Final Report Summary - RISKSUR (Providing a new generation of methodologies and tools for cost-effective risk-based animal health surveillance systems for the benefit of livestock producers, decision makers and consumers)
Executive Summary:
The risk of occurrence of exotic, new (emerging), re-emerging as well as endemic diseases has increased substantially particularly due to changes in consumption of animal source foods and the related intensification of animal production and increase in global trade of animals and their products. Bovine spongiform encephalopathy (BSE), foot and mouth disease (FMD) and avian influenza, as well as human severe acute respiratory syndrome (SARS), influenza A virus subtype H1N1 and enterohaemorrhagic Escherichia coli (EHEC) epidemics demonstrated the immense adverse socio-economic impact of animal and human disease. While the need for effective animal health surveillance is widely recognised for the management of such threats, the currently used systems do not take optimal advantage of recent advances in epidemiological approaches. In addition, investment is being constrained due to significant financial budget reductions in most countries. This can result in a potentially risky situation that demands the implementation of a new approach to surveillance system design.
The overall aim of RISKSUR was to develop decision support tools for the design of cost-effective risk-based surveillance systems that integrate the most recent advances in epidemiological methodologies.
This was achieved by the development and application of design and evaluation frameworks for animal health surveillance systems for three different surveillance objectives: 1) Early detection of animal disease, 2) Demonstration of freedom from animal disease, and 3) Determination of disease frequency and detection of cases of endemic animal disease. Tools for comparative epidemiological and economic evaluation of different surveillance system designs were developed and made available through a web-based decision support tool. The project provided decision-makers with a validated decision support tool tailored to their needs that allows the design of more cost-effective animal health surveillance and thereby reduces direct and indirect impact of animal disease on European Union citizens.
RISKSUR partners represent a unique combination of internationally recognised expertise in animal disease surveillance methodologies and economic evaluation as well as applied experience in the delivery of surveillance programmes in a variety of socio-economic contexts as well as from a national and global perspective. The practical implementation of the tools and the transfer of the knowledge was led by four small and medium enterprises (SMEs) with a strong track record in early adoption and translation of research results.
Project Context and Objectives:
‘Prevention is better than cure’ was formally adopted as the strategic direction of the new EU animal health policy in September 2007. The aim of the new strategy is to reduce the likelihood of animal disease, and to minimize the impact when outbreaks occur. The increase in the occurrence of new (emerging) diseases and the spread of exotic ones along with increasing levels of endemic and food-borne diseases require earlier, better and safer surveillance to achieve this objective.
Furthermore, it is recognised that improved protection of human health will require more effective linking of animal and human health surveillance systems as well as adoption of interdisciplinary approaches, which is one of the aspects of the One Health concept. However, financial constraints of governments in times of public spending cuts have reduced or limited investments in surveillance of animal disease.
In addition, cross-compliance and equivalence issues constrain governments in relation to surveillance strategy design. Efficient and practical solutions taking advantage of novel scientific methodologies, such as risk-based surveillance, are urgently needed, as well as integration of epidemiological approaches with socio-economic and qualitative methods where appropriate. At the same time, governments are required to reduce their expenditure in response to growing fiscal constraints. With respect to animal health surveillance systems, very significant advances have been made in laboratory diagnostic tools whereas the implementation of ‘smart’ methods for collection and analysis of surveillance data has been neglected in the past decades. This is despite of much improved understanding of factors influencing animal disease risks which should provide an
opportunity to design more efficient surveillance systems. In addition there has been a lack of data collection and information driven work in relation to improving our understanding of the changing dynamics of the livestock sector.
This will be of utmost importance as especially the past twenty years have seen the EU experiencing several animal health crises. These crises have had dramatic adverse effects on the livestock sector and public health, and resulted in significant disruptions to markets and the wider economy. Several factors have compounded the risk of such crises – globalisation and the resulting increase in trade, the expansion of EU borders to the East, and changes in the structure of the EU livestock sector.
Recent outbreaks of exotic diseases such as avian influenza (AI), foot and mouth disease (FMD) and bluetongue (BT) in previously unaffected regions of the EU have highlighted the threat posed by emergence and / or spread of infectious agents, and emphasise the need for well-developed and adequately resourced animal health systems, to ensure early detection and rapid containment.
Over the period 1998-2009 outbreaks of avian influenza mainly affected Germany, Italy, and the Netherlands; FMD caused a severe crisis in the UK in 2001 and re-emerged in January 2011 in Bulgaria; several bluetongue outbreaks occurred unexpectedly north of the Alps in countries such as France, Italy, Netherlands and Germany since 2006. Furthermore, there were cases of Schmallenberg (SBV) detected in the Netherlands and Germany between August and October 2011.
The economic implications of animal diseases and animal disease control are severe. In 2008, the cumulative global cost of the H5N1 avian influenza pandemic was estimated to have been at least US$ 25-30 billion (Newcomb, 2008). Given that the epidemic has not been controlled in several countries, the cost will now be much higher than that figure. Another example is the human EHEC outbreak in Germany which while relatively small resulted in significant negative economic and human health impacts and prompted renewed demands for more effective surveillance systems (Turner, 2011).
Evaluation implies a judgment on the surveillance system and recommendations for improvement. This is done in reference to various aspects of surveillance, specifically whether it allows transparent interpretation of outputs, more objective decision making and resource allocation as well as improvements in system design and enhanced acceptance of system outputs by stakeholders at local (e.g. farmers, veterinarians) and national levels (e.g. reference laboratory, veterinarians at central level). This is particularly important given the knowledge gaps in understanding of many diseases, which leads to varying degrees of uncertainty and bias in generated outputs.
The use of economic evaluation in decision making for animal health surveillance policies has been limited so far. When considering the start, end or change of a surveillance programme, policy makers need to know if and how much surveillance is needed and what the most economical options are and how it best integrates with interventions. There are few specific guidelines available for economic evaluation of animal health surveillance, and only a limited number of empirical applications on the economic value of surveillance. It is critical to select appropriate economic efficiency criteria taking into account the evaluation context and viewpoint of the analysis. Priority setting, affordability, sustainability, social acceptance and communication are all issues that policy makers have to consider when designing and implementing disease management policy. These assessments are not easy as the data available to quantify investments in surveillance and monitoring activities are very limited at both national and international levels. It is expected that member states would value access to a tool that would help them with technical and budgetary optimization of resource utilization when defining animal surveillance policy to as part of the national animal health policy.
For these reasons, a need for frameworks and integrated tools that allow the design and implementation of epidemiological and economically optimised animal health surveillance systems that generate outputs in relation to the animal health status of the livestock sector, allowing effective management of any emerging or existing risks, was identified.
The overall aim of RISKSUR was to develop and validate conceptual and decision support frameworks and associated tools for designing efficient risk-based animal health surveillance systems based on an interdisciplinary approach tailored to the needs of EU Member States. The project aimed to develop tools and frameworks targeted at the following surveillance objectives associated with livestock diseases:
1. Detection of incursion of exotic, new (emerging) and re-emerging disease
2. Declaration of freedom from specified diseases and infections
3. Monitoring of endemic diseases (case detection, disease frequency estimation)
The RISKSUR project addressed these three surveillance objectives in different work packages (WP). The first (WP1) develop a conceptual evaluation and decision-making framework for risk-based animal health surveillance system designs based on work to identify decision-makers involved in the resource allocation for surveillance; described decision-making criteria used; characterised surveillance components in place, economic criteria and associated data flows; and identified evaluation attributes. The work in WP1 informed further activities for each of the aforementioned three surveillance objectives in WP2-WP4. Work package 2 developed a framework to support the design of surveillance systems and associated tools aimed at early detection of animal diseases (including exotic, new and re-emerging diseases) by conducting a systematic review; identifying and defining key risk scenarios; describing data requirements, collection and analyses tools, and by applying the framework to case study disease. Work packages 3 and 4 used a similar approach focusing on the development of a decision support framework for surveillance systems and associated tools aimed at demonstrating freedom from diseases and estimating disease frequency and detection of cases of disease, as well as multi-hazard surveillance systems, respectively. As some of the data and methods were used in several WPs, there was extensive collaboration and exchange between partners. The results of WPs 2-4 were evaluated for single and multi-hazard surveillance systems specifically in relation to their efficiency in WP5. The objective of WP5 assessed the technical and economic efficiency of the surveillance designs developed in RISKSUR by developing and applying an evaluation support tool for the integrated epidemiological and economic evaluation of surveillance. The transfer of knowledge and technology to key stakeholders from policy and industry was facilitated through the development of practical and accessible web-based decision support tool enabling decision makers, industry and the broader scientific community to take advantage of the research outcomes in WP6. In WP7, dissemination of new methods and novel applications for surveillance design by sharing information with different shareholder groups, and by providing training to policy makers and surveillance workers, was performed. WP8 covered the project coordination and management.
In summary, the RISKSUR consortium used an interdisciplinary approach to develop novel decision support frameworks and tools integrating novel epidemiological, economic and qualitative research methods for design of efficient risk-based animal health surveillance systems. This will allow evidence-based decision-making by key stakeholders in EU member state veterinary administrations.
References:
Newcomb, J. 2008: Economic Opportunities and Benefits of Implementing One World-One Health Strategies for Health Promotion and Disease Prevention. Bio-Era Associates Report.
Turner, M. 2011: Microbe outbreak panics Europe Nature 474, 137| doi:10.1038/474137a.
Project Results:
LIST of abbreviations for partner institutions
ACCEL- ACCELOPMENT AG
APHA- Animal and Plant Health Agency
CIRAD- Centre de Coopération Internationale en Recherche Agronomique pour le Développement
FAO- Food and Agriculture Organisation of the United Nations
FLI- Friedrich-Loeffler-Institut
GD- Gezondheidsdienst Voor Dieren
RVC- The Royal Veterinary College
Safoso- Safe Food Solutions AG
SVA- Statens Veterinaermediciniska Anstalt
TT- Tracetracker
UCM – Universidad Complutense de Madrid
The RISKSUR project has achieved all Scientific and Technological target objectives, set at the beginning of the project. All 36 deliverables were submitted over the course of the project, between November 2012 and October 2015.
1. Work package 1 - Development of a conceptual evaluation framework
This work package generated the report: ‘Mapping of the main surveillance systems and decision-makers.’ The objective related to Task 1.1 and its associated Deliverable D1.1 was to produce a high level overview of the populations, surveillance systems and decision-making processes currently used in the EU. Primary and secondary data were collected and collated in seven RISKSUR partner countries (France, Germany, Great Britain, the Netherlands, Spain, Sweden and Switzerland) and six non-partner countries (Belgium, Bulgaria, Czech Republic, Denmark, Ireland and Italy). Members from the RISKSUR Consortium countries screened web pages, government reports, scientific and other grey literature and approached relevant contacts with the aim to describe all existing private and public surveillance system components (SSC) in the year 2011 in all animal species of economic value. Both disease-specific as well as health-event related surveillance was considered. Further, for the seven partner countries, data were collated on critical infrastructure and interviews were conducted with 35 decision-makers involved in private and public surveillance. For all EU member countries data on livestock and bee holdings in Europe, human and animal populations, and gross domestic product and farm values were collated from Eurostat. Data on trade was obtained from DG SANCO’S Trade Control and Expert System (TRACES). These data were used to create maps to visualise distributions and differences in human and animal populations and holding densities as well as animal movements across Europe. All data were analysed descriptively and presented in D1.1 report for the seven countries. The data and analysis for the additional 6 countries were presented in a supplement to the D1.1. The research brought together key system attributes of the selected EU countries allowing a comparison between countries and systems, as well as appreciating the impact of the linkages between systems and countries and their consequences for risk management. It was a collaborative effort under RVC lead involving nearly all Consortium partners.
Next within WP1 a systematic review of evaluation methods of surveillance systems and current practices was carried out, submitted as Deliverable D1.2 report. The work was divided into two independent subtasks: a systematic review of economic assessments of surveillance systems, criteria and methods, and a systematic review of the guidelines, framework, methods and tools for the evaluation of surveillance systems. The former led by RVC and APHA, produced an overview of the approaches and techniques available to perform economic assessments of surveillance. Therefore, a systematic literature review of economic evaluations of animal health surveillance systems published in the last 20 years was performed. Most economic evaluations were found to focus on surveillance systems targeting single diseases, even though an economic evaluation of surveillance would ideally address different diseases and cover multiple geographical regions. This review also emphasised the lack of standardised approaches for measuring the effectiveness of surveillance and for conducting cost-effectiveness analyses. The relatively small number of papers published in this area highlighted that economic evaluation of surveillance is rare in animal health. The review concluded by discussing the potential relevance of economic approaches that are still underutilised for evaluating surveillance. The ‘review of surveillance evaluation frameworks and practical decision-making factors’, carried out by APHA, CIRAD and RVC identifies the main advantages and limitations of the existing guides used for the evaluation of surveillance systems, either in public health, animal health or environmental health. It highlights potential ways of improving the evaluation process in order to provide useful recommendations for the development of the RISKSUR decision support tool. Most guides identified were general and provided recommendations for broad evaluations. Several common steps in the evaluation process were identified, namely (i) Defining the surveillance system under evaluation, (ii) Designing the evaluation process, (iii) Implementing the evaluation, and (iv) Drawing conclusions and recommendations. The analysis highlighted a lack of information regarding the identification and selection of methods and tools to assess the evaluation attributes as well as a lack of consideration of economic attributes and sociological aspects. This review emphasized the need to develop a decision support tool for economic evaluation of surveillance systems, which integrates and complements the existing frameworks, methods and tools.
Deliverable 1.3: ‘Prioritisation and relative weighting of surveillance characteristics and evaluation attributes as well as factors relevant to the selection of evaluation criteria’, was led by RVC together with APHA and CIRAD. As a first step towards the development of the RISKSUR evaluation tool, this task and its associated deliverable D1.3 resulted in a better understanding of the surveillance attributes used for the evaluation and the complex inter-relations between them, to identify potential gaps in evaluation attributes and to suggest new approaches for optimising the use of attributes in the evaluation of the effectiveness of the surveillance. This task was divided into 5 Subtasks. The results highlighted the need to work towards a theoretical and unified approach for evaluating effectiveness of surveillance, which is then further investigated in one of the subtasks. Moreover, the findings presented in the report cover many aspects of evaluation attributes from agreeing on definitions to the development of a new measure of effectiveness.
