Nova Approval is Creating a TurboTax for FDA Filings

Founded out of Stanford University, Nova Approval uses artificial intelligence to streamline the FDA filing process. Nova Approval aggregates decades of FDA documentation into a search engine that functions as a digital consultant for FDA submission.

Nova Approval is Creating a TurboTax for FDA Filings

Emily Skahill | June 22, 2018

Filing for FDA approval is costly and inefficient for pharmaceutical companies. Companies spend 4.6 billion dollars each year on regulatory consulting and often wait between six months and three years to have their drugs approved. The lag time between when companies finish developing drugs and receive FDA approval can mean less treatment options for patients in critical need.

Founded out of Stanford University, Nova Approval plans to use artificial intelligence to streamline the FDA filing process. Currently, the company is aggregating decades of FDA documentation into a search engine intended to one day function as a digital consultant for FDA submissions.

We had the opportunity to hear from Michelle Wu, CEO and Founder of Nova Approval, on how the company is bringing regulatory compliance into the 21st century.


What was your initial inspiration for Nova Approval? 

I spent three years as a Business Development and Licensing Manager at Novartis, where I often conducted market research for new treatments. I found it frustrating how long it took to get new drugs approved. It was heartbreaking to look into the eyes of those I was surveying – some of them with Stage IV cancer – and say, “Hang in there. We’ll try get this product approved in two years by the Chinese Food and Drug Administration.”

Now as a business school student at Stanford, I’m seeing that even in the most advanced life sciences market, the lengthiness of the FDA approval process is a persistent obstacle.

There’s urgency in the market to speed up the process by shortening the timeline for clinical trials. This would be very difficult for the FDA, however, because it’s is responsible for public health, and shortening trials is a critical mistake.

The way we bridge this gap is by focusing on reducing the timeline for the review process. It can take from six months to three years for the FDA to complete a review because so much time is spent on manual tasks. Sometimes dossiers and applications can be 2,000 pages long. One regulatory consultant I interviewed had a room filled with papers that she was rewording and formatting by hand.

I saw an opportunity to use machine learning, deep learning, and automation to streamline the review process.

What makes your team especially capable of solving this particular problem?

Our laser focus and our diversity are key.

As a student at the Graduate Business School I’ve had the opportunity to extend my network into the best computer science department in the country and recruit incredibly capable people to my vision. I was careful, however, to craft a team of students whose enthusiasm extended beyond their computational and engineering curriculum. They all hold a strong interest in making an impact in the life sciences, and my job is to point them toward that goal and get out of their way.

As a young team entering an older industry, I want to say we bring a fresh perspective to the problems at hand, but the truth is, with our diversity of backgrounds and cultures, we really bring a multitude of fresh perspectives, and this is how we generate incredible ideas.

The Nova Approval team of Michelle Wu, Eric Loreaux, and Ashwin Ramaswami


How do you plan to use artificial intelligence to expedite the FDA approval process and foster communication between the FDA and industry? 

In assessing the current regulatory market and devising a long term plan, we realized that we were entering a space that has experienced very little technological innovation in a long time.

With some of our more ambitious goals, it felt like designing desk lamps without first securing a source of electricity. Consultants, companies, law firms, investors, all have been bravely operating in this space with very few specialized services, and so we’re first working to empower these people with newer tools to interact with the existing data.

We’re currently in the process of building the first dedicated search engine platform for medical devices, which pulls together all FDA data and acts as an intuitive portal between the community and the FDA’s websites.

Not only does the tool help users find the proper information, but it also generates beautiful visualizations that tell a story beyond that found on the FDA site. Our reasoning for the tool is simple: newcomers and experts alike have trouble determining the proper regulatory path for a new technology, and this becomes a time sink. With this tool, some of that search time can be given back to the patients.

Ultimately, we want to create a tool that redefines what it means to apply for FDA approval.

