CHIP: Commodity based Hazard Identification Protocol for emerging diseases in plants and animals

Report(9.5 Mb)

Summary

Introduction and objectives

In plant health during the last decade, both EPPO (European and Mediterranean Plant Protection Organization) (EPPO 2011) and EFSA (EFSA 2010) have developed pest risk assessment schemes, which are largely comparable. In animal health, guidelines for import risk analysis have been developed by the OIE (OIE 2004), and a risk assessment scheme for vector-borne diseases is being developed in the Netherlands (Vos et al. 2011). These assessment schemes are agent- or organism-based, which implies that they can be applied only to identified pathogens/pests. However, the fact that trade is a major pathway for the entry of exotic plant and animal pathogens/pests increases the need for a protocol to determine the risk profile of commodities.

Methodology

The decision tree was developed in a structured process that consisted of the following stages:

1. Preparatory research. This comprised:

1. The development of a pathway model to enable the risk assessor to collect and structure all necessary information regarding the production, processing, transport and storage processes that affect the likelihood of the entry of pests/pathogens into the country of destination.
2. The compilation of a commodity list. All commodities of plant and animal origin were selected and stored in a database, including characteristics that affect the likelihood of pest/pathogen association with the commodity, survival during transport and storage, and potential contact with local hosts after entry into the country of destination
3. An extensive review of scientific literature on pest/pathogen characteristics that affect the likelihood of commodity contamination, survival of the pest/pathogen during commodity processing and transport, and the likelihood of infecting a local host once introduced. This resulted in the following basic pest/pathogen characteristics to include in the decision support tree:

1. The location where the pest/pathogen survives on the commodity

2. The capacity to withstand treatments designed to eliminate pests/pathogens on commodities.

4. A review of risk assessment schemes, which resulted in a list of risk factors that experts generally agreed upon and can be combined with the results of the systematic review for inclusion in the decision tree.
5. A review of hazard identification protocols to identify best practices for developing a hazard identification protocol in order to prevent pitfalls in the development of a decision support tree for the CHIP project.

2. In the second stage, the decision tree was developed on the basis of the following principles:

1. The use of the decision tree is triggered by a trade signal. Trade signals can be generated by sudden changes in commodity type, country of origin and the volume of the traded commodity. Therefore, the traded volume was not included in the decision tree.
2. The decision tree has a modular structure. This enables the risk assessor to perform a quick analysis and, if necessary, an in-depth analysis, but is not intended to replace a full import risk assessment. Level 0 comprises an assessment on the basis of the traits of the commodity: state of the product and its intended use. Level 1 consists of traits of the production and post-harvest processes that determine the likelihood that the commodity has been infested by any pathogen/pest before it is transported to the country of origin. Level 2 consists of a comparison between the countries of origin and destination and pest/pathogen characteristics that enable the infestation of local hosts.
3. The modules are kept as simple as possible, which means that the traits of commodities and pests are summarized. Addressing them individually would make the decision tree extremely complex and suggest a level of detail that could not be justified. Empirical evidence about the effect of each characteristic of a commodity and a pest/pathogen on the likelihood of entry is lacking (see Appendix 3). Furthermore, the decision tree is used as a tool to prioritize risk assessment. This requires a robust rather than a sophisticated approach: namely an approach that easily and roughly separates the wheat from the chaff.
4. The decision tree is restricted to the likelihood of the entry of pest and pathogens associated with agricultural commodities. Establishment, spread and impact are not included in the protocol. If the likelihood of entry of a certain pest or pathogen is high, the recommendation will be to conduct a full organism-based risk assessment in which those aspects are addressed. When building the decision tree, the following approach was applied. The decision tree integrated the results of the preparatory work as described before. The development of the decision tree was based on expert judgement and, where possible, on the results of the reviews. The characteristics with the most discriminating power were put the highest in the decision tree.

3. The results of this study were presented to and discussed with six international experts in two expert meetings.
4. The decision support tree was tested with six case studies: Pacific oyster from China, poultry from Thailand, sausage casings from Algeria, litchis from Madagascar, tomato seed from Mexico and trees from Canada.

Results

The decision support tree has the following characteristics:

1. It is generic, which means that it can be applied to all commodities of plant and animal origin.
2. The decision tree has a modular structure to facilitate both quick and in-depth assessments.
3. The program, which is in Excel, can be applied freely by stakeholders. The decision tree was described and documented in a user manual and the results were justified in the report.
4. The decision tree is applicable for the rapid screening of commodities in the EU. It considers the pathway starting from the country of origin to entry into the EU.
5. The characteristics with the most discriminating power are highest in the decision tree.
6. Three levels of likelihood are distinguished: high, moderate and low.

The decision tree was programmed in Excel. The case studies resulted in some adjustments. For the likelihood of survival during transport and storage, it soon turned out that this level did not have extra discriminatory power, because the survival of pests/pathogens is mainly determined by the state of the product and the preservation methods used, which stay the same during transport and storage, and therefore the transport and storage time does not have much influence. A few questions were reformulated. The case studies repeated with the final version of the decision support tree show consistency between the results of level 0 and level 2. A manual has been developed to assist the risk assessor to apply the model. All results are summarized in a report containing all questions, answers, scores, conclusions and decision rules.

Discussion

The decision tree is a prototype model. This implies that the emphasis in the development is on the structure rather than the content. The combination of plant and animal pests and pathogens in one model has proved to be possible. The major hurdle was the use of different terminology in each domain when referring to the same thing. The content needs additional research because the results of the systematic review are of limited importance. This is due to the lack of empirical evidence. The number of studies in which this has been studied is limited and they always focussed on certain pest/pathogen categories. However, it became apparent that the likelihood of pest/pathogen survival due to exposure to all potential circumstances during production, processing, storage and transport is the most crucial aspect to be addressed in the decision tree. Those relationships were mainly based on the expertise of the project team and the involved experts. However, this basis is rather limited, since it requires expertise both on all pest/pathogen categories and on the length of the period and intensity of the treatments the commodity is exposed to. Therefore, it is highly recommended to revise the scores on the basis of scientific literature review.

Although in this project the focus was on pests and pathogens that can cause diseases in animals or plants, the structure of the decision tree can also be applied to zoonoses and other biological hazards threatening human health. Level 1 will to a large extent be the same, although it must be noted that biological hazards affecting human health will in most cases not cause problems to plant and animal health. Level 2 will focus on the likelihood that commodities come in direct contact with humans, either by use of the commodity or by consumption. Furthermore a comparable decision support tree can be developed to assess the likelihood that commodities are contaminated with chemical hazards, threatening human health.