The deliverable D1.4: ‘Development of an integrated evaluation matrix’, involving CIRAD, RVC, APHA, ARCADIA and other Consortium members, integrated outputs from previous tasks, and included the development of an evaluation matrix that will help to identify the most suitable epidemiological and economic approaches, criteria and techniques. After defining the context, the RISKSUR Consortium with CIRAD’s lead developed a conceptual model of a decision support tool for the integrated epidemiological and economic evaluation of animal health surveillance system (EVA tool). Relevant information from previous tasks was used to build on existing evaluation frameworks, methods and tools and thereby make best use of already available approaches. The EVA tool provides comprehensive and practical guidance to decision-makers and their technical advisers to plan and conduct evaluations of animal health surveillance systems. It guides the user through several steps and options on what, why and how to measure in order to evaluate animal health surveillance systems. The users have to provide inputs in relation to the evaluation question and the general context (e.g. epidemiological situation, surveillance objectives, and data availability) to describe the evaluation context and select a suitable evaluation question. The tool also helps to select suitable evaluation attributes and corresponding measurement methods. Moreover, it provides detailed guidance on the application of the methods and the interpretation of the results.
The last task and its associated deliverable D1.5 generated in this work package: ‘Define integrated data collection protocols and mechanisms to ensure robust data quality’. FLI lead this task, on which RVC, APHA, SVA, UCM, CIRAD collaborated. Its objectives were:
a) To describe the final design adapted for the mapping and review of surveillance systems,
b) To document the results of consistency checks as well as revised definitions and
c) To discuss the current limitations of information describing surveillance.
The D1.5 report presents guidelines to address these areas and the agreed additional data requirements for the review of surveillance systems. A standardised protocol was developed for the review to standardize data collection amongst project partners, since questionnaire results had highlighted the variability in data completeness and interpretation. The application of consistent terminology is crucial to ensure robust data quality and facilitate comparisons between surveillance systems. Efforts to standardize terminology for animal health surveillance at the International Conference on Animal Health Surveillance (ICAHS) conference in 2011 were used as a basis for the mapping and review task. The data collection served as a validation of these terms across a range of surveillance systems and countries. Some problem areas were identified, and as a result terminology was further developed. Particularly for risk-based surveillance the terminology was not adequate to reflect the diversity of components that could be termed risk-based but vary in their nature and intensity. Hence, it was recommended to further refine the terminology related to risk-based surveillance. The mapping and surveillance system review did provide many examples that can be used for illustration. Early input from external experts was obtained by encouraging general feedback through establishing a glossary on the RISKSUR website. Another area identified with potential for improvement was that surveillance activities were often not well documented, which made it difficult to capture all existing surveillance activities despite considerable efforts made. In addition, regional variation within countries with a decentralized administration was difficult to capture. Registering surveillance activities and their design would allow providing an overview and better coordination of efforts made by the public and private sector.
2. Work package 2 - Detection of exotic, new, or re-emerging disease
The report ‘Review of surveillance systems and methods for early detection of disease’, carried out by UCM, was the first deliverable D2.6 completed during the project period. This report consists of a compilation of information regarding surveillance systems and strategies aimed at the early detection of new, exotic and re-emerging diseases. It looks at different epidemiological methods used in surveillance systems for early detection of diseases, including current methods such as passive notification of suspect cases and active strategies such as sentinel surveillance systems.
The next task and its associated deliverable report D2.7 entitled ‘Definition of key risk scenarios using selected case study diseases’ involved several partners from WPs 2-4. The corresponding tasks in these WPs were grouped together as the ‘Case study selection’ Theme. Consortium members contributing to this theme worked in collaboration with CIRAD, in order to establish a consistent process for selecting case study hazards and scenarios. The selection process was based on three steps: 1) reducing the list of hazards identified as important by both the EU and the Consortium members after assessing presence/absence and initial data availability; 2) selecting a shortlist of hazards based on transmission, surveillance, detailed analysis of data, species representation, selection in other WPs and opinion of the Advisory Boards; and 3) selecting case study countries for selected hazards to obtain the final case studies for WP2 based on data on historical occurrence, host and production types, environmental characteristics and surveillance results. The Advisory Boards supported the arguments that led to the final selection of case studies. For WP2 the following case studies were chosen: avian influenza in the UK; African swine fever in an Eastern European country, possibly Poland; and bluetongue in Spain. The routes of introduction and contact interfaces with wildlife were represented in pathway diagrams and the potential source of data needed was indicated. The selection of the case
studies and the description of pathway diagrams were described in detail in deliverable report D2.7.
This was then followed by the task ‘Development of a framework to inform the design of surveillance systems’, D2.8. Partners involved in WPs 2-4 worked together on this task. The surveillance design framework, available publicly through the wiki link (http://surveillance-design-framework.wikispaces.com/) comprises 14 steps, namely, 1) description of the surveillance design; 2) overview of surveillance components; 3) description of target population; 4) steps that lead to disease suspicion; 5) enhancements; 6) testing protocol in place; 7) study design for active surveillance components; 8) sampling strategy; 9) questions on data generation process; 10) document transfer mechanisms; 11) document data translation process; 12) document epidemiological data and analysis; 13) document how will surveillance be disseminated; 14) last questions on how the review will be performed. The next step in the framework was to evaluate the surveillance system, for which the EVA-tool (developed and applied under WP5) is to be used. Activities from the WP2-specific portion of the framework development ensure that the surveillance for early detection of known hazards is well covered in the final framework. The results from this task are presented in the Deliverable 2.8 report and also influenced the writing of the early detection chapter of the ‘Best Practice Guidelines’.
Finally, this WP and its associated deliverable D2.9: ‘Tools for risk assessment and evaluation’, involved the RISKSUR partners UCM, APHA, CIRAD and GD. The framework was applied to two risk scenario diseases for early detection surveillance: ‘avian influenza in the UK ‘and ‘African swine fever in Poland’. The flexibility and robustness of the framework and tools developed were evaluated. The data required to characterize the population at risk, quantify the pathways and assess the probability of detection were also analysed. New surveillance systems are proposed and designed based on the review of surveillance options, development of data collection and analysis tools. The epidemiological performance of the surveillance systems was evaluated by comparing risk-based against traditional surveillance systems with the new tools developed. The model used to test and analyse the epidemiological performance of the new component for avian influenza passive surveillance in the UK proved that outbreak detection was quicker for the risk-based scenario than the current passive surveillance scheme. For African swine fever we performed a qualitative risk assessment based on the probability to detect field infection in pigs and wild boar and proposed a new risk based surveillance component design based not only on the situation of the bordering countries, but particularly on the current risk of spread within Poland using spatio-temporal assessment tools that consider the distribution of the population at risk and the direction of the outbreaks. The work from this task was presented in Deliverable report D2.9.
The main conclusions are that the framework is very useful for documenting and for training in surveillance, as well as for being able to identify strengths and weaknesses of a surveillance system. However, other approaches to presenting surveillance components are also possible. The separation of surveillance objectives for each surveillance component was not always useful, since often many components together have the objective of identifying or ruling out potential suspects, which is one of the main aims of early detection surveillance, while if considered separately and on a purely theoretical basis, the surveillance objective can be interpreted together with another surveillance objective. Other potential areas of improvement include extending the documentation on suspicion of disease to include findings other than those resulting from passive surveillance; include a multi-hazard surveillance design for early detection; or to improve the standardisation of the final sections of the design so that the different levels and trainings in epidemiology across Europe can eventually be comparable. A very significant benefit of the design framework is that it offers a transparent way of documenting surveillance design that can improve communication between all actors involved in surveillance.
3. Work package 3 - Demonstration of freedom from disease
This work package first produced the ‘Review of surveillance systems and methods used to demonstrate disease freedom’, under the lead of FLI together with APHA, SVA, RVC, UCM. Task 3.1 was divided into two parts: a systematic literature review and a systematic review of surveillance systems. The Deliverable D3.10 report includes the former only. The review covers 131 articles and the deliverable report provides an overview of the development and interrelation of methods related to demonstrating freedom from disease including sample size calculations, scenario tree models, and simulation models. The results show to what extent important methodological aspects have been covered by the reviewed articles. In the discussion, proposals were made in terms of which well-established methodologies could potentially be further advanced or should be explored in further detail to make an adequate assessment. The outputs from this report were used as basis for the development of the design framework and associated tools. For the review of surveillance systems, an inventory of 515 surveillance components was created covering 26 hazards considered for case study selection and nine EU countries. The countries included the three partner countries producing the report (DE, ES, SE), four partner countries contributing to data collection (CH, FR, GB, NL) and two non-partner countries (DK, IT). The analysis considered those variables identified under Task 1.1 as well as 23 additional variables collected for the Task ‘Review of surveillance systems’. Since the additional information includes sensitive data such as the number of sampled herds and animals, results were summarized as an internal report only to inform project specific tasks such as the framework development and case study selection. The systematic review has provided crucial background information for the selection of tools (literature review) and the development of the design framework (literature and surveillance system review). The surveillance system review furthermore highlighted limitations in terminology and the way surveillance is documented.
Task 3.2 of this work package: ‘Development of a framework and associated tools for demonstrating freedom from disease’, was divided into 3 sub-tasks. First: Defining key risk scenarios and associated diseases relevant for freedom from disease, which was carried out by FLI together with APHA, RVC, SVA, Safoso and UCM. The following case study hazards were selected for WP3, based on a systematic selection process: a) classical swine fever; b) bluetongue disease and c) infectious bovine rhinotracheitis. Deliverable report D3.11 summarizes risk pathways for classical swine fever and the data requirements for all three selected case study hazards. Next: ‘Development of a framework for designing surveillance systems for demonstrating freedom from disease, with specific reference to choice of sampling methods’, carried out by FLI with APHA, SVA, RVC, UCM, Safoso and resulted in the associated deliverable report D3.12. Here scientists of the FLI participated in meetings and provided input to the development of the surveillance design framework, which was led by SVA. Furthermore, an inventory of tools to be evaluated under RISKSUR was proposed for WP3. Differences in the interpretation of terminology were identified as a problem during the WP1 ‘Mapping and review of surveillance systems’, as well as during the development of the surveillance design framework. Therefore, a terminology working group led by the FLI was initiated and two topics were covered (Topic 1: Means of data acquisition, Topic 2: Surveillance purpose). Questionnaires were used to obtain inputs from eight senior consortium members in three rounds before summarizing the conclusions. Also an ‘Evaluation of different epidemiological methods for demonstrating freedom from disease’ was carried out by FLI together with SVA, UCM, CIRAD and RVC. The performance of conventional and alternative surveillance strategies was compared in two case studies. The results of the first case study suggested that a deeper insight into the overall utility of a surveillance system could be gained when more than one effectiveness attribute was included in the evaluation. It was concluded that risk-based surveillance strategies should be considered when designing more efficient surveillance systems. This task resulted in Deliverable D3.13 completing the assessment of the framework and tools for certifying freedom from disease.
4. Work package 4 - Prevalence estimation and case detection for endemic disease
This work package first produced a ‘Review of literature on surveillance systems and associated methodologies from an epidemiological perspective for finding cases of endemic disease, including systems used in EU member states and other relevant countries and to generate an inventory of the extent of multi-hazard surveillance activities’. It was carried out by SVA together with FLI, APHA, UCM, CIRAD and GD and divided into two parts: a systematic literature review and a systematic review of surveillance systems. The systematic literature review for WP4 was carried out focussing on surveillance designs and associated methods aimed at finding cases of endemic disease and estimating prevalence. The initial literature search was coordinated between WPs 1-4. The systematic review covers, in all, 69 peer reviewed papers and constitutes D4.14. This deliverable was also further developed into a draft manuscript for publication in a peer-reviewed journal.
In addition to the systematic literature review, a review was conducted covering on-going surveillance in partner and non-partner countries (non RISKSUR Consortium members), using data obtained from the grey literature (national surveillance reports, personal communications etc.). For this review, an inventory of 515 surveillance components covering 26 hazards and nine EU countries has been compiled, in close collaboration with WP2 and WP3. The analysis considered those variables identified under Task 1.1 as well as 23 additional variables collected for the task ‘Review of surveillance systems’. Since the additional information includes sensitive data such as the number of sampled herds and animals, results were summarized as an internal report only to inform project specific tasks such as the framework development and case study selection. In the review part of the report, special attention is given by WP4 to the concept of multi-hazard surveillance. Such data were initially captured in Task 1.1 but were further refined and analysed in the internal report.
As indicated above the review of surveillance systems has had a certain overlap with Task 1.1 but has had a different focus and required additional data collection. The data collection was challenging as there were differences in how surveillance concepts are interpreted even among experts in the area. Therefore, significant effort was invested into retrospectively checking consistency of the data initially collected. Several suggestions for improvement and harmonisation of surveillance terminology were identified, and highlighted the importance of a generic terminology task.
A further task in this WP consisted of ‘Defining key risk scenarios and associated diseases relevant for surveillance aimed at estimating disease frequency and detection of cases of endemic disease’. This was carried out by SVA together with APHA, UCM, FLI, GD and as aforementioned, the work within this task was coordinated between WPs 2-4, under the theme ‘key risk scenarios’, initially lead by APHA and subsequently by GD. Potential diseases that are endemic in some or all EU member countries were identified, and their suitability as case study diseases has been assessed by considering disease characteristics, potential for improving surveillance using risk-based approaches as well as data availability and expertise among partners. In this process, WP4 contributed to the data collection on data availability and expertise for Sweden. The final selection of case study hazards was also informed by the characteristics of the livestock populations in partner countries as well as if the disease was indeed endemic in the countries considered. Opinions of stakeholders about the importance of the hazard in question, and the need for a review of the current surveillance were also taken into account. The hazard eventually selected for the case studies focusing on case finding capacity and disease frequency estimation was Salmonella Dublin in dairy cattle. For the case study on optimisation of multi-hazard surveillance, porcine reproductive & respiratory syndrome (swine) was chosen as the hazard for which the backbone surveillance activity is designed, with classical swine fever, swine vesicular disease and Aujeszky’s disease as hazards for which the same data collection process was utilised. Sweden and the Netherlands were used as case study countries. The process of selecting case study hazards, their nature and risk pathways as well as data needs are described in deliverable report D4.15.
