Most customer-grade AI tools are trained on the premise that finding something is better than finding nothing. To please the end user, the model will present whatever data results it can. But MRO‘s Prodigy tool, which performs clinical data abstraction, will refuse to present search results unless they’re believed to be of high quality. Chief Product Officer Mo Weitnauer, at the AHIMA Annual Conference, says they achieve 97% accuracy in using the new tool for unstructured EHR/EMR clinical data abstraction.
The reason for this rigor is that clinical data abstraction is a “high-consequence area” used for such sensitive tasks as patient recruitment for clinical trials, infection control, and accreditation. Prodigy automates the harmonization of unstructured EHR text into CDISC SDTM format using ontology-based normalization. Context is key in this specific use case of real-world evidence (RWE) curation: for instance, you might be searching not just for patients with a particular disease, but patients who had that disease during a particular time period.
The critical role of context also means that autonomous AI is not a good choice for the task. Instead, Prodigy works with a human abstractor, performing initial searches.
As clinical data abstraction is being used in more and more areas, Weitnauer says, it is also getting more difficult. This is due to the growing complexity of both the input data and the applications for that data. Patients have comorbidities, regulators are asking for more types of data, and the standards that clinicians have to meet are getting more detailed.
Weitnauer mentions that 40% of the abstractor’s time, traditionally, is spent finding and copying data from the EHR. Prodigy does that task, rendering the abstractor more efficient. MRO stresses interoperability and taps into many types of EHRs. In short, MRO wants to improve how data can be used strategically.
Quality is built into Prodigy, Weitnauer explains, by leveraging the expertise of clinicians in building the models. The company doesn’t focus on a single specialty, but has “tons and tons” of data across the field of medicine. Prodigy also tracks when an abstractor corrects its output, using that information to improve its own quality.
Watch our interview with Mo Weitnauer from MRO to gain a deeper understanding of the value of clinical data abstraction and what it takes to do it effectively.
Learn more about MRO: https://mrocorp.com/
Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.
And for an exclusive look at our top stories, subscribe to our newsletter and YouTube.
Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.
MRO is a proud sponsor of Healthcare Scene.
No comments:
Post a Comment