Healthcare Analytics, Population Health Management, Healthcare Big Data

Analytics in Action News

Predictive Analytics with Claims Data Can Identify High-Cost Patients

Predictive analytics using spending history, prescription drug coverage, age, and gender can help identify patients likely to be costly in the future.

Predictive analytics and claims data can identify high-cost patients

Source: Thinkstock

By Jessica Kent

- Payers may be able to identify future high cost patients by employing predictive analytics strategies to examine past claims,, according to a report from the Society of Actuaries (SOA).

An individual’s spending history, prescription drug coverage history, age, and gender are the most significant predictors of whether or not a person is likely to incur high costs in the future.

The report stated that just 17 percent of members included in the Healthcare Cost Institute (HCCI) Database are responsible for nearly 75 percent of all healthcare expenditures, indicating the importance of flagging potential high spenders and intervening where appropriate.

While some of this spending occurs because of unexpected events or illnesses, some of it is due to chronic disease or cancer care. For these more common conditions, stakeholders are interested in building predictive analytics models to identify high-risk patients, as well as to determine future costs for patients and populations.

To determine which characteristics best predict and describe high-cost members, SOA used the HCCI Database, which includes claims data on approximately 47 million members over a seven-year period. Researchers looked at health information from three of the largest health insurers in the US between 2009 and 2015.

The team found that of all the characteristics examined, member cost history has the most significant impact on the probability of an individual being high-cost. If a patient was high-cost the year before, then they are more likely to be high-cost the year after, and this impact increases as prior year costs increase.

Prescription drug coverage and gender are also indicators of future high-cost patients, but SOA found that these characteristics have less significant effects on patient costs than member cost history. Age is also a significant predictor of future high cost patients, with older individuals having a higher likelihood of spending more on healthcare.

With this information, SOA believes that providers will be able to target high-cost patients and populations and improve care for these individuals.

Predictive analytics has shown potential in many areas of healthcare, and the organization said that going forward, applying predictive analytics to HCCI data could provide additional critical insights that could reduce healthcare costs.  

“The HCCI database is among the most robust collection of medical claims data the SOA has ever had access to,” said Dale Hall, FSA, CERA, MAAA, SOA managing director of research.

“The magnitude of the database allows us to further an understanding of national health care cost and utilization trends, which may ultimately help lower health expenditures when combined with the appropriate interventions.”

SOA aims to use the HCCI data to further explore healthcare cost and utilization trends, as well as the financial impacts of preventive care.

“There are many potential avenues for future work. With the HCCI data, it would be very interesting to explore the many relationships among the members (spatially, temporally and hierarchically) in more depth. It would also be interesting to try to quantify the impact of wellness programs,” the report concluded.


Join 25,000 of your peers

Register for free to get access to all our articles, webcasts, white papers and exclusive interviews.

Our privacy policy

no, thanks

Continue to site...