Insights

Predictive Analytics in Insurance Underwriting | Prognos Health

Written by Nick Zgorski | Sep 24, 2021 1:51:21 PM

Predicting risk in health insurance underwriting requires an accurate snapshot of a prospective group’s healthcare needs for the coming year. 

But for decades, lab data has been the missing factor when predicting group risk. 

Why is this an important factor?

According to the CDC, 70 percent of medical decisions depend on lab results.

What is the result?

Deeper clinical insights that allow you to rapidly deliver the most competitive bid with the lowest risk of loss.  

The status quo: Manual rate, prescription histories provide limited view

Current state: 

  • Many healthcare payers rely on manual rates based on past cost experience, using statistics and probability to predict the financial impact of an uncertain future state. 
  • Some payers have incorporated additional outside data (such as behavioral data and prescription histories) to add deeper insights to the underwriting process. 
  • Prescription histories provide valuable information on administered medications. This information can allow a payer to deduce the medical conditions impacting a group, but with limited confidence. 

The Problem: 

  • Traditional sources do not fully account for the current health of the group(s) to be quoted. 
  • Behavioral data does not always directly relate to the clinical health of prospective members.  
  • Prescription data may not provide insight into disease progression or severity, whereas laboratory data can. 
  • Medication histories are also based on prescription claims, that include a latency period (sometimes months) between when a therapy is prescribed and dispensed and when the claim is actually received and processed. Additionally, they provide little insight into an individual’s adherence to the guidance given to them by their care provider.

The Solution: Prognos Health Underwriting Risk Predictor uses lab data to deliver a comprehensive, timely clinical picture

Lab data is a largely untapped resource among health insurance underwriters, but it has two vital qualities necessary to provide the most accurate risk prediction scores possible.

  1. Clinical richness — Lab data provides a level of clinical detail that surpasses the limited medication and payment information available in a prescription history. With 70 percent of medical decisions depending on lab results, this dataset is the best indicator of disease severity and progression. 

    Results:
    • Lab data provides the earliest indicator of emerging conditions.
    • Payers no longer need to deduce diagnosis or disease severity. This specific information can be pulled directly from lab results, providing a clearer picture of member health.
    • Payers can more accurately predict risk with deeper clinical insights.
  1. Timeliness — Lab data is more timely than what’s available in actuarial sources or prescription histories. 

    Results:
    • Lab data provides payers with more up-to-date health information because it isn’t dependent on a claim being processed. It is accessible once a test is complete and the results are available.
    • By reducing latency, lab data helps eliminate gaps in care that might be present when relying on prescription or behavioral data alone. Lab data provides the best insight into current member health  — a must when predicting future healthcare usage.

So why isn’t lab data leveraged more? The main reason is, when purchased in bulk, lab data is not standardized and very difficult to operationalize in its raw form. 

Harness the power of lab data to improve your underwriting accuracy.

Prognos Health has spent years creating the gold standard of lab test results used by global leaders in the payer and pharmaceutical industries. Now, our data scientists have standardized and applied machine learning models to yield clinical insights that can be readily factored into an underwriting risk prediction score.

Prognos Underwriting Risk Predictor applies advanced analytics to its registry of more than 325M de-identified lives to predict the future cost of healthcare — a level of accuracy that can’t be matched by methods based on manual rate, prescription histories, or behavioral data.

See how this powerful tool can deliver measurable results to your underwriting efforts by receiving a custom proof of concept from one of our payer solution experts.