WellPoint's Patrick McIntyre on anti-fraud analytics partnerships
FierceHealthPayer: Anti-Fraud spoke to Patrick McIntyre for expert advice on smart shopping for anti-fraud data analytics and how to work well with vendors. McIntyre (pictured) is senior vice president of healthcare analytics at Indianapolis-based WellPoint. For more than 25 years, he's led business operations, finance, analytics and information technology functions.
FierceHealthPayer: Anti-Fraud: With so many anti-fraud data solutions available, what criteria can insurers use to evaluate products and choose what's best for their business?
Patrick McIntyre: Know what your internal capabilities are and match those to the solution you need. Figure out if you have the internal resources to be able to do components of the overall solution implementation yourself.
For example, we were looking for a solution years ago since we didn't have the right internal expertise to build an anti-fraud analytic capability. So we looked for more of an out-of-the-box solution. If we were making that purchasing decision later, our needs would be different based on a new mix of employees with more advanced analytics expertise.
It's also important to determine if potential business partners have a proven record of delivering services they commit to. The last thing you want is to partner with a company that oversells capabilities you rely on and then find out they can't deliver on these when you reach the implementation stages.
FierceHealthPayer: Anti-Fraud: What are some key features insurers can look for that characterize high-value fraud detection software? For example, some products claim to "blend analytics with rules." Others are marketed as "fully-integrated solutions." What's important from a performance/bang-for-the-buck perspective?
Patrick McIntyre: Flexibility in the offering is important. I believe strongly that you need a blend of built-in analytics and an ability to build your own rules.
We have very specific contracts with providers, for example, and some of the claim edit tools we use affect what claims looks like when they come out. So there are custom detection paths we must build, and we need flexibility for that.
I use an 80/20 rule: If we can buy a product that gives us 80 percent of what we need in packaged analytics, that's good. Down to 70 percent or up to 90 percent is an acceptable range, but I want to be able to create custom rules specific to our needs. I look for that in any data analytics capabilities we purchase.
The other key feature for fraud and abuse detection is whether you can move the capability to the front end to work on a prepayment basis. Capabilities that can do that move to the top of our list, while those that can't generally aren't viable solutions.
FHPAF: Do you recommend doing an information security risk management assessment on potential vendors and/or their products before contracting? Why or why not?
McIntyre: We absolutely must do that. Information security is paramount today. Information service providers that don't have an air-tight information security risk management process aren't viable partners.
Routinely monitoring security procedures for data and information sharing is a minimum requirement we have for external vendors. Besides assessing on the front end if vendors have the right types of security and controls in place, I also recommend ongoing assessments and audits. Do these at least annually, and even more often to ensure that controls still work in view of changes to business, analytic or data processes.
As we enhance these internally, we modify the basic foundational beta structure of core systems. All data feeds going downstream into different analytic environments must be continually monitored to ensure we're still meeting information security requirements.
FierceHealthPayer: Anti-Fraud: Some vendors have been described as "good dates but terrible spouses." How can payers avoid contracting with companies whose service levels drop dramatically once a contract is signed?
Patrick McIntyre: Thorough reference checks are key. Don't just interview potential partners; dig deep to find out if they can and have delivered on what they sell.
For a fraud detection vendor we selected, we met with a couple of their existing customers and in one case spent an entire day on site with them. Without the vendor present, we asked questions about what worked and what didn't. We saw and experienced the end product.
Also, my goal with contracts is to finalize everything we could possibly want in the contract before signing, and then put the document in a drawer and never have to bring it out again.
But if you must do that, it's important to be absolutely clear about what the intent of the parties are. For any analytics vendor, you need a robust service level agreement in place with clear performance metrics you can track and monitor regularly. This keeps conversations with vendors objective rather than subjective. We include performance guarantees or penalties for failure to meet objectives wherever possible.
FHPAF: What's the status of anti-fraud predictive analytics: Are they a dream for the future, or are they available now?
McIntyre: This is still a developing area, though we're getting better at it as an industry. There are predictive analytic methodologies in place today, but many are based on historical patterns. With the proliferation of big data and the ability to get more external information through social media and text mining, I think we'll get better at predictive analytics.
Meanwhile, I tell my folks that web research is critical to fraud detection. It's important to learn what fraud practices are implemented nationwide and what the new fraud schemes are. Some move faster than we can react to, so staying current on trends is critical. Criminals move on to the next fraud practice or the next fraud victim. In either case, you've got to stay on top of your game.
FHPAF: Do you have any other smart shopping or "let-the-buyer-beware" tips not covered in the above questions?
McIntyre: Do your homework on the front end to find organizations with whom you can work closely and truly partner. Remember to look for partnerships; if you settle for a vendor relationship, you're probably in the wrong starting spot.
Look for organizations that want to grow with you and implement what you've learned together in the analytic capabilities you use and build.
Find partners you can work with to improve not only what you do but their products as well. Then use objective criteria to track and monitor partner performance. That's the winning combination for success.
Editor's Note: This interview has been edited and condensed for clarity.