This episode was recorded at the 2025 Western Dairy Management Conference in Reno, Nevada. Dr. Cantor gives an overview of her presentation at the conference, focusing on data from accelerometers and robotic feeders to predict calf sickness. While the correlations are there and we know calves change activity, behavior and feeding behavior before they get sick, there is more work to be done before the technology is ready for wide implementation. When data from both accelerometers and robotic feeders were used, Dr. Cantor’s group was able to find respiratory disease with a 96% accuracy six days before clinical symptoms. (2:36)
This episode was recorded at the 2025 Western Dairy Management Conference in Reno, Nevada.
Dr. Cantor gives an overview of her presentation at the conference, focusing on data from accelerometers and robotic feeders to predict calf sickness. While the correlations are there and we know calves change activity, behavior and feeding behavior before they get sick, there is more work to be done before the technology is ready for wide implementation. When data from both accelerometers and robotic feeders were used, Dr. Cantor’s group was able to find respiratory disease with a 96% accuracy six days before clinical symptoms. (2:36)
Dr. James and Dr. Cantor discuss the use of robotic feeders in the industry and the under-utilization of data collected by the feeders. Dr. James shares observations from a farm he works with about heifers coming in to the milking herd who were raised on robotic feeders compared to those raised in calf hutches. (6:15)
The panel discusses the accuracy, specificity and sensitivity of the predictions from monitoring technologies. They also touch on challenges around deciding what parameters to use to classify an animal experiencing the onset of clinical disease and how that will vary depending on the disease. They go on to share their experiences with training algorithms and how computer scientists have different goals than animal scientists with this type of technology. (11:17)
Dr. James talks about how data collection and using data can be a hard sell on some calf ranches. The panel talks about some of the challenges they have seen with adoption of technology and recordkeeping on dairies of various sizes. (28:30)
Dr. Giordano gives an overview of his presentation on using monitoring technology in fresh cows to predict disease. His group has worked with wearable sensors that monitor rumination time and physical activity. More recently, sensor companies have added eating behavior and body temperature. Variations in these parameters create a health alert to check on that particular animal. (39:08)
He goes on to describe two extremes in dairy farms. One spends little time and effort on looking for sick cows, while the other puts a lot of time and effort into this task. He discusses how bringing technology to these two types of farms benefits them and what drawbacks there are, along with an economic analysis for each. (43:14)
The panel discusses how implementing monitoring technologies require a change in management. Allowing animals the opportunity to express their natural behavior is critical to success. They also talk about how veterinarians view this technology and the target age for calves to best learn how to use a robotic feeder. (48:54)
Panelists share their take-home thoughts. (57:11)
Please subscribe and share with your industry friends to invite more people to join us at the Real Science Exchange virtual pub table.
If you want one of our Real Science Exchange t-shirts, screenshot your rating, review, or subscription, and email a picture to anh.marketing@balchem.com. Include your size and mailing address, and we’ll mail you a shirt.
Scott (00:00:07):
Good evening everyone, and welcome to the Real Science Exchange, the pubcast where leading scientists and industry professionals meet over a few drinks to discuss the latest ideas and trends in animal nutrition. Hi, I'm Scott Sorrell. I'm gonna be your host here today. And in this episode we're gonna be talking about technology, technology both for, uh, adult cattle as well as calves. And joining me as co-host is, uh, Dr. Bob James. Uh, Bob, thank you, uh, for, for joining us. This is not your first time to the pub, so you must like it here. I do
Dr. Bob James (00:00:37):
It's a, it's a wonderful, wonderful program, but a great history.
Scott (00:00:41):
Yeah. Good. So, well, thank you for joining us and, um, talking about Cavs. Today is gonna be Dr. Melissa Cantor from Penn State University. Kind of got the Eastern Scene Board covered, uh, today with Virginia Tech, uh, Penn State and Cornell. And we've got, uh, Dr. Julio Gano and from Cornell, and he's gonna be talking about technology and how it's used in, uh, uh, cows. And both of these folks have given presentations here at the Western Dairy Management Conference, and we're gonna kind of use that as our, as our guidepost, at least in the beginning for these conversations. Um, Melissa, why don't we start with you? Let's, let's, let's hit those calves first. Um, just kind of gimme a brief overview of what, uh, what your talk was about today.
Speaker 3 (00:01:33):
New research is changing everything we thought we knew about Choline's impact on the cow and her calf and top scientists have a lot to say about it. They're presenting new research that supports choline as a required nutrient to optimize milk production choline as a required nutrient to support a healthy transition choline as a required nutrient to improve calf health and growth and choline as a required nutrient to increase colostrum quantity. This new research is solidifying choline's role as a required nutrient for essentially every cow, regardless of health status, milk production level, or body condition score. Learn more about the science that is changing the game and the choline source that is making it happen. ReaShure Precision Release Choline from Balchem visit balchem.com/scientistssay to learn more.
Dr. Melissa Cantor (00:02:36):
Absolutely. Yeah. So my talk today was really to, it described to everyone where we're at with calf management and related to disease detection and related to actually doing something with the alerts once we have them. And what we've actually tried and what has worked and what has not. Um, particularly, we've really kind of found out that the correlations are there. We know that calves change their activity, behavior and their feeding behavior before they get these diseases like diarrhea and respiratory. But we're not quite at the point, while AI shows the potential for disease detection in calves, we're not quite yet to hit the go button. We're still got some work to do. And so that was really the big synopsis. Synopsis.
Scott (00:03:16):
And what technology were you using?
Dr. Melissa Cantor (00:03:18):
Yeah, so in this particular studies we were using accelerometers pedometers that we put on the calves Okay. For the activity. And then for the robotic feeders, we were collecting all the information from those, so that would've been through milk meters and also manometers.
Scott (00:03:32):
Okay. And you were able to get individual, uh, data on from, from that. Okay. Yeah,
Dr. Melissa Cantor (00:03:38):
Absolutely.
Scott (00:03:38):
Yeah. And what can you get from the, I forget what you call the kind of pedometers
Dr. Melissa Cantor (00:03:42):
Yeah, the little, the pedometers, the Fitbits
Scott (00:03:43):
For cows. Yep, yep,
Dr. Melissa Cantor (00:03:44):
Yep. Yeah. So very similar to what we would get on the transition cows, but more just the activity behavior. So things like lying times, step counts, uh, how often the calf gets up and lays down. And even activity index are all things we can use that we know are going to be associated with these diseases that
Scott (00:04:01):
Cavs get. Yeah. And can you measure play and does that have any indicators? Uh, yeah.
Dr. Melissa Cantor (00:04:04):
So there, in terms of health's, great question. So the play behavior, there's one particular study out there that has found that play is, uh, correlated with the particular activity index that we used for the studies. It was, uh, from peacock technologies was the type of pedometer that we were using for those. And, um, but we haven't really looked at plate behavior related to disease yet. Okay. More work to be done.
Scott (00:04:23):
Okay. Very well. Yeah. And then the activity between, between the pedometers and, and uh, consumption. Any correlations there?
