Data Integration in ACOs

Submitted by jonpearce on Sun, 2011-04-17 17:49

The Healthcare IT Guy's recent blog about the analytical systems that ACOs will require does an excellent job of framing the issue of the IT needs of ACOs and other similar organizations.  We would like to highlight several points from this article.

First, these systems will require data integration capabilities of a type not typically seen in hospitals.  Hospital systems typically integrate data though the HL7 standard, designed to allow such systems communicate using the same protocols.  But HL7 is a messaging protocol, designed to transfer data among systems adhering to its standard in near-real time.  By contrast, ACOs will need to absorb historical data from disparate billing, cost accounting, quality reporting and other systems (including some homegrown systems to report on new quality metrics) that don’t comply to any specific standards.  In addition, they are also expected to integrate aggregate and patient-level data from CMS.  This data isn’t “messaged” in the same way as HL7 data – instead it’s transferred in various file structures that need to be loaded, transformed, quality checked and integrated into a format that allows integration and reporting.  The tools used for this process aren’t HL7 tools – they’re the typical data integration tools used by data warehouses in all major industries.  The skills necessary to use these tools may not reside within the ACO’s IT shop, or even within the hospital’s IT departments.  And utilizing them properly in a healthcare setting requires additional healthcare related skills, such as understanding how to deal with CPT code modifiers or understanding the significance of the trailing digits in ICD-9 codes.

This problem can be further complicated as new sources of data from outside of the organization are introduced into the mix.  Singletrack Analytics has worked with organizations that rely on spreadsheets transferred between the organization and PCP offices to report such data elements as blood pressure readings or lab test results for chronic care patients when EHR systems have not yet been implemented.  Because a consistent data structure cannot be guaranteed with a spreadsheet, additional data quality checks must be implemented to assure that users haven’t changed the structure of the spreadsheet or subverted data validation steps. 

These issues point to the need for a comprehensive data acquisition and quality assurance strategy that relies on standard data warehouse techniques, rather than expecting existing hospital systems to accommodate the diverse needs of the organization.  ACOs need to be aware of these complexities and plan their IT strategies accordingly.

This blog discusses the process for getting data INTO the ACO.  A future posting will discuss getting data OUT of the data repository and creating usable information for the clinical, financial and operational managers who will create success or failure for the ACO.