Control of foot and mouth disease for the Malaysia-Thailand-Myanmar peninsula campaign: uses of quantitative risk assessment and epidemiological modeling

Abstract

Livestock industries in the countries of South East Asia (SEA) have suffered from the presence of foot and mouth disease (FMD). Realizing that FMD is transboundary, SEA countries have formed a regional collaborative organization utilizing a progressive zoning approach to assist in the FMD eradication. A campaign involving three countries, including Malaysia, Myanmar, and Thailand, called MTM, has been initiated. To assist the campaign in achieving its ultimate goal, the MTM Tri-state Commission identified the necessity of the risk analysis study. Hence, a quantitative risk assessment (QRA) study was conducted in the MTM peninsula to assess the likelihood of an introduction to FMD through importation of livestock. Patterns of livestock movement were ascertained through reviewing the relevant governmental records, and by interviewing farmers, traders and local officers when the records did not exist. Movement regulations were reviewed and the information was used to construct QRA scenario trees. QRA processes began by identifying model parameters and their values. The parameters were probabilities of livestock having FMD and of FMD infection undetected during import processes. Given current import regulations, the study determined the probability of a livestock accepted for import having FMD. The sensitivity analysis was carried out to determine the effects that each parameter had on the model. The model simulations were run in @Risk® for Excel® software using a stochastic approach. The simulations yielded the most-likely prevalence at 8%. On average, each animal in a quarantine station would have 2.7% chance of having an unapparent form of FMD infection; hence will not be identified as infected. The findings showed that the average probability of an accepted livestock carrying FMD was 2.9%, and the risk was as high as 11%. A compartmental epidemiological model was developed to assess the impacts of the introduction of FMD and FMD control programs in Southern Thailand. The values for the model parameters were ascertained by reviews of literatures and records retrieved from Department of Livestock Development (DLD). Outbreak data was gathered from reports submitted to DLD and from the database of the Southeast Asia Foot and Mouth Disease (SEAFMD) Regional Coordinating Unit (RCU). The value for the transmission rate was acknowledged by comparing observed to expected outbreak data using least square method. Stochastic simulations were also run in the @Risk for Excel® software. The sensitivity analysis for annual cumulative incidence (ACI) was performed; and the effects of vaccination (E), stamping-out policy (K), and sensitivity of the screening processes during quarantine (PS) were examined. The significances of the parameters used in the sensitivity analysis were evaluated using a partial rank correlation coefficient method. The results showed that all parameters revealed negative correlations with ACI, but only PS was considered significant (p-value < 0.0001). This suggested that prevention of the introduction of FMD was more important than dealing with the disease once it enters the areas. Further analyses for the alternative control options were carried out. Five scenarios were created to imitate different FMD control strategies in Southern Thailand. The baseline scenario (BS) was constructed emulating current control strategy used in the south, and was used as the comparison point for all other subsequent scenarios. Eradication Scenario (ES) I was created for the investigation of improved sensitivity during the disease screening process, i.e. assimilation of NSP test during the quarantine. Having created much debate in recent years, the impact of mass vaccination was examined in the ES n, where 80% vaccination coverage was assumed to be implemented randomly in the MTM zones in Thailand. ES III evaluated an integrated use of NSP test and vaccines; and ES IV investigated an addition of the implementation of the stamping-out policy to the ES III. Approximately thirty-five thousands cases were expected from the BS. About 30% and 85% reduction of AGI from the BS were observed for the ES I and ES II, respectively. ES III resulted in the 96% ACI reduction; and slightly more drop in the ACI (98%) was seen in the ES IV. The results from the analyses suggested that the vaccination had more impact on the ACI than the use of NSP test. It was indicative that the use of NSP test reduced the ACI during the peak season, where the vaccination diminished the underlying incidence of FMD. The model indicated that the best mitigation strategy was the integrated and strategic use of control techniques.