Dr. Melissa Cantor (00:04:33):
Yeah, so really the idea is the more information we put together, the better that we can detect disease. So really when we're taking pedometers plus a robot, we can actually find disease with a 96% accuracy six days before that calf has the clinical lung consolidation and the outward signs of pneumonia such as coughing and eye discharge, nasal discharge, those sorts of things. Okay. But obviously,
Scott (00:04:58):
So you reduced activity. Yeah, there's reduced, so they probably activity temperatures or having just feeling malaise.
Dr. Melissa Cantor (00:05:03):
Surprisingly, not all of them. So a lot of the, so we, we suspect that a lot of these were viral infections in calves, um, which don't always necessarily spike a fever. Some did, some didn't. But yeah, we did see, what was really interesting to me is taking the combination of the two technologies together improves the detection accuracy better than a person. Yeah. Um, but of course that's not real reality. Most farms can't afford a robot and pedometers for their calves. So the next step is really looking at just using one. And we still get increased accuracy doing that, but not 96%. Right. Yeah. Yeah.
Scott (00:05:37):
Okay. So you mentioned Fitbit and I'm, I'm probably already know the answer, but you know, my Fitbits measuring my heart rate. You have a Fitbit Yeah. And oxygen levels. Yeah. All the things. I'm gonna assume that that what the technology you have, you're not doing that at this point?
Dr. Melissa Cantor (00:05:51):
Not at this point. We've actually explored looking at at Pulse oximeters Yeah. Which is what you have in your, your Fitbit or your Apple Watch. And um, unfortunately calves are really good at taking those off. Okay.
Scott (00:06:10):
Yeah. Okay. Bob, I think you were at the presentation. What were some of your takeaways?
Dr. Bob James (00:06:15):
Well, you know, I, I come at from it. We, uh, before I retired several years ago, um, we did a, a study of our LA one of our last studies was really looking at, at some field situations. We had, uh, six farms in Virginia. Uh, we collaborated with Sandra Godden and, and, uh, one of her graduate students in doing some work at looking at, um, looking at behavior of calves and the data from the auto feeders pri primarily, and then disease reporting. And, and I think in, in talking to Melissa today, um, you know, we had farms where we had a lot more restricted intake. And, and I think we're just learning so much more, uh, about calf behavior. I, you know, I've written an article in, uh, in popular press and it, it's titled What Auto Feeders Have Taught Me about Feeding Calfs. We've had so many misconceptions about, about what's normal, you know, 'cause what had been normal for us, essentially is starving the poor little things to death.
Dr. Bob James (00:07:13):
And it's a very different behavior. You know, when you look at these animals when they're have satiety, you know, they're just kind of cool. And, and so I think that's the, the really exciting thing. I, I'll end with one last thing. I I, uh, met a person in, in England and, and Melissa May come and talk about this. And she was a business person. She had a dairy beef operation and she had 800 calves on auto feeders and, and really good health records. And she was highly motivated to do better. And I said, I have just the person for you,
Dr. Melissa Cantor (00:08:12):
And every farm does something different. Right. Some don't look at the data at all. Others are making treatment decisions off of that information. And yet, wait
Scott (00:08:19):
A minute, back up. Why are they not looking at the data at all? Well,
Dr. Melissa Cantor (00:08:23):
A lot of them are just feeding, feeding cats. Yeah.
Dr. Bob James (00:08:25):
They did it 'cause it's gonna save labor. And that's the really sad thing. You know, and, and they just don't. And quite frankly, that's one of the early promoters was, oh, it's gonna save labor. And when, you know, I have, I, I did another one, another presentation that's called, are You a Calf Manager Or Calf? Feeder? And historically on dairy farms, um, people were calf feeders. You know, what data did we have? Uh, you know, sometimes treatment mortality. And that was about it. And so, uh, you know, I work with some other operations that are not, um, auto feeder. And getting them to develop data is like pulling eye teeth and to make any kind of management decision. So I just get really excited with the possibilities. Um, Melissa just really did a super job of kind of covering the waterfront far more than just the data, but also where, what some of the other opportunities are.
Scott (00:09:22):
Yeah. I I kinda wanna get off topic here for a second real quick. Sorry.
Dr Julio O. Giordano (00:09:25):
Really? But I got some questions for Melissa.
Scott (00:09:26):
Yeah, yeah, yeah.
Dr Julio O. Giordano (00:09:32):
You told me that you wanted
Scott (00:09:33):
Me to ask
Dr Julio O. Giordano (00:09:46):
Scott (00:09:46):
They are fun. Um, you mentioned behavior and I, I'm remembering back to a presentation that Nina Van Kaiser link gave and, and, and she was talking about more of the cultural behaviors. And I'm wondering, are there any learnings that you've gotten, you, you said you know about feeding, but what about culturally and, and interpersonal relationships, that kinda stuff with cats? Well, one,
Dr. Bob James (00:10:07):
One of the farms that I work with in Virginia is very interesting. They bought one feeder about 10 years ago, and they fed a third of their calves. And this is strictly circumstantial on the auto feeder, the other two thirds of the calf and the calf hutches. And so, you know, four or five years later, they saw these animals coming into the Milken string. And the ones on the auto feeder were, it's a very well managed, or boy, they were, some of 'em were two months younger. They were better grown. And the other thing is they were so much more easier to handle. And I wanna make one comment to kind of serve as the lead in before I forget it. And the other farms I've dealt with who started out with robotic calves, they made much better robot cows. And I think that's where the industry has just flat dropped it. They have not looked at that tie in. And I think that's, that's a critical part of getting that data from the calf all the way through into the cows record. And I think there's a super opportunity, 'cause if we do something with the calf, I wanna know that she's going to last longer in the herd and milk better.
Scott (00:11:12):
Yeah, exactly. Uh, great, great
Dr. Bob James (00:11:14):
Point. I'll try and be quiet now.
Scott (00:11:15):
Well, yeah,
Dr Julio O. Giordano (00:11:17):
So Melissa, um, and I may have missed this from your great talk, by the way. Um, but can you tell us a little bit more, because we are talking about science, right? I mean, like, so yes, the out of feeders and the, uh, activity levels, uh, from the, uh, accelerometers. Uh, but which ones were more, as, you know, out of the parameters, you know, monitor by each one, which ones were, if you have data or if you have a particular thought, the most predictive, right. Like mm-hmm
Dr. Melissa Cantor (00:12:07):
What would it be? Yeah. So, so I can tell you what we currently know, and some of this is my PhD student's work to answer that question that you're asking for diarrhea with the robotic feeder only, not, not accounting for a pedometer milk and intake plus drinking speed seems to be the most valuable information for predicting a calf's about to get clinical diarrhea. And that's a 24 hour window that we usually use for that pneumonia. Big old question mark. We know, we know the correlations are there. We know when use everything together, we have great accuracy. We know that when we restrict to just one of the technologies, we improve accuracy. But which one is best is still a million dollar thing that needs to be answered. And so the University of Guelph and myself are both trying to answer that question in different ways. They're working on the pedometers and I'm working on the automated calf feeder data.
