Latest / Code & Cure / #32 - When Data Isn’t Better: Rethinking Fertility Tracking

Transcript

#32 - When Data Isn’t Better: Rethinking Fertility Tracking · Code & Cure

• 0:00
  SPEAKER_00: It's exciting to see wearables show up for women's health.
• 0:04
  SPEAKER_00: But are they solving a hard problem or dressing up an easy one?
• 0:08
  SPEAKER_00: Let's talk accuracy, fertility windows, and what really works.
• 0:23
  SPEAKER_01: Hello, and welcome to Code and Cure, the podcast where we discuss decoding health in the age of AI.
• 0:28
  SPEAKER_01: My name is Vasant Sarathi, and I'm a cognitive scientist and an AI researcher.
• 0:33
  SPEAKER_01: And I'm with Laura Hagopian.
• 0:35
  SPEAKER_00: I'm an emergency medicine physician and I work in digital health.
• 0:39
  SPEAKER_00: And today we are talking about an application of wearable digital technology and specifically looking at the fertility window.
• 0:51
  SPEAKER_01: Yeah.
• 0:52
  SPEAKER_01: What's the fertility window?
• 0:54
  SPEAKER_00: Thank you for asking.
• 0:56
  SPEAKER_00: Uh the fertility window, people may want to detect it for different reasons, right?
• 1:01
  SPEAKER_00: But it's like, hey, when can you get pregnant?
• 1:03
  SPEAKER_00: Right?
• 1:04
  SPEAKER_00: When are you most fertile?
• 1:06
  SPEAKER_00: And it's usually around the time of like ovulation.
• 1:08
  SPEAKER_00: Yep.
• 1:09
  SPEAKER_00: What before the egg is released and after the egg is released, it's a few days, few day window.
• 1:15
  SPEAKER_00: Okay.
• 1:16
  SPEAKER_00: And you might want to know the fertility window in order to get pregnant, right?
• 1:21
  SPEAKER_00: Because then you're like, okay, this is when I'm going to, you know, have intercourse or do insemination or whatever.
• 1:28
  SPEAKER_00: Or you might want to know the fertility window in order to not get pregnant.
• 1:33
  SPEAKER_00: Right.
• 1:33
  SPEAKER_00: Right.
• 1:33
  SPEAKER_00: And use that as a form of contraception.
• 1:35
  SPEAKER_00: It's not a great form of contraception.
• 1:37
  SPEAKER_00: Yeah.
• 1:37
  SPEAKER_00: It's one of the ones that has like a high error rate.
• 1:41
  SPEAKER_00: Um, but those are the two major reasons either to get pregnant or to not get pregnant.
• 1:47
  SPEAKER_00: Yep.
• 1:48
  SPEAKER_00: That makes sense.
• 1:50
  SPEAKER_00: And so traditionally, there are a few different ways that people can detect the fertility window.
• 1:57
  SPEAKER_00: Um number one is cervical mucus monitoring.
• 2:03
  SPEAKER_00: So this is like basically discharge.
• 2:07
  SPEAKER_00: It would be sort of this like raw egg white consistency, I think is the best way to describe it around the time of ovulation.
• 2:13
  SPEAKER_00: So it's like clear, it's stretchy, it's slippery.
• 2:17
  SPEAKER_00: And that's when you know that you're ovulating.
• 2:20
  SPEAKER_01: And and that's like a like a ground truth definite measure that you're ovulating.
• 2:24
  SPEAKER_00: Well, there's no, I mean, if you want to talk like definite measure, yeah, then that's an ultrasound.
• 2:29
  SPEAKER_01: Okay.
• 2:29
  SPEAKER_00: But nobody's going in to get an ultrasound every time they ovulate.
• 2:32
  SPEAKER_01: Oh, right, right.
• 2:33
  SPEAKER_01: Of course, yeah.
• 2:33
  SPEAKER_00: Right?
• 2:34
  SPEAKER_00: That's a lot of doctors' visits, that's a lot of money, that's a lot of you know, testing.
• 2:39
  SPEAKER_00: But cervical mucus monitoring is actually quite good.
• 2:42
  SPEAKER_00: I mean, it's very accurate.
• 2:44
  SPEAKER_00: Um, it's got good specificity.
