14 Jun The Effectiveness of AI Guidance for Health Systems is Directly Dependent on the Quality of the Data

While there’s a current mad rush in the development of Artificial Intelligence (AI) solutions for the healthcare arena, health payers and health systems may want to step back and truly assess before implementing these kinds of offerings.

As we have seen lately in the mainstream media, Generative AI models, such as ChatGPT, are known to generate mistakes or “hallucinations.”

Just last week, a lawyer got into trouble for using ChatGPT for a court filing that referenced cases that were fake. OpenAI is also being sued for libel in what could be a landmark case around AI and defamation.

Fortunately, there is a rise in new AI detection tools that can detect academic writing that originated from ChatGPT. A chemist at the University of Kansas developed a new tool that detects with 99 percent accuracy scientific text generated by ChatGPT. The source code is also available for others to replicate this tool.

What does all of this mean for health payers interested in exploring AI? The quality of clinical data is paramount. Otherwise any new AI solution could potentially provide insights that are incorrect, and potentially be harmful to patients.

At IMAT, we are exploring benchmarking studies to assess the validity of using AI generated data compared to the data provided by the Perfect Search technology behind our solutions.

The IMAT Payer Solution already takes the unclean, incomplete, raw clinical data that comes out of the EHR eco-system and dramatically improves the overall data quality.

This is achieved by increasing the usability of all source data, providing data quality improvements with a focus on clinical content, as well as leveraging use case inference and algorithm logic for further data usability. It also comes down to the transparent process of getting to fully clean data, and getting access to the source data in the health record. Ultimately, this will prevent scenarios where an AI system will inadvertently create incorrect information.

Our offering also enables clinical and business use cases that truly matter by ensuring that the data is current, and undergoes the required data quality improvements.

Clinical Data Integration (CDI) is the foundation for the IMAT Payer Solution. In fact, IMAT Solutions was recently highlighted as a CDI vendor in Gartner’s “Clinical Data Integration Capabilities and Sourcing Recommendations for U.S. Healthcare Payers” report.

For health payers, it’s good to remember that data quality is of utmost importance – whether when exploring AI solutions or even advancing the current use of health data overall.

Got questions about AI for health payers? Our team of experienced health informatics experts is ready to assist. Contact us today.