The Big Data Revolution—Is It Our Turn?

As we all are reminded each and every day, we are now in the big-data era.

In many respects, the insurance industry was the original big-data industry, relying essentially since its inception on the aggregation of massive amounts of claims and loss data to create underwriting algorithms to insure risk.

That said, we seem to have lagged in some respects in fully participating in the current big-data revolution. New government-created health data collection and aggregation platforms may, however, give us the data resources to explore new opportunities, such as developing tools to help drive plan participants toward more effective and efficient healthcare providers.

CMS Data

At the federal level, the Centers for Medicare & Medicaid Services (CMS)— the largest single payer of healthcare services in the country—has created a publicly accessible Medicare claims database. Who may access what varies by the nature of the requestor and the intended use.

For example, to access data containing beneficiaries’ protected health information (PHI) (which can be used to individually identify them), a requester must plan to use such data for “research purposes,” sign a data use agreement, and be reviewed/ approved by CMS’s Privacy Board. The standard data-use agreement clarifies that, to qualify as “research,” CMS must determine that the use will result in “generalized knowledge” that will “provide assistance to CMS in monitoring, managing and improving the Medicare and Medicaid programs or the services provided to beneficiaries[.]”

However, CMS’s public-use files— which do not contain any data that can be used to identify the beneficiary—may provide our best opportunities for access because these data are readily available to the public for any purpose. The data sets are fairly robust and include information on, among other things:

• Physicians and other suppliers (utilization, payment, submitted charges, place of service)
• Inpatient and outpatient services (utilization, payment, hospital-specific charges)
• Medicare Part D prescribers (number of prescriptions dispensed, total drug costs).

There are, however, practical limitations of the public-use files. First, they are by definition limited to Medicare participants and covered claims (CMS has embarked on a data collection/sharing initiative for Medicare Advantage plans, and access is, at least initially, restricted to researchers).

Second, none of the data are risk-adjusted to account for differences in underlying severity of disease of patient populations.

And third, the public-use files do not contain any information on quality of care (readmissions/follow-up visits/etc.). CMS has separately begun publishing quality of care ratings for group practices on its Physician Compare portal and intends to add ratings for individual physicians and other healthcare professionals in the future.

State All-Payer Claims Databases

All-payer claims databases (APCDs) are state-sponsored, large-scale endeavors that systematically collect and require the submission of medical claims, pharmacy claims and dental claims, as well as additional information about provider and patient demographics from public (e.g., Medicare and Medicaid) and private payers. They are currently being implemented or operated in 19 states: Arkansas, Colorado, Connecticut, Delaware, Florida, Hawaii, Maine, Maryland, Massachusetts, Minnesota, New Hampshire, New York, Oregon, Rhode Island, Utah, Vermont, Virginia, Washington and West Virginia (a 20th—Tennessee—apparently has abandoned its efforts to operate an APCD, at least for now).

APCDs generally were created as a tool to measure healthcare costs, quality and utilization and to support public transparency of that information. The Virginia APCD, for example, was created expressly “to facilitate data-driven, evidence-based improvements in access, quality, and cost of health care and to promote and improve the public health through the understanding of health care expenditure patterns and operation and performance of the health care system.”

APCD operation—e.g., what data are collected, the availability of such data, and who has oversight authority over the program—varies by state.

Some states—e.g., Colorado and Washington—publicize reports and data sources for public consumption via a consumer-facing website. This allows people to assess differences in pricing for common procedures, to check costs, and to effectively evaluate quality performance throughout the state. Other states—e.g., Maryland and Oregon— require interested parties to formally request access via application and/or to sign a data use agreement with the data management authority (which can be a state agency or an external nonprofit) to gain access to a specific data set.

As with the CMS data, limitations also exist in the state system. Notably, over half of states have not yet adopted a health data aggregation system. And some of those states that have begun collecting data have very restrictive access rules. Hawaii and Minnesota, for instance, currently will release their data to government entities only.

We have developed a comprehensive survey that outlines the current state of play for all of the state APCDs (what data are collected, who may gain access, etc.) at ciab.com/healthdata. It may be helpful as you begin to assess whether these might be valuable tools for you. In the healthcare space, we really are at the inception of the big-data revolution. But data sources are out there. And recognition of the importance of such data—for businesses and patients alike—to encourage informed, value-driven purchasing and care decisions is growing.

The new tools that are being created are like first drafts. Together, we will have the opportunity to build upon and perfect them by identifying ways current data collection efforts can be enriched and standardized to maximize usefulness in our space. “We” can only do that, though, to the extent you are engaged—both with the new tools and in sharing your experiences and views of those new tools with us. Let’s get that conversation started.