2 Comments

Fascinating. Longitudinal data in the non-universal healthcare US not so great; VA probably best bet. Huge potential; I actually thought (stupid I know) electronic medical records would help form the basis for improved disease understanding but their purpose is billing. Available to life insurance companies? Health insurance companies? Employers? There's the rub.

Expand full comment

So helpful. Powerful, too -- to see the NOA use case and understand how possibke it is today to develop the large computational models you highlight. One difference between the weather and model and what is needed in healthcare: public access to large data sets. Despite federal law to ensure patient access to personal data, the friction for accessing large healthcare data via silo'd systems is daunting. Might this suggest a different G2M for US "mass computational" efforts, or a need to go to large exUS systems to prove the model (India). Interesting to consider merist of ex-US system, which may offer broader data sets for correlation (UBI, etc.)?

Expand full comment