In an interactive session at the Becker's 11th CEO + CFO Roundtable, Alex Kleinman, global leader of Genpact's healthcare business, discussed the application of analytics and AI in transforming population health and achieving cost savings and other clinical performance goals. Genpact, a professional services firm focused on delivering outcomes that transform business, serves various sectors, including healthcare.
Mr. Kleinman highlighted the importance of data and analytics in healthcare, in areas such as population health management and patient contact work. He also discussed the potential of AI in healthcare, noting that while the buzz around artificial intelligence has increased, its application in healthcare is not new. For example, Genpact has been using AI to help providers access patient insights, improve revenue cycle management, and better manage population health without increasing the burden on clinicians.
He also discussed the challenges in AI application, such as data access and interoperability, validation for clinical purposes and the development of compelling visualizations. Key to overcoming these challenges are strong partnerships, such as the collaborations Genpact has with Health EC, a population health management platform.
4 takeaways
1. AI tools such as machine learning and natural language processing can improve decision-making and cost reductions in healthcare.
"We've been using AI tools in our work for a number of years … to optimize collections and improve call handling in member support contact centers," Mr. Kleinman said.
2. While AI and machine learning have been leveraged in healthcare for years, the success of the technology has been undermined by data-related challenges.
"The problems and challenges that we see in some of the more traditional or older AI applications are related to having access to the right data," Mr. Kleinman said. "Then the challenge is making sure that data is interoperable, clean and normalized."
3. AI — in the form of large language models — can be used to assess disparate patient data to help inform care decisions.
"Large language models are particularly effective at going through multiple medical records and pulling out relevant information, which takes that task off the care coordinator's plate," Mr. Kleinman said.
4. LLMs accelerate the value of machine learning- and AI-powered solutions.
"Historically, you might have to train a machine learning model for six months with
a ton of data to get something useful out of it," Mr. Kleinman said. "LLMs are pre-trained, which really accelerates the timeline to impact. The real transformation they offer is the speed to insight, which is very exciting."