Abstract from 1996 SRA-Europe Annual Meeting

Summary of Meeting Paper

The 1996 Annual Meeting of the Society for Risk Analysis-Europe

An Analysis of the Risk of Introduction of Porcine Reproductive and Respiratory Syndrome to Ireland. Richard O'Flaherty, MVB, MRCVS, Veterinary Officer, Department of Agriculture for Northern Ireland, Crown Buildings, Bridge Street, Newry, County Down, BT34 3LU,Northern Ireland

Introduction

Porcine Reproductive and Respiratory Syndrome, a virus disease which affects pigs, is important for a number of reasons:

The disease has spread across Europe since 1990. To date it has not been introduced to the island of Ireland.
The pig industries in both parts of the island recognise the threat that the disease poses to their profitability.
European legislation does not permit either Northern Ireland or the Republic of Ireland to demand freedom from the disease as an import requirement.
A voluntary code of practice, introduced by pig producers, requires importers all pig to seek certain guarantees of disease freedom before they import any pigs.

The analysis focuses on the threat of the infection being introduced to the island of Ireland through the importation of five pigs from Great Britain and considers the value of the voluntary code as a means of preventing disease entry.

Porcine Reproductive and Respiratory Syndrome

The disease syndrome is caused by a group of viruses known as the PRRS viruses. These are RNA viruses closely related to Equine Arteritis Virus, Lactic Dehydrogenase Virus of mice and Simian Haemorrhagic Fever Virus. They all share the ability to replicate within macrophages (tissue defence cefls) and to cause persistent, often asymptomatic, infections.

There is no specific treatment for the acute effects of the disease. Once introduced to a farm the costs of eradicating the disease are high. Control by means of vaccination is of limited value since new strains of the virus are evolving constantly. National and international attempts to control the spread of the disease have been varied and largely unsuccessful.

While the virus has been shown to be capable of airborne spread and transmission in semen or embryos, the most likely route by which it would be introduced to Ireland is by the importation of asymptomatic, infected pigs. Live pigs may, in theory, be imported into Ireland from any of the EU member states as well as from Third Countries. In practice, only a few pigs are imported, mainly as replacements for the breeding herd. Those pigs which are imported originate almost exclusively in Great Britain.

Table 1 The Major Effects of PRRS:

Economic Costs:

Reduced growth rate

Reduced food conversion efficiency

Increased incidence of secondary disease

Increased abortion rate

Infertility

Diagnostic, treatment, control and prevention costs

Disruption of production flow

Reduced carcass and breeding stock quality

Economic Benefits:

Higher pig prices (due to reduced supply)

Management or housing problems may be rectified

Trade Effects:

The effect on trade in pigs, semen or meat will depend on:

Trade control measures; e.g. may affect the importation of breeding stock

Demand may depend on the infection or immunity status of the herd

Effects on competitors

Business uncertainty increased

Welfare:

Animal welfare aspects of illness and deaths

Procedure

The procedure follows the guidelines laid out in the OIIE review of risk analysis as it is applied to trade in animals. Having initially studied the available literature on the disease, the risk model was developed. Next, data was collected and entered into the model.

Mathematical Model

The model attempts to simulate the movement of an infected pig into a native herd (see Figure 1).

The probability that infection would by-pass each step was defined mathematically. The derivation of these definitions follows the outline given in the OEE review.

The overall risk estimate is the product of the likelihood of evasion at each step. The equation linking these probability estimates takes the following form:

Risk Estimate = {1-[1-(pl x p3 x p4 x p5 x p6)]ⁿ) x p2 x p7

Where,

px is the probability of infection being present after step x.
n is the number of pigs in the consignment to be imported.

Assigning Values

The model demanded data from a variety of sources. These included, characteristics of the infective process, the prevalence of the infection and the sensitivity of the diagnostic blood test. These values were determined by reference to the scientific literature. Other data was obtained from census returns, trade figures, published reports and DANI records.

Where no fixed numerical value could be assigned, the computer software allowed a distribution of acceptable values to be used.

Analysis

The model was created on a Microsoft Excel spreadsheet and analysed using @Risk, a risk analysis software package published by the Palisade Corporation, Newfield, New York.

The risk was estimated in two ways:

1. The risk management steps 2 - 6 were not enforced (i.e. px = 1). This is equivalent to the level of control defined by European legislation. It is referred to as the Unrestricted Risk Estimate.

2. Values for steps 2 - 6 were calculated from input values. This simulates the effect of enforcing the import requirements specified by the industry's voluntary code of practice. It is referred to as the Restricted Risk Estimate.

This allowed the two import policies to be compared.

Results

The simulations demonstrated the effectiveness of the voluntary code in preventing the disease from entering Ireland. The unrestricted risk estimate exceeded the restricted estimate by a factor of 10,000.

@Risk allowed the input values to be correlated against the output distributions. The variables which appeared to have the greatest influence on the restricted risk estimate were the prevalence of infection in an infected herd, the sensitivity of the blood test used, and the likelihood of clinical detection during the quarantine periods.

Discussion

The purpose of animal import risk analysis is as much to provide a framework for discussion as it is to calculate accurate probabilities for the likelihood of agent entry. Assumptions made in the development of the model may indicate areas where more research is required. Similarly, analysis of the sensitivity of outputs to changes in input highlights those variables where the highest degree of accuracy is required.

Where possible, the validity of input data should be checked by referring to multiple sources. This appears to be particularly true of reported disease incidence and international trade records.

Acknowledgements

I am particularly indebted to Dr. Marion Wooldridge for her support and guidance at all stages. I must also thank Mr. John Wilesmith, Dr. Richard Clifton Hadley, as well as the pathologists, virologists and librarians within the Central Veterinary Laboratory in Weybridge who directed me in my research of the subject.

Review literature

Wilson-DW; Banks-DJD. (1993). The Application of Risk Assessment in Animal Quarantine in Australia. Revue Scientifique et Technique, OEF. 1993; 12, 4, 1121-1133.

Meredith-MJ. Review of porcine reproductive and respiratory syndrome (1995). Pig Disease Information Centre, Department of Clinical Veterinary Medicine, University of Cambridge, Cambridge CB3 OES, LTK.