• 2:49
  SPEAKER_00: So that's one of the one of the common methods that people use.
• 2:53
  SPEAKER_00: Um, but not everyone feels comfortable with it.
• 2:56
  SPEAKER_00: Not everybody knows exactly how to how to do it?
• 2:59
  SPEAKER_00: How to I mean, it's just looking at your cervical mucus.
• 3:02
  SPEAKER_00: It's not really that.
• 3:03
  SPEAKER_01: Yeah, yeah, yeah.
• 3:04
  SPEAKER_00: But it it may not feel like, oh, I I know exactly what to do here, right?
• 3:09
  SPEAKER_00: Right.
• 3:09
  SPEAKER_00: What is it?
• 3:10
  SPEAKER_00: What is this raw egg white consistency versus another consistency?
• 3:13
  SPEAKER_00: Or some people may be like, I don't want to touch it.
• 3:16
  SPEAKER_00: I don't know.
• 3:16
  SPEAKER_00: No, that's fair.
• 3:17
  SPEAKER_01: Yeah, yeah, yeah.
• 3:18
  SPEAKER_01: That that makes sense.
• 3:18
  SPEAKER_01: Yeah.
• 3:19
  SPEAKER_00: And so another commonly used method is urine test strips, not for pregnancy, but for a hormone called glutenizing hormone LH.
• 3:29
  SPEAKER_00: Okay.
• 3:29
  SPEAKER_00: And so the LH level will surge, you know, a day or a day and a half before ovulation happens.
• 3:38
  SPEAKER_00: And so this is something you can just like buy over the counter.
• 3:40
  SPEAKER_00: They're cheap.
• 3:41
  SPEAKER_00: Okay.
• 3:42
  SPEAKER_00: Um, and so, you know, once that test strip is positive, you can be like, okay, well, I know I'm gonna ovulate soon.
• 3:49
  SPEAKER_00: This is my I'm in my fertile window.
• 3:52
  SPEAKER_01: It just tells you you're in your window.
• 3:53
  SPEAKER_01: It doesn't tell you if you are at the beginning of the window, at the end of the window.
• 3:57
  SPEAKER_01: It doesn't give you like a range, right?
• 3:59
  SPEAKER_01: It just tells you if you're in or not.
• 4:01
  SPEAKER_00: Well, I you're you're assuming everything's sort of binary here, but it's not really like your fertile window can be like a few days before and then a day after ovulation.
• 4:10
  SPEAKER_00: So your fertile window isn't like I'm only fertile for 30 minutes on the window.
• 4:14
  SPEAKER_00: No, no, that's what I mean.
• 4:15
  SPEAKER_01: No, that's not what I mean.
• 4:16
  SPEAKER_01: I mean like it's a window of days, right?
• 4:18
  SPEAKER_01: Yeah, it's a number of days.
• 4:19
  SPEAKER_01: Right, right, right.
• 4:20
  SPEAKER_00: It's it does predict that.
• 4:21
  SPEAKER_01: Okay.
• 4:23
  SPEAKER_01: It gives you information.
• 4:24
  SPEAKER_01: Okay.
• 4:25
  SPEAKER_00: No, the the strip doesn't say you are fertile from two days ago to one day in the future.
• 4:31
  SPEAKER_00: It doesn't say that.
• 4:32
  SPEAKER_00: It's literally just when the surge, when the LH surge happens, the test strip turns positive.
• 4:38
  SPEAKER_00: And so you know I'm gonna ovulate in about a day or a day and a half.
• 4:42
  SPEAKER_00: Got it.
• 4:43
  SPEAKER_01: Okay, okay.
• 4:44
  SPEAKER_00: So it's not like, you know, it's nice to, I suppose, as a scientist, you're like, well, I want exact data, but it's enough to be able to predict, hey, this is a good time to either have or avoid intercourse or do the insemination or whatever.
• 4:59
  SPEAKER_00: Yeah.
• 5:00
  SPEAKER_00: By using the hormones.
• 5:03
  SPEAKER_00: There are other ways that people try to detect their fertility window too that aren't as good.
• 5:08
  SPEAKER_00: So the cervical mucus method is really good.
• 5:10
  SPEAKER_00: The urine LH testing is really good.
• 5:13
  SPEAKER_00: When you combine them, they're even better.
