Periodic Reporting for period 2 - DECIDE (Data-driven control and prioritisation of non-EU-regulated contagious animal diseases)
Reporting period: 2023-01-01 to 2024-06-30
Farmers, veterinarians and other animal health managers in the livestock sector are currently in need of adequate tools for risk assessment and for prioritisation of control measures for endemic contagious animal diseases.
The DECIDE project develops data-driven decision support tools, which present robust and early signals of disease emergence and options for diagnostic confirmation; and options for controlling the disease along with their implications in terms of disease spread, economic burden and animal welfare.
DECIDE focuses on respiratory and gastro-intestinal syndromes in pigs, poultry, and cattle; and on growth reduction and mortality in salmonids. For each of these species, we will (i) identify the stakeholder needs; (ii) determine the burden of disease and costs of control measures; (iii) develop data sharing frameworks based on federated data access and federated learning; (iv) build multivariate and multi-level models for creating an early warning system. These elements will be included in decision support tools to be integrated in existing farm management systems wherever possible and to be evaluated in several pilot implementations in farms across Europe.
To achieve these goals, DECIDE has assembled a unique multidisciplinary consortium of experts which also includes several representatives of stakeholders with ample access to data, such as national animal health agencies, providers of veterinary services or farm equipment suppliers.
The DECIDE project develops data-driven decision support tools, which present robust and early signals of disease emergence and options for diagnostic confirmation; and options for controlling the disease along with their implications in terms of disease spread, economic burden and animal welfare.
DECIDE focuses on respiratory and gastro-intestinal syndromes in pigs, poultry, and cattle; and on growth reduction and mortality in salmonids. For each of these species, we will (i) identify the stakeholder needs; (ii) determine the burden of disease and costs of control measures; (iii) develop data sharing frameworks based on federated data access and federated learning; (iv) build multivariate and multi-level models for creating an early warning system. These elements will be included in decision support tools to be integrated in existing farm management systems wherever possible and to be evaluated in several pilot implementations in farms across Europe.
To achieve these goals, DECIDE has assembled a unique multidisciplinary consortium of experts which also includes several representatives of stakeholders with ample access to data, such as national animal health agencies, providers of veterinary services or farm equipment suppliers.
In an initial phase it was ensured that sufficient data was available for all parts of the project and data sharing agreements with data providers were made bilaterally. Then, a Joint Controller Agreement was set up by those project partners that will share data from sources outside project, which is meant to facilitate data sharing in the future. Another important element of interoperability is the development of species-specific ontologies for data integration. The DECIDE ontology is stored in an OWL format, documented on https://github.com/decide-project-eu/ontology-decide(opens in new window) and made available on https://bioportal.bioontology.org/ontologies/DECIDE(opens in new window). Federated data access is being implemented for one cattle tool. A workshop was held to discuss the challenges and solutions on data access and re-use in animal health. Guidelines on data management are produced to facilitate data access and re-use.
With respect to the modelling work in DECIDE, progress is made in different case studies. Multivariate dynamic models for early detection of diseases in salmon are developed on Scottish data and will be applied to Norwegian and possibly Irish data. For cattle and pigs, collaborations are established for analyses of longitudinal data for disease detection. Using the EMULSION software package, a model for bovine respiratory disease (BRD) in beef cattle was developed. Software is developed to link monitoring data for early disease detection to the model that simulates disease spread and control options. The model is now validated on field data and evaluated with stakeholders to verify its effectiveness for decision support. The model is also adapted to model interventions for mycoplasma hyopneumonia (Mhyo) in pigs.
In the social science part various approaches were utilised to assess the use of data-driven tools. This includes focus groups and questionnaires for all addressed animal species and for the different stakeholders, i.e. farmers, veterinarians, health managers. Focus groups were held in most partner countries and feedback from the stakeholders were obtained. A survey with consumers as contextual factors was conducted. Further work focusses on tool design and evaluation.
In parallel, work is ongoing on the methodology to determine the multidimensional burden of disease, i.e. considering economics, welfare and medicine use. Frameworks for the biomass and animal health loss calculations are being developed in close collaboration with the GBADs programme (Global burden of animal diseases). An inventory within the consortium showed that a combination of welfare and disease data is rarely available. Only for poultry and veal such data were available and analysed. For pigs, the welfare quality protocol is applied to obtain such data. In addition, expert elicitation is applied for the welfare impact of BRD in dairy calves.
Finally, many activities were carried out to communicate about the DECIDE project and to ensure that its results were widely disseminated among target groups. This included the publication of the first set of practise abstracts to disseminate our results to stakeholders. The project website and newsletters serve as the main sources of up-to-date information about the project’s progress.
With respect to the modelling work in DECIDE, progress is made in different case studies. Multivariate dynamic models for early detection of diseases in salmon are developed on Scottish data and will be applied to Norwegian and possibly Irish data. For cattle and pigs, collaborations are established for analyses of longitudinal data for disease detection. Using the EMULSION software package, a model for bovine respiratory disease (BRD) in beef cattle was developed. Software is developed to link monitoring data for early disease detection to the model that simulates disease spread and control options. The model is now validated on field data and evaluated with stakeholders to verify its effectiveness for decision support. The model is also adapted to model interventions for mycoplasma hyopneumonia (Mhyo) in pigs.
