iiinova — where your digital self keeps evolving

Theodore Pak

@tedpak.bsky.social

488 followers 348 following 54 posts

Assistant Professor @ucirvine.bsky.social. Former fellow @mgh-id.bsky.social, MGH Medicine resident, Sinai MD-PhD, mghwhitebook.app creator. Posts my own. Lab website: tpaklab.org

Posts Replies Media

Theodore Pak @tedpak.bsky.social · 19/12/2024

Board certified! Glad that I'm now officially qualified to wear sunglasses while ranting about doxycycline, the definition of "FUO," and Streptococcus taxonomy controversies 😎

5140

Theodore Pak @tedpak.bsky.social · 01/12/2024

All three viruses (flu, COVID, RSV) are combined for the plotted points. The vast majority of respiratory viruses and hospital-onset cases in those two periods was COVID, so this is why the combined measure changes when the variant changes.

020

Theodore Pak @tedpak.bsky.social · 29/11/2024

So yes. This is why we need to publish data on what happens when universal masking & testing starts and ends, from as many US hospitals as possible (In the absence of controlled trials, which were not practical or ethical during the pandemic, for hopefully obvious reasons) 6/🧵 #IDSky #MedSky

35011

Theodore Pak @tedpak.bsky.social · 29/11/2024

This is part of a trend beginning back in Dec 2022, when prominent voices in infection control started to spread opinion pieces that questioned the value of precautions like admission testing and universal masking. 4/🧵 #IDSky #MedSky

2265

Theodore Pak @tedpak.bsky.social · 29/11/2024

Here is the original study with the data this post refers to (I am the author, AMA) #OpenAccess jamanetwork.com/journals/jam...

000

Theodore Pak @tedpak.bsky.social · 28/11/2024

Of note, last winter respiratory virus season was more evenly split among COVID, flu, and RSV admissions. So we should seek precautions against all three. Having hospital staff mask in the winter is relatively low-cost, and our data suggest it decreases nosocomial spread 11/🧵 #IDSky

180

Theodore Pak @tedpak.bsky.social · 28/11/2024

In January 2024, given the high rates of flu-like illness in New England, our system restarted masking of staff only during patient care interactions. This lasted ~3mos, but we still found a 33% decrease in hospital-onset infections (aRR, 0.67; 0.52-0.85) 10/🧵 #IDSky

1121

Theodore Pak @tedpak.bsky.social · 28/11/2024

As you might suspect, most hospital-onset respiratory viruses in the de-implementation period were COVID infections. Crucially, we were able to check the charts of these cases to see if these were more likely acute cases or "asymptomatic shedders," as others thought 8/🧵 #IDSky

2112

Theodore Pak @tedpak.bsky.social · 28/11/2024

As the plot shows, there was an immediate increase in the ratio of hospital-onset to community-onset respiratory viral infections when masking/testing stopped: without adjustment, 7.6% → 15.5% (13.6-17.4%). In an adjusted model, the rate ratio was 1.25 (1.02-1.53) 7/🧵 #IDSky

2112

Theodore Pak @tedpak.bsky.social · 28/11/2024

Last year, we pointed out that hospital-onset COVID rates increased among hospitals in England and Scotland when NHS stopped universal admission testing at different times in each country. jamanetwork.com/journals/jam... 2/🧵 #IDSky

1131

Theodore Pak @tedpak.bsky.social · 28/11/2024

This Thanksgiving, let's thank hospitals that use masking during winter respiratory virus season. Out in JAMA Network Open: Hospital-onset flu, COVID, and RSV rates when 10 Mass Gen Brigham sites stopped universal testing & masking, then restarted staff masking in Jan 2024. 0/🧵 #IDSky #MedSky

711633

Theodore Pak @tedpak.bsky.social · 28/11/2024

Two amusing cases of asking GPT-4o and Llama3.2 to interpret medical images GPT-4o gives serious consideration to inserting iron tablets into the ears, and reads an X-ray backwards (right vs left) Original source: composio.dev/blog/meta-ll...

030

Theodore Pak @tedpak.bsky.social · 26/11/2024

Solving the "onion problem": where do you point the knife when slicing a half-onion to make the most uniform pieces from all the layers? Answer: 55.7% of the radius below the center of the onion Assuming a spherical onion, and with a whole lot of math: medium.com/@drspoulsen/...

120