HackerTrans
トップ新着トレンドコメント過去質問紹介求人

apyrros

no profile record

投稿

Opportunistic detection of type 2 diabetes using deep learning from chest x-rays

nature.com
1 ポイント·投稿者 apyrros·3 年前·1 コメント

コメント

apyrros
·2 年前·議論
That’s not really true. Insurance companies use intercompany eliminations to bypass the rules. Look at United Healthcare’s record-breaking profit.https://www.axios.com/2021/07/16/unitedhealth-optum-provider...
apyrros
·2 年前·議論
Trust me it isn’t 6%. United is one of the largest healthcare companies by market cap. They have aggressively acquired physicians under Optum, and use an accounting trick called intercompany eliminations to shuffle profits and skirt the law on the medical loss ratio. https://www.axios.com/2021/07/16/unitedhealth-optum-provider...
apyrros
·3 年前·議論
I must mention that I wrote the article titled “Opportunistic detection of type 2 diabetes using deep learning from frontal chest radiographs” (you can find it here: https://www.nature.com/articles/s41467-023-39631-x). While many of these models present promising advantages for population health, there are still many challenges related to bias and execution. However, it’s worth noting that early disease detection can provide benefits, not necessarily cures. The IDEAs study is a good example of this in Alzheimer’s disease.
apyrros
·3 年前·議論
In a remarkable demonstration of the capabilities of large multimodal ResNet models applied to chest X-ray images, we've revealed how the integration of EMR data can enable accurate predictions of Type 2 diabetes. Much akin to the principles of real estate, location is paramount; the area where fat is deposited in the body significantly influences the risk. The beauty of this is that chest radiographs are common and in many cases will precede the diagnosis of diabetes.