Improving Public Health with Artificial Intelligence: What Is The Potential?

Public health is the science and practice behind the protection of the health of a community, and running an effective system requires the collaborative effort of many different sectors. Public health can help prevent the spread of disease by screening and treating it, protect the community from environmental health hazards, promote health education through various channels, and provide social support to ensure a standard of living for everyone. It is everything a nation does together to create conditions in which everyone can be healthy.

Periodically, circumstances arise that cause dramatic shifts in the way public health operates. The COVID-19 pandemic is a prime example of this, where several key lessons were learned that changed the way many health protocols operated, including infection prevention and control, the recognition of epidemics as “standing threats”, and addressing weaknesses in the aged care system. 

Artificial Intelligence (AI) presents another opportunity to transform the public healthcare system positively. AI is a technology that enables computers and other machines to simulate human learning to learn, reason, problem-solve, decision-make, and create, providing unprecedented opportunities in the health sector. For many professionals and those still studying requisite courses such as a graduate certificate in public health, AI represents the future of public health; however, organisations need to consider AI thoughtfully and develop strategies for its implementation rather than simply jumping on the AI powerhouse bandwagon.

How AI can be used in public health

Broadly, three key benefits come from utilising AI in public health: better patient outcomes, greater public health surveillance, and improved healthcare education. 

Patient outcomes

The availability of data for health has increased significantly over the past few decades, providing detailed insights into social, behavioural, and environmental factors on a much grader scale. Data from social media, search engines, community forums, news media, mobile devices, apps, and wearable technology can all be collected to form a much more accurate picture of the state of public health than through older, more traditional methodologies. 

Even environmental sensors detailing the condition of air pollution, water quality, environmental noise, and weather conditions contribute and provide valuable insights. Using AI, this vast amount of data can be analysed in real time, which can then inform public health officials how the right intervention can be targeted to the correct population at the appropriate time to provide better patient outcomes. 

Public health surveillance

Traditionally collected via population health surveys, clinical data, and other public health reporting systems, public health surveillance in the past could be limited in its ability to identify fast-approaching and emerging threats. With access to new data sources, including news media, online data, and mobile device data, and the AI capability to analyse large amounts of data quickly, the opportunity to identify these emerging threats becomes faster while also painting a more detailed picture of the population diseases and risk factor distributions. 

AI can also assist in summarising surveillance data from unstructured sources. This might include identifying free-text information in sources such as death certificates or detecting and tracking infectious disease outbreaks via commercial flight itineraries.

Healthcare education

AI can also help with healthcare education by promoting activities to a targeted cohort more efficiently. For example, in a study on identifying key target audiences for public health campaigns, individuals with mixed opinions on hookah tobacco smoking were identified using a sentiment analysis on Twitter via AI, and a targeted campaign was built and aimed at those who were considered the most receptive. The sentiment analysis allowed researchers to identify ideal populations for public health intervention who would be most receptive. 

Healthcare education can become better targeted with AI integrations. 

Challenges of AI in public health

Despite its many benefits, there are also some challenges, risks, and limitations in using AI for public health. This may include;

• Lack of diversity: Without diverse datasets, there is a risk that AI could reinforce public health inequalities, particularly in rural and underprivileged communities.
• Data privacy: With AI given vast amounts of sensitive health information, the risk of a data privacy breach is higher, while the misuse of health data is already occurring, bringing to light ethical concerns.
• Lack of oversight: A lack of proper oversight when using AI, including appropriate infrastructure and a lack of workforce training, can result in poor model interpretability, inappropriate usage, and further data breach risks. 

Implementing AI responsibly 

Improving the public health system is no easy task, but its importance cannot be overlooked. Implementing AI can have a huge impact on this goal, and while it is clearly the way forward, it must be leveraged responsibly to ensure the best outcomes for everyone involved. 

To mitigate the potential risks associated with AI and implement it responsibly, considerations that should be built into its framework include; 

• Ensuring data systems are upgraded: AI is only as effective as the data it is trained with, so ensuring it is clean, accurate, and diverse is crucial to implementing effective AI processes. Starting with smaller test projects before scaling up is ideal. 
• Continuous monitoring: AI has come a long way, but it still requires human intervention and monitoring to ensure it meets all ethical and operational requirements. 
• Workforce training: Users of AI systems must be appropriately trained to manage the programs to optimise efficiencies, identify issues, and avoid misuse.
• Clear accountability: Before AI is implemented in a workplace, it must receive oversight and sign-off from various teams, including legal, ethical, and technical experts. 
• Real-time feedback: One of the key benefits of AI is its ability to provide data in real time, so clinicians, patients, and other users who may input data must also be working in real time to ensure data is as accurate as possible. 

If the integration of AI into public health systems can be done responsibly, it can truly transform the landscape. The potential to deliver measurable, impactful results across the board necessitates its existence in public health, and with proper implementation and regulation, will help improve the system for years to come.