Dr Julio O. Giordano (00:12:54):
Okay. Yes. Yeah. And then
Dr. Melissa Cantor (00:12:55):
Great
Dr Julio O. Giordano (00:12:56):
Question. So when, uh, you mentioned 96% accuracy and, uh, you know, that's, you know, kind of an oversimplification sometimes of what it is, right? Absolutely. 'cause accuracy is an overall measure of the performance of a test to tell you if something, you know, it's positive, negative, whatever it is. Absolutely. Can you expand a little bit more on, you know, sensitivity and specificity, the predictive values, right. Because you can be highly accurate and, and have really good sensitivity, but you can do really poorly with specificity and, you know Yeah. Both. That is the scenario trade off, right? For farm management, right. So the trade off, so you can comment a little bit more.
Dr. Melissa Cantor (00:13:34):
Absolutely. So, yep.
Dr. Bob James (00:13:36):
Clarify sensitivity and
Dr. Melissa Cantor (00:13:38):
Specificity. I was just about to do that. Good. Yeah. Okay. Yeah. So sensitivity is gonna tell us how many positive animals are actually truly positive outta everything that that text says is positive. And specificity is the opposite. It's gonna tell us the healthy animals that are actually truly healthy, right? And, and because of the way that they're calculated, they're almost inversely related. And there's always this balance, and it's a real trade off in the system. And the particular case of pneumonia, a day de delay for finding that calf is a big deal, right? So I care more about a high sensitivity at the sacrifice of having some healthy calves that we're gonna have to weed out. Um, and so to answer your question, in those particular cases we were shooting for really high 90% or greater sensitivity at the sacrifice of not the best specificity. Yeah. How did,
Dr Julio O. Giordano (00:14:25):
How do, how did it go then? I mean, because you, you know, you can have awesome sensitivity Yeah. You just check everyone, right? And you'll never miss anyone. Right? So how, how did it play out?
Dr. Melissa Cantor (00:14:35):
Yeah. So, so to answer that question, um, and, and, and again with pneumonia, we haven't gotten there yet. Um, we're at the point where we know the machine learning potential is there, but we haven't actually tested it in real time on a group of calves. That's, that's what's Brianna is working on right now, is collecting a huge data set of 2,500 calves from one facility, from 20 different farms all over the country. So we have all the variability in the world in that data set to do exactly what you're saying, because that's great if it flags everybody, but then it's a waste of my time because I'm looking at everybody, right? So there needs to be a labor savings there. Um, but the k and n it, you know, that's the thing I was working with computer scientists and we're fighting interests. Their interest is the best algorithm ever. My interest is the pharm realistic one. Right? So this time I'm doing it my own way with my own data set to see if we can answer some of those questions. Very good question though. I think
Dr. Bob James (00:15:24):
The challenge too is identifying what is pneumonia. Yeah. And that, that's the big challenge I see that on the calf ranches that I work with is, is what's pneumonia? 'cause I'll see repeats on there and you wonder, you know, when is it a new case and when is it a and, and I think there, there tends to be, and I think that's the beauty of the data set here from what I understand with Melissa, is that you have a lot of confidence in the health records.
Dr Julio O. Giordano (00:15:49):
Actually, that was, that's a my follow up question. I figured that. So if you, if you can describe a little bit like, you know, what exactly was done to evaluate the clinical health status of the animal? Yes. Because, you know, and this is a very important thing that I'll emphasize a lot when I talk about the cow side of things. It's all about what you compare it to. And it's all about, you know, we talk a lot about machine learning and all of these things is what do you use to train the algorithms? And, uh, you know, as someone who was, uh, originally trained in repro, and I always tell people, you know, it's like, I don't know why in the world I decided to start working on health. Because health is kind of a health
Dr. Melissa Cantor (00:16:30):
It's what's called health.
Dr Julio O. Giordano (00:16:31):
Yeah. What's diarrhea, what's metritis, what's ketosis. Right.
Dr. Melissa Cantor (00:16:35):
And every farm's different
Dr Julio O. Giordano (00:16:36):
Pregnancy, like the cow is pregnant or not. Yeah. Period. Okay. Uh, but with health is such a challenge. Absolute. And then it's even like, when did the onset of clinical disease start? So, and unless you're looking at animals daily or even more frequently Yes. It's really, really hard. I mean, to interpret any data. So if you could expand on
Dr. Melissa Cantor (00:16:57):
That Excellent question. Yes. So when the machine learning has been done many times with retrospective producer reported records, but the issue is Yeah. The stuff, you know, how consistent is the stuff. Yeah. What I did for my PhD was every single day, seven days a week for the first 90 days of life for 120 calves mm-hmm
Dr Julio O. Giordano (00:18:04):
So maybe an an important question is, if you treat them at that stage, will there still be a positive treatment response? And that's a part that
Dr. Melissa Cantor (00:18:13):
I, everybody except fif, so 15% of them would need to be retreated. We ha we did follow them up for relapse. Um, so we left them for seven days and we didn't do anything because we wanted to look at recovery behavior. And that's in scientific reports. And we found, actually, you can even see who ca which calves get better with the technology, the pedometer was the most useful in that case. So calves that are going to recover, they actually increase their step counts, their activity index, their lying behavior three days as early as three days after the initial treatment with antibiotics. The ones that are going to relapse seven to 10 days later, those are the calves to really show you that's something's not right. And of course, we, for ethical reasons, we had to follow up at that point with antibiotic treatment. A different one. Yeah.
Dr Julio O. Giordano (00:18:55):
I think just following up and I'll, I'll stop
Dr. Melissa Cantor (00:19:00):
Phy, this is great, you
Dr Julio O. Giordano (00:19:00):
Rescued me. But you know, as, as we, as we talk about metrics of performance for technology and, uh, we use terms as accuracy, sensitivity, specificity with health, and I go back to the onset of clinical disease. And then, you know, when, you know when does it end or relapses is, um, when, when do you count a health alert as true, positive or true negative, right? So was it exactly on the day of diagnosis or is there a window of time around clinical diagnosis at, at which you say, well, and I'll use this example, right? If you find the calf sick tomorrow and the algorithm is telling you today that the calf has something going on, was that counted as a false positive or as a true positive? Just that there is a delay to clinical diagnosis. Wouldn't
Dr. Melissa Cantor (00:19:56):
You agree? That's disease specific, right?
Dr Julio O. Giordano (00:19:59):
Yeah. No, even the transition
Dr. Melissa Cantor (00:20:02):
Cows especially. Oh,
Dr Julio O. Giordano (00:20:03):
Absolutely. So we, we are using, using windows be really sensitive. Know it's super, super difficult. But you know, because I mean, you can punish your algorithm, you can make your algorithm look a whole lot better. And you know, the best example is if you use a window of, uh, 10 days after clinical diagnosis and your algorithm tells you eight, nine days later, useless fair. Right? Yeah. Right. It's useless, but it looks very good from an accuracy perspective. Yeah, absolutely. So those windows are matter. So critical, you know, both on how the algorithms are trained and then how they are validated or
Dr. Melissa Cantor (00:20:36):
Absolutely. So in the particular, all the data that we presented, um, except the diarrhea algorithm, so the pneumonia stuff for machine learning day zero, actual day of diagnosis, is it, and that's,
Dr Julio O. Giordano (00:20:46):
That's very tough for the algorithm. I
Dr. Melissa Cantor (00:20:48):
Mean absolutely. But that's why I tough, I tortured myself and followed 120 calves every single day for 15 months straight without
Dr Julio O. Giordano (00:20:55):
A day else. But for the training, you mean, or then for the validation. So 'cause for the training, yes. I mean, you can train them. Yes. This is the day, you know, you put 0, 0, 0, 0, 0 and then one Yes. Sick. Right. But for the testing dataset, there was only one day.
