Janssen's Lessons Learned From Digital Health Implementation
By Ed Miseta, Chief Editor, Clinical Leader
In his role as product line owner for digital health at Janssen, Mark Sapp leads a team that works directly with different therapeutic areas inside the company. The clinical teams identify the data points they would like to capture from an upcoming trial, and the digital health team works with them to determine what technologies they might want to deploy. This team will oversee the implementation from ideation through to deployment and conclusion of the study. The digital initiative at Janssen has been underway for more than six years.
“Our focus is helping our clinical teams get from point A to point B,” says Sapp. “We assist them with taking their idea and scaling it up, identifying the clinical endpoints they hope to achieve, and helping them stay focused on the results they need to generate. We ensure these teams will have a successful technology implementation.”
One of Janssen’s success stories is the Janssen Autism Knowledge Engine (JAKE) started in 2015. JAKE is an application used for trial participants with autism and was just recently launched for a compound study. In some ways, this was the company’s first foray into digital health.
Janssen has another robust digital health application called Revere. The application replaces a clinician-administered test and uses speech recognition to measure cognitive impairment. It is something that Janssen has used in a few different areas. As with many of the company’s digital health programs, Revere started off as a proof of concept and has grown into a full-blown program.
Project Avatar is another proof-of-concept success story. For this endeavor Janssen partnered with another company to build a video game that can potentially identify and longitudinally track depression in people based on how they play the game. Sapp states Project Avatar is a legitimate game, not just puzzles.
Janssen has another ongoing digital health-powered study for bipolar disorder that uses real-time data from patients that help clinicians identify patients who may be at risk of committing suicide.
An Evolution in Development
When Janssen first dipped its toe into the digital space, the company was forced to fit digital solutions inside a predefined box. A protocol was developed which referenced data that needed to be collected. An application then had to be built that would fit that mold. Sapp notes that was not a very forward way of thinking.
“Today the process of incorporating digital devices is more ideal,” he states. “A therapeutic area will start designing an upcoming trial and state what data it hopes to capture. Our digital team will then be asked how it can help, and that’s when the fun begins. The digital health team will roll up its sleeves and start looking at different ways it can use technology to engage the patient.”
The digital team will start any assessment by asking, What is the best data you can possibly have? It then looks at different ways to capture the data. If a possible solution is identified, it may be tested in a conference room study to see if the technology will perform as desired. Sapp also explores whether it is possible to scale the technology.
In an ideal world, the digital team will be able to influence the writing of the protocol. While that is not always possible, Sapp notes a portion of some protocols can be centered around the digital technologies being deployed.
The thinking around digital technologies is also changing. The go-to approach for many is simply capturing data on a mobile app. According to Sapp, that is yesterday’s thinking. He prefers to look at less invasive ways to capture data, and believes biosensors and wearables are better and can lessen the patient burden.
Get All Stakeholders Involved
Sapp states that in many cases it doesn’t make sense to build a fully-validated, GXP-compliant, HIPAA-valid technology solution for an internal pilot program. Doing that would take too much time. His primary concern is to quickly determine whether a technology will be a good fit for a study. Doing so requires his team to have a discrete definition of success.
One of his major challenges is a proof-of-concept study where someone has a list of things they want to accomplish in a short amount of time. Oftentimes, he must tell them that in the designated timeframe, they can only accomplish two or three of those tasks. That will lead to a discussion of what is the bare minimum that can be done and provide the information required. Then a decision can be made on whether to move forward.
A cross-functional team is part of that assessment framework. That team will involve individuals from across the organization, including legal, quality, compliance, security, and procurement; all are valued stakeholders in the development of digital assets.
“We might have a great device and it may be capable of providing us with the data we need,” says Sapp. But our security folks might say, ‘No way are you going to deploy that in the hands of our patients.’ Procurement might say, ‘It’s a great idea, but that can’t possibly scale up to meet our production requirements.’ All of this input helps us determine if the solution will meet our internal requirements and the needs of our stakeholders.”
Technical Versus Clinical
Another challenge for Sapp is technical validation versus clinical validation. The technology team needs to make sure it can deploy a solution and get good data. But what is “good” data? That question becomes a hot potato. The clinical team will ask if the data is clinically validated. The technology folks will say, “I don’t know. You’re the clinician, you figure it out. We’re just collecting the data.”
“These are questions that have to answered early on,” says Sapp. “Everyone has to agree that the data can be collected, that it is valid, and that it will support the study endpoints. I believe in getting the technology issues worked out first. Ensure security and legal are on board. Then you can focus on giving your clinical team the best data possible.
Being able to scale the technology is also important. Sapp has found brilliant technologies that met the needs of the clinical team and patients, only to have issues come up with the manufacturer. Sometimes he will find out the manufacturer’s 510(k) is still in process. In some cases, he found out manufacturing issues would result in a delay of several months.
Sapp also believes that if you’re going to invest a lot of money into getting this type of data from a patient, you should always try to maximize the value of it. That strategy should include reusing the data as much as possible. Years ago, data would be gathered, used for one trial, and then stored away somewhere. Today researchers are looking into ways to store the data so that it can be reused.
“We need to spend more time focusing on how we get the data and how we bring it into the organization,” says Sapp. “We have to apply context to it so that somebody else can look at that data and get some use out of it. The key word here is ‘platform.’ How can we standardize our approach to bringing data in, so we’re focused more on science and less on building internal infrastructure?
A Plan To Properly Manage Devices
Managing deployed assets is the final piece of the puzzle. Shipping a hundred interconnected biosensors around the world and putting them into patients’ hands, sometimes in resource-limited settings, can be a huge task to manage. Devices will break and need replacing, there must be a process in place to minimize downtime, and there must be people available to answer questions.
“We have had to work through all of these issues in the last several years,” states Sapp. “We learned this requires a real people perspective. For example, almost every year in September a new iOS update comes out. When that happens, you must doublecheck every iOS application to make sure they’re going to work with the new update. You need to have a plan in place to manage all of that. Digital health solutions are products and you need to have a true product manager overseeing it with adequate budget.”
Finally, be ready to expect the unexpected. Even if you think you made something fool-proof, someone will come along and prove you wrong. Sapp has seen some interesting data come in, only to find users were not using the application as intended.
“We’ve put some things out in the wild and had some wild things happen,” says Sapp. “Always make sure you’ve got the right focus in the right place from a clinical and technology perspective. And make sure you have the best help-desk support available, even if that means you must outsource it. You could do everything yourself and end up building a very monolithic organization. We have tried that, and it was a nightmare. Doing it all yourself will result in a lot of sleepless nights and gray hair.”