4. Discussion
This study is a first attempt at reporting the uses of and perspectives on animal information systems and risk analysis tools by professionals from all around the world. Similar surveys have been conducted but referring specifically to a single animal health information system (i.e. analysing the WAHIS database only or giving general summary of animal health platforms) (OIE, 2017; OIE, 2020; FAO, 2011). To date, attitudes towards the use of these platforms have not been discussed/analysed. However, this same point highlights a limitation of this study only in that it gives only a general picture of the constraints and attitudes. This is because there is too wide a variability of i) animal health information systems (national and international) as well as ii) risk analysis tools, to be able to provide details. A larger study, specific to these systems or tools would be required.
The response rate to this study could be considered as overestimated given the fact that a snowball strategy was used. The anonymous form of the survey forbids quantification of the number of experts who could have been added to the survey using the network of the original set of identified professional (snowball strategy) (Lupo et al., 2016). Although a low response rate was achieved, this strategy provided a good representation of professionals who used the systems as the survey was specifically sent to focal points responsible for notifying animal diseases. Moreover, the years of experience in the field and age of the respondents were well represented as well as international location and area of professional activities of the respondents. The sample population carried out their professional activities from a broad range of nations, which gives a good general picture of the uses of these systems at an international level. The survey also captured the international activities of the respondents, probably at different levels or at international organizations. The responses were analysed as a group (i.e. without dividing it into subgroups by professionals’ provenance) and did not compare relationship between issues or restraints and regions (i.e. differences in term of animal health institutionalisation or data accessibility). It is important to highlight that the snowball strategy was done using specific focal, thus there may have been differences in issues by regions. The sample would have differed if the snowball strategy for example was used in an institution of a university in the United States. The sample would have represented more the United States universities, which may have arisen other issues or restraints.
Most of the respondents in this survey were employed in governmental institutions, research and universities with animal and public health being the field of competency most represented. Academia and governmental institutions were the places where animal health information systems and risk analysis tools are most used, most likely them having easier access to these tools and a level of understanding of using the tools.
As for the utility of animal health information systems, it is important to consider that an animal health information system is only as good as the data it contains (OIE, 2020). This survey highlighted that the degree of frequency of animal health systems use and the information type found in them was not related. According to the gathered expert opinion, prevention of disease occurrence is more important than treatment. This study highlighted how professionals give important focus to the type of information related to i) disease incursion and ii) epidemiological characteristics of diseases (i.e. information on cases/incidence, the evolution/spread of the disease). The latter informs actions to limit the introduction of a disease into a country free from the infection.
Data sources were more needed than used, showing there is a lower access to data sources than required. However, there were no questions to know if there were issues in having a knowledge (i.e. understanding the known databases, mechanisms of extraction, obtaining information) or due to limitations in technology.
Although, there is accessibility to certain data (e.g. by officially demanding access to international organizations) respondents showed that access to databases on public and animal health, access limitation are still high (Bellet et al., 2012; EFSA, 2020; Humblet et al., 2016). Additionally this access can be hindered by the limited knowledge of computer science (translating PDF or HTLM format to EXCEL or text using text mining) or heavy manual work required by the conversion as stated by Humblet et al. (2016). Although raw tabulated data (e.g. EXCEL and TEXT files) are more appropriate for risk assessment, these sources are not often available and sometimes difficult to access (e.g. restricted or paying access) (EFSA, 2020).
Preferred forms of data where Excel and PDF, but as stated by Humblet et al. (2016), the main forms of data they found were PDF and HTLM files. Although raw tabulated data (e.g. EXCEL and TEXT files) are more appropriate for risk assessment, these sources are not often available and sometimes difficult to access (e.g. restricted or paying access) (Bellet et al. 2012).
Data availability and accessibility are crucial for epidemiological analysis. Availability of the data, more than its accessibility, is the main issue for experts and research scientists/assessors. The data format plays a key role in the feasibility and rapidness of data management and analysis. The HTML format allows easier management of data than PDF files because it is more appropriate for data extraction; PDF data are better adapted to consulting only (Bellet et al., 2012). Additional training skills and collaborations though multidisciplinary disciplines could help in overcoming the issues on accessibility to the right form of data and also its availability.
Harmonization of animal health systems, in regard to data collection and accessibility is encouraged, to provide useful and reliable data, both at the national and the international levels for both animal and human health.