Dr. Melissa Cantor (00:21:11):
Yes. And we, and we also added, um, another dataset from the same farm to, to test that. And that would be the, the greedy one I told you about with the economics. That one was the one where we added more calves to penalize it even more, um, if it got it wrong. And the window for diarrhea, we allowed 24 hours because I think it's really important to at least have a watch list of who you need to go look at. Plus we were trying to test different treatments, so we wanted to have a window before they were actually sick.
Dr Julio O. Giordano (00:21:37):
That's interesting. And maybe something that people will learn today is that is a lot tougher with adult cows. It is not that simple. I mean, uh, uh, it's really hard, you know, to try to like match exactly the day off. Uh, you struggle quite a bit too with the fact that what we are measuring, these behaviors, physiological parameters, I mean they start changing before clinical diagnosis. Right? So, you know, your, your algorithms may start telling you before clinical diagnosis are present, that there's something wrong with the animal. And then it's, it is kind of a conundrum for us as scientists or for people in industry who are trying to bring this to farms is like, you know, how yeah. How do you train the algorithm and do you call it a false positive if it's telling you two days before? And if it tells you one day later or two days later is at the end of the world, does it mean that the algorithm will not be effective for on farm use? So this term of, we call it a window, so a window of time around clinical diagnosis a few days before and a few days after that, uh, you know, you have to give the algorithm a little bit of leeway there because if not, I mean, without the windows, our results for adult cows are disastrous. I mean, like, uh, a really, really bad and you know, again, the overall accuracy looks good, but your, your sensitivity or your specificity suffers. Your positive predictive values are really, really poor.
Dr. Melissa Cantor (00:23:05):
Absolutely. Yeah. Question to follow up on what you're saying. So two, two points. One is, I guess penalizing algorithm would be so disease specific, because obviously if you miss a DA for two days, a cow is probably going to die. Yep. Mm-hmm
Dr Julio O. Giordano (00:23:43):
So for the algorithms that we have been, uh, trying to develop, and in fact your point about the challenges with multiple disorders that may be affecting a cawa at the same time are, are very problematic, right? Because you have more than one thing at a time. We are trying to develop algorithms to identify cows for a clinical exam. Mm-hmm
Dr. Melissa Cantor (00:24:07):
Dr Julio O. Giordano (00:24:07):
Makes sense. That's what we're trying to do. Right? Got it. So that, that means that the algorithm has to identify cows that ha
Dr. Melissa Cantor (00:24:13):
All the things, all
Dr Julio O. Giordano (00:24:14):
The potentially any of the things. So that's the big challenge. Yeah. You know, which may be one way to get around that is like develop individual algorithms, right. That identify hundred percent agree. And then use what is called assemble algorithms, which are algorithms that combine multiple algorithms Yes. To come up with the one because yes, on farm we need very simple answer is yes or no. That's right. Is the cow on the list to look at now at 7:00 AM used to be or not? Yeah. Is that
Dr. Bob James (00:24:45):
Simple? Yeah. We tell people on a practical basis with the auto feeder, you walk the pens first and then you look at the data. Because if you look at the data first and then you ignore the rest of the animals in the pen. Yeah. And I think it, it's the same with, uh, you know, with with the cows. I, I, uh,
Scott (00:25:02):
Maybe a naive question, but, um, AI is advancing so quickly right now. Yes. Is will, will that make it easier, uh, to, to refine your algorithms?
Dr Julio O. Giordano (00:25:14):
Short answer, um, yes. We hope. I think that that's the hope of all of us who are, uh, in this field of science trying to use machine learning or other tools in artificial intelligence within the field of artificial intelligence to just do things better. Um, I'm positive and hopeful, but I'll say that it's not as straightforward as we may want it to be. And I think we still have a lot to learn and we are learning very fast. Just like these things are evolving. Yes. Or understanding, right? I mean, it's evolving.
Dr. Melissa Cantor (00:25:50):
New techniques are coming out all the time too. And you're like, okay,
Dr Julio O. Giordano (00:25:54):
So yes, new techniques, the variety, I mean, we just sent a paper to JDS, we tested, uh, 38 different algorithms on the same data set. So there are even those questions. Which ones do you pick? Yeah. And again, this one is better for this, this other one is better for this other thing. So, you know, we are all in this conundrum of, you know, how do we get to the very practical, simple, useful decision making tools that we need on farms. But again, the straightforward answer is yes, it will. You know, it's just going to take some time, some time to get there. Just like most things
Scott (00:26:30):
Gonna need more data.
Dr. Melissa Cantor (00:26:31):
Yeah, absolutely. And it's a little depressing, I feel like, 'cause I get calls all the time. I'm sure you do too. Yeah. It's awesome from farmers being like, what is the perfect system for pneumonia? And I'm like, we're not there yet. I wish we were. Yeah, yeah.
Dr Julio O. Giordano (00:26:41):
You know, and it's awesome to have such a diversity of technologies and algorithms and all of that, but you know, it's just a little bit overwhelming, you know, with the speed at which these things are evolving and how much we can do good science because, you know, you, you can do things you know, quickly and you know, and then over promise. And then when, when you bring these things to farm, you know, it can be a little
Dr. Melissa Cantor (00:27:06):
Frustrating. I completely, I completely agree with you. And, and, and again, I'm sure you're collaborating with computer scientists also. We have to mm-hmm
Dr. Bob James (00:27:40):
The other, it's between sensitivity one. The other one I think too is, is I look at, at farms that have larger number of calves on auto feeders is tell me which pen I need to go to. You know, and I, I've been talking to the, 'cause the, the companies love to develop technology, but they're the priority for using that technology is not quite there. I mean, it's just not as, as important for them, at least on the calf side of things. And I think, I know Melissa can say that, and you mentioned it very well, tell me what feeder I need to go to and then tell me when I go on that feeder, what calfs do I need to look at? And I don't care whether that's diarrhea or respiratory. You know, you want to, that's when, and I think you showed that in your data, when you combine the, the, the data as well as the observation, there's where you bumped up, you know, particularly from just the observation. There
Dr Julio O. Giordano (00:28:30):
Are a couple of things that you guys keep saying about calves that have me very curious
Dr. Melissa Cantor (00:29:00):
Yeah. So to, to kind of, to kind of say that I'm saying the potential is there, like we, we, we've actually restricted algorithms with constraints surrounding economics that a farmer can actually afford rather than Melissa's perfect lab scenario with 500 things. Um, and it still increases that detection accuracy compared to just paying a really trained person to find them. And if we think about where we are with the prevalence of respiratory disease and diarrhea in our calves on the average dairy, I am not talking about the five top 5% dairies. They don't need the technology. Right. I'm talking about that average farm. There's opportunities. I
Dr. Bob James (00:29:34):
Disagree with you there. You disagree. Even the, even the better farms need it, you know, and, and I think I, again, I come back to saying, tell me, you know, I I just want to, which animals do I need to look at and what are my trends? The other part that I think is important to say is, is looking at this data being part of the, their lifetime records coming into the, into the her value as a cow. 'cause I hope that what I do is far beyond just the value of treatment and response. But does she,
Dr Julio O. Giordano (00:30:03):
That's just one little piece.
