Training AI requires consuming copious amounts of data. Consume bad data and the output will be wrong. In computer science this is called garbage in / garbage out. In some AI models once a bad piece of information is learned it becomes exceedingly difficult to purge it.
Category: Artificial Intelligence and Healthcare
Tuesday Links
- Why does time fly faster as we get older?
- TikTok sale may not be as easy as Trump thinks.
- Microsoft’s new AI system correctly diagnosed patient cases 80% of the time. According to the company, the results from the tool are four times better than the human physicians in the study, who reached the correct answer 20% of the time.
- States receive $9 in federal funds for every $1 of state spending on able-bodied, working-age adults compared to an average of only $1.33 in federal funds for every $1 of state spending on children, pregnant women, seniors, and people with disabilities.
WSJ: Health Insurers Vow to Streamline Prior Authorization
One (blunt) instrument to avoid unnecessary care is prior authorization. Doctors and patients hate prior authorization. Prior authorization is when your doctor wants to prescribe a certain drug or order a procedure but must seek approval from your health plan prior to providing that care. Critics claim that prior authorization is a rationing tool, designed to wear down patients into forgoing care due to the hassle. Congress and state governments occasionally weigh-in and try to reform the use of prior authorization.
Friday Links
- House reconciliation bill creates new subsidies for small business to help employees buy individually owned health insurance.
- How AEI would reform MA plans.
- Does exercise really help cancer survival?
- For diabetes, lifestyle changes are twice as effective as Metformin.
- Should everyone have a glucose monitor?