Document Type

Presentation

Loading...

Media is loading
 

Publication Date

6-28-2024

Comments

Presentation: 4:47

Poster attached as supplemental file below

Abstract

Public health surveillance refers to the detection, and identification of potential threats to public health through data collection, analysis, and interpretation. Data are typically from various entities including healthcare systems, various governing bodies and organizations, and public or consumer data. Through the integration of these disparate datasets, a comprehensive landscape can be created that is reflective of the public health population. Effective public health surveillance is limited today due to data fragmentation, data silos, antiquated systems, data quality and access constraints, and real-time data limitations. The application of new data techniques and technologies like artificial intelligence (AI) have been proposed to bridge these gaps.

In order to assess the viability of using AI to support data integration to improve public health surveillance, a rapid scoping review was conducted. A total of 91 studies were identified in PubMed and Scopus, which after screening, resulted in 4 studies to be included for analysis. All four studies applied various AI methods to combine disparate data sources (including both traditional and non-traditional sources), and successfully created AI models that could accurately predict, forecast, or trend for disease and health outbreaks. Although the potential of AI was demonstrated via these studies, further research is needed around overcoming the limitations of AI, which include the lack of AI model standardization, lack of AI model transparency, and ethical and privacy concerns. Addressing these shortcomings can help advance public health surveillance to be more effective and proactive in its response to emerging public health threats.

Lay Summary

Public health surveillance is the detection and identification of potential health threats or disease outbreaks by collecting, analyzing, and interpreting data. This data comes from many sources like hospitals, insurance companies, government organizations, and even public or consumer information. When you combine all of these different types of data, one can have a more comprehensive picture of the general health of the public.

There are several challenges that limit the effectiveness of public health surveillance today. Data are often faced with challenges around fragmentation, outdated storage systems, varying levels of quality, difficulty in access, and not reflecting real-time activity. New technology like AI can help fix these issues.

A rapid scoping review was used to see if AI can help combine data to improve public health surveillance. Out of 91 studies found in two major databases (PubMed and Scopus), 4 studies were chosen for analysis. All 4 studies used AI methods to combine different types of data, and successfully created AI models that could accurately predict and track disease and health outbreaks.

These studies showed that AI could successfully be used to improve how data can be used in public health surveillance. However, more research is needed to address limitations such as the lack of standardization and transparency in AI models, and concerns around ethics and privacy. Resolving these limitations can help unlock the full potential of AI in improving public health surveillance through data.

Additional Files
202406_Lee_Paul_poster.pptx (375 kB)

Language

English

Download Presentation

Submit Feedback on Presentation

Share

COinS