Next followed the ‘Development of framework and associated methodologies to inform the design of surveillance systems’, led by SVA in collaboration with APHA, UCM, FLI, CIRAD and RVC. It included three subtasks, addressing the different surveillance perspectives covered by WP4; the maximisation of case detection capacity, the estimation of disease frequency and optimisation of multi-hazard surveillance systems. The structure and contents of the design framework were informed by previous tasks in RISKSUR; a literature review (D4.14) the mapping of surveillance systems (D1.1) and a review of existing surveillance systems in the EU (internal report). In particular, the latter highlighted the need for more transparent and standardised documentation of surveillance activities across the EU, and it was therefore concluded that documentation and reporting of surveillance design decisions should become an important feature of the surveillance design framework. The generic construction of the surveillance design framework was approached by decomposing surveillance activities into several attributes that represent the different steps in the planning of a surveillance system. These were subsequently reviewed to define and characterise the ‘building blocks’ of good surveillance design practice. The blocks were organised into a logical flow and grouped into broader surveillance design sections. This general framework was then refined to address the specific surveillance goals of early detection, disease freedom documentation, case detection and prevalence estimation. The framework was subsequently implemented as a questionnaire in a Microsoft Excel® spreadsheet using real data. At various stages of development the framework was tested and revised with the help of surveillance professionals from within the partner institutions. To support the use of the Excel tool, a wiki page has been developed and made publicly available at http://surveillance-design-framework.wikispaces.com/. The wiki content is based on the expertise of the RISKSUR consortium members developing the framework, and also incorporates feedback from the user group workshops (provided via a standardised questionnaire) as well as comments from the wiki members. It also includes a glossary and links to statistical tools (e.g. sample size calculators) that can be used during the process of design. The structure of the wiki is identical to the Excel® tool, making it easy to retrieve the advice needed at each design step. The development of the design framework was performed in close collaboration and coordination with WP5 activities, to avoid duplication and overlap with the development of the evaluation tool (EVA tool). This pertained both to the conceptual development and to the development of the web tools performed in WP6. The design part of the RISKSUR web tool is identical to the Excel tool for the higher levels of hierarchy, i.e. surveillance systems’ level and definition of surveillance components in place. However, only the Excel tool provides an opportunity to produce detailed documentation of different components, and to provide assistance in identifying design steps that could become targets for redesign after an evaluation. The decision to refrain from a full incorporation of the design framework in the web-based tool was taken at a time when the Excel design tool was in place and functional, while the development of the web-based EVA tool was under time and resource constraints. It was therefore decided to allocate the remaining WP6 resources to the completion of the EVA tool. However, the availability of the decision support in form of an Excel tool, in conjunction with the wiki page, has proven to be a robust and flexible format which is, in practice, useful for a broader target group than if it had been exclusively implemented in the online web tool. The framework is described in deliverable D4.16.
The development of the design framework required intensive coordination with WP5, to avoid duplication and overlap. This coordination was instrumental since the generic design framework developed by WPs 2-4 in collaboration and the evaluation tool developed in WP5 are complementary tools. For WP5 the focus was on evaluation of surveillance performance whereas the design framework aims to guide the user through the decisions leading up to an optimal design, taking into account the factors that would subsequently be of interest to evaluate. The final task consisted of the application of the framework to assess the epidemiological performance of surveillance systems aimed at case finding and disease frequency estimation for endemic diseases using different risk scenario diseases, and multi-hazard surveillance. This was carried out by SVA, RVC and GD. The surveillance design framework was subsequently applied to design, redesign and documentation of existing surveillance systems, aimed at assessing whether the structure of the tool was appropriate for the endemic disease surveillance objectives of case detection and prevalence estimation, as well as for multi-hazard surveillance. The case studies conducted in WP4 focused on Salmonella dublin in cattle in Sweden (case finding and prevalence estimation), and the multi-hazard surveillance covering porcine respiratory and reproductive syndrome (PRRS, mother component), Aujeszky’s disease (AD, child component) and classical swine fever (CSF, child component), also in Sweden.
The Salmonella dublin case study focused on alternative designs for a new surveillance component under development, which meant there was limited pre-existing data available. Instead, the case study was approached by adapting a data-driven modelling framework developed by SVA for verotoxigenic E. coli O157. The model simulates the spread of S. dublin in the Swedish cattle population during a 8.5-year period, using real animal movement records from the national cattle registry. Two different surveillance strategies were defined; i) a non-risk based scheme where all herds are sampled once per quarter and ii) a risk-based scheme where herds in high risk areas are sampled quarterly, and all other herds are sampled once per year, during the season with highest probability of detection. Output from the model has also been used as the basis for the economic evaluation conducted in WP5, in parallel with the estimation of the epidemiological performance. It was concluded that a conventional surveillance approach would indeed allow the detection of more infected herds in the short run, but a risk-based surveillance would achieve the same effectiveness over a longer time span, at a much lower cost. Deliverable D4.17 report describes the application of the surveillance design framework to the documentation of surveillance activities with objectives covered by WP4.
5. Work package 5 – Evaluation of epidemiological and economic effectiveness of surveillance systems
The first task carried out was the ‘Specification and further development of frameworks and methods for the evaluation of economic efficiency of surveillance’. It involved CIRAD, ARCADIA, RVC, APHA and FAO and included a review of the surveillance system evaluation frameworks, methods and tools. Also, the main challenges linked to the economic evaluation of animal health surveillance were addressed, notably how to promote economic evaluation, how to provide integrated guidance and how to report on the evaluation results. The outputs of this sub-task were reported in the evaluation Wikispace, in the Best Practice Guidelines document, and through the WP5 deliverables D5.18 D5.21 D5.22 and D5.23. Also, the EVA tool was developed based on the conceptual model produced within the first 18 months of the project. This consisted of the development of the evaluation tool structure and logic (CIRAD, RVC), an Online EVA Web Tool development (structure, logics and content), a Wikispace development (to provide guidance on evaluation concept and use of EVA tool) (structure and content), an evaluation question guidance pathway, attribute selection logics, assessment method selection logics and a review of the important points to transfer results of the evaluation to decision makers and protocol development (ARCADIA, CIRAD, RVC).
A final WP5 workshop was organised to review the status of the EVA tool development, to validate the logic of the tool and to review/finalise the on-going work on economic evaluation case studies. Specific action points and an updated work plan were identified to finalise the tool and the evaluation case study and prepare the final deliverables (D5.22 and D5.23). The next task in this work package consisted of a ‘Standardisation of data collection for economic impact assessment’. Partners were CIRAD as the lead, and ARCADIA, RVC, APHA and FAO. Different protocols for data collection for economic evaluation of animal health surveillance systems were developed and are briefly detailed below. Moreover, a work plan was developed to provide guidance for the evaluation of a surveillance system (CIRAD); this work plan was applied to the RISKSUR economic evaluation case studies (in collaboration with RVC, FLI, SVA and APHA). The full description of those protocols and the related questionnaires has been reported in Deliverable D5.19. This work plan presented all steps that the evaluator should take to perform the evaluation once the evaluation question and methods have been selected, from the validation of the evaluation protocol (evaluation question, selection of attributes and methods) to the reporting of the results of the evaluation to the decision makers.
The final task consisted of the ‘Selection of case studies and application of the evaluation matrix’. This task was led by CIRAD and involved all Consortium partners and resulted in the deliverable D5.20 ‘Case studies selected and described’. The case studies used within WP5 to develop, test and validate the tool by implementation of the economic evaluation are described in Figure 1. The task consisted of the evaluation of the economic efficiency of the novel designs developed in WPs 2-4 and validation of the frameworks and methods elaborated and involved CIRAD, ARCADIA, RVC, APHA and the FAO. The economic evaluation of the RISKSUR case studies was performed in the last 18 months of the project and reported in Deliverable D5.22.
6. Work package 6 – Decision-making tools for implementing risk-based surveillance
The first task of this project was to develop a statistical software package encapsulating functions for design of surveillance systems, in particular involving risk-based sampling. The development of the statistical software package continues to be in progress. It required Implementation of algorithms as an R package. This was led by FLI and included TT, RVC, SVA, UCM, APHA and CIRAD. Fifty-three functions, which can be used for the design and analysis of surveillance systems, were implemented as part of the R package RSurveillance. These functions related to the surveillance objective demonstrating disease freedom (n = 40), including risk-based approaches (n = 12), and prevalence estimation (n = 7). Furthermore, functions were included for combined and parallel testing (n = 6). A further task consisted of the Development of documentation in standard R formats. Partners involved were FLI, TT, RVC, SVA, UCM, APHA and CIRAD. These functions were documented in standard R formats (https://cran.r-project.org/web/packages/surveillance/surveillance.pdf). This was followed by a further task ‘Validation of R functions’ led by TT and involving: FAO, Safoso and RVC. The R functions were used by various Consortium members as part of the case studies to analyse results from the survey of Federal States in Germany and as part of other applications.
The next task consisted of the Publication of an R package for designing surveillance systems with specific reference to risk-based systems. Led by FLI and involving TT and RVC; the R package RSurveillance was published within the open-source software environment R (http://cran.r-project.org/web/packages/surveillance/index.html). This work package also included the task ‘Development of user-friendly accessible web-based decision support tool implementing methodologies for design and evaluation of surveillance systems’. Once the development of the tool was completed it was made available under: webtools.fp7-risksur.eu. Furthermore, a web-team had been in operation to secure several important tasks: ensuring all essential functions are transferred from the WPs 1-5, ensuring all design aspects from the Excel prototype were transferred, acting as QA for ensuring user-friendliness, and acting as a test team for all the above issues. This was crucial to be able to proceed to the next task to implement web-based decision support tool. It was led by TT, FLI, RVC, UCM and CIRAD and the web-based decision support system consisted of five main parts that were developed and implemented: a tool for designing surveillance systems, a tool for designing components within a surveillance system, a tool for evaluation of a surveillance system, a tool for statistical methods to be utilized in a surveillance system (this was reported in Deliverable D6.24) as well as administrative tools for the complete package.
The work package, under the lead of TT and included FLI, RVC, UCM and CIRAD; also had the task to Develop on-line documentation. The online documentation is implemented in the systems at several levels. There is some overlap between the different levels, mainly to assist users in the use of the online tools. The documentation is implemented both to enhance user-friendliness of the tools and to support clarification, consideration and understanding of the decisions the tools are supposed to support: Ease of use (wizards, mouse-over, context dependent information) of the tool itself. Examples, tutorials and best practice guidelines are accessible through links to wikis. The same team then proceeded to the next task, which resulted in deliverable D6.25 and which had to implement a multi-lingual interface translated into the main EU languages. The multi-lingual support was implemented through translation files. Currently, only an English version is accessible and a French language version is under development. The support of other languages will be a relatively straight forward task, by editing a language file.
The last task ‘Mobile and cloud computing tool development’ was subdivided into 2 sub-tasks, all led by TT and involving FLI, RVC and UCM. The task corresponded to Deliverable D.26. The first task was to assess the need for mobile and cloud computing tools. The project consortium explored the value of having this added functionality and concluded that it could not currently be justified considering the additional cost for developing tools on a mobile platform. The next task was to ‘Develop mobile and cloud computing tools for surveillance systems design’. The architecture of the web-tool has taken into account cloud computing and centralized storage capabilities, thus opening up opportunities for designing the user interface for a mobile device, as a frontend to the centralized cloud service, at a later stage.
7. Work package 7 – Training, dissemination and exploitation
The first task generated the ‘Development and maintenance of general dissemination tools’. It was led by ACCEL and included all other partners besides TT and RVC. One sub-task consisted of the setting up and maintenance of a Website (http://www.fp7-risksur.eu/). It also included features such as: terminology/glossary, impact tracking of the peer-reviewed publications, answers to 25 evaluation-related questions, and 13 online training modules. Another sub-task was an ‘Electronic newsletter’, a total of 6 newsletters with an additional final issue, with the overall description of the project major outputs were published by January 2016. The next sub-task, ‘Science to society’, led by Safoso, included a specific RISKSUR LinkedIn group that included 186 members, as of November 2015. It was used to regularly disseminate jobs related to animal health surveillance. The ‘Science to professionals’ sub-task consisted of a dissemination package, which had been prepared by accelopment and Safoso, was shared with the RISKSUR advisory board members and consortium members, for further dissemination through their professional networks. It consists of about 10 summary slides and an updated project fact-sheet.
The next task ‘Scientific dissemination’ was led by the RVC and involved the rest of the team. It consisted of the sub-task ‘Literature warehouse’: a Mendeley group had been created and had 12 members from different Consortium partners. 825 references were stored in the online database. The next sub-task ‘Scientific publications’, involved all partners and consisted of RISKSUR outputs being presented at several conferences and manuscripts being published in peer-reviewed journals. A full list can be found in section 4.2 of this report. The next task consisted of several ‘Research briefs’ and involved the entire RISKSUR Consortium. Topics covered were: mapping of surveillance systems; animal populations; trade flows; critical infrastructure and decision making processes in seven European countries.
Furthermore, a ‘Symposium’ was organised in March 2015, prior to the Annual Meeting of the Society for Veterinary Epidemiology and Preventive Medicine (SVEPM), at Het Pand in Ghent, Belgium. The theme was ‘Animal Health Surveillance 2.0’ and corresponded to Deliverable D7.29. Fifty participants from 17 different countries attended this event. Scientific exchange was promoted through the establishment of formal links and exchange of information on activities through different international projects and institutions (http://www.fp7-risksur.eu/related-networks).
