AI Solutions for Government Healthcare

Artificial intelligence (AI) is a powerful and disruptive digital technology that has revolutionized the way we operate across many industries and also has the potential to transform the way we look at healthcare delivery. According to a report by PwC, AI will contribute an additional $15.7 trillion to the world economy by 2030, and the greatest impact will be in the field of healthcare. AI solutions in Government healthcare bring strong capabilities in optimizing care delivery, reducing care costs, and improving outcomes to enable personalized and effective patient interactions. Automating routine tasks helps to improve overall administrative efficiency.

The US Department of Health and Human Services (HHS) has taken a lead role in incorporating AI solutions in government healthcare, acting in accordance with the Advancing American AI Act. The HHS has been focusing on encouraging AI adoption, enabling HHS-wide AI familiarity and fluency, promoting AI scaling, and sparking AI acceleration across the government healthcare sector. The ambition of HHS AI strategy is:

“Together with its partners in academia, industry, and government, HHS will leverage AI to solve previously unsolvable problems by continuing to lead advances in the health and well-being of the American people, responding to the use of AI across the health and human services ecosystem, and scaling trustworthy AI adoption across the Department.”

Acting as a regulator, investor, conserver, and catalyst, the HHS is set to play the leadership role in innovating health and human services. HHS will meet the dynamic needs of the American people and the healthcare sector through the trustworthy and ethical use of AI innovations.^[4]

Challenges to Implementing AI Solutions in Healthcare

Despite their tremendous potential and ongoing efforts by the federal government to embrace AI technologies, AI solutions in Government healthcare are not without their challenges. Besides the challenges related to data security and privacy, interoperability, legacy infrastructure, skill gaps, and cost considerations, AI solutions also have some methodological and technical shortcomings limiting their implementation and adoption. Some of the most prevalent challenges to implementing AI solutions in Government healthcare include:^[1][2][3]

• Healthcare data is complex, heterogeneous, multi-format, and often exists in aged and siloed legacy systems, which leads to challenges in data access and harmonization to a common format for use in training the AI models.
• Lack of data quality, lineage, and traceability can challenge the reproducibility of AI models for the same use cases over time.
• Algorithmic and data bias can lead to suboptimal performance and prediction errors that can impact end-user acceptance, trust, and compliance.
• Data privacy and security aspects related to dealing with healthcare and patient clinical data.
• AI solutions are most commonly seen as Black Box; hence, the lack of transparency can make it difficult for end users to trust the results.
• Regulatory gaps concerning liability and patient information-sharing issues.
• Building strategic capacity for change management, including resources for AI implementation and strategies for stakeholder collaboration.

The use of complex AI algorithms makes AI recommendations un-explainable, especially when they are to be used in public health decision-making. If left unaddressed, these challenges can impact the outcomes and predictions for community health programs- HIV and other STDs, Cancer, Mental Health, further making healthcare difficult to access. It can also result in inequitable distribution of resources across geography and ineffective clinical trials due to lack of representation and inclusivity.

The US Department of Health and Human Services has been working on overcoming these challenges and promoting the integration of AI technology within the government healthcare sector with an AI strategy that can help foster an AI-driven healthcare ecosystem, largely focused on the protection of patient individuals and fairness of AI systems (risk management) adoption.

Building AI Solutions for Healthcare- A Journey not just a Destination

With huge volumes of multi-modal data (genomics, economic, demographic, clinical, social, and administrative) being available to healthcare organizations coupled with digital technology innovations in mobile, internet of things (IoT), geospatial, and high-performance computing, a moment of convergence between healthcare and technology has happened to transform models of healthcare delivery through AI-augmented healthcare systems fundamentally. Public health systems are no different. But, the challenges – when it comes to incorporating artificial intelligence in healthcare systems – may seem to be daunting. However, a systematic approach – via collaboration between the public, private sector, and academia along with a common operating model– to build effective and trusted AI solutions can help reinforce that AI is here to augment – not replace – human intelligence in healthcare. So, how can we build effective and reliable AI solutions for government healthcare?^[1][5]

The Art of Possible
Draw and Conform to the AI Guidelines

The next step is to evaluate the statistical validity, clinical utility, and economic utility of the AI tool. This can give you insights into the real-world implications of the AI solution within the healthcare ecosystem. The criticality for this phase will be to invoke a human-centric design with a focus on AI-specific validation scenarios like bias, fairness, accountability, robustness, drift, and security.

Prioritizing High-Value Use Cases

The majority of AI solutions in public health remain focused on invoking the right intervention for the right population at the right time, improving speed and accuracy in disease surveillance, outbreak detection and response management, etc. Developing specific AI tools to address them would need special consideration for petabyte-scale custom architecture to gain easy access to the volume, velocity, and variety of data, which needs to be of the highest quality. Last but not least, enhanced security and auditability, with increased regulatory compliance, audits, and policies, will be crucial to model outcomes and practices specific to the use cases.

Monitor and Maintain

Similar to any AI solution, the last step is to continually monitor and maintain the potential risks or adverse events in terms of the performance of the AI model, standard of patient care, and patient safety. AI model lifecycle management with Responsible AI controls and measures will be crucial to create, deploy, train, operate, AI solutions to provide accurate predictions of healthcare outcomes, engagement, experience across race, gender, socioeconomic, and behavioral status to achieve the health equity goals.

Conclusion

AI is not just about adopting a new, forward-thinking technology, tools, and solutions. It’s about bringing the expertise and experience in place by imagining all the future possibilities and challenges. It is organizational responsibility- people, process, technology, and environment.

While AI solutions have tremendous potential to transform the way we look at healthcare systems today and pave the way for a precise and personalized healthcare future, the transformation of healthcare systems to adopt and operationalize them will depend on how well the stakeholders can plan to navigate the challenging landscape.

References

1. https://www.gao.gov/products/gao-22-104629
2. https://emeritus.org/blog/healthcare-challenges-of-ai-in-healthcare/
3. https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-022-08215-8
4. https://www.hhs.gov/sites/default/files/hhs-ai-strategy.pdf
5. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8285156/
6. https://www.hhs.gov/sites/default/files/hhs-artificial-intelligence-select-use-cases.pdf

Author Details

Dr. Suman De, Principal Consultant and Head, Government Healthcare Analytics Solutions, Infosys Public Services

Dr. Suman is head of government healthcare analytics for Infosys Public Services. He has extensive experience in the public healthcare sector and previously worked for the World Health Organization, UNICEF and the Indian Public Health Association.

At Infosys, Dr. Suman leads the area of advanced data science and artificial intelligence-enabled population health, social determinants of health analytics, opioid management, care management, and value-based care. He is a frequent public speaker at various healthcare conferences, forums and at major universities, including the Massachusetts Institute of Technology.

Dr. Suman is based in Hartford, Connecticut. He holds a medical degree from the University of Calcutta and master’s degree in healthcare administration from the Tata Institute of Social Sciences in Mumbai, India.