UpToDate

Um-

Like UpToDate.

Say again.

UpToDate.

Oh, yeah, UpToDate.

Um, UpToDate is really good.

There are...

I, I did a, uh, course at an- another university a year or so ago, and they talked a lot about AI.

I'll give you another example.

So there are big medical systems like UpToDate.

It's one of them.

The other one that just came out with OpenEvidence.

So he jumped in gently.

You guys, Will, you will certainly remember UpToDate, which nobody uses anymore, but was, you know, critical for us during medical school, right?

And I think there's still some people holding out hope

that UpToDate will become more up-to-date.

So, yes, yes.

And they're no longer founder controlled anymore.

So I think things started when I was, way back when, when I was a registrar, and I had this sort of Jerry Maguire moment with my directors of training.

I was re- I was k- turning up to a Wednesday morning teaching as a re- as a reg, and I was getting a bit annoyed, to be honest, that, wondering what the point of me being there was, listening to some pretty average presentations that I could have read the same text in half an hour looking at UpToDate.

And I thought, "Really, there's definitely gotta be a better way than doing this." And so I wrote an email to our directors and said, you know, "Why am I here?" Um-

I didn't know this story

it did not get the best response, shall we say.

Um, but I, you know, I created a Google page for the directors of training that, to, to use.

I suggested some ways that we could be, have, put some rules in place around, you know, presentation style and not just cutting and pasting things from, straight from UpToDate.

And things went pretty well, and I was really try to sort of show by example.

Uh, disclaimers aside, you know, I practice hospital medicine and I'm, I'm very much a pragmatist, so something like an OpenEvidence tool, I'd say a handful of my colleagues use it, but it's not yet to the point where we as clinicians find it particularly useful.

And the way I think about it is something there's this software reference called UpToDate, which you may or may not be familiar with, which started 35 years ago or so, back when we had the external CD-ROM drives and you could just, you know, load the CD cassette on.

Mm-hmm.

And then UpToDate eventually became something that you could install on your computer, and now it's of course available on the internet, uh, and it's a searchable reference, and we use that.

And so I would view OpenEvidence and these other tools as potentially eventually replacing UpToDate once the interface improves and, you know, people have more confidence in that evidentiary basis.

So I'd say for some of those resource tools or reference tools, not fully or widely adapted.

Other things that are quite convenient, for example, AI can comb the medical record and create a summary of the patient's hospitalization, and something like that is super convenient because otherwise I as the physician could literally sit there and review every single progress note, every single specialist note, and pull all of those details and I could spend an hour or two doing that.

So that could be more time with the patient, more time talking with colleagues, and less time using the keyboard as sort of your arbiter of clinical reasoning and clinical thought.

Sachs, the mother of one of my patients attended your talk last night where you claimed that these medications damage the motivational center of the brain.

I'm a professor of child psychiatry at NYU, and I have never heard of any such evidence." So I sent her the bibliography showing all 15 links, which she could have found herself in five minutes at Open Evidence or UpToDate or any of the tools that we physicians use.

She could have found it very quickly and easily.

She didn't bother, 'cause she was arrogant and condescending, and she thought as a professor of child psychiatry, she knew more than I did as a family doctor.

They're integrating it with clinical decision support.

It's not necessarily a new announcement, although they did partner with UpToDate to include all that information.

So that was good But it is that kind of maturing of the things that they were talking about, which is kind of classic Epic, right? Like they announce it, a couple of people get on board early and everyone else waits for it to mature a little bit so that it can be competitive with the other products out there.

And then they kind of switch to it because it's cheaper or it's more integrated.

One of the things I'll often do with youth and parents is speak to both of them together about where they're getting their information online, because inevitably both parents and kids are getting their information online, and I'd like to know what websites they're using.

You could even say, because I use UpToDate so much, that I get my information online.

Because I use this website all the time, especially when somebody will use a terminology that I haven't used before.

And that does happen also.

So all physicians know that the curbside consult is you don't really need the ID doc or the cardiologist to see your patient, review the chart, write a note, but you have a question that you're just not 100% sure of the answer.

And five years ago, you might have looked it up in a tool like UpToDate, but it's not good to give it a whole case, a complex, nuanced case and get an answer.

So those were where I would walk into the cafeteria and I'd see Emma, the ID doc, across the room and I would run literally and say, Emma, can I quickly run this case by you? And she was very nice and she said, sure.

And then I'd give it, you know, give her a case and she'd say, oh, yeah, I'd be worried about Cryptococcus.

We talk about that a lot on The Blind Spot.

When AI surfaces a finding or a risk score that changes what you know about a patient, how does that conversation actually go? Do clinicians find the communication changes when an AI tool has been used?

I'd say for most clinicians, I'm not sure that the majority of clinicians tell patients what the source of a prediction was, whether they got it from up-to-date or reading a textbook or reading an article or reading a guideline from their specialty society or got it from A.I., What I do think is interesting is I remember when UpToDate first came out, so it must have been 25 years ago.

I'm ashamed of this, but I'll tell you.