Why health task and report? 

Prevention is better than curing. The health report or task combines data from available sensors to detect cows with (first) signs of illness and show these in one central overview. This report or task helps you making the right decision at an early stage. Displayed cows always need attention; therefore, it is advised to embed this report in your daily routines. Besides, the generated ‘sick chance’ is a good input for automatic routing. 

How to apply health task and report? 

The smart health report is a standard report and uses the available sensors to generate a so-called ‘sick chance’ per animal.  

The health task or report shows all cows exceeding the sick chance threshold (by default 10). The calculated sick chance indicates the chance the cow is sick ranging from 0 to 100. For each cow on the list, all deviating values are shown including a bar indicating the severity of the deviation. 
If a cow has already a treatment assigned, a first aid box is shown. 

The increase/decrease arrow in the sick chance bar is determined by:

- Increase 5% above max sick chance last 24 hours

- Decrease 5% below average sick chance last 24 hours

The NEW icon in the column sick chance status is determined by:

- Sick chance of the last 14 days is below the threshold of the famer (excluding last 24 hours) and

- Cow is currently not in treatment

You can add a treatment to a cow by clicking on the three dots at the end of the row and select ‘Health treatment’. Note: Select the treatment that suits the condition of the cow. It is recommended to use treatment plans, this document explains more about why we recommend this, and how to create them.

Good to know when interpreting the health report:  
– Milking parameters are only included for cows “in production”  
– Cows below 60 days will get more attention as they are more open for disease 
– Cow over 200 days since insemination will get less attention  
– Cows with a registered heat or insemination or a heat attention will get less attention, because heat can cause abnormal values.  
– Young stock and dry cows have lower levels of attention because there is less data available but are included in the report.

 What parameters can be shown in the health reports  

Herd Correction  
For the parameters, milk drop, temperature, eating time the values are corrected for herd drops. This is done to avoid that a drop of the whole herd causes many extra attentions  

MilkDropKg  
The milk drop is calculated based on 24-hour milk yield data of last visits. To get the moving average, expected milk day production is subtracted with actual milk production registered via the robot. The last day where there is a milking was taken into account. 

WeightTrend and TotWeightLoss  
The weight trend and total loss are coming from weighing floor aka “gravitor” after smoothing the data in the original algorithm the outcomes of those are taken into account. Total weight loss is taken from the time there was a tipping point (typically at calving event) till now for example if cow weighted at calving around 650 Kg and now dropped to 540 her weight drop is 110 Kg. Only negative values are taken into the calculation. Weight trend is evaluated to make sure that cow not quickly uses weight. it is expressed as Kg/Day loss in weight.  
Additionally, sensors are not taken into account twice, for example, acute weight drop and weight trend deviations, or milkdrop and milktrend. Only the parameter deviating most is selected.  

Activity & Activity(rumination) (SCR)  
The activity and rumination values are evaluated by algorithm SCR proposed and which is also used in report 24. (see column index activity & rumination) baseline of algorithm is 100 % is healthy, <= 86 cow is needed attention, lowest values seen are around 20.  
For reporting there is made a differentiation so it is easy to grasp if we are looking at an H-tag or HR-tag.  

Activity & Activity(eating) (NEDAP)  
Activity is comparing activity levels for last 24 hour to values between -1 and -11 days  
Eating is evaluated the same way as activity, for the eatingtime sensors (LD), the same like rumination the values are combined for both activity & eating and are presented to the algorithm as a single parameter. in this single parameter eating is more important than activity.  
Note: for some older 2 hour tags (read out by robot instead of LD) it was found a lot of blocks with activity counter value being 0 , those tags are identified and excluded from algorithm.  

