Health system-scale language models are all-purpose prediction engines.
Jiang LY, Liu XC, Nejatian NP, Nasir-Moin M, Wang D, Abidin A, Eaton K, Riina HA, Laufer I, Punjabi P, Miceli M, Kim NC, Orillac C, Schnurman Z, Livia C, Weiss H, Kurland D, Neifert S, Dastagirzada Y, Kondziolka D, Cheung ATM, Yang G, Cao M, Flores M, Costa AB, Aphinyanaphongs Y, Cho K, Oermann EK. · 2023
Read Health system-scale language models are all-purpose prediction engines. by Jiang LY, Liu XC, Nejatian NP, Nasir-Moin M, Wang D, Abidin A, Eaton K, Riina HA, Laufer I, Punjabi P, Miceli M, Kim NC, Orillac C, Schnurman Z, Livia C, Weiss H, Kurland D, Neifert S, Dastagirzada Y, Kondziolka D, Cheung ATM, Yang G, Cao M, Flores M, Costa AB, Aphinyanaphongs Y, Cho K, Oermann EK. (2023) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc10338337.txt.
Description
Jiang led a team in 2023 that trained a language model on hospital notes. Doctors write free-text notes on each patient in the health record. Older risk tools need neat, coded data, which is slow to build and hard to deploy. The team named its model NYUTron and taught it on raw clinical text.
NYUTron then took on five forecast tasks with strong scores. The tasks were return to hospital in 30 days, death in hospital, illness load, length of stay, and claim denial. Scores ran from 78.7% to 94.9% on a standard scale. The model beat older tools by 5 to 15 points. The team also ran it live in a trial at their own health system.
Catalog record
| CREATOR | Jiang LY, Liu XC, Nejatian NP, Nasir-Moin M, Wang D, Abidin A, Eaton K, Riina HA, Laufer I, Punjabi P, Miceli M, Kim NC, Orillac C, Schnurman Z, Livia C, Weiss H, Kurland D, Neifert S, Dastagirzada Y, Kondziolka D, Cheung ATM, Yang G, Cao M, Flores M, Costa AB, Aphinyanaphongs Y, Cho K, Oermann EK. |
|---|---|
| YEAR | 2023 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |