Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.
Current state of the art for producing high quality patient level data suitable for research is to leverage clinicians / nurses. While this approach can be accurate, it is slow and does not scale to the volume of health care data necessary to answer the most important questions in healthcare today. There are also tasks that humans simply do not perform well on - for instance reading every word of the very large documents that can represent a full patient journey. Prominent real world evidence companies have attempted to address this limitation by leveraging NLP, machine learning, AI, etc. None of the
Our biggest differentiator is Resolve, which can weigh evidence across thousands of patient level documents to identify the most correct value for each data element of interest to a researcher. Other technologies can extract a subset of clinical facts and return them to the user. Mendel goes several steps further to understand each word in the sentence, in context and can summarize a patient record automatically. No other NLP / NLU / AI system can do this. We are the only end to end AI-driven clinical abstraction and research engine.
We have validated our approach both holistically and vs many commonly used technologies from the big 3 cloud companies, Google, Amazon, Microsoft. We compared our OCR technology to Google OCR on a set of 430 documents / images. Our system produced signficaintly higher Precision and Recall scores while showing an error rate of only 6% compared to 25% from the google system. We've also had a third party validate our Redact technology which allows for fully automated redaction of PHI information from unstructured documents. Our system performs at 100% Precision & 99.85% Recall - meaning we perform almost perfectly in redacting all the items that should be redacted and almost done of the items that don't need to be. We are not aware of any other system that can perform at this level. While we're proud of the breakthrough work in these individual components, we're most proud of the impact of our end to end platform. In a recent study involving 50,000 patients, researchers cut the time spent abstracting charts from 27,000 labor hours using primarily-manual methods down to 15 minutes using Mendel. With Mendel, the burden of abstraction no longer dictates how often researchers can iterate on study parameters.