Dr. Bob James (00:30:04):
Sure. Does she just one little piece, does she milk better and so much more last longer in the herd?
Dr Julio O. Giordano (00:30:09):
Yeah. And the thing is that, you know, these are the more tangible things that we have been, you know, tangible as well as, you know, the elephant in the room to start with. You know, let's find sick calves. Let's find sick cows now, you know, so we can treat 'em and all of that. But then there is a whole lot more underneath, right? Oh yeah. Oh yeah, you're right. And data is gonna navel preventative. So cows should not get sick. Yeah. Calves should not get sick. We should rescue and clinically sick. Right? So we, we, you know, we should change management. We, we should implement interventions before these animals go into clinical disease. Right? And, and that's where I think that technology is going to completely transform what we are doing. It's just that if it is a little tough to do what we are doing now, that other piece is even harder. Right. It's very hard, you know, than did
Dr. Melissa Cantor (00:31:00):
The treatment work or your algorithm or what. Right.
Dr Julio O. Giordano (00:31:02):
Because you're even challenging a whole lot more about just like risk. I mean like, you know, once that an animal has a disorder, you know it's there. Right. You know, you have the clinical science, you have something to look at. You know, let's say we are, we are trying to predict, you know, if a cow in the pref fresh pan is gonna have a clinical disorder when she's in the fresh pan three weeks into lactation, you're pretty far out from that. And then all that you're doing is like, what risk, you know, what's the risk of that animal going there? And then when start getting into risks, then you know, you have start having issues with are you gonna implement management strategies just based on risk? Are we going to start managing different subgroups of cows just because we know that there's a risk, but there's an inter you know, there's a confidence interval where that risk, you know, could be a healthy cows.
Dr Julio O. Giordano (00:31:53):
Right? It could be, you know, there are healthy cows just like we, we even have now cows that, you know, have alerts and are okay, there's nothing wrong with look okay to us. Yeah. It's gonna be even more difficult. And then, you know, you have so much time until, you know, when you are measuring where you're measuring and when it might be manifested. A lot of things can happen in between too. Right. And and to your point, Bob, you know, and, and your point about farms not being a lab, I mean, dairy farms are the most dynamic systems I think probably that we have out there with Yeah. Environmental factors and probably the worst labor factors. Yeah. Ask humans, you know, you know, getting in the way of a lot of things. So we may be really good at predicting something, you know, let's say that we predict cow health postpartum before calving and then someone completely messes things up at calving. Right. That completely throws off any predictions that we do before. Right. So there are all these things that, uh,
Dr. Bob James (00:32:49):
Yeah, I wanna come back to the practical side. Yes. Yeah. Because that's, you know, I work with, with kef ranches and I work with, and I, where I really develop the insight with the auto feeders was working with some, several really large herds in Australia that had 800 to 1200 calves. And, uh, and you think it, and then I work with the other ones with the calf ranches. So what data do we collect on calves on most farms mortality? How many of 'em collect treatment data so they have nothing? Yeah. And so anything we do is almost a quantum leap
Dr. Bob James (00:33:40):
And I thought, ah, so in my conventional farms, that was a little report item that I had. And all you had to do was for the morning. And if they don't drink in the morning, they're, they're not doing well. Tell me how many calves drank slowly this morning. And so the manager gets the report the next day and whoa, we went from 10 to 20. Something's happening here. And so I think we're being a little bit tough on ourselves too, particularly with calves. 'cause we started it, nothing. And, and it's still a challenge right now. This farm has, it has a, uh, they carry handheld. So we are recording health, uh, treatments on that. And that goes into the data set. The problem is, you know, how soon we detect, but at least we've got that information now and it's pretty exciting. 'cause you look at having, you know, thousands and thousands of calves every day. And so that gives you a little bit more confidence. But, you know, I, this is something that I really struggle with with farms is getting them to collect anything at all. So that's where I'm a real optimist with the auto feeders. And, and I guess I got real excited, you know, when I went to some of the first robot farms and you look at the data that the robot dairy looks at versus the conventional dairy. And it's a different mindset. It's a different mindset
Dr. Melissa Cantor (00:35:01):
If they're not just feeding their caps. Right. The ones actually look at the data.
Dr. Bob James (00:35:03):
But I said, if you look at the industry, it's, that's
Dr Julio O. Giordano (00:35:07):
Where it's going. Yeah.
Dr. Bob James (00:35:08):
Pretty slim. You know? Sure. I look at the bigger calf ranches and they record arrival and that, no, they record treatments mostly and we're working at getting, uh, birth weights. But that's a struggle. I do record. I've got this farm and we're recording weaning weights. 'cause we can do that electronically and getting to where every time we've gotta move them, they go across the scale. And so now we can get performance. But boy, showing the value is just really hard. Um, on those fac on those, uh, dairy calf and heifer facilities, how
Dr Julio O. Giordano (00:35:40):
Can it be hard? I mean, with all
Dr. Bob James (00:35:42):
The, it's, it's showing them the value of that. And, you know, as I say, I get to work with a lot of people. And, and that's my biggest challenge is recording that stuff, uh, initially. So I look at the auto feeder and boy, it gives me a jump on that, except we don't have it in a report format. It's all on individual information and there's no summary of that information yet. Yeah,
Dr. Melissa Cantor (00:36:04):
That's that's very true. Yeah. You can make it challenging. Um, one thing I can say is I think, you know, a lot of, there's a lot of sensors coming out now, um, for calves and there's quite a few on the market actually. And I think that is what some of these calf ranches are thinking about. Yeah. Is can I see if the whole group is off? Right. Can I see if, you know, everyone's not drinking their milk. Right. Um, so to kind of look, to kind of predict for something like an epidemic, like I know with the bird flu, there's anecdotal information. I don't know if anyone's actually gonna publish it using te ear sensor data from transition cows and other cows to look at, you know, the potential onset of something like bird flu in their herd. Um, and then that, that's something that I don't know. I've heard farms talk about it. And in California and even the tech companies, I don't know if they're gonna publish it, but I do think there is an opportunity even for the really simple calf ranch Oh yeah. To look at things, to see if something is about to go wrong. Like if we think about salmonella Dublin, it's a nightmare for a calf ranch. Sure. You know,
Dr. Bob James (00:37:01):
But you look at, you know, you look at the classic calf ranch, and boy, they're a low cost operation, they're fortunate to work with one that's not that mindset, but to say you're gonna adopt this technology. Now, maybe when labor starts becoming a lot more of a challenge, that may be something. And uh, uh, but boy, just getting data, getting them to, to see the value of the data. And we can talk all we want to, but it's a tough sell. Yeah. Because they're, they're, they're a really low cost model by and large. And I look at the data that comes from a lot of these places and it's, it's pretty minuscule and it's largely just health data and that's it.
