Protecting vital public health programs with AI and Knowledge Graphs | Vanessa Lopez | KnowCon 2020
Science & Technology
Introduction
Introduction
In the final session of the morning before the lunch break at KnowCon 2020, Vanessa Lopez from IBM Research Ireland presented her research focusing on the application of artificial intelligence (AI) and knowledge graphs in the health and social care domain. Vanessa, a research scientist and manager at IBM, shared insights into her team's efforts to address significant societal challenges, particularly in combating fraud, waste, and abuse in public health programs. Her talk emphasized the importance of a collaborative approach with subject matter experts and the use of advanced technologies to improve patient outcomes and optimize resource allocation.
AI and Knowledge Graphs in Public Health
Vanessa discussed her academic background and how it aligns with the goals of the IBM Watson Health team, which is dedicated to transforming healthcare through AI innovations. Healthcare and program integrity are complex fields, and Vanessa highlighted that it is crucial to leverage AI technologies, particularly knowledge graphs, to create interconnected models that aid in understanding and addressing healthcare challenges.
Addressing Fraud, Waste, and Abuse
One of the projects highlighted was aimed at combating fraud, waste, and abuse in public health programs, particularly Medicaid and Medicare in the United States. With billions of dollars lost annually due to fraudulent activities, it is essential to protect vital public health programs. Vanessa described a process where claims submitted by providers are validated against policy rules encoded in large documents. The investigation into claims is labor-intensive and often results in significant errors, leading to lost revenue.
To streamline this process, Vanessa's team utilized natural language processing (NLP) and AI to extract actionable knowledge from policy texts. This results in a knowledge base of benefit rules that can be applied to assess claims for compliance. By converting policy text into machine-readable formats, investigators can more effectively identify claims that may be at risk of fraud or abuse.
Human-AI Collaboration
Vanessa emphasized the importance of human-AI collaboration in this process. By engaging subject matter experts early in development, they created a user-friendly interface for reviewing and validating the extracted rules. This collaboration allows investigators to prioritize investigations better, based on the likelihood of recovering funds.
Achieving Explainability
Another critical aspect of the project is ensuring explainability. Each flagged claim is linked to the benefit rule and the corresponding policy language that outlines the violation. This dexterity in the system helps improve trust and allows investigators to effectively argue cases for recovery.
Knowledge Graphs in Social Determinants of Health
Beyond fraud detection, Vanessa briefly discussed a secondary exploratory project using knowledge graphs to analyze the impact of social determinants on health outcomes. The team focused on understanding trends in social issues, such as food insecurity, exacerbated by the COVID-19 pandemic, and how these factors correlate with various health outcomes.
Conclusion
In conclusion, Vanessa's talk highlighted the transformative potential of AI and knowledge graphs in the health and social care sector. The emphasis on user-centered design complemented by rigorous scientific methods showcases the innovative approaches required to tackle the complex challenges of healthcare today.
Keywords
- AI
- Knowledge Graphs
- Public Health
- Fraud Detection
- Natural Language Processing
- Social Determinants of Health
- Healthcare Innovation
- Expert Collaboration
- Policy Compliance
FAQ
Q: What is the primary goal of using AI and Knowledge Graphs in public health?
A: The primary goal is to protect vital public health programs from fraud, waste, and abuse, ultimately improving patient outcomes and optimizing resource allocation.
Q: How do knowledge graphs aid in fraud detection within healthcare?
A: Knowledge graphs help to convert complex policy texts into actionable rules, making it easier to assess claims for compliance against these rules.
Q: What role do subject matter experts play in the process?
A: Subject matter experts are vital for reviewing and validating the extracted benefit rules, ensuring accuracy and improving the effectiveness of investigations.
Q: What is the significance of explainability in AI systems used for healthcare?
A: Explainability helps increase trust in the system by linking flagged claims to specific policy violations, thereby providing clarity and insight into the decision-making process.
Q: What other applications of knowledge graphs were mentioned in the talk?
A: The talk also covered the use of knowledge graphs to analyze social determinants of health and their impact on health outcomes, particularly during the COVID-19 pandemic.