Another task ‘Best practice recommendations’, led by Safoso and including all other partners besides TT and RVC was subdivided into two sub-tasks. First a ‘Best practice workshop’, Deliverable D7.31 which was held in September 2014 in the Netherlands. The goal and the next task was to subsequently develop best practice guidelines for animal health surveillance tailored to user needs. During the morning, participants reviewed existing standards and discussed its Strengths, Weaknesses, Opportunities and Threats (SWOT analysis). In the afternoon, there was a presentation about the ‘Surveillance landscape in Europe’ (a presentation of the mapping report conducted under the (RISKSUR WP1 work), and this was followed by the presentation and discussion of different surveillance systems being currently used in Europe (e.g. UK, NL). The document ‘Best practice guidelines’, Deliverable D7.32 was coordinated by FAO and Safoso and included contributions from several other consortium members. The original ideas about its contents and structure were developed at the ‘Best Practice Workshop’, and lead to a document that links all the work developed during RISKSUR, providing real life animal health example problems. It covers all steps involved in surveillance system design and implementation, divided into ten chapters as follows: Introduction; Terminology; Prioritisation; Planning; Risk-based design; Implementation; Evaluation and performance monitoring; Economic evaluation; From evaluation to strengthening surveillance and a Conclusion chapter. For the Dissemination of best practice, consortium members were encouraged to disseminate the ‘Best practice guidelines’ among the veterinary services of their countries.
Furthermore, for the Support service for Common Agricultural Policy (CAP) and Common Animal Health Policy (CAHP): Surveillance surgery led by Safoso and involving ACCEL, GD, SVA and RVC, 6 ‘Surveillance Surgery Webinars’ were delivered over the course of the project. They were open to scientists, community and MS policy makers as well as industry. Participation from accession countries and third countries were also encouraged to attend. The discussions were moderated by GD with support from Safoso. Topics included: African swine fever (ASF) surveillance; Looking at society from an animal health surveillance perspective; Surveillance of antimicrobial resistance; Surveillance of Avian Influenza; RISKSUR Surveillance Design Framework; RISKSUR Surveillance Evaluation Framework.
Potential Impact:
After this three-year project now having been completed, policy makers and other stakeholders have access to a set of tools that can be used to inform decision making on the design and evaluation of cost-effective animal health surveillance systems addressing the surveillance objectives ‘early detection of disease’, ‘demonstration of freedom from disease’, ‘determination of disease frequency’ as well as multi-hazard surveillance. The criteria for surveillance design apart from epidemiological considerations included the wider systems context and economic factors. A key epidemiological feature is the explicit inclusion of risk-based approaches to surveillance amongst the possible surveillance methodologies and the tools therefore provide a much-needed scientific evidence base for application of such approaches (often referred to as risk-based surveillance). Also, the tools are supported by appropriate documentation. Associated learning materials have been made available in the public domain to allow widespread access. To ensure that the approaches are based on sound scientific principles, the Consortium produced peer-reviewed publications in relevant scientific journals in the relation to the outputs produced by the project, a process which started off with some the literature reviews and was followed up by further publications
The main dissemination activities and exploitation of results were:
A key feature of the RISKSUR project was the dissemination of information to different stakeholders in an easily accessible format using various communication methods and tools. The coordinator RVC managed the dissemination/exploitation of the scientific and technological results at the project level, and WP6 leader Safoso was responsible for the execution of the actual dissemination/exploitation activities of individual partners. Newly acquired knowledge was continuously being evaluated for protection, dissemination and exploitation of intellectual property rights by each partner.
The main objectives of all dissemination activities were the communication and spread of the project results to the following stakeholders and the promotion of commercial exploitation:
- National government veterinary services: RISKSUR provided recommendation solutions for national governments in the context of specific disease control activities (e.g. foot-and-mouth control in the UK). From the start of the project CVOs did show their interest in using the project results.
- Supra-national government organizations: RISKSUR offers advice to supra-national organizations such as EFSA, FAO and World Bank. The outputs from RISKSUR will be relevant for all these organisations.
- Pharmaceutical and diagnostic industry: These companies are increasingly looking for integrated solutions which include the related surveillance activities. RISKSUR output will inform this stakeholder group.
- Livestock industry: RISKSUR results will be very relevant for this group to assure their efficiency and competitiveness in an international market.
- Scientific community: The results of WPs 1-4 will be of particular interests to scientists in this research area. For this reason, exchange of knowledge and interactions with other projects in this field will be important to RISKSUR researchers.
- General public: Science in society is important to RISKSUR since an informed society can help prevent the transmission of diseases. It is therefore crucial to raise awareness for novel surveillance systems and their potential benefits to society.
All activities targeted to these different stakeholders have been planned in WP7, under the lead of Safoso. Key to the exploitation was a transparent and fair communication and management of both existing information and project results. For this reason, RISKSUR partners had agreed on the following Intellectual Property issues:
All information provided by a partner to other partners within the project was and still is confidential unless:
* It was already known to the partner before the negotiations started, or
* The information provided is public property, or
* It is explicitly specified otherwise by the originator of the information.
Partners had also agreed to use the information provided only for the purposes of conducting the project. Any disclosure of confidential information to a third party requires the explicit consent of the originator of that information. The SME partners Safoso and TT were granted sole ownership of the services (e.g. training courses) and products (e.g. software application) developed in this project. When more than one partner claimed joint ownership of newly produced intellectual property other than those mentioned in point 3 before, the partners involved make provisions to clarify the terms of joint ownership among them. Partners were not restricted in any sense regarding the rights associated with the ownership of any intellectual
property they produce while conducting the project activities.
Impact of dissemination activities
Dissemination activities targeted to the general public and industry:
Web site: ACCEL established and maintained a project website (www.fp7-risksur.eu) with sections for the general public, for surveillance professionals, policy makers, industry and researchers. Relevant news from other sources such as DGSANCO or scientific publications in the field were integrated. Project partners used the restricted domain of the website to share information and work on jointly authored documents. The website was designed such that it can be sustained beyond the lifespan of the project. It also included all available training materials, recordings of webinars, copies of public deliverables and a Glossary.
Electronic newsletter: To ensure that the level of interest in the project among stakeholders did not decline, the pull communication (website) with this permission-based push communication on RISKSUR generated knowledge was combined. The target group of the newsletter was non-scientific, mainly policy makers and industry. ACCEL and RVC jointly elaborated and disseminated this bi-annual newsletter with a target audience of about 5,000 subscribers. Other institutions contributed to the content of the newsletter. Overall 6 newsletters were sent out, a final 7th is to be sent out in January 2016.
Science to society and to professionals: Under the coordination of ACCEL, RISKSUR used current electronic media such as LinkedIn to disseminate information on animal disease and risk surveillance to the general public. The aim was to especially target rural communities, livestock dealers and other associated industries.
Impact: The aforementioned efforts contributed towards increasing the effectiveness of surveillance simply by increasing stakeholder awareness, and generating improved acceptance and trust.
Dissemination targeted to the scientific community:
Scientific publications: All participants published their scientific findings in leading international and national journals in the fields of veterinary epidemiology, disease control, infectious disease, and livestock production. Papers in the national language will reach audiences not familiar with English. By the end of the project in October 2015, 5 papers had already been published in high-impact journals. A minimum of a further 5 were in preparation and to be submitted in January 2016. Additionally, RISKUSR results were presented at national and international scientific meetings such as SVEPM 2013-2015, ICAHS 2014 and ISVEE 2015.
Symposium: A Symposium was organised on 24 March 2015, prior to the Annual Meeting of the Society for Veterinary Epidemiology and Preventive Medicine (SVEPM), at Het Pand in Ghent, Belgium. The theme was ‘Animal Health Surveillance 2.0’. Fifty participants from 17 different countries attended this event. It aimed to disseminate research results, to promote the dialogue between scientists and surveillance users and to facilitate networking within the surveillance community.
Scientific exchange: Exchange with other existing networks was promoted by establishing formal links and by joint meeting and exchange of information on activities. This activity was coordinated by Safoso. EU-funded research projects and networks with a link to RISKSUR were, for example, LINKTADS, MARLON; EPIZONE, DIPNET, MED_REO_NET, PANDA, EVENT, EMPERIE.
Impact: Based on these targeted dissemination activities, RISKSUR communicated the outputs to the science community, thereby provided opportunity for them to be scrutinized so as to ensure a high level of scientific validity. At the same time, it resulted in dissemination of the tools amongst the scientific community for further development, as well as widespread adoption of the underlying methodologies which in turn did also benefit the EU and its member states by allowing safer and more transparent trade and safeguarding animal and human health.
Dissemination targeted to policy-makers:
Best practice workshop: The Best Practice Workshop was held on September 30, 2014, at the Ministerie van Economische Zaken, The Hague, Netherlands. The goal was to subsequently develop best practice guidelines for animal health surveillance tailored to user needs. During the morning, participants reviewed existing standards and discussed its Strengths, Weaknesses, Opportunities and Threats (SWOT analysis). In the afternoon, there was a presentation about the ‘Surveillance landscape in Europe’ (a presentation of the mapping report conducted under the RISKSUR WP1 work), and this was followed by the presentation and discussion of different surveillance systems being currently used in Europe (e.g. UK, NL). Participants had a mixed background in surveillance research and practice as well as policy.
Best practice guidelines: FAO and Safoso coordinated the development of this document, that included contributions from several other consortium members. The original ideas about its contents and structure started building up at the Best Practice Workshop, and lead to a document that links all the work developed during RISKSUR, providing examples with real life animal health problems. It covered all steps involved in surveillance system design and implementation, divided into ten chapters as follows:
1. Introduction
2. Terminology
3. Prioritisation
4. Planning
5. Risk-based design
6. Implementation
7. Evaluation and performance monitoring
8. Economic evaluation
9. From evaluation to strengthening surveillance
10. Conclusion
Dissemination of best practice:
Consortium members were encouraged to disseminate the ‘Best practice guidelines’ among the veterinary services of their countries.
Support service for Common Agricultural Policy (CAP) and Common Animal Health Policy (CAHP): Safoso initiated and moderated surveillance of issues via a widely available web-tool Adobe Connect. Overall 6 so-called ‘Surveillance Surgery Webinars’ occurred over the course of the project. They were open to scientists, community and MS policy makers as well as industry. Participation from accession countries and third countries were also encouraged to attend. The discussions were moderated by GD with support from Safoso. Topics included: African swine fever (ASF) surveillance, Looking at society from an animal health surveillance perspective, Surveillance of antimicrobial resistance, Surveillance of Avian Influenza, RISKSUR Surveillance
Design Framework, RISKSUR Surveillance Evaluation Framework.
Impact: Design of surveillance systems is an extremely complex issue with a variety of factors that influence the decision making process. RISKSUR provided a transparent, validated and user-friendly web-based decision support tool that allows policy makers to consider all possible options of system configurations and then make an informed choice, taking into account cost-effectiveness. There is currently no other such system available, and decision-makers usually still work on the basis of methods, which have not much changed over the last 100+ years. RISKSUR, therefore, for the first time provides, an objective mechanism for justifying and rationalizing the usage and combination of traditional as well as novel surveillance approaches. This also contributes towards enhanced equivalence in approaches and outputs generated to inform trade decisions.
Dissemination to practitioners:
Training material and online case studies: RVC in collaboration with Consortium members developed training materials, each based on a work package, in the form of short online lectures. All 13 lectures were made available via the RISKSUR web site. Communication channels as described above will be used to promote training material.
Impact: RISKSUR will transfer knowledge to technical experts within the EU and elsewhere on the general concepts of risk-based surveillance and the outputs of the project. This resulted in enhanced surveillance capacity in the EU and elsewhere, and also facilitated communication of the approaches used.
Impact of exploitation activities:
Three of the four SMEs involved had a strong commercial interest in the expected results to be developed in RISKSUR. In order to facilitate the exploitation of results, partners had already agreed on the following two principles that will govern the IP terms in this project, and form the basis of the IP terms in the Consortium Agreement:
Solely generated IP will be solely owned by the generating party with a first option to the project and associated partners to a non-exclusive option to license foreground IP for commercialisation.
All intellectual property generated jointly during the course of the Project (‘Joint Foreground IP’) will be jointly owned by the generating parties, and shall be managed in accordance with the terms of a Joint Ownership and Management Agreement (JOMA).
RVC in its role as the Coordinator was responsible for ensuring that a secure and suitable knowledge management system is put in place, which did run as soon as possible after the project has started and that data protection legislation is followed. In the case of conflict of interests, ACCEL assisted and moderated between the commercial partners, with no interest in the commercially exploitable results. This IP management strategy allowed especially our SME partners to fully exploit the potential of RISKSUR. Exploitation will be actively supported by the project’s Advisory Boards. Decision on the use of project results will be made by the Steering Committee according to the decision-making process.
Roadmap for cost-effective surveillance of disease beyond RISKSUR:
RISKSUR had the first development of an integrated decision support tool for design and evaluation of animal health surveillance systems. It was underpinned by sound science-based evaluation and decision support frameworks, which explicitly use cost-effectiveness as one of the outputs. It was important to recognise that the tool needed to be flexible and required review and potentially updating on a regular basis. Assuming that it was able to demonstrate its utility to end users, they were motivated to perform and fund those updates. It is also conceivable that new approaches of delivering the functionality of the tool will become available, and it then needs to be revised accordingly.
National CVOs and other organisations were committed to apply and support the use of the project results and thereby both ensure the sustainability of the project and facilitate more cost-effective surveillance in the future. Additionally, FAO’s wide network helped pave the way for future roads in surveillance of animal diseases.
The frameworks and methodologies developed as part of RISKSUR will thus continue to evolve, but be able to do so more effectively and in a balanced fashion. It is also likely that new surveillance methods will become available, but the frameworks will be sufficiently flexible to take such developments into account.
List of Websites:
Address of project public website and relevant contact details
RISKSUR website: www.fp7-risksur.eu
Contact:
Professor Dirk Pfeiffer
Royal Veterinary College, UK.
