Information Mining in the Treasure Valley

I recently attended the BCBS IMS Symposium in Boise, Idaho. The theme of this year’s conference, “Information Mining in the Treasure Valley,” centered around how every healthcare organization has their own “treasure” (data) to manage in their specific “valley,” or area of impact, be it across their membership, provider networks or internal operations.

The conference lent itself to multiple areas of interest and reoccurring topics in our industry today. These included analytics and big data, predictive analytics and modeling, data governance and utilizing agile/scrum for data-driven products and platforms. Below is a synopsis; enjoy!

Opening speaker Charlene Maher, President and CEO at Blue Cross of Idaho, gave a resonating presentation on the fact that data is at the absolute core of all that we do every day in our industry. She shared how we can utilize data and knowledge to change product development, healthcare operations, provider networks, our culture and focus by using existing data and skills to significantly enhance our daily interactions with healthcare consumers. She noted, “Data drives everything and is where we need to focus. It is at the forefront of changing the entire industry. IT combined with analytics has the ability to change the industry, and it must.” Charlene added that while we are still many years behind other industries in terms of analytics and merging analytics, merging business and IT together as one is a significant step in the right direction emphasizing, “Everyone has to listen to analytics and use the data to bring us together.”

Data governance is a topic we have been discussing in great detail in recent months, and this conference was no exception. It is very important that we understand the big picture of data management, which also means knowing the difference among data owners, data stewards and data custodians. Or, those who actually own the data, manage the data and maintain the data. Better quality data must be shared in a much more practical manner among payers, providers and other stakeholders.

Drew Hobby, SVP health economics from BCBS Idaho, shared his organization’s journey to determine ROI on all programs, including MSK and other specialty plans, wellness, disease management and behavioral health. One specific example was their nurse hotline program, which was sending far too many members directly to the ER, yielding a negative ROI. The four pillars of Idaho’s health economics process are data, design, implement and measure. Drew’s associate, Marc Roberts, director healthcare analytics, also discussed their behavioral health program. First, they embedded their data scientists to fully understand the program. Next, they defined KPIs and a study design via a one-year retroactive study and propensity score matching. They then compared the control and trial groups, before modifying interventions for better ER support, and reassessed the program impact. One key point Drew noted was the importance of the data analyst team to communicate the data and results to the client, not the broker.

The use cases are extensive but common scenarios across multiple plans included risk of hospitalization, risk of ER, risk of 30-day readmission and future high-cost claimants. Mathew Giglia, senior data scientist from Excellus, included rating and underwriting, accountability cost agreements and care management, where a gap in risk score from last month to this month may trigger a case creation. He also shared in his presentation, “Predicted PMPY: Accurately Predicting Individual Member Future Total Cost Using ML and Big Data” their process for creating specific models. Matthew emphasized that we must build the model every month from scratch to “train” it. This build-score-monitor model would then observe the trends to make improvements. For example, if a specific drug indicative of future cost goes from formulary to generic, then this needs to be factored. Finally, they would measure the accuracy. What was the score from the same month, one year ago and was the previous prediction accurate?

Whether working with vendor partners or reorganizing within their own internal teams, many Blues teams are maturing their methodologies and delivery around agile, PODS and hyper-productive teams.  Incorporating a dev-ops driven approach to new development models for data-driven products and features is paying off.

It was exciting to witness the advancements, maturity and collaboration among the Blues, and I am looking forward to next year. See you there!

As vice president, Payer Solutions at emids, Gray Karnes is responsible for leading his team in analyzing and addressing specific healthcare business challenges and providing custom technology-based solutions in conjunction with a dedicated team of health consultants and technologists. Two primary end-to-end service initiatives Gray oversees are application and product engineering and healthcare data services. Be sure to connect with Gray on LinkedIn.

Sign up to receive our latest insights