Mastering Data in Healthcare: Beyond Patients

RHIA,Global Healthcare Industry Ambassador, IBM Information Management

Increasingly, recent conversations with customers around the world have included the need to address multiple types of data, commonly called “multi-domain” in master data management (MDM) circles. This stems from several factors: an increasing emphasis on data exchange; the rise of accountable care organizations (ACOs) in the United States; greater understanding that care coordination is one of the pillars of improving quality and reducing costs; the rise of consumerism in healthcare; and the big data and analytic movements.

In the past, many organizations may have deployed what was called EMPI (enterprise master patient index) technology. Now, as their needs have grown broader, IBM MDM technology can help healthcare organizations meet the challenges of managing data across several domains.

Let’s explore some of these domains and their importance:

Patient/Person – Who is the patient/person being treated despite all the variation in names, addresses, and key demographic data? Answering this question accurately and quickly is foundational to all healthcare needs. The privacy implications should be obvious—matching patient records accurately is essential to avoiding inadvertent disclosures. Plus, with patients as the core of ACOs, matching patient records across the many systems in a healthcare enterprise is essential to aggregating individual data to demonstrate higher quality and reduced costs.

Provider – Who is the provider delivering care? Is it an individual, a corporation, or a small business? The complexity in defining a provider (i.e. individual, pharmacy, home health agency, taxi service, nurse practitioner, chiropractor) makes this domain particularly challenging, with accuracy playing very high stakes in the world of ACOs and integrated care. Financial and privacy implications abound if this domain, like patient, is not mastered well.

Insurer – Payers or health plans will need to master the insured, as well as members and subscribers. Payers are increasingly putting higher emphasis on member service, member satisfaction, and data sharing to create competitive advance. And payers are launching their own ACOs.

Reference data – Mastering enterprise reference data is problematic and costly in many regards, with the challenges increasing as physician practices consolidate or healthcare systems acquire many previously small or independent facilities. Examples include mastering or crosswalking ICD 9 to ICD 10 codes, Snomed, CPT or ICD lab codes to LOINC, and medications to RxNorm. Managing enterprise reference data becomes more important as costs increase and the security, business and regulatory implications become apparent.

Managing multiple domains heightens the importance of a highly scalable, accurate matching and searching technology. Applying rules to potential matches simply doesn’t cut it in this day and age; the stakes are too high. Rules don’t scale, require updating when new data sources are added, and can’t address the variation in data capture that is common in today’s diverse healthcare ecosystem. Rather, organizations that want to master multiple domains should use probabilistic algorithms that will scale, deliver subsecond response times, and be easy to implement and maintain. Plus, these algorithms must allow the customer to create and apply multiple thresholds to matching based upon the use cases for the domains.

Healthcare organizations and jurisdictions would do well to include a multi-domain approach to mastering key data in this effort.

Read more: Lorraine's take on mastering the Location and Organization domains in healthcare

This post was originally published on August 27, 2012 on the Mastering Data Management Blog.