Dear colleagues, first of all, I would like to tell, many thanks to the organisers for this, possibility to share my experience with you this evening. So, The topic of my presentation accuracy to predict the onset of calving in dairy farms by using different precisions, livestock farming devices. The outline of my presentation the following .
After a very brief introduction dealing with the prevalence of steel in different countries, after this, I would like to show you the different methods. How can we predict the set of calving on dairy farms. There are traditional methods like evaluating different clinical signs.
And the other side, there are different devices which are already possible to buy them commercially. We can measure body temperature like rectal temperature, vaginal temperature, ear surface temperature, ventral taba and temperature and rear ruminant temperature. We can detect the behavioural changes like eating, rumination, and activity.
We can also detect errors and behavioural changes, and nowadays, it's also possible to have activity monitoring, video cameras to detect, the onset of carving. And finally, I would like to summarise all of this, knowledge and give you some take home message and further perspectives. Oh, a very brief introduction about the prevalence of stillbirths.
As far as you know, we are, talking about perinatal mortality or stillbirths when we can lose a mature calve during calving and in the 1st 24 hours of life, and the duration of gas station is must be greater than 260 days. There are different countries all over the world . And you can see, especially in hay first, the prevalence rate is quite high, close to 10% among these countries, somewhat better for hayers and cows together, but it is somehow between 5 to 10%.
And other side, there are some data from the Netherlands, from United Kingdom, and from US it's quite high and especially in Norway and Sweden because of the red, Hosten, they have a, a somewhat lower in. And also the mega figures about the prevalence of stillbirth rate in Horten Fries and he first in Sweden, in the Netherlands, and in the US, so it's quite high. We can lose quite a lot of number of of human calves, .
In case of heyer carvings. In, in this respect, I would like to emphasise the importance of predicting the onset of calving in dairy coal. Especially in large dairy farm, it is very difficult, .
Day and night to be there and to find out the exact time of the appearance of the, the amniotic or the Olympics at first and after the amniotic. The traditional methods for us is to use clinical signs during the preparatory stage of cattle. However, it is very important for us to know in each case if we have different breeds, the main gas station lengths of dairy and beef cattle breeds, which is about 2982 days to 292 days again.
And We may use a different clinical signs during the preparative stage of cater for repertuition scoring system or a calving scoring system. We can use different clinical signs like relaxation of the broad pelvic ligaments, secretion of voginum mucus, physic physiological hyperplasia of the other. Edoema of the other, feeling of the tits, relaxation of the tail, edoema of the vulva.
Among these clinical signs, there are two very important signs like relaxation of the broad pelvic ligaments and filling of the teeth, which can give much better . Picture about the exact time of part tuition. It is very important to emphasise that a threshold of 4 or less than 4 calving score points means that calving within the next 12 hours could be ruled out with a probability of about 99.3 in cows and 95.5% ins.
Altogether, the probability is about. 98.6% for Hers and costs to get.
So it means that We cannot, after all, we could rule out. The carving within 12 hours. However, The probability of predicting calvin within 12 or 5% costs was only about 15%.
So it is not accurate. The clinical signs are not accurate. Especially again in large farms.
For for that reason. We can measure the body temperature, rectal, vaginal, ventral surface. Earth surface or vertical ruminant temperature.
As far as you know, to take the rector temperature is, is not a practical thing. It would be very difficult. To check each animal.
Several times during a day. However, This . Results can inform us that it is possible to measure, but not by this old method, by electronic method.
It is possible, but, but not in the, the rectum, but other parts in the body. But it is important to mention first the The rectal temperature changes. Pico even when exhibiting external sizes of imminent parturition such as mammary distension, relaxation of pelvic ligaments, and vulval enlargement.
Were unlikely to carve within the succeeding 12 hours if the rector temperature was above. 38.8 sizes.
