TOP is hosting it’s Demo Day this year on Dec. 10. For 2019, participants gain access to the de-identified precision oncology open data sets from the VA and National Cancer Institute. Oracle’s demo combines its technology with that data to show how it can leverage artificial intelligence and customer experience solutions to connect patients with experimental therapies. The company aims to explore machine learning and AI applications in prospective trail recruitment.
According to Rebecca Laborde, the global lead product strategist for healthcare and precision medicine of Oracle Healthcare Global Business Unit, the current process for matching patients with clinical trials is complex and cumbersome. Healthcare providers search disparate databases to identify potential clinical trials for their patients and must consider several unique factors in recommending patients for these trials.
“All the information needed for patients to find clinical trials is publicly available already, but the problem is that the format of the data and volume of data makes it difficult for patients and caregivers, even physicians, to sort through and find the best fit,” she said. “We spent some time talking to patients and clinicals as well as patient advocacy groups to get an impression of their process today. It’s possible for them to find a match, but they know they are missing out on opportunities because they aren’t seeing all possible trials before making a selection. There is interest by advocates and support networks to find more information as specific to the patient as possible.”
There is also interest by pharmaceutical companies and the industry to more effectively match patients with clinical trials. In some cases, clinical trials that could have great benefit to patients close and don’t progress because they can’t find the correct number of patients to move forward.
Oracle used the Sprint as a research opportunity to use existing Oracle technology and products to put together a data flow and system that connects patients with clinical trials, as well as setting up a format to support communication between patients and other stakeholders. When patients and clinicians currently search for available clinical trials, a lot of the information is non-structured, difficult to mine and not machine readable.
The first thing Oracle did was to take the data that comes to clincialtrial.gov and other sources, and use machine learning and AI to convert it to a machine readable format. Then the company examined communication modalities, including chat bots as digital assistants to help walk patients through the clinical trial process to send their information to a physician, with their consent.
The website would also be made available to the patient’s clinician to enter in the appropriate clinical values in a de-identified format to provide the most detail possible when the patient is considered for eligibility in a clinical trial. Then, using AI, the website could check which clinical trials are available and guide the clinician through the process of entering the patient’s further information.
“You would be able to make the direct matching and allow the clinician to go in and see what factors make the patient eligible or ineligible,” said Ms. Laborde. “It’s not intended in any way to make a clinical decision; it’s all about providing the right data so the clinician and patient can sit in the room for an hour or 30 minutes and have the correct information in front of them so they can make an informed decision about what is best for the patient’s care.”
One of the biggest challenges that patients communicated to the team was the extreme stress they are under when making decisions about treatment and clinical trials. As a result, the Oracle team aimed to sharpen patient education and communication to ease as much of the stress as possible.
“The main thing that came up over and over again from patients and caregivers was the fact that this activity is happening in the most stressful point in their entire lives and their ability to process information is not at it’s peak because of the emotional toll they are going through,” said Ms. Laborde. “The way their information needs to be displayed has to be in an intuitive fashion and easy to navigate.”
The Oracle Health Sciences team spoke with multiple oncologists, practitioners and nurses during the development process as well. Clinicians wanted the process to be easy for patients, but also prioritized patient data privacy. The conversation around data privacy is evolving, and Ms. Laborde said the Sprint allows them to participate in the ongoing discussion.
“We really view this as a way to understand the space better and participate in the conversation so we know moving forward how to build products that support the process but adhere to the standards around privacy,” she said. There are multiple ways data could be collected and stored, including on the cloud or within the firewall of an institution. During the Sprint, Oracle showed that there is flexibility in how the data can be stored and accessed, as well as which tools could interact with it.
“That is important because it means we can adjust the system to adhere to the policies around data use,” she said. “Anything that was too rigid couldn’t evolve fast enough to keep up with the evolution of data standards.”
More articles on health IT:
Houston Methodist’s Dr. Roberta Schwartz on the ‘struggle’ of balancing data privacy with innovation: ‘It comes with friction’
The biggest IT threat to health systems goes beyond cyberattacks
How CIO Myra Davis defined a culture of innovation at Texas Children’s Hospital