Email: pfeiffer@rvc.ac.uk
The risk of occurrence of exotic, new (emerging), re-emerging as well as endemic diseases has increased substantially particularly due to changes in consumption of animal source foods and the related intensification of animal production and increase in global trade of animals and their products. Bovine spongiform encephalopathy (BSE), foot and mouth disease (FMD) and avian influenza, as well as human severe acute respiratory syndrome (SARS), influenza A virus subtype H1N1 and enterohaemorrhagic Escherichia coli (EHEC) epidemics demonstrated the immense adverse socio-economic impact of animal and human disease. While the need for effective animal health surveillance is widely recognised for the management of such threats, the currently used systems do not take optimal advantage of recent advances in epidemiological approaches. In addition, investment is being constrained due to significant financial budget reductions in most countries. This can result in a potentially risky situation that demands the implementation of a new approach to surveillance system design.
The overall aim of RISKSUR was to develop decision support tools for the design of cost-effective risk-based surveillance systems that integrate the most recent advances in epidemiological methodologies.
This was achieved by the development and application of design and evaluation frameworks for animal health surveillance systems for three different surveillance objectives: 1) Early detection of animal disease, 2) Demonstration of freedom from animal disease, and 3) Determination of disease frequency and detection of cases of endemic animal disease. Tools for comparative epidemiological and economic evaluation of different surveillance system designs were developed and made available through a web-based decision support tool. The project provided decision-makers with a validated decision support tool tailored to their needs that allows the design of more cost-effective animal health surveillance and thereby reduces direct and indirect impact of animal disease on European Union citizens.
RISKSUR partners represent a unique combination of internationally recognised expertise in animal disease surveillance methodologies and economic evaluation as well as applied experience in the delivery of surveillance programmes in a variety of socio-economic contexts as well as from a national and global perspective. The practical implementation of the tools and the transfer of the knowledge was led by four small and medium enterprises (SMEs) with a strong track record in early adoption and translation of research results.
Project Context and Objectives:
‘Prevention is better than cure’ was formally adopted as the strategic direction of the new EU animal health policy in September 2007. The aim of the new strategy is to reduce the likelihood of animal disease, and to minimize the impact when outbreaks occur. The increase in the occurrence of new (emerging) diseases and the spread of exotic ones along with increasing levels of endemic and food-borne diseases require earlier, better and safer surveillance to achieve this objective.
Furthermore, it is recognised that improved protection of human health will require more effective linking of animal and human health surveillance systems as well as adoption of interdisciplinary approaches, which is one of the aspects of the One Health concept. However, financial constraints of governments in times of public spending cuts have reduced or limited investments in surveillance of animal disease.
In addition, cross-compliance and equivalence issues constrain governments in relation to surveillance strategy design. Efficient and practical solutions taking advantage of novel scientific methodologies, such as risk-based surveillance, are urgently needed, as well as integration of epidemiological approaches with socio-economic and qualitative methods where appropriate. At the same time, governments are required to reduce their expenditure in response to growing fiscal constraints. With respect to animal health surveillance systems, very significant advances have been made in laboratory diagnostic tools whereas the implementation of ‘smart’ methods for collection and analysis of surveillance data has been neglected in the past decades. This is despite of much improved understanding of factors influencing animal disease risks which should provide an
opportunity to design more efficient surveillance systems. In addition there has been a lack of data collection and information driven work in relation to improving our understanding of the changing dynamics of the livestock sector.
This will be of utmost importance as especially the past twenty years have seen the EU experiencing several animal health crises. These crises have had dramatic adverse effects on the livestock sector and public health, and resulted in significant disruptions to markets and the wider economy. Several factors have compounded the risk of such crises – globalisation and the resulting increase in trade, the expansion of EU borders to the East, and changes in the structure of the EU livestock sector.
Recent outbreaks of exotic diseases such as avian influenza (AI), foot and mouth disease (FMD) and bluetongue (BT) in previously unaffected regions of the EU have highlighted the threat posed by emergence and / or spread of infectious agents, and emphasise the need for well-developed and adequately resourced animal health systems, to ensure early detection and rapid containment.
Over the period 1998-2009 outbreaks of avian influenza mainly affected Germany, Italy, and the Netherlands; FMD caused a severe crisis in the UK in 2001 and re-emerged in January 2011 in Bulgaria; several bluetongue outbreaks occurred unexpectedly north of the Alps in countries such as France, Italy, Netherlands and Germany since 2006. Furthermore, there were cases of Schmallenberg (SBV) detected in the Netherlands and Germany between August and October 2011.
The economic implications of animal diseases and animal disease control are severe. In 2008, the cumulative global cost of the H5N1 avian influenza pandemic was estimated to have been at least US$ 25-30 billion (Newcomb, 2008). Given that the epidemic has not been controlled in several countries, the cost will now be much higher than that figure. Another example is the human EHEC outbreak in Germany which while relatively small resulted in significant negative economic and human health impacts and prompted renewed demands for more effective surveillance systems (Turner, 2011).
Evaluation implies a judgment on the surveillance system and recommendations for improvement. This is done in reference to various aspects of surveillance, specifically whether it allows transparent interpretation of outputs, more objective decision making and resource allocation as well as improvements in system design and enhanced acceptance of system outputs by stakeholders at local (e.g. farmers, veterinarians) and national levels (e.g. reference laboratory, veterinarians at central level). This is particularly important given the knowledge gaps in understanding of many diseases, which leads to varying degrees of uncertainty and bias in generated outputs.
The use of economic evaluation in decision making for animal health surveillance policies has been limited so far. When considering the start, end or change of a surveillance programme, policy makers need to know if and how much surveillance is needed and what the most economical options are and how it best integrates with interventions. There are few specific guidelines available for economic evaluation of animal health surveillance, and only a limited number of empirical applications on the economic value of surveillance. It is critical to select appropriate economic efficiency criteria taking into account the evaluation context and viewpoint of the analysis. Priority setting, affordability, sustainability, social acceptance and communication are all issues that policy makers have to consider when designing and implementing disease management policy. These assessments are not easy as the data available to quantify investments in surveillance and monitoring activities are very limited at both national and international levels. It is expected that member states would value access to a tool that would help them with technical and budgetary optimization of resource utilization when defining animal surveillance policy to as part of the national animal health policy.
For these reasons, a need for frameworks and integrated tools that allow the design and implementation of epidemiological and economically optimised animal health surveillance systems that generate outputs in relation to the animal health status of the livestock sector, allowing effective management of any emerging or existing risks, was identified.
The overall aim of RISKSUR was to develop and validate conceptual and decision support frameworks and associated tools for designing efficient risk-based animal health surveillance systems based on an interdisciplinary approach tailored to the needs of EU Member States. The project aimed to develop tools and frameworks targeted at the following surveillance objectives associated with livestock diseases:
1. Detection of incursion of exotic, new (emerging) and re-emerging disease
2. Declaration of freedom from specified diseases and infections
3. Monitoring of endemic diseases (case detection, disease frequency estimation)
The RISKSUR project addressed these three surveillance objectives in different work packages (WP). The first (WP1) develop a conceptual evaluation and decision-making framework for risk-based animal health surveillance system designs based on work to identify decision-makers involved in the resource allocation for surveillance; described decision-making criteria used; characterised surveillance components in place, economic criteria and associated data flows; and identified evaluation attributes. The work in WP1 informed further activities for each of the aforementioned three surveillance objectives in WP2-WP4. Work package 2 developed a framework to support the design of surveillance systems and associated tools aimed at early detection of animal diseases (including exotic, new and re-emerging diseases) by conducting a systematic review; identifying and defining key risk scenarios; describing data requirements, collection and analyses tools, and by applying the framework to case study disease. Work packages 3 and 4 used a similar approach focusing on the development of a decision support framework for surveillance systems and associated tools aimed at demonstrating freedom from diseases and estimating disease frequency and detection of cases of disease, as well as multi-hazard surveillance systems, respectively. As some of the data and methods were used in several WPs, there was extensive collaboration and exchange between partners. The results of WPs 2-4 were evaluated for single and multi-hazard surveillance systems specifically in relation to their efficiency in WP5. The objective of WP5 assessed the technical and economic efficiency of the surveillance designs developed in RISKSUR by developing and applying an evaluation support tool for the integrated epidemiological and economic evaluation of surveillance. The transfer of knowledge and technology to key stakeholders from policy and industry was facilitated through the development of practical and accessible web-based decision support tool enabling decision makers, industry and the broader scientific community to take advantage of the research outcomes in WP6. In WP7, dissemination of new methods and novel applications for surveillance design by sharing information with different shareholder groups, and by providing training to policy makers and surveillance workers, was performed. WP8 covered the project coordination and management.
In summary, the RISKSUR consortium used an interdisciplinary approach to develop novel decision support frameworks and tools integrating novel epidemiological, economic and qualitative research methods for design of efficient risk-based animal health surveillance systems. This will allow evidence-based decision-making by key stakeholders in EU member state veterinary administrations.
References:
Newcomb, J. 2008: Economic Opportunities and Benefits of Implementing One World-One Health Strategies for Health Promotion and Disease Prevention. Bio-Era Associates Report.
Turner, M. 2011: Microbe outbreak panics Europe Nature 474, 137| doi:10.1038/474137a.
Project Results:
LIST of abbreviations for partner institutions
ACCEL- ACCELOPMENT AG
APHA- Animal and Plant Health Agency
CIRAD- Centre de Coopération Internationale en Recherche Agronomique pour le Développement
FAO- Food and Agriculture Organisation of the United Nations
FLI- Friedrich-Loeffler-Institut
GD- Gezondheidsdienst Voor Dieren
RVC- The Royal Veterinary College
Safoso- Safe Food Solutions AG
SVA- Statens Veterinaermediciniska Anstalt
TT- Tracetracker
UCM – Universidad Complutense de Madrid
The RISKSUR project has achieved all Scientific and Technological target objectives, set at the beginning of the project. All 36 deliverables were submitted over the course of the project, between November 2012 and October 2015.
1. Work package 1 - Development of a conceptual evaluation framework
This work package generated the report: ‘Mapping of the main surveillance systems and decision-makers.’ The objective related to Task 1.1 and its associated Deliverable D1.1 was to produce a high level overview of the populations, surveillance systems and decision-making processes currently used in the EU. Primary and secondary data were collected and collated in seven RISKSUR partner countries (France, Germany, Great Britain, the Netherlands, Spain, Sweden and Switzerland) and six non-partner countries (Belgium, Bulgaria, Czech Republic, Denmark, Ireland and Italy). Members from the RISKSUR Consortium countries screened web pages, government reports, scientific and other grey literature and approached relevant contacts with the aim to describe all existing private and public surveillance system components (SSC) in the year 2011 in all animal species of economic value. Both disease-specific as well as health-event related surveillance was considered. Further, for the seven partner countries, data were collated on critical infrastructure and interviews were conducted with 35 decision-makers involved in private and public surveillance. For all EU member countries data on livestock and bee holdings in Europe, human and animal populations, and gross domestic product and farm values were collated from Eurostat. Data on trade was obtained from DG SANCO’S Trade Control and Expert System (TRACES). These data were used to create maps to visualise distributions and differences in human and animal populations and holding densities as well as animal movements across Europe. All data were analysed descriptively and presented in D1.1 report for the seven countries. The data and analysis for the additional 6 countries were presented in a supplement to the D1.1. The research brought together key system attributes of the selected EU countries allowing a comparison between countries and systems, as well as appreciating the impact of the linkages between systems and countries and their consequences for risk management. It was a collaborative effort under RVC lead involving nearly all Consortium partners.
Next within WP1 a systematic review of evaluation methods of surveillance systems and current practices was carried out, submitted as Deliverable D1.2 report. The work was divided into two independent subtasks: a systematic review of economic assessments of surveillance systems, criteria and methods, and a systematic review of the guidelines, framework, methods and tools for the evaluation of surveillance systems. The former led by RVC and APHA, produced an overview of the approaches and techniques available to perform economic assessments of surveillance. Therefore, a systematic literature review of economic evaluations of animal health surveillance systems published in the last 20 years was performed. Most economic evaluations were found to focus on surveillance systems targeting single diseases, even though an economic evaluation of surveillance would ideally address different diseases and cover multiple geographical regions. This review also emphasised the lack of standardised approaches for measuring the effectiveness of surveillance and for conducting cost-effectiveness analyses. The relatively small number of papers published in this area highlighted that economic evaluation of surveillance is rare in animal health. The review concluded by discussing the potential relevance of economic approaches that are still underutilised for evaluating surveillance. The ‘review of surveillance evaluation frameworks and practical decision-making factors’, carried out by APHA, CIRAD and RVC identifies the main advantages and limitations of the existing guides used for the evaluation of surveillance systems, either in public health, animal health or environmental health. It highlights potential ways of improving the evaluation process in order to provide useful recommendations for the development of the RISKSUR decision support tool. Most guides identified were general and provided recommendations for broad evaluations. Several common steps in the evaluation process were identified, namely (i) Defining the surveillance system under evaluation, (ii) Designing the evaluation process, (iii) Implementing the evaluation, and (iv) Drawing conclusions and recommendations. The analysis highlighted a lack of information regarding the identification and selection of methods and tools to assess the evaluation attributes as well as a lack of consideration of economic attributes and sociological aspects. This review emphasized the need to develop a decision support tool for economic evaluation of surveillance systems, which integrates and complements the existing frameworks, methods and tools.
Deliverable 1.3: ‘Prioritisation and relative weighting of surveillance characteristics and evaluation attributes as well as factors relevant to the selection of evaluation criteria’, was led by RVC together with APHA and CIRAD. As a first step towards the development of the RISKSUR evaluation tool, this task and its associated deliverable D1.3 resulted in a better understanding of the surveillance attributes used for the evaluation and the complex inter-relations between them, to identify potential gaps in evaluation attributes and to suggest new approaches for optimising the use of attributes in the evaluation of the effectiveness of the surveillance. This task was divided into 5 Subtasks. The results highlighted the need to work towards a theoretical and unified approach for evaluating effectiveness of surveillance, which is then further investigated in one of the subtasks. Moreover, the findings presented in the report cover many aspects of evaluation attributes from agreeing on definitions to the development of a new measure of effectiveness.
