Data analytics has become vital in today's health insurance environment. It not only offers insurers the opportunity to harness lots of data to effectively deter fraud, waste and abuse, it also helps them identify areas to improve quality and reduce cost.
Recognizing that, Boston Medical Center HealthNet Plan (BMCHP) has leveraged analytics to monitor emerging trends and track cost savings opportunities. The 350,000-member health plan has been investing in enhanced analytics tools to keep up with the demand for information to make good business decisions, Lisa Feingold, vice president of Clinical Informatics at BMCHP, told FierceHealthPayer in an exclusive interview.
"The more people who are looking at information, as opposed to data, the smarter the decision-making process will be," Feingold said.
FierceHealthPayer spoke with Feingold to learn more about how BMCHP uses analytics as well as the need for self-service analytics.
FierceHealthPayer: Can you describe the health plan's focus on analytics?
Lisa Feingold: My department is within the finance area. We're the corporate analytics department, and we focus on quality measurements, including HEDIS and program evaluation for clinical quality programs we run. We support high-risk case identification in building a registry for the care management team and identification of high-cost cases.
Within other parts of the finance area, corporate analytics also supports cost and utilization trending, identification of outliers and opportunities for cost savings. We do a lot of pharmacy analytics as well, in just looking at existing trends. There are actuaries within the department who look at model cost and utilization, so when it comes time to contract with different state entities, we've got the most robust and up-to-date information.
We also try to support individual provider groups, by giving those provider groups we have risk deals with as much as information as possible, in terms of high-risk patients if they're doing care management directly and information about how they're performing on various quality care metrics.
FierceHealthPayer: How has enhanced analytics tools helped the organization improve care and lower costs?
Lisa Feingold: For the most part, we find ourselves in a typical situation: We have fairly extensive data warehouse and a fairly large group of analysts, and a lot of the work we do is more ad hoc in nature. For example, somebody wants to know about new drugs. A couple months ago nobody really had heard about Sovaldi, and now we're seeing it's an issue and we're seeing trends going up. If people wanted to get into the details, they needed to go into the warehouse and pull all of the information in very fine detail.
We found our analysts were spending about 75 to 80 percent of their time managing data, leaving very little time to do true analysis. We wanted to turn that upside down so that getting to the data was much quicker and easily accessible and allow the analysts to do some in-depth analytic work about what was going on.
Around the ad hoc analytics, we wanted to be able provide an opportunity for more self-service. We only have so many sources and there are tons of questions that our CMO or medical director comes up with during the course of a week. But we needed to put the questions in a queue until somebody freed up to answer it.
It may be something that just pops in their head and they're curious, which at the present time we say ''we're too busy, I can't event look at it right now.' Or it may be something that is important, in which case it will get scheduled.
We wanted to avoid that whole scheduling process so people could sit at their desktops, whether it's the CMO, medical director, head of care management or head of the quality area, and be able to take a look and see what's actually going on and drill into the data to answer all the questions they had, that don't need a high-paid analyst sitting there to get the details for them.
FierceHealthPayer: How can payers use analytics to identify high-risk members?
Lisa Feingold: We're using it to identify members based on risk scores, which are primarily based on their diagnoses over a one-year period. We take that as one piece of what we're doing. We're actually looking into the claims data to identify 15 to 20 different areas and opportunities we feel there are members who can become part of our complex care management system or our standing care management and disease management activities.
We're using utilization data, ambulatory sensitive use of the emergency department, hospital readmissions, patients who have various medication possession ratios that aren't as high as we would like to see. We're using as much information that exists within the claims system to support identifying the best subset of members we think can benefit from care management programs.
FHP: What emerging trends has the use of advanced analytics revealed?
Feingold: All health plans at this point are keeping a close eye on the increasing ED utilization related to ambulatory sensitive care. We're going to be looking at the same per inpatient admission and admissions that may have been preventable if better or more accessible treatment was available.
At the pharmacy end of it, a small population of members using an expensive drug can have a significant increase on your pharmacy trend pretty quickly. I used the Sovaldi example, the new hepatitis C medication approved in December by the FDA. It's one pill for 12 weeks, and the cost of that 12-week treatment is $84,000. As you can imagine, looking at the cost and utilization trend you quickly can see what's popping up.
FHP: What advice would you give to payers looking to leverage analytics to improve financial and clinical outcomes?
Feingold: You can never have enough information, so the key is trying to find a way a way in which you can expeditiously get access to that information.
We always have annual brainstorming sessions about trying to identify opportunities to improve quality and reduce cost. As a result of that brainstorm, we may come up with 20 different areas where we think we can do things better. We used to then take those 20 things back and build the data to answer each and every question--'is there an opportunity,' 'is there sufficient volume,' 'what's the process that's in place to manage this.'
Now we anticipate we'll be able to sit in the brainstorm meeting with the data in front of us and actually do those queries on the fly. We can then rule in or rule out things much more efficiently than having to meet, brainstorm and analyze over the course of several months and then come back and say 'these are worth pursuing, and these are no longer worth pursuing.'
Editor's Note: This interview has been condensed and edited for clarity.