Research Project Database
Code: EMIDA2
Title: Early detection data - Improving early detection of emerging vector borne diseases by using existing production and diagnostic data.
Country: Netherlands
Funding Organisation: French National Research Agency (ANR)
Veterinary and Agrochemical Research Centre (CODA-CERVA)
Netherlands Food and Consumer Product Safety Authority
Animal Group:  
Disease: Vector-borne diseases
Research Organisation: Veterinary and Agrochemical Research Centre (CODA-CERVA)
Animal Health Service Deventer
Number of Research Staff (FTE):  
Principal Investigator (PI): Paul Wever, Animal Health Service Deventer
Cost (Euros): 468794
End Date (dd/mm/yyyy): 31-12-2013
Duration (months): 25
Project objectives and deliverables with estimated delivery dates for each deliverable (if possible): Increased trade and subsequently movement of animals as well as climate change may lead to (re-) establishment of several vector-borne diseases in Europe. The recent outbreak of BTV in North-western Europe in 2006 and 2007 highlights this concern and requires effective surveillance systems for early detection of vector-borne diseases. It is important to consider current capabilities and future needs to effectively address surveillance, preparedness and response strategies for vector-borne diseases that emerge and may become prevalent.
At several stages in the animal production chains data are recorded for several purposes such as management, quality controls, breeding and animal health care. Apart from their primary purpose for collection, these data could also be used for signalling changes in the epidemiological situation, which is hardly being performed as a tool in animal disease control. Signal theory methods and (real-time) time series analysis, could be adapted to such non-specific repeated data to detect a disease emergence in the absence of a previously identified pathogen. Vector-borne diseases are likely to exhibit a spatio-temporal distribution pattern of detected signals associated with the possible distribution and spread of vectors.
The aims of this project are to develop and evaluate a monitoring and early detection system for emerging vector-borne diseases in cattle, based on indicators derived from existing data, such as production records and diagnostic data. The project will also aim for cross-border cooperation regarding monitoring and surveillance of animal health.
The project will deliver statistical and modelling methods for existing production and diagnostic data to detect deviations from trends in time and space as indicators for emerging vector-borne diseases. These models will be used to design a surveillance system for early detection of emerging vector-borne diseases based on existing data. The surveillance systems for early detection of vector-borne diseases will be validated by using BTV as a case-study. The monitoring and surveillance systems (components) in the participating countries will be evaluated with scenario tree modelling and, where possible, economical (cost/benefit) impact will be determined. Eventually, a framework for an optimal mix per country and cooperation between the participating countries in a joint surveillance system for early detection of emerging vector-borne diseases will be proposed.

Executive summary from final progress report:

The aim of the project was to evaluate syndromic surveillance, based on pre collected data, as a tool for early detection of vector borne diseases in cattle. The bluetongue epidemic in 2006/2007 and the epidemic of Schmallenberg virus were used as events for validation of the methods developed and evaluated in the project.

An inventory was made of vector-borne diseases in cattle. The relevance of the diseases was evaluated based on several criteria such as likelihood of introduction in the participating countries (Netherlands, Belgium and France), presence of a capable vector, effect on production (morbidity and mortality), zoonotic potential, etc. The expected signals per disease, such as production effects, and possible differential diagnoses with endemic diseases were determined.

Milk production data, fertility data and mortality data were collected from relevant data owners in each country. Data from each country was analyzed separately. Building up a single and complete dataset and analyzing data cross-border was hampered by a number of issues.
1. Data in some cases were owned by several organizations lacking harmonized data collection.
2. Negotiating data transfer with data owners sometimes was time consuming and some issues (privacy and blinding data) could not always be solved in the project period.
3. Data owners in some cases did not allow to share data cross-border because of business interest.
4. Laboratory data not always were consistent and comparable over a time span of several years, due to implementation of new programs and/or legislation.
5. Combined analysis of data cross border was not feasible because there were intrinsic differences between the datasets (e.g. different animal categories, different levels and different patterns and different trends in mortality rates), which could not be accounted for in the models.
Initially data were collected from 2005 until the end of 2010. As for the Netherlands data from 2011 and the first half of 2012 were available, it was decided to use data from that period as well, in order utilize the SBV outbreak for validation purpose.

Five reproductive indicators based on the occurrence of return-to-service or gestation length were proposed and evaluated. Other indicators evaluated were milk production and mortality rates in ear tagged cadavers and non-ear tagged cadavers.

The results show that syndromic surveillance, based on the use of data collected for other purposes (mainly day to day farming practice), can be a useful tool for disease monitoring. This applies specifically to vector-borne diseases having a more or less gradual radial spread. Without any other surveillance tool in place, we would have been able to detect changes in indicators that could be attributed to vector-borne disease events. Short gestation and milk production were the most promising indicators regarding BTV and SBV, taking into account specificity, sensitivity and timeliness, although further investigations would be required to determine whether a (vector borne) disease event really was the cause of a signal. A substantial number of false alarms would have to be investigated, pointing out that a balance needs to be found between the sensitivity and the specificity of such a system. However, alarms from the syndromic surveillance can be supportive of the interpretation of those from passive clinical surveillance and vice versa.

The project team recommends as follows:
1. For policy makers it seems worthwhile to consider implementing syndromic surveillance on pre-collected data as a tool for early detection of vector-borne diseases. Because the signals will not be specific, it would have to be incorporated in general surveillance to allow signals derived of it to be further investigated in relation to other surveillance components. In general, implementing syndromic surveillance will have more added value when passive surveillance is less efficient.
2. When implementing, focus should be on the logistic features of syndromic surveillance as well as assuring data quality on the country level rather than on achieving cross-border databases and analyses.
3. When implementing syndromic surveillance, further development of methods is advised. While operating systems can be done on the national level, sharing data definitions, development and updating of the statistical methods should be a joint action so that results will be comparable between countries.
4. Operators of surveillance systems in neighboring countries should share and discuss signals cross border within an organized framework.
FilesNo files are attached.