Dr. Melissa Cantor (00:37:41):
Yeah. I I, I, I could totally agree with you that some of the larger ranches aren't thinking about it yet, but I can tell you a lot of medium to a thousand cow herds are calling me. Yep, yep. And they're asking very different, very different. They wanna know. Yeah. They're like, what can I buy for my calves? Which is shocking. 'cause five years ago people laughed at me that I was doing this type of
Dr. Bob James (00:37:58):
Research. No, I think that's the perfect model is the 500 to thousand, maybe even 2000 cow mm-hmm
Dr. Melissa Cantor (00:38:17):
Well, they have the herd size to make that work. Right, right. I should clarify.
Dr. Bob James (00:38:19):
Right. They have the people and they people, the age difference is, well, but you look at the literature and some of the European literature said, oh, don't have more than 10 calves in a pen. What? That's economic ridiculousness. I mean, you can't afford it.
Dr. Melissa Cantor (00:38:32):
Well, it goes back to what Julio was saying about risk. Right. Because there are some farms that can handle that increased risk, increasing the group size. But the reality is that if you're gonna make a universal blanket statement, oh yeah, you should go for the smaller group because everyone can do it covering yourself too. Yeah. Yeah. Exactly. You may recommend the 20 calf group. You better have an average maximum age between calves of two days Yeah. To make that work there. There are no absolutes, no
Scott (00:38:57):
Julio, can you talk a little bit about the presentation you gave today? I mean, I'm sure we've already kinda interspersed some of that in there, but
Dr Julio O. Giordano (00:39:03):
A little bit, a little bit of
Dr. Melissa Cantor (00:39:05):
Yeah. You should talk about the economic part. 'cause I think that's really cool. Okay.
Dr Julio O. Giordano (00:39:08):
Very good. So, well, I talked obviously about adult cows and, and primarily, uh, most of our work has been done on identifying, uh, fresh cows with health disorders. And, you know, the reasons are obvious. That's when cows, you know, have the highest risk of having, uh, clinical health disorders of interest other than mastitis, which happens throughout the entire lactation, metabolic and digestive disorders. Metritis, uh, happen, uh, very early on. And, um, so, uh, today I briefly, you know, uh, summarize what we have learned in terms of, uh, again, how the patterns of the, uh, parameters measured by the sensors change. And again, like Melissa said, for caps tons of data. Yeah.
Scott (00:39:54):
So what sensors were in technology were you using?
Dr Julio O. Giordano (00:39:56):
Very good. So, um, we have done a lot of work with, uh, wearable sensors that monitor rumination time and physical activity. And that combine rumination time and physical activity to create these health alerts. Which by the way, uh, maybe it's worth clarifying a little bit. I mean, a health alert is basically a tool that tells you, again, it's this yes no thing. Or at least some is like a gradient is like, uh, green, yellow, red, or a number that, you know, below a certain threshold or above a certain threshold, someone should look at the individual because something may be happening with the animal. So anytime that I say a health alert, that's what I mean. It's like a yes no thing that tells a farm worker to go and look at that animal or not look at the ones that don't have the alert. So rumination physical activity, we have used quite a bit, uh, daily meal weights. So drops in milky, which is one of, if not the oldest, the oldest parameter that dairies have been using to identify cows with health disorders. Which by the way doesn't work that well. Yeah.
Dr. Melissa Cantor (00:41:11):
Conductivity doesn't work either.
Dr Julio O. Giordano (00:41:13):
Scott (00:41:15):
Again? Were you measuring by quarter?
Dr Julio O. Giordano (00:41:17):
No, no. Okay. Just cow level milk yield, cow level milk yield. Uh, no. Uh, for milk kill conductivity, no. Yeah. You see, yeah, you can do quarters in robots or you can do a, a whole comite sample in, uh, in the, in any case. Um, so, and then, uh, the para, so eating behavior as well and body temperature are more recent. So rumination physical activity are the ones that we had had for the longest time. And more recently, companies added eating behavior and which is not feeding intake, eating behaviors, eating behavior, you know, is the cow eating at the pump, maybe typically the cow with her head down at the feed bank. There is, by the way, at least one system now that, uh, does measure intake, which we're working with, uh, with the company. And um, is that the AFA milk one? Yeah, the Africa, we had them in a, in a podcast.
Dr Julio O. Giordano (00:42:11):
Oh, you have? Yeah. So we, we collaborate with afi. We have been collaborating for, for quite a while, and we're working with them on understanding their data and putting it to use and doing a little bit of validation as well. So, um, those are the parameters. And you know, we have done a lot of research, my group as well as many other groups, uh, in the US and overseas that have done work and parameters measured by sensors change when cows have a clinical health disorder. So that was a simple message. Now how do they change? So the magnitude of the change and then the timing of the change varies by the parameter being measured by the sensor, by the sensor itself. Different companies have different algorithms and different ways of measuring and reporting things. But I think that the most important part of my presentation was, um, what's the value proposition of these technologies for integration on day-to-day operation of dairy farms?
Dr Julio O. Giordano (00:43:14):
And, uh, one thing that I, um, always do in my presentations, I, I have to make the case going back to the whole area of health and what we were just talking about. Um, and I think that we are still lucky that the dairy industry is so diverse that not two dairy farms are the same. And, and they have completely different management styles and even goals and interests. That there are two types of farms, the ones that, um, spend little time and put little effort or don't have a systematic approach to look for sick cows. That's one extreme. And then you have the other extreme, the ones that put a lot of time and effort on finding cows with health disorders because they don't wanna miss anyone. And, um, we have done research trying to understand how bringing technology to those two type of farms can help them and what the drawbacks are.
Dr Julio O. Giordano (00:44:12):
And, uh, in an oversimplification of all the research that we have done, and we can discuss things in more detail. Uh, so farms that currently do little to nothing to identify cows with health disorders, those are the ones that have the opportunity to, to find more cows with conditions, help them through treatments and improve their performance in. So we have done one study, uh, that we observe a gain in daily milk yield for the first 21 days in milk by using automation versus very simple visual observation. And that's the one Melissa was, uh, talking about for, uh, the economics. So we accounted for, uh, all the extra costs of the technology. So the colors, the time, the labor to put the colors on cows, all the extra treatment, all the extra clinical exams and accounting for all of that, the value of the extra milk and the reduction in replacement costs offset completely the extra expenses. And it was even positive, you know, there's a difference of almost $20 on average per cow within the first a hundred days in milk by incorporating automation in a farm like that. So a farm. So again, it's like, you know, a farm that is not finding as many cows as they probably should and treat as many cows as they should, there is potential to help cows that are going undetected. You
Dr. Melissa Cantor (00:45:45):
Know what's really interesting about that is that was the farm that the repro technology for the sensors for heat detection also really benefited was those farms that were not really actively looking for heats. Well, and so it's very interesting that you found the same thing for transition, right? Yeah.
