T 1785/14 (System and method for predicting physiological state of animal/LATTEC I/S) 07-08-2019

1. The appeal is admissible.

Main request

2. Novelty (Article 54(2) EPC)

2.1 In the appealed decision, the opposition division decided that none of the documents D1, D2 and D8 was novelty-destroying for the subject-matter of the independent claims 1 and 15 of the main request. The board concurs with this conclusion for the following reasons.

2.2 Document D1

2.2.1 Document D1 discloses automated methods and systems for determining progesterone in raw milk of dairy animals. The method comprises multiple measurements of progesterone concentration of individual cows by using a biosensor connected to a computer. In this way, the course of the concentration of progesterone in raw milk, spread in time, can be determined per animal, stored and further processed. On the basis of the time-dependent concentration curve, it can be accurately determined when the cow is in heat, whether the cow is in-calf, or whether the cow possibly has a defect of its reproductive apparatus (page 3, line 15, to page 4, line 4). A general representation of the course of the milk progesterone concentration in time is shown in Figure 4. By signalling the time at which the concentration sharply decreases, the time of oestrus can be accurately determined. When the concentration does not decrease again around the twenty-sixth day, the animal is either gestating or possibly has a defect of its reproductive apparatus. Since it is known that the progesterone concentration will remain high for about two weeks after a change in the determined amount of progesterone in time has reached a minimum, the progesterone content of the milk of the cow is not determined during this time. As shown in Figure 4, in general only after the fourteenth day, when the animal is not gestating and/or does not have a defect of its reproductive apparatus, will the concentration decrease again. Only after 14 days, therefore, is it again of interest to start monitoring the change over time of the amount of progesterone in the milk (page 5, line 20, to page 6, line 11).

2.2.2 It is thus undisputed that D1 discloses "a system for observing and predicting a physiological state of an animal" (oestrus in D1), which comprises "a computer comprising a processor and being operatively connected to a database, at least one sample providing device for repetitively providing at least one sample of a body fluid of the animal, an analysis apparatus for analysing the at least one sample, so as to obtain at least one sample value of at least one parameter of the body fluid, a data interface for repetitively entering the sample value of the at least one parameter in the database, wherein the database is adapted to store multiple database entries representing the sample value of the at least one parameter at various points in time". Moreover, D1 also discloses that "the processor is programmed to perform at least one mathematical analysis of the at least one sample value, the mathematical analysis being a statistical analysis which is a univariate analysis of the database entries" (e.g. the determination of a minimum of the progesterone concentration) and that "on the basis of the at least one mathematical analysis, the point in time for providing a subsequent sample and performing a subsequent analysis of said subsequent sample for at least one of the parameters" is selected (namely 14 days after the minimum is measured). However, D1 does not disclose that "a first set of data representing expected sample values of at least one of the parameters at future points in time" is obtained by mathematical analysis. The appellant failed to indicate any passage in D1 where such a feature is disclosed, and the board disagrees that such a disclosure may be considered implicit. While, as argued by the appellant, the system of D1 is based on the prediction that, after a concentration dip, the progesterone concentration will remain high for at least the next 14 days, this prediction is not equivalent to creating a data set of predicted values.

2.2.3 Hence, the subject-matter of claims 1 and 15 of the main request is novel over D1 (Article 54(2) EPC).

2.3 Document D2

2.3.1 Document D2 is an international patent application published after the priority dates of the present patent but before its filing date. During the proceedings before the opposition division, it was concluded that the priority was not valid for the subject-matter of claims 1 and 15 of the main request, and this finding was not contested during appeal. Hence, document D2 is considered prior art citable under Article 54(2) EPC.

2.3.2 Document D2 discloses an automated system for optimising the production performance of a milk producing animal herd, which comprises "an apparatus for analysing a plurality of compounds or parameters in a milk sample of an individual member of a milk producing animal herd, said apparatus comprising: (i) separate means for analysing individual compounds or parameters in the milk sample, each of said separate means being capable of generating a detectable signal in the presence of an individual sample compound or parameter, (ii) means for directing a part of the milk sample to each separate analysing means, said directing means being controlled by means for storing data for the physiological and nutritional state of each individual herd member, including data indicating point in time in the reproduction and lactation cycles of said herd member, such that the directing means is only activated at pre-selected points in time or at pre-selected time intervals in the production or lactation cycles of the individual herd member" (page 8, lines 19 to 33).

