Zoom on Deep6.ai: The AI clinical company

deep6ai main

Clinical trials are the bedrock of medical innovation. Without clinical trials, new drugs, therapies, and medical devices would never end up in a patient’s hands. But clinical trials can be costly and time-consuming, and one of the major reasons for that is patient enrollment. It can take many months to identify and enroll enough eligible patients into a clinical trial, partly because of how long it takes to sift through medical records by hand.

Deep6.ai is aiming to change all of that. Their software uses artificial intelligence (AI), natural language processing (NLP), and machine learning to make identifying patients a process that takes “minutes, not months.”

Who is Deep6.ai?

Deep6.ai is a health and technology startup that is applying artificial intelligence to patient recruitment into clinical trials. They have developed a platform that can be used by researchers, medical centers, and more to find patients for clinical trials by using AI to search through complex clinical documents for eligibility criteria. They are also a platform through which medical device companies can connect with medical centers via an online marketplace.

How does it work?

One of the main reasons why clinical data can take so long to sift through in the search for eligible clinical trial patients is that a lot of it is in free text form. Some medical data is structured, such as ICD10 codes, which are accessible to input and filter through. The majority of patient data, however, is unstructured. It comes in the form of paragraphs of texts and notes contained in documents like doctor’s notes and pathology reports. This type of text, until now, can only be sorted through by hand. And this can not only take a very long time, but it is also labor-intensive.

Deep6.ai applies natural language processing (NLP) to search through this clinical data. Natural language processing is a form of computational linguistics that uses statistical machine learning to quickly, accurately, and efficiently extract data from large datasets of language. In the case of Deep6.ai natural language processing is used to extract tens of thousands of clinically significant data points from unstructured clinical notes. These data points can include symptoms, diagnoses, treatments, and more. They can be processed against clinical trial eligibility criteria to identify those patients that can be enrolled quickly.

The software can also make inferences about medical conditions based on symptoms that a patient may have. This is important because not all patients will have the necessary diagnosis written out in their notes in a recognizable way. However, they may still be eligible for a clinical trial based on their cluster of symptoms.

Why is Deep6.ai important?

This is a very specific application of artificial intelligence, so why is it important? First, using manual methods of identifying patients that meet the eligibility criteria for clinical trials can take many months, and, even then, more than 80% of clinical trials fail to enroll at the estimated time, and up to 40% of clinical trials fail to meet their enrollment goals altogether.

This doesn’t just mean that clinical trials end up costing more money than predicted. Although it does, it can also have a quantifiable impact on patients. For example, the life-saving or life-enhancing drug, therapy, or medical device tested in the clinical trial will take longer to market because of enrollment delays for its clinical trials. This means that patients with the condition will have to wait longer for potentially effective treatment than they would have done had enrollment taken less time.

Clinical trials can also be necessary for the patients enrolled in them. Doctors will often refer patients to a clinical trial hoping that they will receive treatment during the trial that will help them, and often these patients have already reached a brick wall at that point in terms of treatment options. Delayed enrollment in clinical trials means that these patients identified by their doctor as potentially able to benefit from the treatment being tested will also have to wait longer.

Deep6.ai aims to make cutting-edge life-saving and life-changing drugs and treatments accessible to patients more efficiently and quickly. Reducing the time that enrollment for clinical trials takes is a measurable way of doing so.

One study that looked at Deep6.ai’s software examined enrollment in a clinical trial related to Crohn’s disease. Using traditional manual methods to search through clinical data, 5,024 potentially eligible patients were initially found, and this dropped to 30 after months of manual record verification. In contrast, Deep6.ai both found and validated 36 eligible patients in less than 45 minutes.

This is one of many examples where Deep6.ai’s software has dramatically cut down the time that it takes to identify eligible clinical trial patients, and there is every possibility that this method of sifting through clinical data will become the go-to method for clinical trial enrollment. If it does, it has the potential to drastically change the speed at which we can get drugs and treatments out to patients. 

Importantly, Deep6.ai can do all of this without having direct access to patients’ sensitive health information. Data privacy and security is, understandably, a significant concern when it comes to using this type of software in a medical setting. The Deep6.ai software sits on top of the health system data and doesn’t directly access the files themselves, so there is no way to link individual patients to the data it is mining. 

To wrap up

Deep6.ai has the potential to revolutionize how eligible patients are identified and selected for clinical trials. Faster clinical trials mean faster access to potentially life-saving drugs and treatments both for the patients involved in the trial and for other patients with the same condition around the world. In addition, its use of artificial intelligence and natural language processing to parse out the clinically relevant information from complex clinical notes can be used by medical centers and researchers the world over. 

Was this article helpful?