The deliverable D1.4: ‘Development of an integrated evaluation matrix’, involving CIRAD, RVC, APHA, ARCADIA and other Consortium members, integrated outputs from previous tasks, and included the development of an evaluation matrix that will help to identify the most suitable epidemiological and economic approaches, criteria and techniques. After defining the context, the RISKSUR Consortium with CIRAD’s lead developed a conceptual model of a decision support tool for the integrated epidemiological and economic evaluation of animal health surveillance system (EVA tool). Relevant information from previous tasks was used to build on existing evaluation frameworks, methods and tools and thereby make best use of already available approaches. The EVA tool provides comprehensive and practical guidance to decision-makers and their technical advisers to plan and conduct evaluations of animal health surveillance systems. It guides the user through several steps and options on what, why and how to measure in order to evaluate animal health surveillance systems. The users have to provide inputs in relation to the evaluation question and the general context (e.g. epidemiological situation, surveillance objectives, and data availability) to describe the evaluation context and select a suitable evaluation question. The tool also helps to select suitable evaluation attributes and corresponding measurement methods. Moreover, it provides detailed guidance on the application of the methods and the interpretation of the results.
The last task and its associated deliverable D1.5 generated in this work package: ‘Define integrated data collection protocols and mechanisms to ensure robust data quality’. FLI lead this task, on which RVC, APHA, SVA, UCM, CIRAD collaborated. Its objectives were:
a) To describe the final design adapted for the mapping and review of surveillance systems,
b) To document the results of consistency checks as well as revised definitions and
c) To discuss the current limitations of information describing surveillance.
The D1.5 report presents guidelines to address these areas and the agreed additional data requirements for the review of surveillance systems. A standardised protocol was developed for the review to standardize data collection amongst project partners, since questionnaire results had highlighted the variability in data completeness and interpretation. The application of consistent terminology is crucial to ensure robust data quality and facilitate comparisons between surveillance systems. Efforts to standardize terminology for animal health surveillance at the International Conference on Animal Health Surveillance (ICAHS) conference in 2011 were used as a basis for the mapping and review task. The data collection served as a validation of these terms across a range of surveillance systems and countries. Some problem areas were identified, and as a result terminology was further developed. Particularly for risk-based surveillance the terminology was not adequate to reflect the diversity of components that could be termed risk-based but vary in their nature and intensity. Hence, it was recommended to further refine the terminology related to risk-based surveillance. The mapping and surveillance system review did provide many examples that can be used for illustration. Early input from external experts was obtained by encouraging general feedback through establishing a glossary on the RISKSUR website. Another area identified with potential for improvement was that surveillance activities were often not well documented, which made it difficult to capture all existing surveillance activities despite considerable efforts made. In addition, regional variation within countries with a decentralized administration was difficult to capture. Registering surveillance activities and their design would allow providing an overview and better coordination of efforts made by the public and private sector.
2. Work package 2 - Detection of exotic, new, or re-emerging disease
The report ‘Review of surveillance systems and methods for early detection of disease’, carried out by UCM, was the first deliverable D2.6 completed during the project period. This report consists of a compilation of information regarding surveillance systems and strategies aimed at the early detection of new, exotic and re-emerging diseases. It looks at different epidemiological methods used in surveillance systems for early detection of diseases, including current methods such as passive notification of suspect cases and active strategies such as sentinel surveillance systems.
The next task and its associated deliverable report D2.7 entitled ‘Definition of key risk scenarios using selected case study diseases’ involved several partners from WPs 2-4. The corresponding tasks in these WPs were grouped together as the ‘Case study selection’ Theme. Consortium members contributing to this theme worked in collaboration with CIRAD, in order to establish a consistent process for selecting case study hazards and scenarios. The selection process was based on three steps: 1) reducing the list of hazards identified as important by both the EU and the Consortium members after assessing presence/absence and initial data availability; 2) selecting a shortlist of hazards based on transmission, surveillance, detailed analysis of data, species representation, selection in other WPs and opinion of the Advisory Boards; and 3) selecting case study countries for selected hazards to obtain the final case studies for WP2 based on data on historical occurrence, host and production types, environmental characteristics and surveillance results. The Advisory Boards supported the arguments that led to the final selection of case studies. For WP2 the following case studies were chosen: avian influenza in the UK; African swine fever in an Eastern European country, possibly Poland; and bluetongue in Spain. The routes of introduction and contact interfaces with wildlife were represented in pathway diagrams and the potential source of data needed was indicated. The selection of the case
studies and the description of pathway diagrams were described in detail in deliverable report D2.7.
This was then followed by the task ‘Development of a framework to inform the design of surveillance systems’, D2.8. Partners involved in WPs 2-4 worked together on this task. The surveillance design framework, available publicly through the wiki link (http://surveillance-design-framework.wikispaces.com/) comprises 14 steps, namely, 1) description of the surveillance design; 2) overview of surveillance components; 3) description of target population; 4) steps that lead to disease suspicion; 5) enhancements; 6) testing protocol in place; 7) study design for active surveillance components; 8) sampling strategy; 9) questions on data generation process; 10) document transfer mechanisms; 11) document data translation process; 12) document epidemiological data and analysis; 13) document how will surveillance be disseminated; 14) last questions on how the review will be performed. The next step in the framework was to evaluate the surveillance system, for which the EVA-tool (developed and applied under WP5) is to be used. Activities from the WP2-specific portion of the framework development ensure that the surveillance for early detection of known hazards is well covered in the final framework. The results from this task are presented in the Deliverable 2.8 report and also influenced the writing of the early detection chapter of the ‘Best Practice Guidelines’.
Finally, this WP and its associated deliverable D2.9: ‘Tools for risk assessment and evaluation’, involved the RISKSUR partners UCM, APHA, CIRAD and GD. The framework was applied to two risk scenario diseases for early detection surveillance: ‘avian influenza in the UK ‘and ‘African swine fever in Poland’. The flexibility and robustness of the framework and tools developed were evaluated. The data required to characterize the population at risk, quantify the pathways and assess the probability of detection were also analysed. New surveillance systems are proposed and designed based on the review of surveillance options, development of data collection and analysis tools. The epidemiological performance of the surveillance systems was evaluated by comparing risk-based against traditional surveillance systems with the new tools developed. The model used to test and analyse the epidemiological performance of the new component for avian influenza passive surveillance in the UK proved that outbreak detection was quicker for the risk-based scenario than the current passive surveillance scheme. For African swine fever we performed a qualitative risk assessment based on the probability to detect field infection in pigs and wild boar and proposed a new risk based surveillance component design based not only on the situation of the bordering countries, but particularly on the current risk of spread within Poland using spatio-temporal assessment tools that consider the distribution of the population at risk and the direction of the outbreaks. The work from this task was presented in Deliverable report D2.9.
The main conclusions are that the framework is very useful for documenting and for training in surveillance, as well as for being able to identify strengths and weaknesses of a surveillance system. However, other approaches to presenting surveillance components are also possible. The separation of surveillance objectives for each surveillance component was not always useful, since often many components together have the objective of identifying or ruling out potential suspects, which is one of the main aims of early detection surveillance, while if considered separately and on a purely theoretical basis, the surveillance objective can be interpreted together with another surveillance objective. Other potential areas of improvement include extending the documentation on suspicion of disease to include findings other than those resulting from passive surveillance; include a multi-hazard surveillance design for early detection; or to improve the standardisation of the final sections of the design so that the different levels and trainings in epidemiology across Europe can eventually be comparable. A very significant benefit of the design framework is that it offers a transparent way of documenting surveillance design that can improve communication between all actors involved in surveillance.
3. Work package 3 - Demonstration of freedom from disease
This work package first produced the ‘Review of surveillance systems and methods used to demonstrate disease freedom’, under the lead of FLI together with APHA, SVA, RVC, UCM. Task 3.1 was divided into two parts: a systematic literature review and a systematic review of surveillance systems. The Deliverable D3.10 report includes the former only. The review covers 131 articles and the deliverable report provides an overview of the development and interrelation of methods related to demonstrating freedom from disease including sample size calculations, scenario tree models, and simulation models. The results show to what extent important methodological aspects have been covered by the reviewed articles. In the discussion, proposals were made in terms of which well-established methodologies could potentially be further advanced or should be explored in further detail to make an adequate assessment. The outputs from this report were used as basis for the development of the design framework and associated tools. For the review of surveillance systems, an inventory of 515 surveillance components was created covering 26 hazards considered for case study selection and nine EU countries. The countries included the three partner countries producing the report (DE, ES, SE), four partner countries contributing to data collection (CH, FR, GB, NL) and two non-partner countries (DK, IT). The analysis considered those variables identified under Task 1.1 as well as 23 additional variables collected for the Task ‘Review of surveillance systems’. Since the additional information includes sensitive data such as the number of sampled herds and animals, results were summarized as an internal report only to inform project specific tasks such as the framework development and case study selection. The systematic review has provided crucial background information for the selection of tools (literature review) and the development of the design framework (literature and surveillance system review). The surveillance system review furthermore highlighted limitations in terminology and the way surveillance is documented.
Task 3.2 of this work package: ‘Development of a framework and associated tools for demonstrating freedom from disease’, was divided into 3 sub-tasks. First: Defining key risk scenarios and associated diseases relevant for freedom from disease, which was carried out by FLI together with APHA, RVC, SVA, Safoso and UCM. The following case study hazards were selected for WP3, based on a systematic selection process: a) classical swine fever; b) bluetongue disease and c) infectious bovine rhinotracheitis. Deliverable report D3.11 summarizes risk pathways for classical swine fever and the data requirements for all three selected case study hazards. Next: ‘Development of a framework for designing surveillance systems for demonstrating freedom from disease, with specific reference to choice of sampling methods’, carried out by FLI with APHA, SVA, RVC, UCM, Safoso and resulted in the associated deliverable report D3.12. Here scientists of the FLI participated in meetings and provided input to the development of the surveillance design framework, which was led by SVA. Furthermore, an inventory of tools to be evaluated under RISKSUR was proposed for WP3. Differences in the interpretation of terminology were identified as a problem during the WP1 ‘Mapping and review of surveillance systems’, as well as during the development of the surveillance design framework. Therefore, a terminology working group led by the FLI was initiated and two topics were covered (Topic 1: Means of data acquisition, Topic 2: Surveillance purpose). Questionnaires were used to obtain inputs from eight senior consortium members in three rounds before summarizing the conclusions. Also an ‘Evaluation of different epidemiological methods for demonstrating freedom from disease’ was carried out by FLI together with SVA, UCM, CIRAD and RVC. The performance of conventional and alternative surveillance strategies was compared in two case studies. The results of the first case study suggested that a deeper insight into the overall utility of a surveillance system could be gained when more than one effectiveness attribute was included in the evaluation. It was concluded that risk-based surveillance strategies should be considered when designing more efficient surveillance systems. This task resulted in Deliverable D3.13 completing the assessment of the framework and tools for certifying freedom from disease.
4. Work package 4 - Prevalence estimation and case detection for endemic disease
This work package first produced a ‘Review of literature on surveillance systems and associated methodologies from an epidemiological perspective for finding cases of endemic disease, including systems used in EU member states and other relevant countries and to generate an inventory of the extent of multi-hazard surveillance activities’. It was carried out by SVA together with FLI, APHA, UCM, CIRAD and GD and divided into two parts: a systematic literature review and a systematic review of surveillance systems. The systematic literature review for WP4 was carried out focussing on surveillance designs and associated methods aimed at finding cases of endemic disease and estimating prevalence. The initial literature search was coordinated between WPs 1-4. The systematic review covers, in all, 69 peer reviewed papers and constitutes D4.14. This deliverable was also further developed into a draft manuscript for publication in a peer-reviewed journal.
In addition to the systematic literature review, a review was conducted covering on-going surveillance in partner and non-partner countries (non RISKSUR Consortium members), using data obtained from the grey literature (national surveillance reports, personal communications etc.). For this review, an inventory of 515 surveillance components covering 26 hazards and nine EU countries has been compiled, in close collaboration with WP2 and WP3. The analysis considered those variables identified under Task 1.1 as well as 23 additional variables collected for the task ‘Review of surveillance systems’. Since the additional information includes sensitive data such as the number of sampled herds and animals, results were summarized as an internal report only to inform project specific tasks such as the framework development and case study selection. In the review part of the report, special attention is given by WP4 to the concept of multi-hazard surveillance. Such data were initially captured in Task 1.1 but were further refined and analysed in the internal report.
As indicated above the review of surveillance systems has had a certain overlap with Task 1.1 but has had a different focus and required additional data collection. The data collection was challenging as there were differences in how surveillance concepts are interpreted even among experts in the area. Therefore, significant effort was invested into retrospectively checking consistency of the data initially collected. Several suggestions for improvement and harmonisation of surveillance terminology were identified, and highlighted the importance of a generic terminology task.
A further task in this WP consisted of ‘Defining key risk scenarios and associated diseases relevant for surveillance aimed at estimating disease frequency and detection of cases of endemic disease’. This was carried out by SVA together with APHA, UCM, FLI, GD and as aforementioned, the work within this task was coordinated between WPs 2-4, under the theme ‘key risk scenarios’, initially lead by APHA and subsequently by GD. Potential diseases that are endemic in some or all EU member countries were identified, and their suitability as case study diseases has been assessed by considering disease characteristics, potential for improving surveillance using risk-based approaches as well as data availability and expertise among partners. In this process, WP4 contributed to the data collection on data availability and expertise for Sweden. The final selection of case study hazards was also informed by the characteristics of the livestock populations in partner countries as well as if the disease was indeed endemic in the countries considered. Opinions of stakeholders about the importance of the hazard in question, and the need for a review of the current surveillance were also taken into account. The hazard eventually selected for the case studies focusing on case finding capacity and disease frequency estimation was Salmonella Dublin in dairy cattle. For the case study on optimisation of multi-hazard surveillance, porcine reproductive & respiratory syndrome (swine) was chosen as the hazard for which the backbone surveillance activity is designed, with classical swine fever, swine vesicular disease and Aujeszky’s disease as hazards for which the same data collection process was utilised. Sweden and the Netherlands were used as case study countries. The process of selecting case study hazards, their nature and risk pathways as well as data needs are described in deliverable report D4.15.
