Role of Artificial Intelligence in Zoonotic Disease Surveillance
Shivika Gupta1, Aman Singh2, Abhishek Verma2, Ankit Shukla2, Shivangi Tripathi2
1PhD Scholar, 2MVSc Scholar
Department of Veterinary Medicine, DUVASU, Mathura
Presenting Author- Shivika Gupta
email: shivikagupta975@gmail.com
Abstract:
The increasing interaction between humans and wildlife, coupled with factors such as climate change and global travel, has heightened the risk of zoonotic outbreaks. Traditional surveillance methods, while effective to some extent, often struggle with the speed and accuracy needed to detect and respond to these diseases promptly. Artificial Intelligence (AI) offers a transformative potential in enhancing zoonotic disease surveillance, providing tools for rapid detection, accurate prediction, and efficient response. AI-driven predictive models leverage extensive datasets to predict disease outbreaks and transmission patterns, thereby facilitating proactive public health responses. Ultimately, harnessing the power of AI to combat zoonotic diseases holds promise for improving global health outcomes, reducing morbidity and mortality, and preventing future pandemics.
Keywords: Artificial Intelligence, zoonotic diseases, global health, surveillance, detection
Introduction
Zoonotic diseases are infectious diseases that can be transmitted between animals and humans and pose a significant threat to global public health. The emergence and re-emergence of zoonotic diseases, such as coronavirus disease 2019 (COVID-19), Ebola virus disease (EVD), and monkeypox, highlight the need for innovative approaches to enhance disease prevention, early diagnosis, and effective treatment.
Artificial intelligence (AI) refers to the ability of computerized systems or computer-controlled robots to execute tasks typically associated with intelligent beings. The integration of AI techniques with conventional disease control strategies offers novel prospects for understanding, predicting, and mitigating the impact of zoonotic diseases. By leveraging advanced algorithms and machine learning (ML) models, AI can analyse vast amounts of complex data from diverse sources, ranging from environmental factors to genetic sequences, enabling researchers and public health authorities to make more informed decisions and implement proactive measures. In recent years, artificial intelligence (AI) has emerged as a powerful tool in the field of healthcare and has shown great potential in addressing these challenges. Furthermore, AI has revolutionized early disease diagnosis by improving the speed and accuracy of disease detection. AI can be used to analyse intricate patterns and biomarkers found in medical imaging, genetic data, and clinical records to identify potential zoonotic infections in their early stages. Early diagnosis not only enhances patient outcomes but also aids in containing the spread of diseases by enabling prompt isolation and treatment.
Figure 1: Diverse sources that provide data to Artificial Intelligence-powered health systems.
Application of AI in epidemiological surveillance
Epidemiological surveillance involves systematic gathering, analysis, interpretation, and sharing of health data, with the goal of preventing and controlling diseases. As per surveillance, AI can be leveraged in determining the ability to detect pathogens, helping research scientists and veterinarians prioritize samples or cases with the potential of being positive, thereby managing resources and laboratory capacities and maintaining focus on relevant samples. AI can enhance traditional surveillance systems by improving the speed and accuracy of data collection and analysis. Automated systems equipped with AI can monitor wildlife populations, track animal movements, and detect anomalies that may indicate the presence of zoonotic pathogens.
Anthrax is an acute infectious disease affecting humans and animals, caused by the bacterium Bacillus anthracis. Early detection of anthrax outbreaks is crucial for minimizing the number of cases, deaths, and the risk of disease spread. A recent study aimed to develop a disease prediction model using ML techniques to forecast anthrax outbreaks in livestock in Karnataka. The goal was to achieve early detection of anthrax cases. The authors employed an ML model developed with version 3.1.3 of the R statistical software. They combined various data mining regression and classification models, including generalized linear models (GLMs), generalized additive models, multiple adaptive regression splines, and flexible discriminant analysis (FDA). Anthrax occurrence data from the Animal Husbandry Department in Bangalore, Karnataka, India, were utilized.
Leptospirosis, a zoonotic disease influenced by weather and environmental changes, can be detected early through the analysis of diverse data sources using AI algorithms. These sources include weather patterns, environmental factors, and animal populations. Rahmat et al. conducted a study that analyzed, captured, and predicted the occurrence of leptospirosis by combining data mining and ML techniques. Their specific focus was on the relationships between the disease and temperature, rainfall, and relative humidity. The study began with exploratory data analysis (EDA) using graphical methods to ascertain the optimal time lag for rainfall analysis. In contrast, non-graphical methods were employed for temperature analysis.
AI technology has demonstrated its invaluable role in predicting potential hosts for EBOV, aiding researchers and public health authorities in understanding and mitigating the risk of EBOV transmission from wildlife to humans. This proactive approach played a crucial role in preventing Ebola outbreak. Monkeypox is a zoonotic disease characterized by fever, rash, and swollen lymph nodes, with potential health consequences including skin lesions and scarring. Recently, this has become a prominent topic in zoonotic disease research. One study proposed a prediction method for monkeypox using ML techniques.