In the social science part various approaches were utilised to assess the use of data-driven tools. This includes focus groups and questionnaires for all addressed animal species and for the different stakeholders, i.e. farmers, veterinarians, health managers. Focus groups were held in most partner countries and feedback from the stakeholders were obtained. A survey with consumers as contextual factors was conducted. Further work focusses on tool design and evaluation.
In parallel, work is ongoing on the methodology to determine the multidimensional burden of disease, i.e. considering economics, welfare and medicine use. Frameworks for the biomass and animal health loss calculations are being developed in close collaboration with the GBADs programme (Global burden of animal diseases). An inventory within the consortium showed that a combination of welfare and disease data is rarely available. Only for poultry and veal such data were available and analysed. For pigs, the welfare quality protocol is applied to obtain such data. In addition, expert elicitation is applied for the welfare impact of BRD in dairy calves.
Finally, many activities were carried out to communicate about the DECIDE project and to ensure that its results were widely disseminated among target groups. This included the publication of the first set of practise abstracts to disseminate our results to stakeholders. The project website and newsletters serve as the main sources of up-to-date information about the project’s progress.
The results of DECIDE will lead to improved decisions on disease control to increase animal health and welfare and protect human health and the food chain in Europe and beyond.
Data science:
The legal documents developed in DECIDE, e.g. the Joint Controller Agreement, can serve as an example for other data-driven projects. The existing “DECIDE” ontology will be extended to species specific ontologies and will be continuously updated. Notebooks and a workshop will be created to illustrate the use of ontologies in statistical analysis and decision support tools. Example code for federated access to data is being developed. Guidelines for data access and re-use of animal health data aim to facilitate and improve good data management. All material will be publicly available for others working in the animal health field.
Modelling:
The improved multivariate dynamic linear models (DLM) on salmon mortality is integrated in dashboards and used in the salmon industries in Norway, Scotland, and potentially Ireland. The DLMs will also be applied to longitudinal data from cattle, pigs and poultry with the purpose of early detection of disease. The model in EMULSION for bovine respiratory disease will be validated on field data from other countries and different calf rearing facilities. EMULSION is also used to develop a model for infectious diseases in pigs. The transmission and the effectiveness of animal health interventions will be modelled.
Decision support tools:
The tools that are developed in the DECIDE project are initially focusing on a specific animal species in specific countries. However, several of the tools are already adapted such that they can be used in other and for other countries and species. The deliverables can serve as a framework for the design and evaluation of data-driven decision support tools with the users that manage health of any animal species or maybe even plant or public health.
Burden of animal disease:
The methodology for determining the burden of animal diseases can be applied to any animal species and any disease. In DECIDE we translate the methodology for global figures to information that supports farm or population level decisions. The multidimensionality of including welfare and medicine use in the burden of disease is an innovative approach that can lead to more sustainable management of animals.
Social Sciences:
The insight gained on psychological and social preconditions of the use of data tools and the quantitative approaches to investigate stakeholders’ drivers, barriers and willingness-to-implement and -use data tools can be used for other animal species but also for users of such tools outside veterinary medicine.
Data science:
The legal documents developed in DECIDE, e.g. the Joint Controller Agreement, can serve as an example for other data-driven projects. The existing “DECIDE” ontology will be extended to species specific ontologies and will be continuously updated. Notebooks and a workshop will be created to illustrate the use of ontologies in statistical analysis and decision support tools. Example code for federated access to data is being developed. Guidelines for data access and re-use of animal health data aim to facilitate and improve good data management. All material will be publicly available for others working in the animal health field.
Modelling:
The improved multivariate dynamic linear models (DLM) on salmon mortality is integrated in dashboards and used in the salmon industries in Norway, Scotland, and potentially Ireland. The DLMs will also be applied to longitudinal data from cattle, pigs and poultry with the purpose of early detection of disease. The model in EMULSION for bovine respiratory disease will be validated on field data from other countries and different calf rearing facilities. EMULSION is also used to develop a model for infectious diseases in pigs. The transmission and the effectiveness of animal health interventions will be modelled.
Decision support tools:
The tools that are developed in the DECIDE project are initially focusing on a specific animal species in specific countries. However, several of the tools are already adapted such that they can be used in other and for other countries and species. The deliverables can serve as a framework for the design and evaluation of data-driven decision support tools with the users that manage health of any animal species or maybe even plant or public health.
Burden of animal disease:
The methodology for determining the burden of animal diseases can be applied to any animal species and any disease. In DECIDE we translate the methodology for global figures to information that supports farm or population level decisions. The multidimensionality of including welfare and medicine use in the burden of disease is an innovative approach that can lead to more sustainable management of animals.
Social Sciences:
The insight gained on psychological and social preconditions of the use of data tools and the quantitative approaches to investigate stakeholders’ drivers, barriers and willingness-to-implement and -use data tools can be used for other animal species but also for users of such tools outside veterinary medicine.