• 5:15
  SPEAKER_00: Yeah.
• 5:16
  SPEAKER_00: Right.
• 5:16
  SPEAKER_00: So you can have an idea of when the fertile window is.
• 5:20
  SPEAKER_00: There are methods that are not as good that are still used because maybe people don't want to buy, you know, the LH tests are actually pretty inexpensive.
• 5:27
  SPEAKER_00: I was just on Amazon looking this up.
• 5:28
  SPEAKER_00: You can get 50 tests for 15 bucks.
• 5:31
  SPEAKER_00: You can also pay more money for tests if you want them to like hook up to a fancy app that talks to your phone, but you don't need that.
• 5:36
  SPEAKER_00: I mean, you literally just like pee out a strip and let it develop.
• 5:39
  SPEAKER_00: You don't need a fancy test for it.
• 5:41
  SPEAKER_01: Yeah, yeah.
• 5:42
  SPEAKER_01: Um yeah, but I also imagine that it gets a little gray once you're further away from the ex exact from the data population, right?
• 5:48
  SPEAKER_01: So, like, depending on when you take the test, it could be different gradations.
• 5:52
  SPEAKER_01: Like, yes, it can tell you if you're going to in the next few days, but if you had taken the test a week before that, maybe it would say no, or would it say, Well, there's a chance that, you know, it's like yes, maybe kind of thing.
• 6:05
  SPEAKER_01: And and then as you get closer and closer, it becomes more and more clear that it it's like if you kept taking the test every day, how do you think that's a good one?
• 6:10
  SPEAKER_00: Well, that's so around the time of ovulation, you are taking the test, not you know, every morning, for example.
• 6:17
  SPEAKER_00: You're taking it.
• 6:17
  SPEAKER_00: So you're not just taking it once, you might be taking it every day until you see the LH surge come through and you know, okay, this is it, this is the time.
• 6:25
  SPEAKER_01: Okay, got it.
• 6:26
  SPEAKER_00: So you might take four or five of them over the course of a month, or or maybe not, right?
• 6:30
  SPEAKER_00: Yeah.
• 6:31
  SPEAKER_00: Um, and other hormones change around the fertility window too, but LH is the one that is easily available over the counter.
• 6:38
  SPEAKER_01: Yeah.
• 6:39
  SPEAKER_00: Like cheap and simple to detect.
• 6:43
  SPEAKER_00: But people do use other methods.
• 6:45
  SPEAKER_00: Um, people, you know, very basic, but they do use the calendar method.
• 6:49
  SPEAKER_00: If your cycle is regular and you know, okay, it comes every 24 days or every 28 days or every 30 32 days, then you can kind of march out and try to predict when you're going to ovulate.
• 7:03
  SPEAKER_01: It's okay.
• 7:04
  SPEAKER_01: It's not great.
• 7:05
  SPEAKER_01: But again, when combined with the others, maybe it becomes more powerful.
• 7:08
  SPEAKER_00: Right.
• 7:08
  SPEAKER_00: But a lot of times people use this on their own.
• 7:10
  SPEAKER_00: If they don't want to do you know, if they don't want to check their cervical mucus or if they don't have a strip, a test strip.
• 7:17
  SPEAKER_00: Yeah.
• 7:17
  SPEAKER_00: Um, and you know, when you talk about using that as a form of birth control, it's like, it's not great.
• 7:23
  SPEAKER_00: People get pregnant, yeah, you know, by mistake when they use this, because it's just, it's okay, but it's not great.
• 7:30
  SPEAKER_00: Um and then another method that people use, and this I think is important for the discussion we're about to have about wearable digital technology, is tracking basal body temperature.
• 7:43
  SPEAKER_00: And so you can check your temperature because as your hormones shift, your basal body temperature shifts just a little bit too.
• 7:51
  SPEAKER_01: Okay.
• 7:51
  SPEAKER_00: And so a sustained rise in your temperature actually happens after ovulation.
• 7:57
  SPEAKER_01: Okay.
• 7:58
  SPEAKER_01: Interesting.
• 7:59
  SPEAKER_00: It's interesting because it's like retrospective.
• 8:01
  SPEAKER_00: It's not good prospectively, right?
• 8:04
  SPEAKER_00: It's like, oh, my temperature rose.
