In the healthcare industry, the strides from machine learning and artificial intelligence have been exciting. And because the nature of underwriting in healthcare relies so heavily on member information and huge volumes of data, the potential of leveraging AI and machine learning in determining underwriting risk cannot be ignored. What could make the equation for a better risk score even more compelling? The answer lies in clinical member data that has been enhanced with technology and expert clinical analysis, then made into actionable insights.
Group health plan underwriting can be conducted and risk-assessed with medical and pharmacy claims, especially if three years of data are readily available. However, the results begin to get murky when underwriting is applied to groups where claims history is not available. In this case, underwriters use demographics, actuarial tables, prescription transaction history and self-reported data for pricing health plans, understanding that these data sets may have their limitations. There is risk, if you will, of premium being mis-matched to risk – and this can affect profit greatly.
The Benefit of Incorporating Clinical Member Insights
Lab data in aggregate can be an impossibly cumbersome asset because in its raw state there is no standardization. In fact, there is no standard even within the same lab testing organization. It’s why Prognos saw an opportunity and took on the effort of bringing together clinical data sets from numerous lab testing organizations. We standardize large amounts of disparate, fragmented and inconsistent data, then apply AI, machine learning and deep domain expertise at scale to produce meaningful analytics solutions. The insights are tailored to use cases across healthcare and life sciences. In the case of underwriting risk, the opportunity to enhance member insights with their actual clinical history and likely health trajectory cannot be overstated. It’s a change well worth exploring to improve risk accuracy and better match risk to premium pricing.
How Clinical Lab Insights Are Predictive
The Society of Actuaries reported that, “as healthcare costs have continued to escalate over the past decades, tools that can be used to predict, explain or understand these costs have become correspondingly more important.”
This brings us back to the notion that AI-driven insights can greatly enhance and streamline analytics while also offering predictive capacity. Incorporating analytics-ready clinical member data can:
- Eliminate the need for simplistic linear regressions and average-driven cost allocations
- Account for non-obvious, non-linear intersections and insights (geographies, comorbidities, disease state progressions)
- Incorporate more recent, definitive facts about individuals’ health status as well as a thorough retrospective view to better predict state of health and trajectory
As we standardize and enrich clinical diagnostic data, we’ve also identified the opportunity to support underwriting by predicting group risk. We’ve developed a secure and cloud-based Underwriting Risk Predictor solution being tested by some of the top five payers to more accurately price group health plans without prior claims history. After we receive a de-identified employer census, we match it to our clinical registry of more than 250 million lives. Predictive analytics can be applied again to produce a mid-year risk score prediction for each group and per-member-per-month cost.
See also: Untapped Potential of Artificial Intelligence
You focus on producing the most accurate risk assessment to deliver a profitable bottom line and to drive better outcomes for your members. AI can provide a predictive solution that may propel your efforts and deliver measurable ROI.