Next followed the ‘Development of framework and associated methodologies to inform the design of surveillance systems’, led by SVA in collaboration with APHA, UCM, FLI, CIRAD and RVC. It included three subtasks, addressing the different surveillance perspectives covered by WP4; the maximisation of case detection capacity, the estimation of disease frequency and optimisation of multi-hazard surveillance systems. The structure and contents of the design framework were informed by previous tasks in RISKSUR; a literature review (D4.14) the mapping of surveillance systems (D1.1) and a review of existing surveillance systems in the EU (internal report). In particular, the latter highlighted the need for more transparent and standardised documentation of surveillance activities across the EU, and it was therefore concluded that documentation and reporting of surveillance design decisions should become an important feature of the surveillance design framework. The generic construction of the surveillance design framework was approached by decomposing surveillance activities into several attributes that represent the different steps in the planning of a surveillance system. These were subsequently reviewed to define and characterise the ‘building blocks’ of good surveillance design practice. The blocks were organised into a logical flow and grouped into broader surveillance design sections. This general framework was then refined to address the specific surveillance goals of early detection, disease freedom documentation, case detection and prevalence estimation. The framework was subsequently implemented as a questionnaire in a Microsoft Excel® spreadsheet using real data. At various stages of development the framework was tested and revised with the help of surveillance professionals from within the partner institutions. To support the use of the Excel tool, a wiki page has been developed and made publicly available at http://surveillance-design-framework.wikispaces.com/. The wiki content is based on the expertise of the RISKSUR consortium members developing the framework, and also incorporates feedback from the user group workshops (provided via a standardised questionnaire) as well as comments from the wiki members. It also includes a glossary and links to statistical tools (e.g. sample size calculators) that can be used during the process of design. The structure of the wiki is identical to the Excel® tool, making it easy to retrieve the advice needed at each design step. The development of the design framework was performed in close collaboration and coordination with WP5 activities, to avoid duplication and overlap with the development of the evaluation tool (EVA tool). This pertained both to the conceptual development and to the development of the web tools performed in WP6. The design part of the RISKSUR web tool is identical to the Excel tool for the higher levels of hierarchy, i.e. surveillance systems’ level and definition of surveillance components in place. However, only the Excel tool provides an opportunity to produce detailed documentation of different components, and to provide assistance in identifying design steps that could become targets for redesign after an evaluation. The decision to refrain from a full incorporation of the design framework in the web-based tool was taken at a time when the Excel design tool was in place and functional, while the development of the web-based EVA tool was under time and resource constraints. It was therefore decided to allocate the remaining WP6 resources to the completion of the EVA tool. However, the availability of the decision support in form of an Excel tool, in conjunction with the wiki page, has proven to be a robust and flexible format which is, in practice, useful for a broader target group than if it had been exclusively implemented in the online web tool. The framework is described in deliverable D4.16.
The development of the design framework required intensive coordination with WP5, to avoid duplication and overlap. This coordination was instrumental since the generic design framework developed by WPs 2-4 in collaboration and the evaluation tool developed in WP5 are complementary tools. For WP5 the focus was on evaluation of surveillance performance whereas the design framework aims to guide the user through the decisions leading up to an optimal design, taking into account the factors that would subsequently be of interest to evaluate. The final task consisted of the application of the framework to assess the epidemiological performance of surveillance systems aimed at case finding and disease frequency estimation for endemic diseases using different risk scenario diseases, and multi-hazard surveillance. This was carried out by SVA, RVC and GD. The surveillance design framework was subsequently applied to design, redesign and documentation of existing surveillance systems, aimed at assessing whether the structure of the tool was appropriate for the endemic disease surveillance objectives of case detection and prevalence estimation, as well as for multi-hazard surveillance. The case studies conducted in WP4 focused on Salmonella dublin in cattle in Sweden (case finding and prevalence estimation), and the multi-hazard surveillance covering porcine respiratory and reproductive syndrome (PRRS, mother component), Aujeszky’s disease (AD, child component) and classical swine fever (CSF, child component), also in Sweden.
The Salmonella dublin case study focused on alternative designs for a new surveillance component under development, which meant there was limited pre-existing data available. Instead, the case study was approached by adapting a data-driven modelling framework developed by SVA for verotoxigenic E. coli O157. The model simulates the spread of S. dublin in the Swedish cattle population during a 8.5-year period, using real animal movement records from the national cattle registry. Two different surveillance strategies were defined; i) a non-risk based scheme where all herds are sampled once per quarter and ii) a risk-based scheme where herds in high risk areas are sampled quarterly, and all other herds are sampled once per year, during the season with highest probability of detection. Output from the model has also been used as the basis for the economic evaluation conducted in WP5, in parallel with the estimation of the epidemiological performance. It was concluded that a conventional surveillance approach would indeed allow the detection of more infected herds in the short run, but a risk-based surveillance would achieve the same effectiveness over a longer time span, at a much lower cost. Deliverable D4.17 report describes the application of the surveillance design framework to the documentation of surveillance activities with objectives covered by WP4.
5. Work package 5 – Evaluation of epidemiological and economic effectiveness of surveillance systems
The first task carried out was the ‘Specification and further development of frameworks and methods for the evaluation of economic efficiency of surveillance’. It involved CIRAD, ARCADIA, RVC, APHA and FAO and included a review of the surveillance system evaluation frameworks, methods and tools. Also, the main challenges linked to the economic evaluation of animal health surveillance were addressed, notably how to promote economic evaluation, how to provide integrated guidance and how to report on the evaluation results. The outputs of this sub-task were reported in the evaluation Wikispace, in the Best Practice Guidelines document, and through the WP5 deliverables D5.18 D5.21 D5.22 and D5.23. Also, the EVA tool was developed based on the conceptual model produced within the first 18 months of the project. This consisted of the development of the evaluation tool structure and logic (CIRAD, RVC), an Online EVA Web Tool development (structure, logics and content), a Wikispace development (to provide guidance on evaluation concept and use of EVA tool) (structure and content), an evaluation question guidance pathway, attribute selection logics, assessment method selection logics and a review of the important points to transfer results of the evaluation to decision makers and protocol development (ARCADIA, CIRAD, RVC).
A final WP5 workshop was organised to review the status of the EVA tool development, to validate the logic of the tool and to review/finalise the on-going work on economic evaluation case studies. Specific action points and an updated work plan were identified to finalise the tool and the evaluation case study and prepare the final deliverables (D5.22 and D5.23). The next task in this work package consisted of a ‘Standardisation of data collection for economic impact assessment’. Partners were CIRAD as the lead, and ARCADIA, RVC, APHA and FAO. Different protocols for data collection for economic evaluation of animal health surveillance systems were developed and are briefly detailed below. Moreover, a work plan was developed to provide guidance for the evaluation of a surveillance system (CIRAD); this work plan was applied to the RISKSUR economic evaluation case studies (in collaboration with RVC, FLI, SVA and APHA). The full description of those protocols and the related questionnaires has been reported in Deliverable D5.19. This work plan presented all steps that the evaluator should take to perform the evaluation once the evaluation question and methods have been selected, from the validation of the evaluation protocol (evaluation question, selection of attributes and methods) to the reporting of the results of the evaluation to the decision makers.
The final task consisted of the ‘Selection of case studies and application of the evaluation matrix’. This task was led by CIRAD and involved all Consortium partners and resulted in the deliverable D5.20 ‘Case studies selected and described’. The case studies used within WP5 to develop, test and validate the tool by implementation of the economic evaluation are described in Figure 1. The task consisted of the evaluation of the economic efficiency of the novel designs developed in WPs 2-4 and validation of the frameworks and methods elaborated and involved CIRAD, ARCADIA, RVC, APHA and the FAO. The economic evaluation of the RISKSUR case studies was performed in the last 18 months of the project and reported in Deliverable D5.22.
6. Work package 6 – Decision-making tools for implementing risk-based surveillance
The first task of this project was to develop a statistical software package encapsulating functions for design of surveillance systems, in particular involving risk-based sampling. The development of the statistical software package continues to be in progress. It required Implementation of algorithms as an R package. This was led by FLI and included TT, RVC, SVA, UCM, APHA and CIRAD. Fifty-three functions, which can be used for the design and analysis of surveillance systems, were implemented as part of the R package RSurveillance. These functions related to the surveillance objective demonstrating disease freedom (n = 40), including risk-based approaches (n = 12), and prevalence estimation (n = 7). Furthermore, functions were included for combined and parallel testing (n = 6). A further task consisted of the Development of documentation in standard R formats. Partners involved were FLI, TT, RVC, SVA, UCM, APHA and CIRAD. These functions were documented in standard R formats (https://cran.r-project.org/web/packages/surveillance/surveillance.pdf). This was followed by a further task ‘Validation of R functions’ led by TT and involving: FAO, Safoso and RVC. The R functions were used by various Consortium members as part of the case studies to analyse results from the survey of Federal States in Germany and as part of other applications.
The next task consisted of the Publication of an R package for designing surveillance systems with specific reference to risk-based systems. Led by FLI and involving TT and RVC; the R package RSurveillance was published within the open-source software environment R (http://cran.r-project.org/web/packages/surveillance/index.html). This work package also included the task ‘Development of user-friendly accessible web-based decision support tool implementing methodologies for design and evaluation of surveillance systems’. Once the development of the tool was completed it was made available under: webtools.fp7-risksur.eu. Furthermore, a web-team had been in operation to secure several important tasks: ensuring all essential functions are transferred from the WPs 1-5, ensuring all design aspects from the Excel prototype were transferred, acting as QA for ensuring user-friendliness, and acting as a test team for all the above issues. This was crucial to be able to proceed to the next task to implement web-based decision support tool. It was led by TT, FLI, RVC, UCM and CIRAD and the web-based decision support system consisted of five main parts that were developed and implemented: a tool for designing surveillance systems, a tool for designing components within a surveillance system, a tool for evaluation of a surveillance system, a tool for statistical methods to be utilized in a surveillance system (this was reported in Deliverable D6.24) as well as administrative tools for the complete package.
The work package, under the lead of TT and included FLI, RVC, UCM and CIRAD; also had the task to Develop on-line documentation. The online documentation is implemented in the systems at several levels. There is some overlap between the different levels, mainly to assist users in the use of the online tools. The documentation is implemented both to enhance user-friendliness of the tools and to support clarification, consideration and understanding of the decisions the tools are supposed to support: Ease of use (wizards, mouse-over, context dependent information) of the tool itself. Examples, tutorials and best practice guidelines are accessible through links to wikis. The same team then proceeded to the next task, which resulted in deliverable D6.25 and which had to implement a multi-lingual interface translated into the main EU languages. The multi-lingual support was implemented through translation files. Currently, only an English version is accessible and a French language version is under development. The support of other languages will be a relatively straight forward task, by editing a language file.
The last task ‘Mobile and cloud computing tool development’ was subdivided into 2 sub-tasks, all led by TT and involving FLI, RVC and UCM. The task corresponded to Deliverable D.26. The first task was to assess the need for mobile and cloud computing tools. The project consortium explored the value of having this added functionality and concluded that it could not currently be justified considering the additional cost for developing tools on a mobile platform. The next task was to ‘Develop mobile and cloud computing tools for surveillance systems design’. The architecture of the web-tool has taken into account cloud computing and centralized storage capabilities, thus opening up opportunities for designing the user interface for a mobile device, as a frontend to the centralized cloud service, at a later stage.
7. Work package 7 – Training, dissemination and exploitation
The first task generated the ‘Development and maintenance of general dissemination tools’. It was led by ACCEL and included all other partners besides TT and RVC. One sub-task consisted of the setting up and maintenance of a Website (http://www.fp7-risksur.eu/). It also included features such as: terminology/glossary, impact tracking of the peer-reviewed publications, answers to 25 evaluation-related questions, and 13 online training modules. Another sub-task was an ‘Electronic newsletter’, a total of 6 newsletters with an additional final issue, with the overall description of the project major outputs were published by January 2016. The next sub-task, ‘Science to society’, led by Safoso, included a specific RISKSUR LinkedIn group that included 186 members, as of November 2015. It was used to regularly disseminate jobs related to animal health surveillance. The ‘Science to professionals’ sub-task consisted of a dissemination package, which had been prepared by accelopment and Safoso, was shared with the RISKSUR advisory board members and consortium members, for further dissemination through their professional networks. It consists of about 10 summary slides and an updated project fact-sheet.
The next task ‘Scientific dissemination’ was led by the RVC and involved the rest of the team. It consisted of the sub-task ‘Literature warehouse’: a Mendeley group had been created and had 12 members from different Consortium partners. 825 references were stored in the online database. The next sub-task ‘Scientific publications’, involved all partners and consisted of RISKSUR outputs being presented at several conferences and manuscripts being published in peer-reviewed journals. A full list can be found in section 4.2 of this report. The next task consisted of several ‘Research briefs’ and involved the entire RISKSUR Consortium. Topics covered were: mapping of surveillance systems; animal populations; trade flows; critical infrastructure and decision making processes in seven European countries.
Furthermore, a ‘Symposium’ was organised in March 2015, prior to the Annual Meeting of the Society for Veterinary Epidemiology and Preventive Medicine (SVEPM), at Het Pand in Ghent, Belgium. The theme was ‘Animal Health Surveillance 2.0’ and corresponded to Deliverable D7.29. Fifty participants from 17 different countries attended this event. Scientific exchange was promoted through the establishment of formal links and exchange of information on activities through different international projects and institutions (http://www.fp7-risksur.eu/related-networks).
