Improving Data Integration & Quality by Leveraging the HIE

By VITL Chief Technology Officer Michael L. Gagnon, CPHIMS

I recently had the opportunity to present at a webinar sponsored by Health Data Management and co-hosted by our technology partner Health Language. The title of the webinar was “Meaningful analytics: Improving data integration and quality by leveraging the HIE.” Health Language provides VITL with a capability to “normalize” local medical terminologies to national standard codes. This terminology services engine is becoming a key component of VITL’s Health Data Management Infrastructure (HDMI) which provides the tools and services to improve the quality of data used for population health.

Vermont has a very high adoption rate of electronic health record systems (EHRs) in its health care organizations, which is great. However the data coming from these EHRs is not always ready for use in population health analytics. As health care moves towards value-based reimbursement models, having useful clinical data that is ready for analysis will be one key to success.

Vermont is moving forward with health care reform as evidenced by the recent approval by both the state and federal governments of the Vermont All-Payer Accountable Care Organization Model Agreement. The all-payer model will create one advanced payment model in the state in the form of a single accountable care organization (ACO) which is being named the Vermont Care Organization (VCO). The VCO would include beneficiaries from Medicare, Medicaid and commercial insurers. Health care organizations would have the option of joining the VCO or striking out on their own under the Merit-Based Incentive Payment System (MIPS). MIPS right now only covers Medicare but it seems likely that Medicaid and commercial will follow suit.

So how does all this change to the payment system impact VITL and the health information exchange? From its inception in 2005, VITL has focused on providing services for health care organizations. From the very beginning we recognized that a centralized data repository would be required at some point for any data analysis. About two years ago we began to purchase and implement the infrastructure to process, collect, and store data as well as analyze data quality for use in population health. While we continued to build interfaces to EHRs to collect data we also recognized that much of the data was in free text or described in local terminology. We could collect the data but not always discern the meaning of it. Being able to understand both the structure of the data and the meaning of it is called semantic interoperability. To help us achieve semantic interoperability we purchased a license for the Health Language terminology services engine to map local terminology to standard code sets. More recently we have been working with Health Language to implement their Natural Language Processing (NLP) solution to extract medical terms from free text in transcribed reports and care summary documents.

With Health Language, the Vermont Care Organization, the Vermont Blueprint for Health and our health care organization clients we are moving towards greatly improving the quality of the data we collect. This high quality data can then be used to ensure that population health solutions will work to improve the care provided to all Vermonters!

Click this link to watch the recorded webinar