So these results come from the 60s, so it's a long time ago, it was already declared that The changes in rectal temperature. May be useful to predict carving. And the other problem was that the rector temperature exhibited circadian rhythm with a minimum temperature in the morning and the maximum temperature in late afternoon.
And on the other hand, windy, rainy weather conditions, heat stress, and the effectiveness of cooling methods may influence body temperature and the pattern of sickle the arrhythm. So it, by this way, it is more complicated. And accuracy of predicting calving by detecting temperature drop was only.
44%, about 44%. A decrease in vector temperature measured in the morning or in the evening over a 24 hour period of 0.3 elus could predict calving within 24 hours.
So in, in contrast with DOCA, As far as you know, that, The, the decrease in the temperature, rector temperature in dogs just before helping, it can be 1.5 °C, but in the course, there is only about 3.3 °C.
But in this respect, it's very difficult. To use . Not an automated methods for measuring rector temperature or other temperature.
So here you can see the rector temperature in the last 5 days before carving them. The morning is the first room, the lower values, and the evening is the higher values. And During the last 24 hours when the temperature is 0.3, says use or greater than 0.3, says use the sensitivity can be between 44 to 70%.
And the specificity is about 86 to 88%. So it means that It is not real accurate. But to measure the rector temperature.
For that reason, We can measure other. Places as well like the oina temperature, like this is a, a valve phone. Which can measure.
The vaginal temperature. And By this new method, it is possible to, to send the, the, the values to our . Hand phone and we can see the actual working that temperature in each case.
And in this, in this study, you can see the decrease in vaina temperature as well. And again When we are, when calving will occur within 24 hours, the, the cutoff value again, 0.3 cas equal or higher than 0.3.
Somewhat higher, the sensitivity because it is between 62 to 71% specificity is about 81 to 80, 87%. So it is still not enough. And here you can see the DUI of walk a temperature comparing the last 48 hours.
The, the last test row and between 49 to 112 hours before coming. The upper line. So we can see this the rhythm.
Right. And this figure may show us the daily boina temperature changes before calving in Nipparus and multiparus host andres and cows. You can see the difference between a and and multiparus cows, and this is even a significant difference can be until .
They minus -3. And after, there is no difference between Between parities, but there is a, a decrease in the temperature. Ricci at all has suggested to use the intravaginal temperature 38.2 °C as a cutoff value to pre the carvin within 24 hours because it can be more accurate.
The sensitivity can be 80. A 6% versus with 66%, then the 0.21 says decrease during the last 24 hours before coming.
We found also that maybe the the mean value. It's better to use this mean value because we had the same, almost the same 38.12 sus between 0.12 hours before carving.
So maybe When the system will be changed for this level cutoff value, maybe at that time, we can have a, a better sensitivity to predict carving. It is very important to mention that walking our temperature were not affected by the gender of the calf. And there was no deal on a variation in body temperature from 48 to 8 hours before calving in beef cows.
On the other hand, parity, dystonia season and length of gas station did not affect the bogina temperature from 60 hours before and up to calming. In contrast, According to our examination, it was significantly affected by parity, season, summer versus autumn time of day, 8 a.m.
Versus 8 p.m. And the 6 hour time intervals.
Whereas gender, birth weight of the calf, twinning, gas station lengths, fertile presentation, dystosia, and occurrence of retained fertile membranes did not affect it significantly. These results can be explained by the deal nari up to 0.5 °C in the Waina temperature during the last 112 hours before calving.
While others could not conform this pre-calving Dal variation. It's very important to mention this as well. The other possibility for us to measure the ventral tail base surface temperature.
The SSM is coming from Japan. A sensor was attached. To the, to the tail.
The Database And by this way, we can measure the ventral taba surface temperature. Which can be approximately 1 °C lower than Waina temperatures throughout the pre-part two days, and this shows a significant correlation with Waina temperature. So we may use this technique as well.
Here you can see that How the ventral tape surface temperature can change, but it is very important to mention. That we have to exclude the effect of circadhythma. On Their base surface temperature as well.
