Just out in @techreview: I look at the companies using AI to measure how much pain patients are in based on everything from involuntary facial movements, to heart rate, to peripheral temperature changes. Will this oust the classic self-reported 1-10 scale?

Internet access seems likely to get significantly more important over time given LLMs. Fortunately, the trend seems to have strongly pulled in that direction by default -

@KelseyTuoc @otis_reid

Cool piece from @KelseyTuoc on whether LLMs are biased by the language the asker is using — mildly surprising to me that the answer was 'no'! (h/t @otis_reid)

Wrote a companion article to my @techreview piece about companies using AI to measure how much pain people are in. This one focuses on the implications: what happens when we quantify something we don’t fully understand, and what clinical judgment might we lose in the process?

Just out in @techreview: I look at the companies using AI to measure how much pain patients are in based on everything from involuntary facial movements, to heart rate, to peripheral temperature changes. Will this oust the classic self-reported 1-10 scale?

Read the full thing here:

And that over-indexing can also result in deskilling. There's already early evidence of this happening in as little as three months of using an AI tool, in the case of doctors identifying important findings in colonoscopies. https://t.co/XM02GumXeM

Of course, we do that with the classic pain scale, too! But clinicians have a notable "automation bias" and tend to take outputs from machines a lot more seriously than they might warrant. This can result in an over-indexing on outputs that makes overall accuracy worse.

There's always a risk of confusing the map for the territory, particularly when we're compressing something that is both complicated and that we don't fully understand into a single metric https://t.co/dLgzVlJgAH

Wrote a companion article to my @techreview piece about companies using AI to measure how much pain people are in. This one focuses on the implications: what happens when we quantify something we don’t fully understand, and what clinical judgment might we lose in the process?

How AI-powered tools like PainChek, an app that scans a person's face for tiny muscle movements, are helping healthcare providers better assess patients' pain (@deenamousa / MIT Technology Review) https://t.co/csnJ7cauCQ https://t.co/un6NSBzxOJ

And thanks to the @PainChek team for talking to me for this!

Gift link:

PainChek Adult was offered de novo FDA clearance this week, and engineers are now adapting the code for the very youngest patients. PainChek Infant targets babies under one year, whose grimaces flicker faster.

Data shows a ~25% drop in anti­psychotic use and, in Scotland, a 42% reduction in falls from use. One clinician mentioned that residents who had skipped meals because of undetected dental pain began eating again, and those who were isolated due to pain began socializing.

Linking the scan to a human‑filled checklist was, they admit, a late design choice. Initially, they thought AI should automate everything but found that hybrid use yielded better results...

There are several devices in clinical use today, like PainChek, a smartphone app that scans the facial expressions of people who have dementia and uses AI to output an expected pain score to inform their care. They also record data for the patient and facility over time.

Full paper:

A new paper in the Lancet finds that doctors got *worse* at finding precancerous growths during colonoscopies on their own if they had just spent three months using an AI assistive tool Worrisome sign that deskilling may happen a lot faster than we'd expect.