cell transfusion is a well-studied intervention with numerous randomized trials to guide management, yet practice remains highly variable. Objective In this study, we explored hos...

020

Raza @transfusion.health · 26/10/2025

For those wrestling with applying MINT trial and new AABB RBC MI guidelines results with nuance, I think this figure strikes a good balance for navigating the evidence base at the bedside doi.org/10.1016/j.cj... #transfusion #medsky #cardiology

010

Raza @transfusion.health · 20/10/2025

A lot of people in healthcare are asking, "I want to use AI for [my field]" Higher yield: Let’s make a list of problems worth solving, then discuss if any are best addressed with AI (most won't be) (e.g. studies aren't typically motivated by, "I want to use retrospective cohort design")

010

Raza @transfusion.health · 18/10/2025

I think most people in the ML/AI space understand that, for most real-world uses, LLMs are merely the gateway drug to entice authority figures to invest in resources that will enable more reliable analytical ML systems. Kind of like how the WWF uses pandas to fund saving other endangered species

000

Reposted by Raza

Brian D. Adkins @path-brian.bsky.social · 06/10/2025

How are you diagnosing and treating mixed AIHA? pmc.ncbi.nlm.nih.gov/articles/PMC...

011

Raza @transfusion.health · 11/10/2025

Profiting from selling unproven therapies to desperate patients is as old as time. Why does it persist? Co-authors discuss antecedents, consequences, and contemporary examples of medical grifting from #transfusion, #hematology, and biotherapies onlinelibrary.wiley.com/doi/10.1111/...

onlinelibrary.wiley.com

Confronting medical grifting: Fraudulent and unproven products and interventions in apheresis, transfusion and biotherapies

A basic overview of medical grifting for haematologists with practical guidance in identification and tips for management.

010

Raza @transfusion.health · 08/10/2025

I was a programmer before I was a data analyst, and so have a toxic trait of jumping to nested loops for problems where dataframe transformation would be far more elegant. I can't possibly be alone? #stats #ML

250

Raza @transfusion.health · 22/09/2025

Classic teaching is that no post-1 hr increment points to immune refractoriness. One important exception is device-related failures, e.g. plts going into a clotted line, a faulty dialysis/apheresis circuit, an ECMO machine, etc! #hemesky #transfusion

040

Raza @transfusion.health · 18/09/2025

What is the current state of AI/ML research in transfusion medicine? Co-authors Na Li, @ruchika_goel1, et al, have compiled a critical summary of research applying AI/ML in transfusion, including some glossary tables lay readers will helpful Read here: www.sciencedirect.com/science/arti...

sciencedirect.com

Artificial Intelligence and Machine Learning in Transfusion Practice: An Analytical Assessment

Transfusion medicine is vital to healthcare and affects clinical outcomes, patient safety, and system resilience while addressing challenges such as b…

010

Raza @transfusion.health · 08/09/2025

My progression as a reader of scientific literature can be summarized by the first sections I read when I pick up an article High school: Abstract → Conclusion Undergrad: Abstract → Introduction Medical school: Abstract → Discussion Postgraduate: Abstract → Results Now: Abstract → Methods #medsky

111

Raza @transfusion.health · 07/09/2025

Can we improve the way we transfuse plasma? Yes. Big gaps. E.g the ineffectual practice of 1-2 plasma unit dosing is widespread, accounting for 4/5ths of all orders Study from our group now out in @aabbupdates.bsky.social journal onlinelibrary.wiley.com/doi/10.1111/...

onlinelibrary.wiley.com

Plasma transfusion practice: A five‐year audit of plasma transfusion at 23 hospitals

Background There is limited high-quality evidence to guide plasma transfusion, and plasma transfusion practices remain variable. Study design and methods This is a retrospective cohort study that ...

010

Raza @transfusion.health · 04/09/2025

Fancy health AI tool/models have to produce *some* real-world healthcare benefit on one of 5 core health outcomes: 1. Death 2. Disease = occurrence/severity/complications 3. Discomfort = phys/emotional 4. Disability = functional impact 5. Destitution (cost to individual/system) #ai #mlsky #medsky

001

Raza @transfusion.health · 29/08/2025

I believe when working with big data for causal inference, you HAVE to pre-specify your analyses (esp within a project). Do it for yourself, not for reviewers. As Feynman said, “[in science] the first principle is that you must not fool yourself and you are the easiest person to fool” #stats

061

Reposted by Raza

REBEL EM @rebel-em.bsky.social · 26/08/2025

🧪 ICARUS-ED: Early albumin in sepsis? 📉 Minor gains in numbers, no real outcome benefit 💰 30x the cost vs crystalloids ⚠️ Outcomes > optics 🔍 Full breakdown: wp.me/pdrP8b-5oc ❤️ Like 🔁 Repost 🔔 Follow for more #FOAMed #MedX #MedTwitter #MedEd

031

Raza @transfusion.health · 26/08/2025

“This trial adds to the growing body of evidence suggesting no clinically meaningful benefit to albumin administration in the resuscitation of septic patients”

011

Raza @transfusion.health · 18/08/2025

We need trials of AI showing better patient important outcomes (I'm not aware of any) #AI #ML #medsky

051

Raza @transfusion.health · 14/08/2025

Seeing a lot of PCC use in liver disease lately. What why how #hemesky #medsky

000

Raza @transfusion.health · 14/08/2025

More seriously, the 'why' seems to be fear of volume overload with plasma. However, liver disease is multi-factor "balanced coagulopathy" state, and giving just 4-6 factors does not make sense to me (nor is there good evidence, to my knowledge?)

000

Raza @transfusion.health · 01/08/2025

Why do most hospitals have vending machines full of junk food but not a gym? #medsky

000

Raza @transfusion.health · 26/07/2025

You can tell if the person in your machine learning course has a math, stats, or compsci background by the quirky way they write code #ML #stats

000

Raza @transfusion.health · 18/07/2025

I agree, it’s an interesting thing to ponder which I haven’t before either. In essence, monkey typing patterns may lack degrees of freedom needed for Shakespeare due to some hard constraint. But: if the probability were anything other than zero, infinity is long enough for it to happen eventually

110