unstructured data: the star waiting in the wings


Data-filled electronic health records hold the promise of predictive, personalized medicine and improved population health. Up until now, much of the data retrieved from records is structured data, information recorded in specific data fields. However, the total amount of structured data accounts for as little as 20% of the EHR with as much as 80% of EHR data unstructured. For example, all MD and other health care provider notes anhd all written reports accompanying examinations (e.g., radiologic data) is unstructured data.

Barriers to retrieving and analyzing unstructured data are many. (Quoting from @Clinithink) It’s messy, dispersed/disconnected, complex, non uniform/non standardized, varied, and inconsistent — all features that make retrieval, filtering, and analysis difficult. Historically, this data has been accessed manually — an inefficient, time consuming process.

Technological projects to get at this elusive data are, however, ongoing and increasing in number, with partnerships between various data analytic firms and hospital systems. A few examples include:

  • IBM’s Watson & Memorial Sloan Kettering looked using unstructured data to help improve decisions for lung cancer treatment. Results have been positive, cutting pre-approval times from 72 hrs to nearly real time.
  • Clinithink CLiX CNLP (Clinical Natural Language Processing) and Swansea University’s Farr Institute (UK). The results improved population health outcomes with reduced cost.

Additionally, unstructured data analysis has been employed by health plans and ACOs to calculate participant profiles to better anticipate and accurately assess risk, and by researchers to help identify clinical trial candidates.

We discussed the topic of unstructured data during the #HITsm chat. Below are the topics. If you were unable to participate in the chat, please leave your thoughts on the topics in the comments!

  1. How will the analysis of Unstructured Data aid in healthcare delivery: for MDs/Providers/Patients/Payers?
  2. How can you, your firms, HIT in general, help in the retrieval & analysis of Unstructured Data?
  3. If you have worked on a project with Unstructured Data-please describe pros/cons of the experience.
  4. Given the base data is text, can you think of creative presentations of the analyzed data that may especially engage MDs/Providers/Patients/Payers?

View the full TweetChat transcript, courtesy of Symplur.

See original posting site here.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s