• 8:06
  SPEAKER_00: That means I just ovulated.
• 8:08
  SPEAKER_01: Right.
• 8:09
  SPEAKER_01: But you can follow if you followed the temperature for a longer period of time over multiple cycles, then you could sort of see the the trend, right?
• 8:16
  SPEAKER_01: Because it's gonna rise and kind of level off, right?
• 8:18
  SPEAKER_00: Right.
• 8:19
  SPEAKER_00: And you're sort of kind of talking about combining that with the calendar method almost, right?
• 8:22
  SPEAKER_00: Yeah, yeah.
• 8:23
  SPEAKER_00: And so you can use these methods in in conjunction, but I think the main point about the body temperature method is like, hey, it once it's happened, you're like too late.
• 8:33
  SPEAKER_01: Yeah.
• 8:34
  SPEAKER_01: Right.
• 8:35
  SPEAKER_01: Right.
• 8:35
  SPEAKER_00: So I mean, all these methods can work, they're all helpful.
• 8:39
  SPEAKER_00: I think in this paper, what they were trying to figure out is, hey, can if people are using wearables, can that can that do it too?
• 8:49
  SPEAKER_00: Can that detect the fertility window?
• 8:51
  SPEAKER_01: Yeah.
• 8:51
  SPEAKER_01: And and from the body temperature perspective, you know, it's sort of the inverse effect, right?
• 8:55
  SPEAKER_01: So if you're trying to track contraception time time windows, then the time window for contraception, rate, contraceptive uh sort of setup would be I mean, if I'm understanding this correctly, would be when the temperature rises, right?
• 9:09
  SPEAKER_00: Or if you're using it as a form of contraception.
• 9:12
  SPEAKER_00: Yeah, I mean, there's lots of different reasons you would want to detect the fertility window.
• 9:16
  SPEAKER_00: I these methods of as a way of contraception are not good.
• 9:21
  SPEAKER_00: So I don't I I don't I want to put that out there.
• 9:23
  SPEAKER_00: Like they're just not as good as other methods of contraception.
• 9:25
  SPEAKER_00: Got it.
• 9:26
  SPEAKER_00: Okay.
• 9:26
  SPEAKER_00: Like it, you know, using an IUD is better, using a condom is better, using pills is better, etc.
• 9:34
  SPEAKER_00: Um, but it they can be used.
• 9:37
  SPEAKER_00: And so I think the thing that they tried to do here is they said, okay, well, we have all this data.
• 9:43
  SPEAKER_00: People are just using wearables anyways, at least some people are.
• 9:46
  SPEAKER_00: We have all this data on them.
• 9:48
  SPEAKER_00: Can we use the information we already have that we're already tracking and use it for a different purpose?
• 9:54
  SPEAKER_00: And so they looked at a bunch of different things.
• 9:58
  SPEAKER_00: The main one that was correlated with being able to predict their fertility window was dun-dun-dun, take a guess.
• 10:06
  SPEAKER_00: Um, I don't know.
• 10:09
  SPEAKER_00: Um the strip?
• 10:11
  SPEAKER_00: The they didn't do the strip.
• 10:13
  SPEAKER_00: The strip is not part of your wearable.
• 10:14
  SPEAKER_01: Oh, right, right.
• 10:15
  SPEAKER_01: You know, the strip is part of wearing it.
• 10:17
  SPEAKER_00: Yeah, there you go.
• 10:18
  SPEAKER_00: There you go, there we go.
• 10:19
  SPEAKER_00: Oh, I almost tricked you there.
• 10:21
  SPEAKER_01: Yeah, never mind.
• 10:22
  SPEAKER_00: Um so basal body temperature was the main one.
• 10:25
  SPEAKER_00: But they looked at other things like someone's heart rate, their heart rate variability, their respiratory rate, et cetera.
• 10:30
  SPEAKER_00: And when they added those in, it made the model a little bit better too.
• 10:34
  SPEAKER_00: Sure.
• 10:34
  SPEAKER_00: But it was mostly the basal body temperature that they were tracking.
• 10:38
  SPEAKER_00: So it's like nothing really super novel there.
• 10:42
  SPEAKER_00: And not the best method, as we know, uh, because cervical mucus and LH test trips are are better.
