The recent HIMSS24 conference in Orlando was abuzz with discussions on Generative AI (Gen AI) in healthcare. With its ability to streamline administrative processes, Gen AI is being hailed as a game-changer in various healthcare applications. However, as promising as it may be, there are questions about how to effectively implement and scale Gen AI models.
To shed light on the state of Gen AI, we had the opportunity to speak with Chas Busenburg, Product Development Manager at MedInsight. In this interview, Chas shares his observations at HIMSS. He also highlights key considerations healthcare organizations should bear in mind as they approach including Gen AI capabilities into their workflows.
1. What specifically caught your attention at HIMSS and why?
At the HIMSS conference, I noticed a strong emphasis on Gen AI. It was intriguing to dig more into this topic and understand how organizations are obtaining data for their Gen AI initiatives. Although some organizations were experimenting with different approaches, there was a lack of transparency about their data acquisition methods. This underscores the need for further work in effectively implementing Gen AI in healthcare settings.
On a positive note, I was impressed by industry leaders who prioritized the data itself and took a more thoughtful approach to adopting Gen AI. These organizations focused on enabling others to effectively use their own data, rather than imposing a one-size-fits-all approach. I found this approach to be resonant as it recognized the unique nature of each organization’s data and the importance of empowering them to unlock its true potential.
2. When it comes to healthcare organizations entering the realm of data and AI/Gen AI, what are opportunities and challenges they face?
One significant opportunity healthcare organizations can take is establishing a framework for data collection and integration within their organizational processes. This will enable them to effectively employ Gen AI use cases in the near future.
However, it is important to acknowledge the challenges that come with the relative immaturity of the field and the rapid pace of Gen AI development. While finding investment opportunities can be difficult, it’s never too late to begin preparing for the utilization of Gen AI. Even if you have limited resources, the opportunity to adopt Gen AI will eventually become more feasible within time. Whether you decide to take the lead or rely on third-party services, we anticipate seeing an increase in options within the next 12 months that will make adopting Gen AI more feasible. For those organizations that are uncertain about the extent of their investments, the availability of third-party providers will continue to grow and provide those organizations with more guidance and support.
3. As organizations approach Gen AI capabilities, what are the key considerations they should keep in mind? Are there any potential pitfalls they should be aware of?
Acknowledging the significance of data in Gen AI models is a first step. The ownership of unique data by an organization holds immense value and can confer a competitive advantage. While not necessarily a challenge or opportunity, it is vital for organizations to strategically and responsibly leverage their data. Moreover, they should consider the distinctiveness and identifiability of their data when collaborating with external entities.
Secondly, organizations should recognize that tackling this endeavor alone is unrealistic. The staffing and financial resources needed to keep up with the rapid pace of change can be unattainable. Therefore, trusting another organization to some extent is essential for enhancing their technological stack. Whether it involves using open-source ML models or opting for a SaaS implementation, forging strong partnerships is a critical aspect of an organization’s Gen AI journey.
As for potential pitfalls, there is one aspect to be aware of in the current state of Gen AI. When AI technology can produce impressive results, we have not yet reached a point where we can completely rely on it without human review. Within MedInsight, we are taking an approach that prioritizes data trustworthiness to support human decision-making. This position is particularly important in healthcare, where the sensitivity and restrictions surrounding data and the potential for ‘hallucinations’ or erroneous conclusions carry serious implications. As the industry continues to address this, our solution offering continues to employ a reliable review process that incorporates both AI and human intelligence.
4. Could you provide insights into how MedInsight is assisting healthcare organizations in preparing to enter the field of data and AI/Gen AI? What unique features or benefits does MedInsight offer in this regard?
Sensitivity is a major concern when it comes to Gen AI, which is why it’s crucial to exercise caution when considering full automation without human involvement. At MedInsight, our current focus is to avoid complete automation and instead augment an organization’s existing capabilities. This empowers health professionals to effectively leverage data and technology according to their needs and to make timely decisions with confidence.
Another strength is our deep experience in handling data, particularly in the areas of claims and electronic health records (EHR). We uniquely meet users where they’re at in their journeys and understanding of their data. For advanced users, our Data Science Portal and integration with Databricks‘ AI assistant are proving to be incredibly helpful in working with the data mart at a granular level.
In the future, we plan to support users in creating views within Query Express or our Cube Browser through free-form text input, providing them with even more flexibility and control. Our primary focus is to continue helping organizations become more self-sufficient and efficient in their data-related tasks.
Watch our AI panel discussion
Learn more about the future of AI in healthcare analytics by watching our panel discussion. Tune in to our webinar titled “Navigating the Future of Healthcare: AI Trends and Predictions for Payers and Providers” to gain valuable insights from industry experts. Our panel includes representatives from Milliman MedInsight, AmeriHealth Caritas, and Databricks, who explore the transformative power of AI in the healthcare industry.