Investing in longitudinal data mining can help payers achieve the triple aim of lowering costs and drive healthier outcomes for its members.
Overpayment of healthcare providers has always been an area of concern for health payers. But it has never held the urgency it does today.
Rising costs and the ever-increasing complexity of the healthcare delivery and payment system can cause overpayment losses to health plans in the millions of dollars. Payers are then forced to raise premiums to the point where some Americans will no longer be able to afford health insurance, and many others will be under-insured, creating an even larger financial crisis around healthcare.
While some progress has been made – the Centers for Medicare and Medicaid Services (CMS) reports that improper payments under Medicare Advantage dropped from 14.1 percent in 2010 to 9.5 percent in 2015, that is still 9.5 percent of a total expense of $210 billion. For the math-challenged, that means nearly $20 billion was wasted rather than going toward care for those who need it.
There can be many reasons for overpayment, including data problems, contracting issues, and opportunistic bad actors. But once those overpayments are identified, providers are generally required to return them within 60 days. In most cases it is up to payers to make providers aware of the issue, creating an extra level of administrative burden as well as potential friction between payers and providers.
Payers have several levers they can pull to ensure payment integrity and avoid overpayments. One that is overlooked, however, is an updated approach to data mining. That’s a shame, because health plans that update their programs to incorporate longitudinal data mining, which replaces point-in-time, individual claims analysis with technology that can identify trends and patterns across larger datasets, are likely to realize significant gains without the provider abrasion that can be associated with other tactics.
Beyond the snapshot
The strength of longitudinal data mining is that it gets beyond the snapshot to help payers capture repeated errors by the same provider, mistakes at the same point in the adjudication process, or some other pattern that is only apparent when looked at over time. Payers can use this information not only to recover overpayments but also to uncover their root causes so they can be addressed.
One of the most important keys to the success of longitudinal data mining is having a unified strategy that can connect massive amounts of data from disparate sources and stratify it at high levels such as by system, specialty, or facility type. Approaching data this way allows additional patterns to emerge. Payers can then benchmark their findings against their peers so they understand how they compare, and where additional work is required.
Improving care transitions
One critical area where longitudinal data mining is extremely helpful is discovering overpayments that can occur when members move from one care site to another. The ability to link and track claims from site to site enables payers to uncover questionable coding and billing patterns more easily, such as inappropriate referrals, or hospitals and skilled nursing facilities both billing for the same rehabilitative care.
These errors are generally unintentional, but if they are not caught they can be costly. Finding and correcting them before payment is made helps control costs while minimizing provider abrasion.
Adding value through hybrid analytics
While the technology behind longitudinal data mining alone holds huge benefits, those benefits can be increased exponentially through what’s known as a hybrid analytics approach. This combination of data mining, machine learning, and human intellect takes the best of each to create a greater whole.
The data-driven process begins with a clear understanding of the business problem it intends to solve. Patterns in the data – which surface through a unified data strategy and longitudinal data mining – inform hypothesis generation and concept development. Testing validates concepts, after which they are applied at scale.
Machine learning is then used to further identify patterns, and the models that are developed are trained to understand the data so they can be improved. As data models are optimized, additional patterns and insight emerge. These insights are then fed back into the models for further optimization. The hybrid analytics continue to “learn,” creating a virtuous cycle of automatic improvement that traditional analytics, or human-driven inspection of individual claims, cannot begin to duplicate.
As models are refined and scaled up, overpayment recovery and root cause analysis both gain efficiency even as false positives are minimized. The result is reduced provider abrasion (because fewer overpayments must be recovered), a seamless process that requires far fewer internal resources (enabling them to be shifted to more mission-critical work), and most importantly more money available to dedicate to ensuring all who need care receive it.
Achieving the Triple Aim
Overpayments are a source of frustration for everyone in healthcare. But with the right approach, the pain can be reduced considerably.
In a challenging market in which excess costs translate to higher premiums that could put care out of reach, investing in longitudinal data mining can help identify and reduce the problems at the source. This will help payers achieve the Triple Aim of lowering costs and improving member satisfaction while ultimately driving healthier outcomes for members.
Lalithya Yerramilli is Vice President of Analytics at SCIO Health Analytics. She has 15 years of experience in analytics in insurance, healthcare, and life sciences industries working with customer info-base, transactional, physician level, patient level, claims, and longitudinal datasets.