Dr Julio O. Giordano (00:46:00):
It's the exact same thing.
Dr. Melissa Cantor (00:46:01):
Yesss like sending the same message
Dr Julio O. Giordano (00:46:03):
The case. So now with the other type of farms, the, the other case that I make is like, we, people are really good at finding sick cows. We are really good at finding cows in heat. What's the problem? We don't have the people or people's time, quality time to dedicate to those tasks anymore. So that's where technology, you know, is, is filling that key gap. So, uh, then that brings me to the other extremes. So for farms that, uh, have place if a systematic intensive program to look for sick cows, when we compare that to technology, to automation, we found even a fewer sick cows. So, you know, fewer sick cows with automation than with the traditional intensive monitoring program.
Dr Julio O. Giordano (00:46:58):
But that difference in cows detected did not lead into more cows sold or dead, and no differences in milk killed or in reproductive performance. So the, the messages, um, a slightly fewer cows detected, which I'll have to discuss that in a moment. Um, but without negative consequences on her performance outcomes, which is what most farms care about. So, you know, I, I always tell 'em my presentation is like, well, you know who Julio tells me that, and I'm a dairy man. I said, you know, why in the world do I want any technology? If I'm not finding any more sea cows, I don't make more milk. Is labor is labor reductions and cow time. So la human time is critical. A cow time is also critical, right? So time to lay down in a stall to eat when the cow wants to express her natural behaviors versus being locked up for, I was actually talking to a couple of farmers that came to talk to me after my talk and uh, it resonated with them a lot.
Dr Julio O. Giordano (00:48:09):
What, what we did. And I asked him, how long do you, uh, keep your cows? Because he said, I have an intensive program to look at my sick cows, so I want to go away from that and I see your research and that's encouraging to me. And um, I said, how, how long do you keep your cows locked up and said an hour and a half to two hours per day. That's a lot of time, especially on cow, these fresh cows. Cows, right. For cows that need to be eaten. They don't wanna be, you know, standing in, you know, in headlocks. Yeah. Should be a or when we sort cows, right? So there is that value. I mean, so it's reducing cow manipulation, reducing the amount of time that cows spend away from doing the things that they can and or want to do, and then people, you know, which is becoming such a
Dr. Bob James (00:48:54):
Critical thing for Yeah. Don't you think that has a negative impact on some of the things that you pick up with technology? If they're locked up like that, they don't have an opportunity to show some of those
Dr Julio O. Giordano (00:49:03):
Behaviors? Well, but, but the idea is that they don't, right? So I mean, we use technology, so yeah, sure. Technology is watching cows and telling us which are the ones that need attention, and we only focus on those ones that, that's the value proposition of technology saying these are the, uh, 10 cows that I have to look at today and my other 9,990 do what they have to do. Right. So in fact, you know, it's to go away from doing that. Yeah.
Dr. Bob James (00:49:34):
Because with, with calves, I think that's, that's really the, the thing is, and, and that that doesn't come across quite strongly is to succeed with this, you gotta feed 'em a lot more milk. And, and people that try to adopt the technology and still don't change the behavior, you know, you look at, at calves even fed a gallon a day, they're ready to drink again in four hours. And it's a, it's such a huge difference. And I think that's something that people have a hard thing, hard time grasping about some of that technology is because it's a, it's a, you've gotta change your management a whole lot more to really see the, the benefits of those things. It's
Dr. Melissa Cantor (00:50:12):
Very true. 'cause when, like, kind of to comment on what you just said, when you start to restrict the cow's natural behavior by putting her in a headlock or in the case of calves, you're restricting what they naturally would go do. Right. And the robot is gonna allocate that a certain amount. The calf doesn't know that it's credit, you know, it has to learn that. And a lot of them don't figure it out honestly. Um, and so again, that was, that, that's kind of the big message here is that if you want to get the best bang for your buck, you should try to use the technology to be a passive watcher on that animal so you're not modifying their behavior, whether it's a calf or a
Dr. Bob James (00:50:44):
Calf. One of, one of the big problems we have with, you know, with calf management, and you're not a veterinarian, right? I am. He
Dr. Melissa Cantor (00:50:50):
Is a veterinarian. Okay. Okay.
Dr. Bob James (00:50:51):
So you're a veterinarian. That's
Dr. Melissa Cantor (00:50:53):
Why he talks about
Dr Julio O. Giordano (00:50:54):
He's
Dr. Bob James (00:50:55):
But no, no, no. But the traditional veterinarian doesn't want these calves touching each other and they don't wanna put 'em on the feeder until they're two weeks old and two weeks old. They've learned to drink twice a day. And it's a very different, it's a very different, uh, animal that, you know, if, if you're waiting two weeks before they put on the auto feeder, it takes 'em another week to 10 days before they realize, oh, I can get more feed. And so, you know, how we've managed calves traditionally has been so different than I think some of the key things with the auto feeder. And that's a, that's a huge problem.
Dr Julio O. Giordano (00:51:30):
If dairies are considering using technologies, they, they have to grasp the idea that they're gonna have to change their management. Yeah, absolutely. You adopt technology to change management. That's right. To make the best out of technology. Right. And to improve your management system and deal with bottlenecks or find opportunities that are not, you know, we cannot take, um, advantage of without technology. So
Dr. Bob James (00:51:56):
How does the veterinary progression profession adopted, you know, some of the stuff that you're doing? Do you see any challenges with that at all?
Dr Julio O. Giordano (00:52:04):
So in general, you know, there are very few veterinarians that are on fresh pans checking cows, right? So veterinarians have taken a different role, you know, in health monitoring and, you know, it's either, you know, training, so training those that are, uh, in the pan looking for cows or designing the protocols, designing the, the protocols that people are doing both things at the same time. Uh, so in general, positively, I, I would say at least, you know, the veterinarians that I work with, uh, in fact, a lot of the research that, uh, that we do is in collaboration with veterinarians. They are eager to learn on how they can use technology to, uh, improve, you know, uh, again, the protocols that they develop and, and the training that they do with people. And, you know, as farms get larger and larger, I mean, the idea that we're gonna be able to spend the time required for a thorough clinical exam on every cow. I mean, that, you know, that, that that's the thing from the past. No small dairies you could argue. But, you know, even in small dairies, I mean, you don't wanna manipulate the cow. You don't want to alter their behavior. So in general, I think it's, it's positive. And, um, you know, they're eager to learn about what
Dr. Bob James (00:53:15):
Technology can do. I think that's one of the problems with calves. 'cause how many, how many DVMs really spend time with the calf program and, and a pretty small proportion of those that, that do that. And so, you know, it's a, it's a real educational challenge that I've had and I've spoken at veterinary meetings and, and boy, you know, the first thing is, is man, I don't want those calves. And the best animal, and I think you would agree with this, is first of all, you have to phenomenal maternity and excellent colostrum. And I want 'em there two or three days of age. And they learn like that. And it's just really phenomenal and a just a tremendous difference in behavior. And, and, and that's another confounding factor that you see in, in some of the data that we've had on farms. 'cause you know, we never even looked at, at that on our field studies. And I know that had an impact on it probably. And yours is probably heavily controlled on that. But when do they come on the feeder?