Rift Valley fever (RVF) is a zoonotic viral disease that can cause mass die-offs in livestock and has a high fatality rate in affected human populations. Despite its significant impact, little is known about the occurrence of RVF and factors influencing its transmission. One study aimed to conduct data-driven epidemiological modelling of RVF using neural network techniques. By integrating landscape archaeology, historical evidence, climate data, and human behavioural evidence collected through ethnoarchaeological research, the study proposes a human-animal paleopathology application framework. This framework analyses the inherent connections between diseases and ecological and social factors, aiding in addressing the threat of zoonotic diseases resulting from climate warming.
Viral zoonotic diseases pose serious threats to human and animal health. Choubey et al. used an enhanced backpropagation ANN (EBP-ANN) algorithm to predict and mitigate the adverse impacts of viral zoonotic diseases on human health. Viral datasets were collected and pre-processed using Z-score normalization. Subsequently, they extracted viral data features using a dynamic angle projection pattern and employed GAs to select more accurate feature data.
In addition, a GIS-based adaptive neuro-fuzzy inference system was employed to explore the spatial distribution patterns of human brucellosis (HB) in Mazandaran Province, Iran. A linear regression model revealed that these parameters influencing HB incidence were independent of each other and could only account for 25.3% of the total variation in the HB incidence logarithm. Pearson’s correlation analysis showed the strongest positive correlation between vegetation and the logarithm of population size and the number of HB cases.
Solutions to improve response to zoonotic disease outbreaks
Robust response systems are crucial during zoonotic disease outbreaks. These include early warning systems, communication, information sharing platforms, point-of-care diagnostics (POC), telemedicine for remote care, and digital contact tracing.
Early warning systems play a crucial role in the prevention and control of zoonotic diseases. The early detection and warning of poultry diseases significantly improve animal welfare and reduce losses. Current early warning technologies offer the capability to continuously and automatically monitor the health status of chickens in a non-invasive manner. Development of Data mining methods and the Dempster-Shafer evidence theory to develop an intelligent poultry disease warning device. The device uses a fast fourier transform and discrete wavelet transform to process the sound signals of chickens in both the frequency and time-frequency domains. It can differentiate between Newcastle disease, infectious bronchitis virus, and avian influenza, and monitor chickens infected with the viruses within two days, achieving an accuracy rate of 91.15% and serve as an early warning system.
Figure 2: Overview of the various applications of artificial intelligence (AI) technology in zoonotic diseases.
Challenges and Ethical Considerations
Implementation of modern technologies and solutions to enhance surveillance and response systems for emerging zoonotic diseases presents several challenges. These challenges range from logistical and technical difficulties to financial constraints and regulatory hurdles. While AI offers significant advantages in zoonotic disease surveillance, it is not without challenges and ethical considerations. Data privacy and security are paramount, as the integration of diverse data sources can lead to concerns about the misuse of sensitive information. Ensuring the accuracy and reliability of AI models is also critical, as erroneous predictions or misinterpretations of data can have serious consequences.
Ethical considerations include the potential for AI to reinforce existing biases in data collection and analysis. It is essential to develop AI systems that are transparent, fair, and accountable, with mechanisms in place to address any unintended biases or errors. Collaboration between AI experts, epidemiologists, public health officials, and policymakers is crucial to address these challenges and ensure the responsible use of AI in zoonotic disease surveillance.
Conclusion
AI has the potential to revolutionize zoonotic disease surveillance, offering tools for rapid detection, accurate prediction, and efficient response. By enhancing traditional surveillance systems, providing predictive analytics, and supporting rapid response efforts, AI can significantly reduce the impact of zoonotic diseases on global health. However, the successful integration of AI in this field requires addressing technical challenges, ethical considerations, and ensuring collaboration across disciplines. As AI technology continues to advance, its role in zoonotic disease surveillance will likely become increasingly critical in safeguarding public health and preventing future pandemics.
References
Anjaria, P., Asediya, V., Bhavsar, P., Pathak, A., Desai, D.; Patil, V. (2023). Artificial Intelligence in Public Health: Revolutionizing Epidemiological Surveillance for Pandemic Preparedness and Equitable Vaccine Access. Vaccines. 11, 1154. https://doi.org/10.3390/vaccines11071154.
Dhama, K.,Chakraborty,S., Kapoor,S., Tiwari,R .,Kumar, ., Deb, A.,Rajagunalan, S.,Singh, R., Vora, R., Natesan, S, (2013) One World, One Health – Veterinary Perspectives. Adv. Anim. Vet. Sci., pp. 5-13.
Pillai, N., Ramkumar, M., Nanduri, B. (2022). Artificial Intelligence Models for Zoonotic Pathogens: A Survey. Microorganisms. 10 (10):1911. doi: 10.3390/microorganisms10101911.
Zeng, D., Cao, Z., Neill, D.B. (2020). Artificial intelligence–enabled public health surveillance—From local detection to global epidemic monitoring and control. In Artificial Intelligence in Medicine; Academic Press: Cambridge, MA, USA, pp. 437–453.
