emtelligent is a Vancouver, BC-based company that partners with healthcare institutions, companies, and organizations to strategically structure their unstructured medical data, helping them increase safety, operating efficiency, and quality of care. Pulse 2.0 interviewed emtelligent Chief Growth Officer Kim Perry to learn more.
Formation of emtelligent
How did the idea for the company come together? Perry said:
“The inspiration for emtelligent came from the frustration of cofounder, Timothy O’Connell, M.D., over the inability of traditional natural language processing (NLP) to provide clinicians accurate and well-organized patient information at the point of care. As an emergency radiologist, Dr. O’Connell always wanted more patient data to guide his clinical recommendations and decision-making, but the existing NLP software couldn’t extract information from unstructured data, which comprises 80% of all medical data. Dr. O’Connell also had long experience in computers, so he worked with another cofounder, Anoop Sarkar, Ph.D., to create a medical-grade NLP platform capable of extracting and normalizing unstructured data. emtelligent launched in 2016.”
Core Products
What are the emtelligent’s core products and features? Perry explained:
“Our proprietary, enterprise-scale emtelliPro NLP engine is purpose-built to reveal key insights buried within unstructured medical text. Utilizing advanced deep-learning models and a custom medical annotation framework, the emtelliPro engine ensures greater accuracy across healthcare contexts. Built by medical experts to mimic their own understanding of medical language, emtelligent’s models transcend traditional NLP capabilities to enable an in-depth understanding of clinical context and complex relations.”
“emtelliSuite is our collection of medical apps that demonstrate how emtelligent’s medical AI solution improves the efficiency and effectiveness of care providers, researchers, and administrators. Apps include a specially trained search engine, a smart summary app, a follow-up detector, an EMR data analyzer, a trainee scoreboard, and a tool that highlights abnormal findings and follow-up recommendations in text-based diagnostic reports.”
Challenges Faced
What challenges has the team faced in building the company? Perry acknowledged:
“A major hurdle for us when we started was that the market wasn’t ready for our product, which really is a medical-grade AI platform. That was because customers didn’t understand that what we were doing was vastly different and superior to what they had seen from traditional NLP in healthcare until then. And what they had seen wasn’t very impressive.”
“As I mentioned, traditional NLP hasn’t been able to extract unstructured data, so most healthcare organizations that tried it were underwhelmed. Convincing them that emtelligent’s medical-grade AI is a quantum leap beyond traditional NLP has been challenging because it requires them to be educated about the capabilities of large language models (LLMs), and that doesn’t happen overnight. Still, we’ve made – and continue to make – great progress in explaining how emtelligent technology can transform healthcare.”
Customer Success Stories
Upon asking Perry about customer success stories, she explained:
“We helped one of the largest U.S. payers build its clinical data mart. This payer was already sitting on a tremendous amount of clinical data from the care delivery assets it owned. The early use cases needed to glean the insights from the unstructured text in the payer’s Continuity of Care Documents (CCDs) to enable the healthcare economics team to create more accurate risk models and to better inform care management and quality teams. A pipeline of new use cases was created and prioritized.”
“Other success stories include increasing clinician productivity through our platform’s ability to summarize a patient’s clinical history and easily enable the creation of notes. We also have improved revenue and profit margins for radiology groups that missed about 70% of the patient follow-up recommendations.”
Total Addressable Market
What total addressable market (TAM) size is the company pursuing? Perry assessed:
“Because our software is built from the ground up to understand all medical text, the emtelligent platform is a true enterprise tool that enables clinical use cases across the entire healthcare industry. We are focused on the payers, health systems, and pharmaceutical companies as the three verticals with priority. Another emerging vertical is life insurance.”
“Within these verticals, the largest and most mature companies are on a journey to build enterprise capabilities. These organizations are our enterprise clients. For the organizations that don’t have the time, resources, or talent to build a platform themselves, they rely on a vendor community to aggregate and cleanse data for use.”
“Additionally, there are thousands of vendors that optimize a single use case or business process (think risk adjustment, prior authorization, revenue cycle, clinical research, real-world evidence). The vendor community that supports these industries also needs to deploy NLP to further enhance their product offerings.”
Opinion Using Generative AI For This Use Case
Can generative AI work for this use case? Perry pointed out:
“At this point, generative AI algorithms such as ChatGPT are too imprecise and prone to error, which makes them less than clinical-grade. Not only will generative AI train a firehose of keyword-driven data on the user, there’s a good chance some of that data will be fabricated. That’s because generative AI currently has problems with “hallucinations,” or invented facts. This limits clinical use as the technology continues to mature. In contrast, our platform has the ability now to understand, contextualize, prioritize, and summarize real-world data from patient records, including unstructured data. Further, all data presented to clinicians cites the source data in the patient record, making hallucinations impossible.”
Differentiation from the Competition
What differentiates the company from its competition? Perry concluded:
“Many companies have built NLP applications for specific use cases in healthcare, and some of those apps offer real value. What differentiates emtelligent from other companies is that our AI-powered platform provides the tools to structure medical data so it can be accessed by NLP apps across multiple use cases and medical specialties.”
“The other thing that makes our platform different is that it was built for clinicians, with input from clinicians. Developers of traditional NLP software for healthcare generally lack the medical expertise to know what type of information would be useful to a clinician or what questions they need answered. Some NLP programs, for example, may offer clinicians a diagnosis based on patient symptoms detected in medical data when what clinicians actually want is a relevant medical history of the patient so they can make the diagnosis. They’re not looking for software to do that for them.”
“Finally, our ability to leverage machine learning and medical expertise allows the emtelligent platform to attain the optimal balance of recall and precision, giving us an accuracy rate of 95% and higher.”
