In our series of blogs on the topic of population health management, we have discussed how to define populations in "Analytics for population health management," and we have looked at the types of analytic tools available for analysis of populations in "Identifying appropriate metrics for population management." In this blog, we will focus on types of data, both traditional and nontraditional, that can assist in the performance of a robust population health analysis.

Many healthcare organizations have access to a wealth of data. However, this data has not historically been brought together in an organized manner. An emerging concept in data management is "big data," an approach where all data relevant to an individual's healthcare—including those data that may exist outside an individual organization's walls—are gathered and linked for analysis.

To effectively gather and link these data for population health management, the enterprise data warehouse needs to be flexible enough to accept data from traditional and nontraditional sources. Data sources that can be accessed and linked for robust population health analysis include:

• Insurance administration data (e.g., claims, pharmacy, enrollment data): Insurance administration data is by far the most commonly used data for analysis.
• Clinical data (e.g., electronic medical records, lab results, registries): Clinical data increases the richness of data available about an individual and a population. This type of data is starting to make its way into the enterprise data warehouse.
• Medical management data (e.g., health risk assessment, authorization, disease/case management data): Most health organizations have medical management data readily available, but most have not yet brought this data into an enterprise database.
• Provider administrative data (e.g., physician practice management, hospital billing, admissions discharge, and transfers data): Provider administrative data is generally available in a more timely manner, allowing for data analysis that is nearer to real-time.
• Public data (e.g., state discharge data sets, immunization registries): Public data has not typically been accessible in an enterprise data warehouse, yet it can provide additional insight for population health analysis.

Why incorporate all these data?

Access to a wide variety of data allows for broader population health analysis, which can lead to new and earlier insights to help identify improvements, efficiencies, and effectiveness of the healthcare delivery system. A population health management strategy with a robust data warehouse can help answer the question, "Who will benefit from which programs?"

This article first appeared at Milliman MedInsight.