As accountable care organization (ACO) contracts gain in popularity among health care systems, the need for a different type of data analytics is growing for provider organizations.  Most health care systems are usually very adept at mining their electronic medical record (EMR) systems to support quality improvement and care management programs.  Reporting of this nature is patient-centric and can be supported by EMR data that typically includes only a partial picture of the medical encounters a person has.  However, the reporting demands of most successful ACO contracts require population-centric metrics.  Many provider-based decision support systems face a number of challenges to meet population-based analytic demands. These include: 

• EMR systems lack a complete medical encounter history (e.g. system leakage)
• Patient enrollment data lacks a complete enrollment/eligibility history
• Pharmacy data is incomplete or not available
• Analytic solutions such as population risk adjusters and episode of care groupers are not available

 Get the Most from Your EMR Data

There are several methods that ACOs can employ to maximize the insight and resluting value they can achieve from their EMR data:

• Develop a member month proxy table based on the patient information included in the EMR.  This method will require gap filling techniques to span periods of no member activity. 
• Leverage all EMR data sources within the system.  Of particular importance is mining the data included in office and clinic settings of primary care practitioners.
• Utilize risk adjustment software that is based on diagnosis only and avoid risk adjustment methodologies that require procedural information as an input.
• Explore the accessibility of All Payer Claims Databases (APCD) in your region/state.  APCDs, while typically member de-identified, can provide a complete episode of care for members who seek services within your system.  Issues such as efficiency of care by episode and system leakage can be identified – highly valuable pieces of information.

 The Ideal Solution

Rather than rely entirely on EMR data, the ideal solution is to deploy and leverage an analytic infrastructure that is designed specifically for population-based healthcare analytics.  Several key characteristics of a successful solution of this nature include: 

• Detailed enrollment information from each ACO contractor.
• Detailed claim information, including allowed amounts, from each ACO contractor.  If allowed amount is not available, encounter data is still useful as RVUs can be used as a proxy for allowed amount.
• Pharmacy data for all members, including NDC and member ID.
• Provider and member matching logic to link data between the various data sources and the systems EMR data.
• Provider attribution logic to assign members to a PCP as well as specialty and facility attribution for episodes of care.
• Classification systems such as DRG assignment, service and utilization count assignment, episodes of care assignment and member risk score assignment.
• Evidence-Based Measures to calculate industry standard quality and care metrics (e.g. HEDIS, PQI, NYU ED Algorithm, etc.)
• Benchmark information from either an APCD or other external source.

 Open the Door to Success

As your organization forms, and then begins executing, its plan for the development of a population-centric decision support system, the first and most important step is to make sure your ACO contracts entitle you to receive detailed enrollment and claim information for the populations you will be taking risk for.  Solve this challenge, and the other elements of a population-centric decision support system will fall into place.