And the reason, how can we do this, we have to calculate the residual. Their base surface temperature. Which is 1, excuse me, 1 with the actual Harley.
The base surface temperature minus mean TBST for the same or. On the previous 3 days. So, My this may We can exclude the effect of the circle on this .
The surface temperature and by this way, we can see a nice degrees, decrease in this temperature. However, Again, the prediction of calving within 24 hour and two farms were examined. The sensitivity was not so bad, but not so good still, about 85% .
At farm and about 82% at farm B. The specificity was about 72% or 68%, which was also not so good. On the other hand, it has an advantage of using this ventral service temperature measurement because It can be continued without interruption after coming, and it can be used to predict different diseases associated with body temperature changes.
For example, retained fertile membranes, matritis, masitis in a timely manner. So in this respect. This method has some advantage as well.
We can measure the Earth's surface temperature. The surface temperature is greatly dependent on the ambient temperature. It is suggested to group the animals in the hotter between May and September and the colder months between October and April, according to the highly measured earth surface temperature into two median temperature groups in a season, high temperature, medium high temperatures, medium low temperatures, and low temperatures.
This surgeons coming from the, the US. So, here we can see that. Among the different temperature zones, we can see a decrease in the temperature around carving.
So in this respect, we may use this earth surface temperatures as well to predict calving. However, I have to mention that there is no data about the sensitivity and specificity of this method, not yet. And then the, the next possibility for us to measure the vertical ruminant temperature, but a Rumen device.
However, the problem of this device is that it's quite expensive. Because this smart tech, for example, can measure the pH value. As well as the, the fluid in the rumen.
And The half life of this, bus is relatively very short, and on the other hand, it's relatively very expensive. So, maybe we can use this, method. Today, we can use this method only for, for, scientific, research work, but not for the daily practise.
The radical room and temperature due to the heat production of, of radical ruin microorgans and around. 0.5 °C higher than the body temperature.
This is very important to know. The beef lower nutritional intake being redundant. But it is true for the, for the, the dairy animals.
On the other hand, water intake depending on the temperature and its volume may cause a transient reduction in vertical or room temperature, and for that reason, it is suggested not to use these values below 37.7 °C. So we have to The, the system, the computer system must rule out this very low temperature.
We have made some research work, dealing with this reticuluminal temperature in daily course with normal and distalic carvings around carving. And here we can see that During normal calming, the, the Practically ruminant temperature was a higher decrease. Because it was about 39 cells it was comparing in case of dystopia.
When it was about 30, 39.2. So by this way, we may not differentiate between normal carving and dystopic carvings.
Temperature sensing, radical room bolus we can use here. Only the 0.2 cutoff value.
And by this way, we can pretty calming within 24 hours, the sensitivity is about 69%, the specificity is also 69%. Coming within 12 hours is a little bit higher. Oh, it's equal sensitivity, 69 and specificity, even a little bit less.
And the other possibility for us. To check behavioural changes. There are neck-mounted accelerometers.
Which can be used. Already a lot of, you know, a lot of dairy farms. And by this way, we can measure the daily eating times, of course, in temperature groups, different temperature groups.
This that also coming from the US showing us that the daily eating times in each temperature circumstances, there is a, a decrease around calving. The same is regarding the daily rumination times as well, different, different temperature circumstances, we can see a nice decrease in the rumination times as well, just before calving. And we also check the rumination time using this bolus and between normal and distalic carvings and here you can see that the, in the In the distosic cases, the decrease in rumination time starting earlier in case of distosia than in normal carvings.
So by this way, we may distinguish between dystosia and normal carvings by measuring the rumination time. We can measure the daily resting times as well. And here you can see that there is an increase in each cases.
Regarding different temperature circumstances. The same for the, the daily active times of calls in different temperature groups, there is an increase in activity. They used the, they try to predict the calvin by using an ear sensor.
