Thread by @WhitneyEpi

1/ Good overview of challenges of doing “big data” health care research in US. For instance, nice description of basic pros and cons of insurance claims vs #EHR health care records

HT @ABatemanHouse⁩ technologyreview.com/2021/06/21/102…

2/ But the article has a prematurely triumphant tone at the beginning (and in the default Twitter tagline). N3C is great but limited, as the article does make clear

3/ And it’s not totally novel. But the novelty to me is that it’s

a non-“federated”* model

*not* run by for-profit companies,

so the data are accessible to the public for analysis

*this term is explained in article

4/ Big companies are actually hosing up a LOT of raw healthcare data, but there’s a catch: they own the data!

“We’re willing to give all our data to a commercial entity and let them sell it back to us, but we’re unwilling to pay for the most basic public health infrastructure”

5/ p.s. The part of Tweet 1 article where I nodded most emphatically was re: frustration of #EHR data cleaning, where health systems sent variables like body weight (ok, nice basic quant variable) - but no units of measurement! 🤯

6/ This example resonated 💯

The *hours* my team spent trouble-shooting our calculated BMI variable before realizing that body weight was stored in OUNCES 🤦🏾‍♀️

We only figured it out trial & error bc system had no units label or variable for units! This is #epitwitter madness!