The Keenan Blog

Blog

Information of Value: Mining for Insights

• Insights gleaned from health claims data over time or by type help focus benefit design to bend the cost curve and enhance the health of your participants.
• A plan year is an ideal sample from which to identify important utilization trends, but it is also valuable to see whether those trends develop over time or if something was simply out of expectations for one year.
• A longitudinal analysis is particularly useful in understanding how health care costs are changing from year to year.

In our earlier blog article, we discussed how an organization’s health claims data may contain keys to understanding where health care spending is most effective or ineffective. Methods for analyzing health claims data involve progressively aggregating related elements within the database of claims transactions, then rearranging them by utilization performance metrics to determine if unexpected patterns can be found.

Successful analytics rely on a statistically significant sampling of claims experience. Individual transactions and even episodes of care do not provide a lot of insight for the overall management and design of health plan benefits because the sample is too small. So, the data in claims analytics are first de-identified to preserve patient confidentiality, then aggregated to develop an overall picture of the whole group’s utilization in an adequately sized sample. Depending on how many participants are in the group, it may be possible to analyze the claims from a single plan year and produce meaningful results. Smaller populations could require aggregating data over multiple plan years to make the data statistically significant for decision making.

Because the population remains fairly consistent, a plan year is an ideal sample from which to identify important utilization trends. But it is also valuable to see whether those trends develop over time or if something was simply out of expectations for one year. When analytics are used to look at trends over time, it is called a longitudinal analysis. A longitudinal analysis is particularly useful in understanding how health care costs are changing from year to year. It may also be helpful in seeing the effects of an aging population. However, there could be significant shifts in the population from one year to the next because employees select different plans during open enrollment, or an employer has a high degree of turnover. In a longitudinal analysis of two or more plan years, these demographic changes need to be accounted for so that we can correctly understand the utilization changes.

Examination of a single plan year typically involves a process called stratification. This simply means separating out the aggregated data by similar characteristics. Typical kinds of stratification may include:

• By type of service (e.g., inpatient, laboratory, physician, etc.)
• By diagnostic category (e.g., circulatory, digestive, infectious diseases)
• By patient age segment (depending on group size, meaningful segments may vary from 5- to 15-year brackets)
• By point of service (e.g., hospital, physician office, pharmacy, etc.) or by provider
• By amount of charges (large vs. small claims)
• By geographic location (e.g. 3-digit zip code, county, employer facility, etc.)

Again, depending on the size of the group being analyzed, stratification of these categories can be more, or less, detailed. For instance, it might only be possible to look at the highest level of a diagnostic category; or there could be enough to look at diabetes separately from all endocrine disorders. The more granular detail may be valuable in pinpointing a need for a specific kind of condition management program. High levels of analysis are important for decisions about plan design and financial considerations such as employee contribution strategies.

Insights gleaned from health claims data over time or by type are valuable. They help focus benefit design to bend the cost curve and enhance the health of your participants.

About Ju Anderson
Ju Anderson, Vice President, works with health care clients to create an effective strategy to address the health and welfare benefit needs including the financial analysis, implementation, efficacy and performance management of the health plans and other vendors.