By the ear sensor, it was possible to measure the activity, rumination, feeding, and then they were able to measure the temperature as well. And about 12 hours before calving, the sensitivity was less than 60%. And the specificity was close to 100%.
On a daily basis, it was the sensitivity. It was less than 40%. So.
By this way, we may say that . Measuring activity, rumination, feeding and temperature by ear sensor. It is not accurate to predict.
Carving before about 12 hours on a daily basis. Because when they use this whole basis 6 hours, 3 hours, 1 hour before. Before carving the accuracy was much lower than when the 12 hour was examined.
Again, using an ear sensor, we can measure activity, rumination, and line time. And about one hour before expected calving. We can pretty carving.
With about 50% accuracy and non-calving about higher than 90%. So the 1224 hours, 12 hours, 6 hours and 3 hours give a lower sensitivity and specificity. So it is not not useful.
This System as well. Again, your sensor was examined them and they measured the rumination time on the only basis of 1222 hours before carving, 12 hours before carving, 6 hours before carving, and here you can see the sensitivity and specificity. So it is not so high again.
We can use a a sensor on the right hind leg, and we can measure the lying boots on an hourly basis or again is 22 hours, 12 hours, 6 hours. But again, it is not so high, the accuracy to predict carving. We can measure the line time with this hind leg sensor.
Somewhat better, the sensitivity and specificity, but it's still not good enough. Again, and the hind leg sensor was examined and by this way the standing and line time and standing boots could be detected and here is the sensitivity, . Close to 80% and the specificity is close to 80% as well during the last 24 hour period.
So, the accuracy is somewhat better but still not so good. An insect electronic feed and water intake system was also examined during the last 24 hour period to pre the calving. And they were able to measure the dry matter intake, feeding time, and water intake, but The sensitivity changed between about 62% to 80%.
And the specificity is also 80% to We are about 55%. So again, it is not so, so accurate. We can use a neck collar sensor to, to detect neck activity and rumination.
And the sensitivity and specificity during the The last 24 hour period was about 70%. And in another experiment, they used the 5 hour window before expected carving. And using this neck collar sensor, checking rumination, eating activity.
And here you can see that the Rumination is about 70% of activity is only about 67%. But eating is only about less than 60% for the sensitivity. Sensitivity and specificity for carving detection using different sensors.
In this experiment, they use neck accelerometers, like accelerometers, and they were able to check the localization as well. And here you can see the sensitivity is quite low. The specificity was not so bad after all.
When they combined the, the data from the neck accelerometer and leg accelerometer. Was hired the sensitivity. The neck localization again higher leg localization and plus localization, about 78.
When all sensors were used at the time, they were able to reach 85% the sensitivity to detect carving. So by this way, they were able to measure line time, number of steps, ruminating time and travel distance. The other possibility for us terror raising and behaviour, detecting terror raising and behaviour changes.
We have also used this sensor on the tail, but . Our problem, what was that? We were able to get about.
The mean value is about 13 messages before carving. And the majority of these messages. Was false positive.
So this is one problem that there are a lot of false positive. When we are using this device. The other problem that because we have to fix this device on the, on the tail.
And this cause severe injuries on the tail as well. So in this respect, we may use only 4. 1 or 2 days maximum 3 days before expected carving to get some data, but .
So, finally, our results were not, not so promising. However, this tail sensor were used in another study using Quite a lot of number of animals, 110 animals, and they use this 5 hour window. They're raising the accuracy to predict carving was about 78%.
And not predicting carving about 83%. So, this is also showing us that this is, this is not so accurate as well. Another study.
They evaluated on the hourly basis and evaluated. 24 hours, 12 hours and 4 hours before Carving. And in this case, a little bit Better sensitivity was, specificity was reached because it's about 90%.
However, the sensitivity was very similar in this period, but 2 hours and 1 hour before carving. The sensitivity was lower. After all, we can use A combination of sensors to pretty carving.