• 10:50
  SPEAKER_00: Um, but it was something that they were able to say, okay, um, it's it works.
• 10:55
  SPEAKER_00: It's it's pretty accurate, it's pretty sensitive, it's pretty specific.
• 10:59
  SPEAKER_01: I mean, again, this paper wasn't actually doing first hand experiments.
• 11:03
  SPEAKER_01: This is a meta-analysis paper, which means it looks at other papers and other studies and then combines the results together to share people share with people what and the overall state of affairs is for um, in this case, wearables being used for this purpose, right?
• 11:19
  SPEAKER_00: Yeah, exactly.
• 11:20
  SPEAKER_00: Yeah.
• 11:21
  SPEAKER_00: And I think my main question when I read this was like, why are we doing this?
• 11:25
  SPEAKER_01: Well, that's always I I I think we can pause for a second there, because that is a great question to ask anytime you apply any piece of technology slash AI to a societally relevant problem.
• 11:36
  SPEAKER_01: Because it's very easy to say, like, oh, cool, we can do this with AI, or oh cool, we have data for this.
• 11:41
  SPEAKER_01: We should do something with it.
• 11:42
  SPEAKER_01: Well, maybe the right answer is not, right?
• 11:44
  SPEAKER_01: Maybe the right answer is that you already have tools that are actually really good.
• 11:49
  SPEAKER_01: And it comes up in other weird scenarios too, not just in the in in the setting you have here, but like there's a lot of settings where we have very specific human expertise, which um exists already, and there's no reason, and or we have specific equations of motion and equations in physics and so on, where there's no reason to have an AI system learn from data.
• 12:09
  SPEAKER_01: Like the whole point of learning from data is when you don't have enough understanding of the underlying phenomena to be able to use the data to maybe get some correlations out, right?
• 12:18
  SPEAKER_01: That's kind of one of the big benefits.
• 12:20
  SPEAKER_01: Um, or you don't have the right tools that already do a pretty good job with you know doing the prediction or the whatever task you want it to do.
• 12:28
  SPEAKER_01: This is an example to me, it seems like, okay, why why are they doing this?
• 12:32
  SPEAKER_01: They have the data, right?
• 12:33
  SPEAKER_00: Right.
• 12:34
  SPEAKER_00: It's like, yeah, sure, you have the data.
• 12:35
  SPEAKER_01: And there's enough stuff.
• 12:36
  SPEAKER_00: You can do machine learning and you can make all these fancy plots and you can add more data in and you know, get a prediction.
• 12:44
  SPEAKER_00: But you all you already had really good ways to predict this.
• 12:47
  SPEAKER_01: Yeah, yeah.
• 12:48
  SPEAKER_01: I mean, I think the hope is that you discover, I think the yes, I mean, I think it's a negative result in that sense, that nothing new was found.
• 12:56
  SPEAKER_01: But I that is still good science.
• 12:57
  SPEAKER_01: You still want to do things that maybe don't pr maybe are not always groundbreaking.
• 13:01
  SPEAKER_01: But it also makes you stop and think for a second is why do we need this?
• 13:06
  SPEAKER_01: Um, but I think you know, part of it is they were hoping that maybe there was a correlation, maybe there was an interesting connection that they could discover, right?
• 13:12
  SPEAKER_01: Between some other variables that weren't previously thought to be true.
• 13:15
  SPEAKER_00: Yeah, totally.
• 13:16
  SPEAKER_00: And and to and I want to step back and say, like, I'm excited to see more research happening in women's health.
• 13:22
  SPEAKER_00: Yes, yes, it is definitely an under-researched area.
• 13:25
  SPEAKER_00: Yeah.
• 13:26
  SPEAKER_00: So that's like on the one hand, super exciting to see research in women's health.
• 13:30
  SPEAKER_00: On the other hand, I'm like, I don't know that I would do anything differently with this.
• 13:35
  SPEAKER_00: Like, it's really, you know, they they were saying, oh, well, you could use this because it's so difficult to tell it's like not difficult to detect with the LH test.
• 13:44
  SPEAKER_01: I mean, presumably wearing a watch that tells you when it's when you're ovulating and when you're not is going to be easier than actively doing something like take a test, right?
• 13:53
  SPEAKER_01: You're just sitting, it's just there.