Dr. Melissa Cantor (00:54:09):
Uh, are you talking about the study where we're following the
Dr. Bob James (00:54:12):
2000 hours? Well, in general In In general. Oh,
Dr. Melissa Cantor (00:54:14):
In general. That varies so much by far. Yeah. It varies so much by,
Dr. Bob James (00:54:18):
Well, it influences the research.
Dr. Melissa Cantor (00:54:20):
Two days of age is when I see the great, honestly, three days, this three, two to three days of age calf is suckling really well, very motivated. Yep. Wants to go on the feeder, put the calf on the feeder. Yep. Don't keep 'em in a pen. Um, there's something to be said. I think there's some neuroplasticity that's going on in the calf brain really early in life that we haven't looked at. Yep. Where they're a lot more flexible with their learning. Yeah. And then you learn as they get, and they've done this research in social research with calves that, you know, as you keep them in one particular system, they become less flexible about being able to unlearn things like
Dr. Bob James (00:54:51):
People
Scott (00:54:54):
Julio circle back with you real quick on that second group of dairymen where they were doing a good job of, uh, you know, finding the disease. But, uh, the, the technology didn't help with that. But what it did do was allow them to reduce labor costs and give the cows back some more time. Does that then pay for the technology? Did you look at that economics
Dr Julio O. Giordano (00:55:14):
In, in some of the, um, not so detailed analysis that we have done? Uh, for the most part it does. Of course it varies dramatically by how much time and and effort they're putting, how many people are involved. Uh, some, some of the analysis we have done, you know, within three to five years. I mean, most, most of the time, at least some of the technologies that we have looked at, again, colors for rumination and physical activity, um, they do break even or are positive depending on labor costs, number of people involved, and, uh, monitoring cows. So for the most part, yes. And as labor costs, I mean labor costs, at least in New York state, I mean, you know, it's increasing at pretty dramatic rates even, you know, faster than the cost of technology is increasing, you know, because technology also goes up.
Scott (00:56:09):
Yeah. Yeah. Great answer. You know, guys, I could go on all day with it. This has been fascinating, but I think they're getting ready to turn the lights out on us, so, okay.
Speaker 3 (00:56:48):
Tonight's last call question is brought to you by NitroShure Precision Release Nitrogen. NitroShure delivers a complete TMR for the room and microbiome helping you feed the microbes that feed your cows. To learn more about maximizing microbial protein output while reducing your carbon footprint, visit alchem.com/nitro. Sure.
Dr Julio O. Giordano (00:57:11):
So, um, in general, I think that automated health monitoring technologies, uh, can help most farms now. How they help, how much they help and what are the challenges for farms? Varies a lot by farm, and it depends a lot on what I described earlier, right. You know, the two extremes. And then as I showed today, everything, there's everything in between. So I think that our science, our research has shown that there is value. It's a matter of identifying the, uh, value for an individual farm, right. And, um, as I said too, I think that we are just scratching the surface with technology. We have focus on, you know, the, the, some of the big elephants in the room or dealing with the most immediate things, whether it's tremendous value on what we can learn about cows, what we can learn about how we manage cows, how the environment affects cows, and then develop decision making tools based on the data generated by technology. And this is what the most excited about, you know, for the future. Right. And, you know, whether it's AI or very simple ways to classify cows in groups with just some data, I think that there's tremendous potential to do better, you know, be more efficient, more sustainable, and improve the lives of people and cows.
Scott (00:58:49):
Yeah. Well said. Well said. Thank you. Melissa, how about a couple things from a calf perspective?
Dr. Melissa Cantor (00:58:54):
Yeah, absolutely. I, I probably won't be as erudite as him. He had such a great thing to say, but I, I think what we know currently is that diarrhea in calfs can be detectable with things like accelerometers, robotic milk, fader, feeder data, so can pneumonia. And there is value to using machine learning algorithms to find those animals. Even when we constrain budgets and think about how much money we have, however, we are not at the point yet where we can say, okay, calf 55, 55 is ready to go. I think that the transition cal research is a little bit farther ahead of us, to be honest. We get inspired from what they're doing to find these diseases. However, we have mercy because diarrhea happens at a very different time than pneumonia. So we're able to get a lot better predictive algorithms, I think, um, because of that.
Dr. Melissa Cantor (00:59:43):
Um, but the other opportunity I think that we didn't talk about, we could always do another webinar later, is there's opportunity to do something with those animals at that point, right? So for finding a calf before she's clinical, what do we do? And that's the million dollar research question. I mean, people are looking at things like colostrum, but colostrum's expensive. Um, we've looked at things like painkillers, like nsaid. I know Bob had talked about ine before we got recorded. Going in with something like that and figuring out what that is, I think that's a huge research career opportunity. Um, we just have more work to do with algorithms before we start doing that. Yeah,
Dr. Bob James (01:00:18):
I think there's one thing that we haven't talked about with the calves, and that's the welfare. These animals are very, very different. They, uh, are, they're, they make better animals as older animals. They adapt to new situations. The research is there from University of British Columbia. You
Dr. Melissa Cantor (01:00:35):
Mean socially have
Dr. Bob James (01:00:36):
Calves Socially. Socially, yes. Not technology calves, no. Socially. And, and when they're in groups and they adapt to new feeds, they adapt to new situations better. And the consumer loves to see these calves in a group pen. I have a four second video that I start my, my presentations with and it's a bunch of calves running around the pen with their tails to cup. And you just look at it and say, wow, that's really nice. And I think that's a component that, that's really important.
Scott (01:01:04):
As I said, enjoyed it. Thank you so much. Appreciate guys sharing your, your experience and, uh, absolutely. Uh, with us to our loyal audience. Uh, as always, thank you for joining us once again. We do hope you learned something. We hope you had some fun and we hope to see you next time here at Real Science Exchange, where it's always happy hour. Thank you. And you're always among friends.
Speaker 3 (01:01:24):
We'd love to hear your comments or ideas for topics and guests. So please reach out via email to anh.marketing at balchem.com with any suggestions and we'll work hard to add them to the schedule. Don't forget to leave a five star rating on your way out. You can request your Real Science Exchange t-shirt in just a few easy steps, just like or subscribe to the Real Science Exchange. And send us a screenshot along with your address and t-shirt size to anh.marketing at bal chem.com. Balchem’s real science lecture series of webinars continues with ruminant focused topics on the first Tuesday of every month. Monogastric focused topics on the second Tuesday of each month, and quarterly topics for the companion animal segment. Visit balchem.com/realscience to see the latest schedule and to register for upcoming webinars.