key: cord-1039884-dblvj87c
authors: Renault, Véronique; Damiaans, Bert; Sarrazin, Steven; Humblet, Marie‐France; Lomba, Marc; Ribbens, Stefaan; Riocreux, Flavien; Koenen, Frank; Cassart, Dominique; Dewulf, Jeroen; Saegerman, Claude
title: Classification of adult cattle infectious diseases: A first step towards prioritization of biosecurity measures
date: 2018-08-21
journal: Transbound Emerg Dis
DOI: 10.1111/tbed.12982
sha: df897e80df9dbba16145a03ee7c4de18a13ada7d
doc_id: 1039884
cord_uid: dblvj87c

An emphasis on biosecurity in the cattle industry was made over the years to improve animal and public health. Nevertheless, the level of implementation of biosecurity measures (BSM) remains largely insufficient due to certain constraints. It is therefore necessary to prioritize the different BSM to be applied in accordance with the individual context and the main infectious diseases affecting cattle. Previous prioritization exercises of infectious diseases were neither specific to Belgium nor based on an exhaustive list of diseases. This study aimed at classifying the most important infectious diseases affecting cattle in Belgium. A list of 74 cattle infectious diseases reported in Europe was compiled based on a literature review. Through an online survey, Belgian rural veterinary practitioners (RVP) were asked to assign a score to each disease according to their frequency (question 1), their trends estimated between 2013‐15 (question 2), and finally to list the five most important diseases for adult cattle (question 3). Respectively, 107 and 93 RVP answered the first two questions and the last one. Results of the survey were used to classify the diseases based on their frequency, trends, and importance through an additional weighting system and a subsequent regression tree analysis. Belgian laboratory databases and previous disease prioritization exercises were also analysed and taken into account as additional data sources. For the most important diseases identified (those ranked as important by the three data sources), a literature review was performed in PubMed to identify their related risk factors and BSM. A total of 48 infectious diseases were classified as important in Belgium with six of them considered as important from the three data sources: bovine respiratory diseases (BRD), bovine respiratory syncytial virus (BRSV), bovine viral diarrhoea (BVD), infectious bovine rhinotracheitis (IBR), Q fever, and salmonellosis. Their related BSM should be prioritized in terms of BSM implementation.

Cattle farming is one of the main food-production species in Belgium. Over the last few years, a shift from curative towards preventive medicine has been observed in the livestock sector and represents a key element of the European Union Animal Health Strategy since 2007 (European Comission, 2007 . Nevertheless, several surveys highlight a low implementation level of biosecurity measures (BSM) by the farmers with different constraints expressed such as cost, usefulness, workload, and lack of clarity on the measures (Brennan & Christley, 2013; Gunn, Heffernan, Hall, McLeod, & Hovi, 2008; Hoe & Ruegg, 2006; Kristensen & Jakobsen, 2011; Nöremark, Frössling, & Lewerin, 2010; Sarrazin, Cay, Laureyns, & Dewulf, 2014; Sayers et al., 2013) . The rate of implementation of BSM seems even lower in cattle farms versus pig or poultry production facilities (Sarrazin et al., 2014) . To better advise cattle farmers and increase their level of implementation, it is essential to prioritize the biosecurity measures, according to the most important infectious diseases affecting or threatening Belgian cattle.

Based on the need to prioritize the infectious diseases (further referred to as diseases only) to address in terms of disease surveillance, control and eradication programs, many prioritization, or categorization exercises were conducted over the last few years. Given the lack of prevalence data for most cattle diseases, most of them The Delphi method based on a consensus approach has many advantages (e.g., no need of scientific evidence as it relies on experts' opinion which can be modified through debates and avoids personal and political influence as a consensus is needed) and is recognized by the scientific community worldwide since its development by the RAND Corporation in the late 1960's. The recent prioritization exercises identified in the literature (ANSES, 2012; Ciliberti, Gavier-Widén, Yon, Hutchings, & Artois, 2015; DISCON-TOOLS, 2016; Havelaar et al., 2010; Humblet et al., 2012; McIntyre et al., 2014) were quantitative, semiquantitative, or qualitative and based on the Delphi method with the exception of two. One of them was based on the H-index (McIntyre et al., 2014) and the second one on a literature review with a scoring and weighting system applied and validated by a panel of experts (Humblet et al., 2012) (Supporting Information Table S1 ).