• 13:54
  SPEAKER_01: It's like an alarm thing that goes off on your phone or something.
• 13:56
  SPEAKER_01: You know, it's like similar to that, where there's no active involvement for you to do anything.
• 14:00
  SPEAKER_01: It just doesn't matter.
• 14:01
  SPEAKER_01: Sure.
• 14:01
  SPEAKER_00: So it's like passive, it's frictionless.
• 14:03
  SPEAKER_00: But it's interesting in this in this scenario.
• 14:06
  SPEAKER_00: So I started off by saying, hey, the best methods are cervical mucus checks and LH surge tests, right?
• 14:13
  SPEAKER_00: Which is just peeing on a stick.
• 14:15
  SPEAKER_00: They they're not hard to do, they're not expensive to do.
• 14:18
  SPEAKER_00: Um that's not what they compared this to.
• 14:23
  SPEAKER_00: So they compared the wearable technology to how well like the calendar method and the body temperature method got it.
• 14:30
  SPEAKER_00: The worst methods were.
• 14:31
  SPEAKER_00: So, like, sure, it was better than those methods, but if you look at how good cervical mucus and urine LH testing are, like maybe this isn't any better than that.
• 14:42
  SPEAKER_01: Maybe not.
• 14:43
  SPEAKER_01: Right.
• 14:44
  SPEAKER_01: Right.
• 14:45
  SPEAKER_01: Yeah.
• 14:46
  SPEAKER_01: I mean, that's maybe the wearables don't have all the things.
• 14:52
  SPEAKER_01: Maybe the the, you know, yes, it's it's interesting to think about wearables as doing all of this, right?
• 14:56
  SPEAKER_01: But maybe they're not suited for every purpose.
• 14:58
  SPEAKER_00: And some, I mean, some people might want to use it.
• 15:00
  SPEAKER_00: I'm I don't want to say nobody would want to use this because you're right, it's frictionless.
• 15:04
  SPEAKER_00: Maybe people are already have like the ring or the fitness watch or whatever on.
• 15:08
  SPEAKER_00: But for all these people who don't own them or don't want to use them or don't want to use the data for the this reason.
• 15:15
  SPEAKER_00: Yeah.
• 15:15
  SPEAKER_00: They don't want their fertility tracked or whatever.
• 15:20
  SPEAKER_00: You don't need to bother.
• 15:21
  SPEAKER_00: Yeah.
• 15:22
  SPEAKER_00: Like there's already good solutions available for this.
• 15:25
  SPEAKER_00: Like, this is sort of in my mind, I'm like, was this a problem that needed to be solved?
• 15:29
  SPEAKER_00: I'm not sure this was a problem that needed to be solved.
• 15:31
  SPEAKER_01: I mean, this reminds me of blood blood uh sorry, uh cuffless blood pressure monitoring, which is also the similar idea that you use wearables for measuring your blood pressure, when in fact a cuff does it way better, right, than any wearable can.
• 15:47
  SPEAKER_01: And uh because of all the other issues with wearables, it's really challenging to get it to do it right and so on and so forth.
• 15:52
  SPEAKER_01: It's kind of the same idea.
• 15:53
  SPEAKER_01: But I I guess there, and there are other examples of when wearables are useful, right, for tracking certain sure.
• 15:58
  SPEAKER_01: I think they use it for tracking epilepsy uh you know events and things like that.
• 16:02
  SPEAKER_00: They've even used it for like you know, back when COVID was a big deal.
• 16:07
  SPEAKER_00: It was like, okay, we can detect like a few days ahead of time your heart rate's going up, your temperature's off, whatever.
• 16:13
  SPEAKER_00: You're not sleeping as well.
• 16:15
  SPEAKER_01: Yeah.
• 16:16
  SPEAKER_00: So there's there's definitely like a lot of data there.
• 16:21
  SPEAKER_00: And the data can be used for different purposes.
• 16:24
  SPEAKER_00: And and fertility is one possible application.
• 16:28
  SPEAKER_00: Yeah.
• 16:29
  SPEAKER_00: But but you're right.
• 16:30
  SPEAKER_00: At the same time, you're like, well, we already have good tools to do this, and they're not that hard.
• 16:34
  SPEAKER_00: Like taking someone's blood pressure, really not that hard.
