Atlanta-based data insight company OMNY Health has added 4 billion unstructured clinical notes to its network.
The data was collected in collaboration with provider organizations, including academic medical centers and integrated delivery networks, as per the company. In an interview with Drug and Device World, Dr Mitesh Rao, CEO and Founder of OMNY Health, stated “democratizing access to clinical notes for researchers and health systems will play a major role in advancing health equity initiatives and the advancement of AI”.
Adding, “When patients tell their doctors that they had to stop taking medication because they could not afford it or missed treatment because they didn’t have a way to get there, we gain more awareness of the social determinants of health. Only by truly understanding the challenges underrepresented populations face can the healthcare industry develop effective solutions to address care gaps.
The more clinical data available to AI models, the more diverse and intelligent they will become. By incorporating additional data points into our ecosystem, we ensure that the AI companies utilizing our technology are training their models on the best available information. These efforts will specifically help expedite things like drug development and medical device trials.”
Unstructured data insights
Dr Rao noted that structured data formats like Electronic Health Records (EHRs), refer to information organized in a standardized, easily searchable, and analyzable format, like patient demographics, lab results, or medication lists, which are typically stored in predefined fields
“In contrast unstructured data, means the data points don’t follow a set structure and sit in pre-defined fields. An example of unstructured data would be the notes a doctor jots down during a visit—they’re full of rich information about a patient and visit but often lack standard formatting. When a researcher or clinician then zooms out to analyze them on a broader scale, it becomes significantly more difficult to pull insights”, added Dr Rao.
Mark Townsend, Chief Clinical Digital Ventures Officer at Bon Secours Mercy Health & Accrete Health stated: “Unstructured data is compelling on multiple levels, beginning at the ‘point of entry’. Using discrete fields for data-entry is onerous, inefficient, and contributes to clinician burnout.
With the advent of ambient documentation, data is increasingly being captured in an unstructured format. AI can be used to ‘file’ the data into specific fields. Ambient documentation will likely lead to growth in the volume of unstructured data as clinicians are no longer compelled to also serve as data-entry technicians, entering data into specific fields.”
Potential uses
Dr Rao noted that the use cases for “leveraging unstructured clinical data are limitless”. “While we didn’t target certain disease groups organically, by pulling these insights, we are able to offer more insight into certain disease groups, like metabolic health and dermatology”, added Dr Rao.
“Take this recent study on generalized pustular psoriasis poses as an example. Historically, researchers faced many challenges in understanding how flares and disease severity affect healthcare resource usage. By democratizing access to clinical data, OMNY Health helped researchers utilize real patient experiences, allowing for a better understanding of how this disease impacts patients beyond just their physical health”, said Dr Rao.
Adding, “this same sort of clinical data was also used to gather information from patient experiences to better understand what drove patients to discontinue the use of GLP-1s, such as financial concerns, pregnancy, and drug contradiction.”
Orr Inbar, CEO and Co-Founder of QuantHealth, one of the OMNY’s pharma customers, said: “Leveraging the information found in unstructured data will be a game changer for us as we use AI to simulate clinical trials. Historically, so much has been lost in a provider’s notes about how drugs are impacting patients, especially as it relates to underrepresented communities.
Now, by having access to these insights, we can unlock this information and have it help inform the next generation of pharmaceuticals with our clinical simulator. As an AI company, our top priority is ensuring our model is trained on the highest-quality, diverse data available. It is refreshing to work with a partner like OMNY that continues to uplevel the quality of data available for health tech and health systems alike.”