Nevertheless, these scoring systems rely solely on expert's opinion and results will vary depending on: initial list of diseases to be assessed, criteria used, ranking methodology proposed, objective of the prioritization exercise, and available resources (e.g., time and quality of the expert panel involved). In addition, most of them did not consider multipathogen diseases such as mastitis, respiratory diseases, and diarrhoea, which are usually a major concern for both animal and public health and should not be automatically omitted.

The objectives of this study are to (a) identify major diseases of concern for Belgian cattle holders and their related BSM using a prioritization methodology based on the outcomes of a veterinary survey, the analysis of 3-year laboratory databases and the review of previous prioritization articles and (b) summarize BSM related to the six most important diseases of concern, i.e., the only diseases defined as important by the three data sources following the classification process described in Figure 1 .

An initial list of infectious cattle diseases was established based on several sources. The list provided by the Center for Food Security and Public Health, Iowa State University (http://www.cfsph.iasta te.edu/DiseaseInfo/index.php) was used and completed by the review of five reference books on cattle diseases (Andrews, Blowey, Boyd, & Roger, 2008; Francoz & Yvon, 2014; Institut de l'Elevage, 2000; Kahrs, 2001; Scott, Penny, & Macrae, 2011) , prioritization articles (Ciliberti et al., 2015; McIntyre et al., 2014; Phylum, 2010) EFSA-ECDC, 2015) and OIE websites and last reports (FAAV/WIV/ CODA-CERVA, 2015). The diseases for which the occurrence or existence in Europe or Belgium was not specified in those sources, a literature review was performed based on a web search in PubMed with the following combinations of terms: "name of the disease" or "name of the pathogen" and "Belgium" and/or "Europe" to complete the information. A list of 90 diseases was established with their occurrence in Europe and in Belgium, their OIE status in Belgium and basic epidemiological data (last occurrence in Belgium and zoonotic character) (Supporting Information Table S2 ).

In order to maintain the length of the questionnaire addressed to the rural veterinary practitioners (RVP) to the minimum, 31 diseases were excluded from the initial list of 90 diseases (16 diseases with no occurrences in Europe and 15 diseases with no occurrences in Belgium).

The RVP were contacted on line through the two regional animal health organizations of the country, i.e., Association Régionale de Santé et d'Identification Animale (ARSIA) in Wallonia (southern part of the country) and Dierengezondheidszorg Vlaanderen (DGZ) in Flanders (northern part of the country) with monthly reminders over 4 months. The questionnaire was anonymous, available in French and Dutch version and could only be filled once by the same IP address.

The number of persons included in the mailing list of the two organizations are respectively of 1876 and 1356 including both rural and small animal's practitioners as it was not possible to identify the part of RVP within these mailing lists.

In Wallonia, the RVP workforce (534 veterinarians having a rural practice out of 1876 veterinarians), was provided for each of the five provinces, by the Board of Veterinary Practitioners. A chi square test has been performed to assess that the sample of responding RVP is not unbalanced from one province to another. The counterpart workforce for Flanders was not available; indeed, in that region, veterinary practitioners have no obligation to provide details on their practices to the Regional Board.

The survey was pretested by four veterinarians before its final validation and included three questions in order to assess the frequency, 3-year trend and the importance of each disease for the Belgian cattle sector. In the first question (Q1), RVP had to assign a score to each disease related to their average frequency based on the following scoring system: (a) never suspected, (b) suspected but never confirmed, (c) several times a year/occasionally, (d) at least once per quarter, and (e) several times a month.

In the second question (Q2), RVP were asked if the disease trend over the last 3 years was decreasing (score of 0), constant (score of 1), or increasing (score of 2).

The third question (Q3) was an open question where RVP were asked to list, in decreasing order of importance, the five main diseases affecting adult cattle; that information would help triangulating the information and identifying eventual diseases of importance omitted in the initial list. Each disease was assigned a score of 1 to 5, depending on its position in the list: (1) fifth disease listed, (2) fourth disease listed, (3) third disease listed, (4) second disease listed, and (5) first disease listed.

The answers to Q1 and Q2 were respectively used to calculate an average frequency score (af) and average trend score (at), for each disease. A global score per disease (GS) was then calculated by adding both averages.

A regression tree analysis based on the GS of the different diseases identified and classified the most important diseases to consider, from the RVP's perspective. The regression tree methodology is a nonlinear and nonparametric test increasingly used by the scientific community in public and animal health. It divides the population (in our case, the diseases) into different subgroups in relation to the GS with minimal within-variance by using cross-validation (Lemon, Roy, Clark, Friedmann, & Rakowski, 2003; Saegerman, Porter, & Humblet, 2011; Salford Systems, 2001) .

