Ian Nason

Working at the intersection of ML and health services research. PhD @Harvard. (he/him). 🇨🇦

[Ian Nason @NasonIan]

Today I realized my last name, Nason, is obviously derived from 'Son of NA'. I can only assume one of my long lost ancestors was missing data.

[MIT Economics @MITEcon]

Considering applying for an economics PhD? The student-led MIT and Harvard Application Assistance and Mentoring Program helps students from underrepresented groups connect with a grad student mentor. Applications are open until July 17! Learn more at https://t.co/ITeBb9dFmw

[𝘽𝙪𝙛𝙛𝙖𝙡𝙤 𝙏𝙧𝙖𝙣𝙨𝙞𝙩 𝙀𝙣𝙟𝙤𝙮𝙚𝙧 @HockeyYearning]

Of course that's your contention. You're a first year hockey twitter account. You just finished figuring out one type of chart -- RAPM probably. You're gonna be running around citing that until next month when you discover HockeyViz finishing stats, and then you'll say that Jesse

[Isaac Gibbs @iswgibbs]

Conformal Prediction With Conditional Guarantees

Excited to share my work with John Cherian (@jjcherian) and Emmanuel Candès https://t.co/Ot41Xrapp7. It is well-known that pointwise conditional coverage is impossible without strong assumptions.

But how close can we get?
(1/6)

[Dan Ly @meddly]

Hi all! Wanted to share a 🧵 on a new paper; this one published in @JAMAInternalMed. We examine what happens when your previously generic medication has a 🚨16-fold🚨 price increase. https://t.co/rXcDtQO4H1

[Irina Degtiar @IDegtiar]

My 2nd dissertation paper is finally out, with @timothyjlayton, @jwswallace, and @sherrirose! Take a look to see how to combine randomized and observational data while addressing potential unmeasured confounding in the observational study: https://t.co/BgF2MYhxLB

[Alex Izydorczyk @aleksizy]

Every healthcare provider in the US receives a unique ID (NIPS) and is made publicly available -- you get their (business) names, addresses, specialties, and places of business. It's a great example of an ostensibly public dataset that isn't easily available or accessible.

[Ian Nason @NasonIan]

@AJHolmgren Why don’t you just send this to their inbox

[Ian Nason @NasonIan]

@AJHolmgren Just looking at this hurts my teeth

[Ian Nason @NasonIan]

@meddly @dtrug Congratulations!!

[𝐒𝐞𝐧𝐝𝐡𝐢𝐥 𝐌𝐮𝐥𝐥𝐚𝐢𝐧𝐚𝐭𝐡𝐚𝐧 @m_sendhil]

🧵1/ New paper: We argue machine learning should be used to generate truly novel hypotheses.

The application: judge decision-making. We find a single factor is heavily related to who gets jailed: the defendant’s mugshot.

What are judges looking at?

Our method unpacks this...

[Ian Nason @NasonIan]

@AJHolmgren Let’s gooooo

[Adrianna McIntyre (@adrianna.bsky.social) @onceuponA]

Horrified looks from everyone in the room https://t.co/y2ThoJppw7

[Lucy Chen @chenlucy8]

As a med student, I’ve looked up to & learned from chief medical residents. What if we better understood the knowledge that docs have about each other’s quality - and harnessed it for quality improvement? New study in @JAMAInternalMed w/ @JMichaelMcW 1/🧵 https://t.co/R6eMNZGuGs

[Toronto Raptors @Raptors]

🗣 INCOMINGGGGGG @chrisboucher

[Ian Nason @NasonIan]

@ekoreen I think Simmons would be a perfect fit for the raptors.

[Sukhi @sukhigulati]

! Recruiting Study Participants !

My lab @ MIT is looking to talk to software engineers and policy experts to learn more about policy compliant code. If you know anyone who might be interested in helping, please reach out.

Study details:

[Ian Nason @NasonIan]

@ThePromoguy123 Twice in one week on Lebron related boosts 🙏

[Ian Nason @NasonIan]

@scottedelstein @huggingface HuggingEgg

[Ian Nason @NasonIan]

Great thread on risk-adjustment. I think the medical / diagnostic AI field can learn a lot from this literature. It’s rare that the loss function of your algorithm lines up perfectly with the desired outcomes you want to optimize for. Especially when inputs are biased.

[Michael McWilliams @JMichaelMcW]

Interested in risk adjustment and health equity in population-based payment models (ie MA or ACOs)?

You may find this paper useful

w/ Chima Ndumele @jwswallace Gabe Weinreb & Lin Ding in @Health_Affairs:

https://t.co/cVF0PWPJnS

1/(this could run long)