Anil Makam
@AnilMakam
UCSF Hospital Medicine Physician Scientist at SFGH. Think about evidence, clinical medicine, outcomes, health services, policy. https://hopelab.ucsf.edu/people/
🚨 New Study in Annals🚨 GLP1ra & SGLT2i are the only diabetes meds that reduce heart attacks & death But can't work if can't prescribe TLDR 40% Medicaid enrollees have restricted access to GLP1 & 25% to SGLT2i much state/plan variability GLP access plateaued in '22 thread
A study from @UCSFMedicine finds that Medicaid often restricts access to cardioprotective type 2 diabetes medications. bit.ly/3Y72K2z @ClarivateIP
history is usually reliable & best test >90% of time minority of patients I would recommend peth to distinguish
I have a good test. Tells you about alcohol intake over the last 3 weeks. Super sensitive. Super specific. Can reclassify the etiology of liver disease. These sorts of tests used to clean up. Total game changer. But remains underutilized. What gives?
why we need truth to power & not platform people who promote disinformation continually this is pseudoscience x.com/AnilMakam/stat…
A large study on the optimal number of steps-per-day of physical activity that's associated with favorable health outcomes. Study pegs it at around 7,000 steps, with some plateau of health benefits seen at higher amounts of activity. Published in @TheLancet h/t @EricTopol…
1 year too late. don't need the retroscope either. bad decision at the time, before 3 deaths fewer drug approvals are not a bad thing if not supported by evidence & have worrisome harm signals that are not corroborated
Exclusive: $SRPT would have to conduct new studies to get back on market, FDA official says - CBER review staff are 'unanimous' that Elevidys should never return to market endpoints.news/sarepta-would-…
Doesn't matter if theres 1000 cohort studies confounded designs are confounded data which is not knowledge this is pseudoscience activity is good is it 7k vs 6k or 8k? we won't know until someone randomizes step thresholds Be active to extent you feel healthy & realistic
Yes, its largely useless, yes
MAP > 65 is a myth Lactate isn’t intrinsically bad 30 ml/kg fluid bolus was invented by a shadow government
the entirety of the evidence is compelling against any meaningful benefit and now with serious harm signal I know it sucks and we want treatments but let's not disguise it as gold
For many Duchenne families, halt to Sarepta $SRPT gene therapy treatments is heartbreak upon heartbreak statnews.com/2025/07/24/duc… My colleague @Jasonmmast with more reporting on the patient/physician perspective.
supreme court sides with specialists over generalists by ruling USPSTF as "inferior officers" j/k good run down of Braidwood separately, colorectal cancer screening for 45-49 year olds is not just legally dubious but medically dubious too no new evidence
And here's the @NEJM piece (paywalled, unfortunately): nejm.org/doi/abs/10.105…
wtf how was this study approved ethically by the IRB how was it allowed to be published? glad I now know not to inject endotoxin into my patient
Inflammation can spike depression This is a 𝘳𝘢𝘯𝘥𝘰𝘮𝘪𝘻𝘦𝘥 𝘵𝘳𝘪𝘢𝘭 of placebo vs. endotoxin (which is what makes certain bacteria inflammatory – it's in the membrane) in older adults Endotoxin spiked depressive symptoms – but only in those with insomnia
we are not sure it is helpful broadly (including addiction) studies highly confounded. need trials but we are sure it helps with all things metabolic: weight loss, diabetes, heart disease, heart failure, kidney disease, sleep apnea, arthritis, back pain, fatty liver disease
Today's essay is three big questions, asked and answered: 1. Are we sure GLP-1 drugs are this broadly effective? 2. Literally, how can one drug do "everything"? 3. If the drugs actually do "everything" at what point should we all be on versions of them? derekthompson.org/p/why-does-it-…
The elite research university: Where the Vice Dean of Nothing makes three times what you do and runs the “Office of Scientific Excellence” with no science inside.
you can remove "for people with diabetes" activity is better than no activity for health
For people with diabetes, even a little activity every week is tied to better health trib.al/yavyyop
Great thread closer to truth than using artificial benchmarks IMO hardest part of AI for clinical care isn't technology, but is lack of gold standard we have no good way of judging good doctoring Kappa scores poor & based on chart review
We saw significant relative error reductions: history-taking (-32%), investigations (-10%), diagnostic (-16%), treatment (-13%) for clinicians with vs. without AI. At Penda alone, AI Consult would avert diagnostic errors in 22k visits and treatment errors in 29k visits yearly.