By this sensors, for example, by the neck sensors. We can measure rumination time, neck activity. And we can measure by using the hind leg, sensors, a number of steps, lying to and lying boots.
And The sensitivity to to pretty carving within 24 hours. Before carving. Can reach 100%.
At least according this study. And the specificity was about 82%. However, When checking the 8 hour before carving at that time, the, the sensitivity and specificity were lower.
Again, another experiment use a combination of sensors, neck sensors. Detecting eating rumination and lying time. And the front leg sensors, they were able to detect a number of steps, standing, walking, and like time.
And the sensitivity. On an hourly basis was around. 85%, but when they use this 24 hours, 12 hours or 60s before expected carving, it was quite, quite good, almost 100% accuracy.
The 3 hour and 1 hour before the calving was not so accurate. We may use activity monitoring video cameras as well for trying to detect the onset of carving. We can check behaviour activities, lying, standing number of changing position between lying down and standing up, holding up the tail, turning the head to the side.
It is very important to mention that they use the They extracted. From recorded video sequences with a hidden mark of model to pretty caring. And by this way, they were able to reach a sensitivity about 91%.
And the pres about 93%, however, they used only 10 hay. This study was done as well in Japan. Accuracy to detect the exposure of the sensor at the beginning of the 2nd stage of calming in, in daily courses.
This is another possibility for us. Because If we are inserting the, the sensors into the vagina or on the Volvo at that time when the, the Olympice will appear. In the vulva.
At that time, the sensor will be forced out from the, from the, from the vagina. And the wolverlips will be opened and this can generate a signal and by this way, we can detect the exact time for the onset of second stage of labour. This idea is coming from the, the equine practise because they inserted this device on the The, the lips lips and This is called as full alert.
And then the The rule was opened. The connection was lost and this generated a signal for the, the, the different equipments and nowadays, the signal can appear in the, the mobile phone as well. The same can be used in the call as well.
However, it is very important to mention that we have to fix this device to the Volvo and it is time consuming. To do it in each case. And, and, you know, equine practise is no problem because there is only 11 or 2 horses, and it is very useful, but in a daily practise when there are a lot of animals, it's not so practical method.
However, they use this method in different studies, and you can see that the sensitivity in each case was 100%. So this is a, a very good good method. On the other hand.
There is another possibility for us to insert the sensor into the, the, the vagina. This is a physical sensor. And when the exe will appear, this sensor will be removed as well.
And here you can see that again 100% accuracy. Was able to reach when they used the. There was only one study.
The accuracy was somewhat lower because there was some problem with the, the Communication and. Radio communication in that area. The other possibilities intravagina.
Device, measuring the temperature. And here you can see that again, 100% accuracy was most of the cases. The middle one is our studies, we examined the 257 cows and we had no false results at all.
There is the other method. This is called as I. That device .
We have to insert this device into the, the vagina. The problem with this, we can do is In a cone, the problem that the, the sensitivity is not so I took Experiments were done. With these devices and the sensitivity was about 80% only.
It's very interesting. But we have to mention that This device can cause severe injuries in, in the vagina. And for that reason, maybe it's not a good idea to use this at all.
Finally, I would like to, to summarise all of these, results and to give you some take home messages and future purposes as well. We have to say that presently we cannot predict accurately the onset of calving in daily calls by evaluating the external preparatory clinical signs for the onset of calving and by using different sensors to detect the decrease in core temperature, vaginal ventral table surface. Earth surface or reticle room temperature or evaluating behaviour signs by using a sing or combination.
Of sensors. However, by using a combination of sensors, we can increase the accuracy of predicting the onset of caring. It is very important to mention that at the same time Santagus.
Emphasise by examining about 3000 carvings. In the Dutch commercial dairy farms. That the accuracy of predicting carvin in a commercial setting based.
On behaviour variables, steps, head movements, sharp at stands, eating, lying, standing, walking. Measured by smart tax, neck and leg could not be improved. So one One part of the studies saying that the combination may help us to increase the accuracy, but the other Literature data saying that We cannot improve it.