• 16:37
  SPEAKER_00: Peeing on an LH test strip, really not that hard.
• 16:40
  SPEAKER_00: Checking your cervical mucus, really not that hard, is is a wearable, uh, all is also really not that hard.
• 16:48
  SPEAKER_00: Yeah, I suppose.
• 16:49
  SPEAKER_00: But it's just like I guess it could add to your list of tools.
• 16:52
  SPEAKER_00: But I it does come back to like just because you like can do it doesn't mean that it's something that's that's super that's additive or super useful.
• 17:01
  SPEAKER_00: Right, right, right.
• 17:03
  SPEAKER_00: Either.
• 17:04
  SPEAKER_00: Yeah.
• 17:04
  SPEAKER_01: I mean, it's also uh I think a bit of a cautionary note in terms of just like, you know, as we as we we learn about these um, you know, new new pieces of technology, new, new, new uh sensors that come in with these things, you know, um, you know, companies are very quick to come out and say we can do all these new things as a sales pitch, right?
• 17:24
  SPEAKER_01: Because they want and sell it to you for more money.
• 17:27
  SPEAKER_00: Look, your wearable can do this, get this add-on to the app.
• 17:30
  SPEAKER_00: Right.
• 17:30
  SPEAKER_01: And the point is, A, it may not do it as well as you think it might.
• 17:34
  SPEAKER_01: And B, um it yeah, and and and that sort of might be the same point, which is that the other tools that you have to do the thing are actually better and maybe even as easy as the as the wearable is.
• 17:46
  SPEAKER_01: So, you know, it's it's it it it's sort of I'm I'm hoping that we can you know have people think about this a little bit and and sort of integrate it into their choices when they have to buy these devices, right?
• 17:55
  SPEAKER_01: I mean, that's kind of where we're getting at with this, is not not every piece of d data or wearable feature is necessarily a good thing or necessarily gonna be better.
• 18:07
  SPEAKER_00: Yeah, I I think people assume that it's gonna be better, right?
• 18:10
  SPEAKER_00: Because it's like, oh, it took in all this data, or there's AI behind it.
• 18:13
  SPEAKER_00: There's some machine learning algorithm that's included here, and that's why I'm gonna pay$50 a month for this add-on, you know, to my wearable device.
• 18:21
  SPEAKER_00: It's like, well, more data isn't always better.
• 18:24
  SPEAKER_01: Right.
• 18:24
  SPEAKER_00: I'm not sure in this case.
• 18:26
  SPEAKER_00: Yeah.
• 18:27
  SPEAKER_00: And I do think in general, it's like when it comes to deciding on research problems, it's like, well, how big is this problem that we're trying to solve?
• 18:34
  SPEAKER_00: Is it really a problem?
• 18:36
  SPEAKER_00: Right.
• 18:36
  SPEAKER_00: And in this case, like, I'm like, I don't know.
• 18:38
  SPEAKER_00: This was this was like really a problem because there were already solutions that are really good for it.
• 18:43
  SPEAKER_00: Yeah, I agree.
• 18:45
  SPEAKER_00: So it's uh it's a it's a good topic.
• 18:48
  SPEAKER_00: I'm glad that we got to discuss it today.
• 18:51
  SPEAKER_00: And um, you know, AI has lots of really cool applications.
• 18:57
  SPEAKER_00: And I think asking questions and also like you pointed out, having studies that show, okay, like it it did fine.
• 19:04
  SPEAKER_00: We didn't find anything extra.
• 19:06
  SPEAKER_00: It it found that body temperature was a thing that correlated pretty well, and that's what we kind of knew already.
• 19:12
  SPEAKER_00: We just detected it in a different form via wearables.
• 19:15
  SPEAKER_00: Like, okay, it's not it's not necessarily additive, but it is, you know, data that they were able to sift through and publish and see if there was anything else to add.
• 19:26
  SPEAKER_00: And there really wasn't much else.
• 19:27
  SPEAKER_00: And I'm glad to see that they published something on it still.
• 19:30
  SPEAKER_00: And I am very glad to see women's health being represented in this space.
• 19:35
  SPEAKER_00: So I think we can end here.
• 19:37
  SPEAKER_00: We will see you next time on Code and Cure.
• 19:39
  SPEAKER_01: Thank you for joining us.