Q3 was analysed separately. The analysis excluded noninfectious diseases such as foreign bodies and metabolic disorders, as mentioned by the RVP. The list of diseases was standardized in terms of disease denomination and consolidated. A disease index was then calculated for each disease by adding all its scores based on RVPs' ranking. In order to identify the most important diseases, the 66 th centile of the disease indexes was used as a threshold (index above 66 th centile).

From the veterinary survey (DS 1), a score of "1" was attributed to (a) all diseases with a GS classified as high or important in the regression tree analysis (Table 1) , and (b) all diseases having a disease index above the calculated 66 th centile).

F I G U R E 1 Selection criteria for the most important diseases to consider RENAULT ET AL.

| 1993

Due to the subjective character of the veterinary survey (DS1), the risk of underestimating some important diseases was not to be neglected, e.g., (re)emerging diseases with no occurrence in Belgium, and major zoonoses with a slight impact on cattle. These diseases were initially identified through the analysis of laboratory databases (DS2) provided by two regional animal health organizations, i.e., ARSIA in Wallonia analysed. Diseases were considered equally important and attributed a score of "1" if, in at least one laboratory, one of the following arbitrary conditions was recorded: (i) >100 tests performed, (ii) >25% positive results, (iii) increasing number of tests requested over the period of concern (>66 th centile), or (iv) increasing number of positive results (>66 th centile). The tests linked to specific research projects were excluded from the analysis but the tests related to the official sampling scheme have been included.

As a third data source (DS3), six recent prioritization exercises (Supporting Information Table S1) 

An overall score (OS) was calculated by adding the scores of the three DS (veterinary survey, laboratory databases, and prioritization exercises) (Table 3) . Following the process of disease selection ( 

For the most important diseases identified (those ranked as important by the three data sources), a literature review was performed in PubMed to identify their related risk factors and biosecurity measures. The keywords used for the search were as follow: "name(s) of the disease" or "name(s) of the pathogen" and "cattle or bovine or cow or beef or calves or dairy" (if disease affecting multiple species only) and "epidemiology" or "pathogenesis" or "control" or "risk".

T A B L E 1 Classification of diseases, per category, based on the Regression Tree analysis of global score (GS), according to participants' responses for questions 1 and 2 (N = 107)

High GS (score = 4) Significant GS (score = 3) Moderate GS (score = 2) Low GS (score = 1) Among the articles selected, only those articles mentioning an analysis and/or the identification of disease-specific risk factors or BSM were fully read.

The After analysing Q3, three diseases showed a high disease index (>66 th centile) but without a high or significant GS: the two diseases not listed in the initial list of diseases, i.e., metritis/endometritis and secondary infections, and IBR, not classified as important by the regression tree analysis. These three diseases were thus classified as important. (Table 3 ). 

Results of the final classification after application of the different filters are summarized in Table 3 for the 48 diseases considered as important. Six of them were identified as important by the three DS:

BRD, BRSV, BVD, IBR, Q fever, and salmonellosis.

Fourteen diseases came out as important from at least two DS.

Finally, 28 diseases were revealed by only one DS: 15 by prioritization exercises, 11 through the veterinary survey, and two based on laboratory databases (Bo-HV4 and BEL).

As a reminder, the initial list of diseases included 77 items (74 diseases initially listed and three diseases added during data analysis), thus 29 of them were not classified as important at the end of the process. They are listed in Supporting Information Table S2, along with the diseases with no occurrence in European countries.

A total of 76 articles were reviewed: 6 for BRSV, 17 for BRD, 11 for BVD, 13 for IBR, 15 for Q fever, and 14 for salmonellosis (Supporting Information (Geurden, 2007) .

The review of previous disease prioritization exercises led to include additional diseases, also considered as important, in the list. Notes. a Includes BRSV, mycoplasmosis, pasteurellosis, para influenza virus 3, and other respiratory diseases.. Coding: "2" for measure listed in literature review either as addressing a specific risk factor or BSM; "1" for measure not found as such during the review, but should have an effect on the disease prevention and management due to its different transmission pathway; "0" for measure without influence on the disease. (Shaapan, 2016) . Its inclusion in the list depends on the objectives and foreseen usage of the disease classification exercise.

Out of the 48 diseases, 25 are nonnotifiable but of major importance in Belgium due to their economic impact and/or high occurrence. Nineteen of them were not considered as important by the previous prioritization exercises while relevant in Belgian adult cattle.

This additional list could guide the decision makers for future control programs as these diseases are a major concern for cattle holders.

