The overall objective of this endeavor was to develop and apply new tools and algorithms to swine diagnostic data as a means for enhancing the existing systems of monitoring the health of Iowa and US swine.
The number of porcine submissions to the ISU VDL and respective disease diagnosis information based on diagnostic codes (Dx codes) were recovered from the ISU VDL laboratory information and management system (ISU LIMS). Data on pathogen detection reported to the Swine Disease Reporting System (SDRS) from the state of Iowa were recovered and used to monitor agent detection in Iowa. These databases were analyzed and different algorithms were developed and applied for each purpose. To scan the historical number of porcine submissions received at ISU VDL each month, and to forecast the expected number of porcine submissions to be made each month for the coming year, an additive Winters model with logistic transformation was used. To monitor disease diagnosis at ISU VDL, an EARS-C1 algorithm was applied to the database organized by a physiological system, and by pathogen. Moreover, SDRS data of cases submitted from Iowa was monitored using a cyclic regression model for weekly proportion of PCR-positive cases, and to forecast the expected upcoming weekly results for Porcine Reproductive Respiratory Syndrome Virus (PRRSV), Porcine Epidemic Diarrhea Virus (PEDV), Porcine Deltacoronavirus (PDCoV) and Mycoplasma hyopneumoniae (MHP).
Throughout the course of the project period, a novel and greatly improved system and associated software application for incorporating Dx codes into tissue-based (sick pig) case records was developed. Each Dx code is composed of four individual components (i.e., System, Insult type, Tissue or Lesion, and Disease or Etiology). One to any number of Dx codes can be assigned per case. Dx codes are subjective (professional) assessments provided by the diagnostic pathologist responsible for coordinating the case. Each component of a Dx code is a discrete categorical data point, making such data well suited for subsequent summary and analysis. This new system of coding was retrospectively applied to historical tissue-based (sick pig) case submissions, and used to facilitate the monitoring of disease diagnosis objective of this project. Going forward, this new Dx coding system, coupled with the use of state of the art business intelligence (BI) software for summarizing and visualizing such data, will greatly enhance the ability to monitor historical, emerging, and remerging trends in Iowa and US swine.
Overall, the algorithms were able to capture changes in the pattern of porcine submissions to the ISU VDL over time. The scanning of disease diagnosis was able to capture changes in the number of diagnoses by different systems or agents at different weeks during the year of 2019. The cyclic regression model, used to scan state of Iowa information present in the SDRS database for the cyclic pattern of agent detection, was able to characterize a clear seasonal pattern of detection of PDCoV, PEDV, PRRSV, and MHP. During different weeks of 2019, signals were issued for increased detection of PDCoV, PEDV, and PRRSV.
This project led to the development and application of new tools and algorithms to monitor diagnostic information for signals of emerging and/or re-emerging pathogens in Iowa and US swine. The algorithms and new web-based systems being established aim to provide a platform for Iowa veterinarians and producers to have state-level information, and use it to compare with agent detection at a national level reported in the SDRS project. Veterinarians and producers can compare this aggregate level data (state or nation-wide) to observations and/or existing datasets of their own production system or veterinary clinic. The models and tools developed will continue to be evaluated, updated, and improved over time. When facing signals for increased agent detection, veterinarians and producers can reinforce biosecurity compliance measures as a way to prevent further spread of the agents. The detected signals were periodically discussed with the SDRS Advisory Council during the year of 2019 and the relevant findings were reported in the SDRS monthly written reports, audio reports, and video reports. Monthly SDRS reports are currently available at the Swine Health Information Center (SHIC) webpage https://www.swinehealth.org/domestic-disease-surveillance-reports/. Additionally, the PDF, audio, and video reports can be accessed at the SDRS project webpage https://www.fieldepi.org/sdrs. Interested individuals can sign up to receive the monthly PDF and audio reports by e-mail, by submitting a request to the SDRS project.