Another task ‘Best practice recommendations’, led by Safoso and including all other partners besides TT and RVC was subdivided into two sub-tasks. First a ‘Best practice workshop’, Deliverable D7.31 which was held in September 2014 in the Netherlands. The goal and the next task was to subsequently develop best practice guidelines for animal health surveillance tailored to user needs. During the morning, participants reviewed existing standards and discussed its Strengths, Weaknesses, Opportunities and Threats (SWOT analysis). In the afternoon, there was a presentation about the ‘Surveillance landscape in Europe’ (a presentation of the mapping report conducted under the (RISKSUR WP1 work), and this was followed by the presentation and discussion of different surveillance systems being currently used in Europe (e.g. UK, NL). The document ‘Best practice guidelines’, Deliverable D7.32 was coordinated by FAO and Safoso and included contributions from several other consortium members. The original ideas about its contents and structure were developed at the ‘Best Practice Workshop’, and lead to a document that links all the work developed during RISKSUR, providing real life animal health example problems. It covers all steps involved in surveillance system design and implementation, divided into ten chapters as follows: Introduction; Terminology; Prioritisation; Planning; Risk-based design; Implementation; Evaluation and performance monitoring; Economic evaluation; From evaluation to strengthening surveillance and a Conclusion chapter. For the Dissemination of best practice, consortium members were encouraged to disseminate the ‘Best practice guidelines’ among the veterinary services of their countries.
Furthermore, for the Support service for Common Agricultural Policy (CAP) and Common Animal Health Policy (CAHP): Surveillance surgery led by Safoso and involving ACCEL, GD, SVA and RVC, 6 ‘Surveillance Surgery Webinars’ were delivered over the course of the project. They were open to scientists, community and MS policy makers as well as industry. Participation from accession countries and third countries were also encouraged to attend. The discussions were moderated by GD with support from Safoso. Topics included: African swine fever (ASF) surveillance; Looking at society from an animal health surveillance perspective; Surveillance of antimicrobial resistance; Surveillance of Avian Influenza; RISKSUR Surveillance Design Framework; RISKSUR Surveillance Evaluation Framework.
Potential Impact:
After this three-year project now having been completed, policy makers and other stakeholders have access to a set of tools that can be used to inform decision making on the design and evaluation of cost-effective animal health surveillance systems addressing the surveillance objectives ‘early detection of disease’, ‘demonstration of freedom from disease’, ‘determination of disease frequency’ as well as multi-hazard surveillance. The criteria for surveillance design apart from epidemiological considerations included the wider systems context and economic factors. A key epidemiological feature is the explicit inclusion of risk-based approaches to surveillance amongst the possible surveillance methodologies and the tools therefore provide a much-needed scientific evidence base for application of such approaches (often referred to as risk-based surveillance). Also, the tools are supported by appropriate documentation. Associated learning materials have been made available in the public domain to allow widespread access. To ensure that the approaches are based on sound scientific principles, the Consortium produced peer-reviewed publications in relevant scientific journals in the relation to the outputs produced by the project, a process which started off with some the literature reviews and was followed up by further publications
The main dissemination activities and exploitation of results were:
A key feature of the RISKSUR project was the dissemination of information to different stakeholders in an easily accessible format using various communication methods and tools. The coordinator RVC managed the dissemination/exploitation of the scientific and technological results at the project level, and WP6 leader Safoso was responsible for the execution of the actual dissemination/exploitation activities of individual partners. Newly acquired knowledge was continuously being evaluated for protection, dissemination and exploitation of intellectual property rights by each partner.
The main objectives of all dissemination activities were the communication and spread of the project results to the following stakeholders and the promotion of commercial exploitation:
- National government veterinary services: RISKSUR provided recommendation solutions for national governments in the context of specific disease control activities (e.g. foot-and-mouth control in the UK). From the start of the project CVOs did show their interest in using the project results.
- Supra-national government organizations: RISKSUR offers advice to supra-national organizations such as EFSA, FAO and World Bank. The outputs from RISKSUR will be relevant for all these organisations.
- Pharmaceutical and diagnostic industry: These companies are increasingly looking for integrated solutions which include the related surveillance activities. RISKSUR output will inform this stakeholder group.
- Livestock industry: RISKSUR results will be very relevant for this group to assure their efficiency and competitiveness in an international market.
- Scientific community: The results of WPs 1-4 will be of particular interests to scientists in this research area. For this reason, exchange of knowledge and interactions with other projects in this field will be important to RISKSUR researchers.
- General public: Science in society is important to RISKSUR since an informed society can help prevent the transmission of diseases. It is therefore crucial to raise awareness for novel surveillance systems and their potential benefits to society.
All activities targeted to these different stakeholders have been planned in WP7, under the lead of Safoso. Key to the exploitation was a transparent and fair communication and management of both existing information and project results. For this reason, RISKSUR partners had agreed on the following Intellectual Property issues:
All information provided by a partner to other partners within the project was and still is confidential unless:
* It was already known to the partner before the negotiations started, or
* The information provided is public property, or
* It is explicitly specified otherwise by the originator of the information.
Partners had also agreed to use the information provided only for the purposes of conducting the project. Any disclosure of confidential information to a third party requires the explicit consent of the originator of that information. The SME partners Safoso and TT were granted sole ownership of the services (e.g. training courses) and products (e.g. software application) developed in this project. When more than one partner claimed joint ownership of newly produced intellectual property other than those mentioned in point 3 before, the partners involved make provisions to clarify the terms of joint ownership among them. Partners were not restricted in any sense regarding the rights associated with the ownership of any intellectual
property they produce while conducting the project activities.
Impact of dissemination activities
Dissemination activities targeted to the general public and industry:
Web site: ACCEL established and maintained a project website (www.fp7-risksur.eu) with sections for the general public, for surveillance professionals, policy makers, industry and researchers. Relevant news from other sources such as DGSANCO or scientific publications in the field were integrated. Project partners used the restricted domain of the website to share information and work on jointly authored documents. The website was designed such that it can be sustained beyond the lifespan of the project. It also included all available training materials, recordings of webinars, copies of public deliverables and a Glossary.
Electronic newsletter: To ensure that the level of interest in the project among stakeholders did not decline, the pull communication (website) with this permission-based push communication on RISKSUR generated knowledge was combined. The target group of the newsletter was non-scientific, mainly policy makers and industry. ACCEL and RVC jointly elaborated and disseminated this bi-annual newsletter with a target audience of about 5,000 subscribers. Other institutions contributed to the content of the newsletter. Overall 6 newsletters were sent out, a final 7th is to be sent out in January 2016.
Science to society and to professionals: Under the coordination of ACCEL, RISKSUR used current electronic media such as LinkedIn to disseminate information on animal disease and risk surveillance to the general public. The aim was to especially target rural communities, livestock dealers and other associated industries.
Impact: The aforementioned efforts contributed towards increasing the effectiveness of surveillance simply by increasing stakeholder awareness, and generating improved acceptance and trust.
Dissemination targeted to the scientific community:
Scientific publications: All participants published their scientific findings in leading international and national journals in the fields of veterinary epidemiology, disease control, infectious disease, and livestock production. Papers in the national language will reach audiences not familiar with English. By the end of the project in October 2015, 5 papers had already been published in high-impact journals. A minimum of a further 5 were in preparation and to be submitted in January 2016. Additionally, RISKUSR results were presented at national and international scientific meetings such as SVEPM 2013-2015, ICAHS 2014 and ISVEE 2015.
Symposium: A Symposium was organised on 24 March 2015, prior to the Annual Meeting of the Society for Veterinary Epidemiology and Preventive Medicine (SVEPM), at Het Pand in Ghent, Belgium. The theme was ‘Animal Health Surveillance 2.0’. Fifty participants from 17 different countries attended this event. It aimed to disseminate research results, to promote the dialogue between scientists and surveillance users and to facilitate networking within the surveillance community.
Scientific exchange: Exchange with other existing networks was promoted by establishing formal links and by joint meeting and exchange of information on activities. This activity was coordinated by Safoso. EU-funded research projects and networks with a link to RISKSUR were, for example, LINKTADS, MARLON; EPIZONE, DIPNET, MED_REO_NET, PANDA, EVENT, EMPERIE.
Impact: Based on these targeted dissemination activities, RISKSUR communicated the outputs to the science community, thereby provided opportunity for them to be scrutinized so as to ensure a high level of scientific validity. At the same time, it resulted in dissemination of the tools amongst the scientific community for further development, as well as widespread adoption of the underlying methodologies which in turn did also benefit the EU and its member states by allowing safer and more transparent trade and safeguarding animal and human health.
Dissemination targeted to policy-makers:
Best practice workshop: The Best Practice Workshop was held on September 30, 2014, at the Ministerie van Economische Zaken, The Hague, Netherlands. The goal was to subsequently develop best practice guidelines for animal health surveillance tailored to user needs. During the morning, participants reviewed existing standards and discussed its Strengths, Weaknesses, Opportunities and Threats (SWOT analysis). In the afternoon, there was a presentation about the ‘Surveillance landscape in Europe’ (a presentation of the mapping report conducted under the RISKSUR WP1 work), and this was followed by the presentation and discussion of different surveillance systems being currently used in Europe (e.g. UK, NL). Participants had a mixed background in surveillance research and practice as well as policy.
Best practice guidelines: FAO and Safoso coordinated the development of this document, that included contributions from several other consortium members. The original ideas about its contents and structure started building up at the Best Practice Workshop, and lead to a document that links all the work developed during RISKSUR, providing examples with real life animal health problems. It covered all steps involved in surveillance system design and implementation, divided into ten chapters as follows:
1. Introduction
2. Terminology
3. Prioritisation
4. Planning
5. Risk-based design
6. Implementation
7. Evaluation and performance monitoring
8. Economic evaluation
9. From evaluation to strengthening surveillance
10. Conclusion
Dissemination of best practice:
Consortium members were encouraged to disseminate the ‘Best practice guidelines’ among the veterinary services of their countries.
Support service for Common Agricultural Policy (CAP) and Common Animal Health Policy (CAHP): Safoso initiated and moderated surveillance of issues via a widely available web-tool Adobe Connect. Overall 6 so-called ‘Surveillance Surgery Webinars’ occurred over the course of the project. They were open to scientists, community and MS policy makers as well as industry. Participation from accession countries and third countries were also encouraged to attend. The discussions were moderated by GD with support from Safoso. Topics included: African swine fever (ASF) surveillance, Looking at society from an animal health surveillance perspective, Surveillance of antimicrobial resistance, Surveillance of Avian Influenza, RISKSUR Surveillance
Design Framework, RISKSUR Surveillance Evaluation Framework.
Impact: Design of surveillance systems is an extremely complex issue with a variety of factors that influence the decision making process. RISKSUR provided a transparent, validated and user-friendly web-based decision support tool that allows policy makers to consider all possible options of system configurations and then make an informed choice, taking into account cost-effectiveness. There is currently no other such system available, and decision-makers usually still work on the basis of methods, which have not much changed over the last 100+ years. RISKSUR, therefore, for the first time provides, an objective mechanism for justifying and rationalizing the usage and combination of traditional as well as novel surveillance approaches. This also contributes towards enhanced equivalence in approaches and outputs generated to inform trade decisions.
Dissemination to practitioners:
Training material and online case studies: RVC in collaboration with Consortium members developed training materials, each based on a work package, in the form of short online lectures. All 13 lectures were made available via the RISKSUR web site. Communication channels as described above will be used to promote training material.
Impact: RISKSUR will transfer knowledge to technical experts within the EU and elsewhere on the general concepts of risk-based surveillance and the outputs of the project. This resulted in enhanced surveillance capacity in the EU and elsewhere, and also facilitated communication of the approaches used.
Impact of exploitation activities:
Three of the four SMEs involved had a strong commercial interest in the expected results to be developed in RISKSUR. In order to facilitate the exploitation of results, partners had already agreed on the following two principles that will govern the IP terms in this project, and form the basis of the IP terms in the Consortium Agreement:
Solely generated IP will be solely owned by the generating party with a first option to the project and associated partners to a non-exclusive option to license foreground IP for commercialisation.
All intellectual property generated jointly during the course of the Project (‘Joint Foreground IP’) will be jointly owned by the generating parties, and shall be managed in accordance with the terms of a Joint Ownership and Management Agreement (JOMA).
RVC in its role as the Coordinator was responsible for ensuring that a secure and suitable knowledge management system is put in place, which did run as soon as possible after the project has started and that data protection legislation is followed. In the case of conflict of interests, ACCEL assisted and moderated between the commercial partners, with no interest in the commercially exploitable results. This IP management strategy allowed especially our SME partners to fully exploit the potential of RISKSUR. Exploitation will be actively supported by the project’s Advisory Boards. Decision on the use of project results will be made by the Steering Committee according to the decision-making process.
Roadmap for cost-effective surveillance of disease beyond RISKSUR:
RISKSUR had the first development of an integrated decision support tool for design and evaluation of animal health surveillance systems. It was underpinned by sound science-based evaluation and decision support frameworks, which explicitly use cost-effectiveness as one of the outputs. It was important to recognise that the tool needed to be flexible and required review and potentially updating on a regular basis. Assuming that it was able to demonstrate its utility to end users, they were motivated to perform and fund those updates. It is also conceivable that new approaches of delivering the functionality of the tool will become available, and it then needs to be revised accordingly.
National CVOs and other organisations were committed to apply and support the use of the project results and thereby both ensure the sustainability of the project and facilitate more cost-effective surveillance in the future. Additionally, FAO’s wide network helped pave the way for future roads in surveillance of animal diseases.
The frameworks and methodologies developed as part of RISKSUR will thus continue to evolve, but be able to do so more effectively and in a balanced fashion. It is also likely that new surveillance methods will become available, but the frameworks will be sufficiently flexible to take such developments into account.
List of Websites:
Address of project public website and relevant contact details
RISKSUR website: www.fp7-risksur.eu
Contact:
Professor Dirk Pfeiffer
Royal Veterinary College, UK.
Email: pfeiffer@rvc.ac.uk