The six diseases identified as important by the three data sources are covering the different diseases transmission pathways, therefore the proper implementation of their related BSM (Table 4) In order to ensure the acceptability of the BSM to be prioritized by the farmers a participative approach in recommended in order to take into account the farmers opinions, perceptions, and expertise on the topic.

Due to their possible impact on the economy, it is important to raise the level of awareness of the herders regarding emerging and exotic diseases. Nevertheless, starting by addressing the farmer's priority issues is a key strategy for them to adopt the biosecurity measures on a long-term perspective. Identifying the most important diseases affecting cattle farms is therefore necessary in order to initiate the process of change. Specific measures related to public health purposes could be introduced easily afterwards. Future researches should focus on the assessment of the level of implementation of the BSM related to the most important diseases to be targeted (six in the case of Belgian cattle herds), as well as the possible constraints and factors affecting their adoption by the farmers in order to be able to prioritize the most effective BSM to be promoted.

The methodology proposed and relying on the outcomes of a veterinary survey, the analysis of the laboratory databases over the past 3 years and the review of previous prioritization exercises, allowed identifying the diseases of major concern for cattle holders.

The proposed methodology represents a practical tool for other users who could easily adjust the selection criteria to their specific objectives, needs, and context. That makes possible the future development of a biosecurity tool useable at the national level.

This study was supported by the Belgian Federal Public Service for Health, Food Safety, and Environment (Contract RT 15/4 BOBIO-SEC1). The implementation of the survey and further analysis were possible; thanks to the collaboration with ARSIA, DGZ and CODA-CERVA who supported the veterinary survey process and shared their 3-year databases. We would also like to thank all the rural veterinary practitioners and the heads of Pathology Departments of both Gent and Liege universities who took the time to answer the online survey. We do hope that the outcomes of this study and further studies will be helpful to their practices.

The authors declare no conflict of interest.

Saegerman http://orcid.org/0000-0001-9087-7436

Situation zoosanitaire en Belgique

Bovine medecine: Diseases and husbandry of cattle

Avis de l'ANSES relatif à "la hiérarchisation de 103 maladies animales présentes dans les filières ruminants, équidés, porcs, volailles et lapins en France métropolitaine

Is there utility in the transtheoretical model? British

Cattle producers' perceptions of biosecurity

Multiple serotypes of bluetongue virus in sheep and cattle

Prioritisation of wildlife pathogens to be targeted in European surveillance programmes: Expert-based risk analysis focus on ruminants

Laboratory findings suggesting an association between BoHV-4 and bovine abortions in Southern Belgium

Implication du BoHV-4 comme agent étiologique d'avortements chez les bovins

Congenital jaundice in bovine aborted foetuses: An emerging syndrome in southern Belgium. Transboundary and Emerging Diseases

The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2013. European Food Safety Authority -European Centre for Disease Prevention and Control (ECDC)

A new Animal Health Strategy for the European Union (2007-2013) where "Prevention is better than cure

Trends and sources 2012-2013, report on zoonotic agents in Belgium

Manuel de médecine des bovins

Cryptosporidium and Giardia in calves in Belgium. Doctoral thesis

Measuring and comparing constraints to improved biosecurity amongst GB farmers, veterinarians and the auxiliary industries

Prioritizing emerging zoonoses in the Netherlands

Opinions and Practices of Wisconsin Dairy Producers About Biosecurity and Animal Well-Being

Multidisciplinary and evidence-based method for prioritizing diseases of food-producing animals and zoonoses

Maladies des bovins (3ème Editi)

Viral diseases of cattle

Danish dairy farmers' perception of biosecurity

Classification and regression tree analysis in public health: Methodological review and comparison with logistic regression

Climate change beliefs, risk perceptions, and adaptation behavior among Midwestern U.S. crop farmers. Climate Risk Management

A quantitative prioritisation of human and domestic animal pathogens in europe

Theories and models of behaviour and behaviour change

Application of routines that contribute to on-farm biosecurity as reported by Swedish livestock farmers

Listing and categorisation of priority animal diseases, including those transmissible to humans -Mission report. Colomiers, France: World Organisation for Animal Health (OIE)

Stages and processes of self-change of smoking: Toward an integrative model of change

The use of modelling to evaluate and adapt strategies for animal disease control

CART: Tree-structured non-parametric data analysis

A survey on biosecurity and management practices in selected Belgian cattle farms

Implementing biosecurity measures on dairy farms in Ireland

Cattle medecine

The common zoonotic protozoal diseases causing abortion

Setting priorities in communicable disease surveillance