InActive Behaviour (Nedap Smarttag all-in-one) Rumination + Eating  
For the smarttag all-in-one (the tag which supports cowlocator also)  
Attention is comparable to Nedap velos we are checking the time in day cow is inactive. InActive = 24 hours – (Rumination + eating)  
Inactiveness is compared with reference period, once increased significantly cow is expected to be sick. Instead of showing this percentage we are expressing this vaule as hh:mm  

Acute Inactive 
Acute inactive is described as a total/ sudden absence of any activity measured by rumination and or eating sensors, acute inactive is represented as hh:mm. 
Nedap:  looks at the most recent 5 hours when data was received, checks if rumination and/or eating is below treshold and if so the algorithm continuously goes back in time till when time spent on rumination / eating was below treshold. This time is represented as the time inactive. 
SCR-DLL:  Uses a value of inactive minutes which is calculated only 10 days before expected calving and 5 days after calving in the case of absence of rumination minutes. (also known as calving– distress) 
SCR-Sensehub:  Sensehub checks for all cows the inactiveness of animals for both rumination and eating, and presents it to Horizon as distress minutes. This value is show 

General: 
Any Inactiveness below 5 hours is ignored. Also any tag without decent reference values for rumination and / or eating is ignored (< 150 minutes), this to make sure we don’t get alerts for unproperly attached / malfunctioning tags. There are other reports for this. 
Also Inactiveness on lactation day 0 is ignored. So user can get alerts before calving is entered but not after, however normal activity continues to monitor the cow here. 
For Acute inactiveness, which is found to be severe, other ‘healthy’ parameters are removed from the algorithm. For single activity sensor (Nedap Eating or SCR Rumination tags) the threshold value of severe is higher than the combined tags (Nedap Smarttag or SCR Sensehub tag), which can detect inactiveness earlier as they cover more of the cow’s activities. 

Conductivity  
Last milkings conductivity values are evaluated and only highest teat is taken into account. As conductivity data is already normalized around 70, these values are used in an absolute way. The level where they (can) show up on the report is minimum 83.  
Note: for reporting we show the highest teat, + all teats with an attention, to make sure we don’t let the farmer miss the other quarters.  
Note: When last milking has failed to milk certain teat and last successful milkng had a conductivity for that teat than teat will be taken into account.  

Conductivity Chronic  
For teats where the number of times that teat had an attention is over 50% in last 14 days (excluding last 3 days) the conductivity is judged as “Conductivity chronic” this means value still expressed as conductivity but severeness is less as this case seems to be a chronic case.  

Temperature  
In the MQC temperature is measured on 4 quarters, the highest temperature is sent to T4C in the milk result.  
Healthy cows have an expected temperature around 38.6 for sick cows it’s on average an degree higher. Temperature is measured absolutely, as the sensors are factory calibrated.  
For situations where it is really hot, the algorithm looks at location average temperatue and if this is above 38.8 then it sets its ‘healthy’ levels higher to prevent false positives due to high outside temperatures  
Note: differences between robots are taken away in T4C.  
Note: Temperatures from MQCII are included only. MQCII is installed since A3Next and A4.  

TimeAway  
Time away is evaluated in %, comparing current interval (last milking till now) with moving average (7days) milking frequency. For reporting we present as hh:mm  

SCC Indication  
Last cell count measured is taken into account only and compared against absolute levels.For the logarithmic behavior, values are first log transformed than evaluated, and transformed back for reporting purposes.  

FatProteinRatio  
The ratio is calculated by dividing fat over protein. Only values from cows between 5 and 100 day in lactation are taken into account.  

MQC color: Watery Mastitis Blood Abnormal  
Values are evaluated as true/false per parameter  
Note: sometimes both blood & mastitis appear on the same teat & milking, in the algorithm both are evaluated and shown. In T4C cow card reports you might not find the blood attention because we only show single value per teat. Note: even though the same attention might appear on multiple quarters it will get the same weight in the algorithm.  
Note: When last milking has failed to milk certain teat and last succesfull milkng had a colourcode for that teat than the colourcode for that teat will be taken into account.  

MilkTrend  
Based upon herd curve which is available for primiparous and multiparous cows, each cow is compared against average cow in the farm in the first 3 weeks of lactation. If the cow yield is significantly low it will be giving an alarm.  
As this parameter is evaluating milk yield, it will be compared with milkdrop and the worst one of the 2 will be maintained.  

FeedRest %  
If a cow did not visit the robot, or Cosmix enough and she has more than 50% of the credit left of the total credit she got the last 24 hours, she will generate rest feed. This term should not be confused with the amount of feed left in the feed bin after she leaves the robot. For more insight in the calculation, visit this page.
Note: Rest feed is not taken into account in the health score below 3 days in lactation.