(US & Canada) VIDEO | PHI Is Both a Blessing and a Challenge With Gen AI — BCBS Massachusetts CDAO

Himanshu Arora, Chief Data and Analytics Officer at Blue Cross Blue Shield of Massachusetts, speaks with Sue Pittacora, Chief Strategy Officer at Wavicle Data Solutions, in a video interview about leveraging AI in healthcare, the different aspects that benefit from AI, internal-facing use cases, and measuring ROI of organizational initiatives.

At the outset, Arora reflects upon the concept of everyday AI versus game-changing AI having one access and being internal facing. By internal facing, he means capabilities available internally for associates and employees, versus external capabilities targeted towards customers.

There are additional complexities in the healthcare sector, in terms of the players involved such as care providers, health insurers, pharmaceuticals, and equipment manufacturers, who must work in alignment. Therefore, bringing in the game-changing AI would require more than one sector to adapt.

Highlighting the health aspect, Arora stresses the importance of Protected Health Information (PHI) as the health data is secured. Now, with generative AI, PHI is both a blessing and a challenge, he adds.

With massive untapped knowledge in health data, it can be game-changing, but it is challenging to keep the privacy and security aspects first, says Arora. Also, he continues, there have been regulatory developments on this front combined with customer excitement and concern.

Speaking of AI use cases, Arora reiterates that the organization applies AI capabilities internally with how data security and privacy are managed and for internal knowledge encapsulation.

Arora shares that associates spend time and energy answering questions from the knowledge compendium, whether it is for customer service or other factors. Therefore, the organization is figuring out how to use generative AI in this area while considering PHI as a contributing factor.

The legal and procurement processes will potentially benefit from generative AI, and the organization is working to engage with LLMs in a cost-effective way, affordable for members. However, Arora maintains that the space is treaded with caution and excitement, given the risk of exposing crucial data.

When asked about advancing with internal use cases, Arora states that the organization is focused on building the employee knowledge base with AI. With an active assistant available to them, the employees have a knowledge compendium to rely on, while answering questions.

Then, he mentions that the benefits aspect in healthcare can be complex to understand, with baseline benefits and variations provided by employers to deliver differentiated services. Consequently, it can be challenging for customer service representatives to answer questions related to benefits.

Therefore, another area of leveraging AI is in overlaying benefit design packages and internal benefit plans that the representative can leverage. This helps the representative to answer in a way that further helps in assessing the member better.

Next, the organization is looking into speeding up procurement processes, entity engagement processes, and legal reviews. While there is no PHI or PII in any of this, there is business proprietary information to be protected and organizations must engage with LLMs accordingly.

Commenting on measuring the ROI on initiatives, Arora states that some of it is a pull forward from measuring the impact of capabilities like advanced analytics, data science, and machine learning.

Discussing the revolution from the advent of the Internet to Web 2.0 has been about creating efficient frontiers from a task breakdown structure perspective. Thus, he notes that the attribution from an efficiency perspective can be more at task level, then leading to the ROI produced by the task.

It can be in the form of fraud prevention, better clinical care, health outcomes, higher revenues, or sales, says Arora. Arora further notes that task efficiency can move to knowledge efficiency.

He says that it is still in the early stages of streamlining the knowledge component in a way that goes beyond efficiency and contributes to business outcomes. The organization still uses the traditional measurement framework that is hardwired into business goals and outcomes in compliance with the business plan.

Furthermore, Arora mentions the adoption and usability perspective to measure how far the solutions are adopted and used across the company. In conclusion, he highlights having internal measures to improve capabilities step by step and value the knowledge encapsulation piece.

CDO magazine appreciates Himanshu Arora for sharing his insights with our global community.