We intend to do this through a series of feature releases on the platform we are developing. Many of these features will involve AI, and harness powerful newer technologies in fields such as natural language processing and decision support. Some of our features will be much simpler, and may help users build application documents quickly and without mistakes. Many of these tools we’re developing will not only empower the applicants but the regulatory agents as well. By organizing these two groups around a set of seamless technologies, the cross communication can replace our current system of official letters and correspondences.  This system is one of the primary causes of drawn out application times.

How do you generate revenue? 

We will only charge for our products when we know they are generating comparable value for our users.

Our search engine is in its early stages, and we are not yet sure what our users will want to get out of it, so we will be relying heavily on the community for feedback. Throughout this period the service will remain entirely free, and may continue to remain so indefinitely. We have a ton of ideas for valuable features to build on top of our initial platform, and if the value provided by these features is sufficient, we may be able to support a permanently open search and analytics ecosystem.

As these more advanced features develop, we would expect to introduce a premium membership to provides access to these tools for a monthly fee. It is too soon to speculate on what these fees might look like, and how flexible they may be, suffice to say that users will find these services to be well worth their cost.

How much traction are you seeing among customers? Any particularly insightful feedback?

It will be hard to gauge traction until our product is in the hands of users, which will finally be happening next week. In terms of product interest, we’ve seen an overwhelming amount, and that gives us confidence in releasing our tool. The community is on our side – they want better tools.


Do you have any major development goals moving forward?

Our plan is to release our search engine next week, and begin intensively working with community partners to improve its functionality and match the industry’s needs.

This work will continue through the summer. Over this time we will also begin work on our next two projects, one AI project and one application-building project. Think Turbotax for FDA applications. We plan to release a tool along these lines sometime in late 2018.

What is your long-term objective with Nova Approval, and what do you see as your greatest hurdles to achieving this?

We’re trying to disrupt a system that spans the entire globe and involves every single government on Earth. There is nothing about our goals that is not ambitious, but if we didn’t think we could succeed we wouldn’t be doing it.

In order to go down in history as one of the first companies to introduce government regulation to the 21st century, we must do just that, and therein lies our most difficult hurdle.

The government has a reputation for being slow and resistant to new things, but technology is rewriting that story. Industries that once evoked images of people, paper, and tedious work – banking, insurance, and finance to name a few – have all undergone massive makeovers and now evoke images of high speed networks and advanced information systems. The same is inevitable for government regulation, and while it may not seem obvious now, our job is to make it so. We, the patients, owe it to ourselves to invent a regulatory system of the future that delivers faster results for all of us.

Tell us about your planned product launch.

We’re really excited to be releasing this tool for the community. Our team is full of perfectionists, and early on we realized that if we let that side take over, the product would fall into a never ending loop of iteration. We elected instead to release what we had at the beginning of the Summer and  finetune the product using actual customer feedback and insight.

The search engine resembles more of an open source project than a company product at the moment, and we like it that way. Real organic development will be crucial for Nova Search – only the users know what they want out of the tool. It’s our job to figure out how we can get it to them.

How do you plan to expand your portfolio globally?

The US life sciences market, regulated by the Food and Drug Administration, remains the premiere market for health technologies. It also remains one of the most stringent regulatory systems on earth because it operates on a case-dependent basis and reviews applications in depth.

Neither of these qualities scale very well, and companies outside the US who don’t have an internal point of contact are at a disadvantage. These companies are looking for better ways to translate ideas and information into the US market in order to secure approval. US companies who are looking to expand globally will also be looking for ways to translate ideas and information across borders. If we can succeed at introducing some form of automation and optimization into this process, we can provide useful tools to these international companies and facilitate the spread of new and exciting technologies.

Our goal for now is to develop tools to aid companies targeting the US market. Once we have succeeded in this regard, other markets will be much easier in comparison, and so we can leverage our experiences with the FDA and our connections abroad to develop outbound channels for US companies expanding globally. We ultimately aim to become a “one-stop-shop” for global applications, offering quick translation and reformatting between markets.

Message for our readers?

We’re still looking for talent. If people are passionate about saving lives, enabling medical innovations to reach patients faster, or have contacts at the FDA, we’d love to talk to them.

Share This Post