Risk assessment plays an important role in in risk of introduction of animal diseases. These are mostly carried out based on available data and an animal health information system is only as good as the data it contains (FAO, 2011; Humblet et al., 2016). However, most of the data required to fully evaluate the extent of a health issue, are generally not available or non-existent. Owing to the lack of relevant data and the very short period of time usually allowed to assess animal health risk on particular topics, many institutions use a qualitative risk method for evaluating animal health risks or crises (Dufour et al., 2011). For this reason, qualitative risk assessment is more in use as reflected in the answers of the respondents. There was no difference between the quantitative and semi-qualitative approach used, which is to be expected, as the semi-qualitative approach could be considered as quantitative.
The risk assessment question following the definition by Dufour et al., (2011) was divided into 3 categories: release assessment (estimation of the likelihood of a hazard being introduced in a particular zone); exposure assessment: estimation of the likelihood of susceptible humans or animals being exposed to the hazards) and consequences assessment: describing the results of the release and exposure of the hazard for humans and animals (health and/or economic consequences). Most of the respondents worked in those 3 categories which combined produce a risk estimation (Dufour et al., 2011). This is consistent with the three most important features they require of a risk analysis platform: a spread assessment, pathways of introduction of a disease up to the border and a quick risk assessment. Further, this corresponds to the fact that scientific panels must often make their assessment over a very short time period, from a couple of days to a few weeks (De Vos et al., 2019; Sharma & Baldock, 1999). Moreover, most commonly, risk assessments are developed to assess the risk for a single disease and risk introduction pathway (De Vos et al., 2019).
As to the perceptions on using animal health platforms or risk analysis tools, the experts survey showed that data accessibility is key, which was also the main issue encountered by the respondents. Difficulties in understanding the page could be due to the fact that a page was not adapted to the respondents’ countries’ needs and there may have been a language barrier.
The feature that they most looked for was again data accessibility and availability and being able to extract this information and its results. Comments stated that certain platforms do not allow for ease of data downloading (e.g. the data had to be copied from the page and pasted in Excel which is time consuming and prone to errors). The display of the information and its extraction is the main limiting feature
As previously mentioned, experts’ location and the one from where they carried out their professional activity both widely affected the efficiency of their interaction with the platform. No assumption on the reason of such a limitation per country could be made from this survey. However, both limited internet connection and knowledge on numerical technologies can be listed. It would benefit future research to compare the functionality of different national health systems. Experts could be asked what the constraints of their own national health systems are and if they know how different it can be from other national systems. Also, they could be asked if they think that standardisation of made efforts can help to improve the effectiveness of such systems. “One Health” is now a goal for the scientific community. However, non-standardised efforts, surveillance systems and collected dataset are still highly limiting.
The user’s satisfaction for using platforms remains high suggesting that the platform choice is not only related to the required information. The platform functionality per se also attracts the user. A focus on increasing the platform functionalities and customising its interface can therefore lead to a higher usage. Providing user friendliness remains one of the most important points to be addressed. A suggestion could be to add to the platform a good online training course.
Global animal health information systems were the most mentioned during the survey. The main one was the World Animal Health Information System (WAHIS) (OIE, 2019) which is the global animal health information system operated by the OIE to handle disease notification and reports from member Countries. It is mandatory for members to report disease events from the notifiable diseases list to the OIE through this system. The second one was EMPRES-i (FAO, 2014) developed by FAO and available for public access followed by Pro-MED which is hosted by the International Society for Infectious Disease and is also publicly available. National animal health information systems were also mentioned, but not specified. Although not on a global scale, these are as important as the international ones. Sharma and Baldock, (1999) described them as:” the complete system responsible for handling information about the health of livestock on a country”. Therefore, if there were better access to these animal health information systems, it would be very useful for research professionals in non-government institutions who would not normally see these data due to governmental restriction or privacy settings on its access. In addition, animal health information systems should also be used to handle information about non-production domestic animals (such as pets) and wildlife. This question was not asked, but for future works, it would be interesting to know if there is such data and how accessible the information is. This situation though can only be applied in countries where a good surveillance system is in place and data is collected and collated into a computerised system. Not all countries have such infrastructure, which makes professionals rely on global systems, particularly WAHIS.
The preference of platform does not improve the level of satisfaction. This could be because the choice of platform is mostly focused on the information available on it, more than finding the platform extremely good.