2.3.3 Similarly to D1, document D2 thus discloses all features of claims 1 and 15 of the main request, with the exception of the feature that "a first set of data representing expected sample values of at least one of the parameters at future points in time" is obtained by the mathematical analysis. Again, while the sampling time is under control of the processor and depends on the values obtained for the measured parameters, there is no disclosure in D2 that a data set of predicted values is created, on the basis of which the sampling time is decided upon. The passages on pages 14 (second paragraph) and 17 (second paragraph) teach that the progesterone measurements are used to predict periods of expected oestrus, but this is not equivalent to creating a set of data representing expected future sample values of the progesterone concentration.

2.3.4 Hence, the subject-matter of claims 1 and 15 of the main request is novel over D2 (Article 54(2) EPC).

2.4 Document D8

2.4.1 Document D8 discloses a system and method for monitoring the physical condition of a herd of livestock, that comprise using errors between values predicted in accordance with a time-series model and corresponding measured values to determine a confidence interval for each animal individually. Figure 2 depicts a flow chart showing how the system is implemented, the algorithm according to the flow chart being preferably repeated at each milking (page 7, lines 7 to 10). Measurements of different parameters are taken at each milking and for each individual, identified animal, and the measured values are read by the data processing structure. On the basis of earlier measurements and predictions, or at the first milking of a lactation as an initial set of values and parameters, status data which determine the prediction for each next value to be measured and characterise the distribution of errors in previous predictions are stored in a memory of the data processing structure for each individual, identified cow (page 7, lines 24 to 30). If at least one of the measured values is outside the confidence interval or if a combination of errors symptomatic of a particular condition of an animal occurs, an attention signal is generated (page 9, lines 2 to 6). After an attention signal has been generated, the model on the basis of which the predictions are being made is not reliable anymore for the respective cow, in particular if the attention signal indicates that a value outside one of the wider confidence intervals has been measured. Therefore, in step 30 of Figure 2, the monitoring of a cow on the basis of the collected data is as a rule stopped in response to an attention signal regarding the respective animal, or at least in response to an attention signal above a certain confidence level regarding the respective animal. If it is decided that no attention signal or no attention signal above a predetermined confidence level is to be generated, the status data for the respective individual cow are updated using two of the following three sets of data: the latest measured values, the latest predictions and the latest errors between the predictions and the corresponding measured values (step 31 of the flow chart). If after an attention signal has been generated verification by the farmer or by a veterinarian reveals that the attention signal was unjustified, the measured values are preferably replaced by the predicted values, so the monitoring of the checked animal can be continued on the basis of the previously collected data and the data entered instead of the latest set of measured values (page 10, lines 2 to 31).

2.4.2 Document D8 thus discloses "a system for observing and predicting a physiological state of an animal" (health status in D8), which comprises "a computer comprising a processor and being operatively connected to a database, at least one sample providing device for repetitively providing at least one sample of a body fluid of the animal, an analysis apparatus for analysing the at least one sample, so as to obtain at least one sample value of at least one parameter of the body fluid, a data interface for repetitively entering the sample value of the at least one parameter in the database, wherein the database is adapted to store multiple database entries representing the sample value of the at least one parameter at various points in time". Moreover, D8 also discloses that "the processor is programmed to perform at least one mathematical analysis of the at least one sample value, the mathematical analysis being a statistical analysis which is a univariate analysis of the database entries to obtain a first set of data representing expected sample values of at least one of the parameters at future points in time" (e.g. the calculation of predicted values and confidence intervals). The system of D8 thus comprises all features of the system of claim 1 of the main request, with the exception that there is no selection of a next time point for sampling on the basis of the predicted data set. In fact, D8's system is programmed for continuous sampling at each milking, and this sampling is only interrupted if there is an error falling outside the confidence interval. Should the error be still within the confidence interval, sampling is continued normally. There is thus no selection of a time point for further sampling but merely a decision to stop or to continue sampling. Contrary to appellant's arguments, this is not equivalent to deciding on a specific time point for the next sampling to take place.