Martinator reported that we might increase the accuracy of predicting the onset of calving by detecting the external preparatory clinical signs for calving and using a sensor on the tail. However, it would be labour and time consuming in a dairy farm, especially in large farms, it's impossible to, to, to check the, the clinical signs for each animal. The onset of calving can be predicted accurately.
The sensitivity can be greater than 97% if in the divides into the vagina some days before the expected calving, which will be removed from the vagina during the opening of the Alan to Korean exec in the wool were just at the beginning of the second stage of labour. So the only possibility for us today to use this Wainna devices because at that time we can reach an acceptable. Accuracy for predicting the exact time for the second stage of labour.
According to our our . Examinations of. Here we gave the prevalence of stillbirths.
You can see the, the, the difference, excuse me. Difference between the control group at the same farm. The prevalence rate was about 10 10% for the control groups for the nip pulse and multi-parse costs comparing about 2% when we use this device.
But it is very important to mention that even the prevalence of retain foetal membranes was significantly less when we use this device compared with the control group. And the prevalence of clinical matritis was also lower than we use this device. So it is important to mention that the device must not irritate the vagina.
And cause discomfort to the animals such as in the case of the I device. So for that reason, We cannot recommend this either device at all. Similarly, the vulva magnetic sensors can be used with high accuracy.
However, this is an invasive method because it needs a vein to suture the transmitter and the magnetic device to the vulva skins. Which limits their use in large dairy farms. So it is impossible to do this with 100s or 1000 animals at all.
Future purposes with the newly discovered sensors or a correct combination of the sensors. Or just improving the algorithms used for evaluating the sensor data, we can increase the accuracy of detecting the onset of scarring to further decrease the delayed obstetric assistances which may cause dystopia and stillbirths. It was recently confirmed that if we can do the appropriately timed obstetrical assistance, appropriately timed obstetrical assistance means that we can do the, the assistance 70 minutes after opening the amniotic sect or within 65 minutes after appearing the fatal hoofs in the vagina.
After the onset of the second stage of labour, we can significantly decrease the prevalence of dystopia, stillbirths, retain fertile membranes, and vvo vagina laceration compared with the inappropriate time obstetrical assistance. So I would like to emphasise this is also our results that I would like to emphasise that it's not enough. To predict the exact time for the second stage of labour, it is also very important to be patient and to do the obstetrical assistance.
About 70 minutes after opening the, the omnitics. You can find some data in the literature that they try to do this . Assistance earlier, but it is very important to mention that in these cases, they use a huge amount of of fluid In order to be able to help the, the obstetriical assistance.
Future, perspective, the behaviour changes like feed intake, rumination, timeline boots, state elevation may be associated with dystopia and can be promising in the early detection, of course, with a higher risk of dystopia. It would be also very important to detect in time the prevalence of dystopia, but maybe this can help us. But there are not so many data in this respect in the literature.
Differences in behaviour between assisted and unassisted calls can be detected by video surveillance as well, or decreasing radicular room in temperature may occur 12 hours earlier in calls with dystopia. Then in auto cause this is our results. Oh, thank you very much for your attention and I may suggest my, review article, the accuracy to predict the same type of, .
Of this workshop, accuracy to predict the onset of coming in dairy farms by using different piece livestock farming devices. I was . I try to summarise all of the literature data in this respect.
And I may suggest our book, maybe you didn't. Not here About this book bovine Prenatal Perinatal. Now this, these books deal with the, the, the beginning of .
Of fertilizationvening. This whole period is discussed in this book. And My co-authors is John May from Ireland, Urich Blau from Switzerland, and Marca Tavena from the Netherlands.
So thank you very much for your kind attention. If you have any question, You are very welcome. Even you can write an email letter to me, and if you have any question.
Thank you very much. Thank you so much, Doctor Atto, and thank you for all the people who joined. See you in another webinar.
Bye-bye.