2.4.3 Hence, document D8 does not disclose all the technical features of claim 1 and claim 15 of the main request. Claims 1 and 15 of the main request are thus novel over D8 (Article 54(2) EPC).

3. Inventive step (Article 56 EPC)

3.1 Documents D1 or D2 as closest prior art

3.1.1 The present patent is directed to a system and a method for observing and predicting the physiological state of an animal, as defined by the features of claims 1 and 15 (for the exact wording, see section III).

3.1.2 Document D1, which also discloses a system and method for observing and measuring the physiological state of an animal (in this case, the fertility status of a dairy animal), is the closest prior art. As discussed above under novelty, the distinguishing feature between the system and method of D1 and those of the claimed invention is that D1 does not disclose that a first set of data representing expected sample values of at least one of the parameters at future points in time is obtained by mathematical analysis, on the basis of which the point in time for a subsequent sampling is selected. The technical effect associated with this feature is that the system can be adapted to each individual animal. The technical problem can thus be formulated as the provision of a system and method which are improved in terms of being more individualised and therefore allowing better accuracy in determining the sampling time. The solution is the subject-matter as claimed and the board is satisfied that it plausibly solves the technical problem.

3.1.3 With the aim of solving the technical problem, the skilled person would have considered modifying the system and method of D1 to render it more individualised. For this purpose, they would have used the data collected for each individual animal to define its particular oestrus cycle, thereby being able to decide, for each animal, what the best point in time for a subsequent sampling would be. Hence, the skilled person would have arrived at a solution to the technical problem, however, not necessarily at the claimed solution, which requires that a data set of predicted values be produced. The board fails to see any incentive or suggestion in D1 or in the remaining prior art to produce a set of predicted values in the context of the present invention or that the common general knowledge would have prompted the skilled person to do so. The passage of page 3, lines 28 to 30, cited by the appellant, only suggests further processing of the data, without teaching how. It is true that the skilled person would have considered the curve of Figure 4 as an over-simplistic model and would thus have been motivated to use the data actually measured for each cow to define a more realistic, individualised curve, but they would not necessarily have been motivated to create a data set of predicted values. Hence the claimed subject-matter is inventive over D1 (Article 56 EPC).

3.1.4 The same reasoning applies also when starting from document D2 as the closest prior art. As determined above under novelty, the distinguishing feature between the system and method of D2 and those of the claimed invention is the same as in relation to document D1 and, again, D2 does not provide any suggestion or incentive, alone or combined with common general knowledge, pointing to the claimed invention. Hence, the claimed subject-matter is also inventive over D2 (Article 56 EPC).

3.1.5 A further issue which was discussed in the context of the distinguishing feature over the systems of D1 and D2 was whether there was indeed an associated technical effect. As the claims do not state explicitly that the selection of the time point for subsequent sampling is on the basis of the data set of predicted values but rather that it is on the basis of mathematical analysis, it could be interpreted that the system did not necessarily require the use of the obtained data set, in which case this feature did not have a technical effect. However, the board agrees with the respondent's arguments that this interpretation would not make technical sense in the context of the claim. The mathematical analysis clearly results in the creation of the data set of predicted values and thus the use of the mathematical analysis implicitly means that the created data set is used.

3.2 Document D8 as closest prior art

3.2.1 Document D8 is further away from the presently claimed invention because it is concerned with monitoring the animal's health status rather than determining physiological states. Starting from document D8, the distinguishing feature between the system and method of D8 and those claimed is that, although there is generation of predicted data values, there is no selection of a time point for subsequent sampling based on these predicted data values. Rather, the predicted data values are used to define confidence intervals which allow determining whether a given measured value falls within the intervals and, accordingly, whether the measurement model is still reliable or has to be reset. The technical effect associated with the distinguishing feature is that a more accurate, animal-tailored sampling scheme, which is particularly useful for measuring parameters which change with the different physiological states of the animal, is provided. The technical problem can thus be formulated as the provision of an improved system and method for assessment of the animal's physiological state. There is no suggestion or incentive in D8 or in the remaining prior art that would have prompted the skilled person, taking their common general knowledge into account, to change the system and method of D8 in this way for solving the technical problem. Hence, the claimed subject-matter is also inventive over document D8 (Article 56 EPC).