By: Eyal Gura
Before the weekend, we announced the launch of AI1, a new way of making our artificial intelligence (AI) algorithms – current and future – available to healthcare providers for $1 per scan. We’re so happy to have seen this social-impact business model covered in by Forbes , Engadget, Venturebeat, and other notable publications. among others.
Clearly, we have a bias toward the significant impact AI will have on the healthcare field for years to come. However, over the past year there have been many conversations within institutions and publicly at conferences and through the media about how quickly and easily AI may be applied. Given our experience in developing the Zebra Medical Vision platform and testing it in more than 50 hospitals globally, we thought we’d share some of our thoughts on how to best approach integrating AI successfully.
Scale is Everything to Global Healthcare
We’ve analyzed over a one million patient scans over the last couple years to develop and test our platform, with just a handful of people. This was while they were working their day jobs as engineers and coding the platform or working as professional academics at different institutions globally. In any average radiology team, this scale would take full-time effort and much more time. As we look at a growing population – and specifically a growing body of middle class and elderly patients – in need of more medical attention, our healthcare solutions must be able to scale more effectively. There is only so much more – if any – benefit that can be derived from shorter hospital visits and trying to pump more medical professionals through their schooling. This particularly isn’t a solution while rates of entry amongst radiologists and other medical professionals are falling.
AI provides the scale of problem solving critical to keeping pace with a growing population as well as delivering on any hope we have of better diagnosing today’s greatest ailments, like cancer. However, to properly integrate AI and its potential, we will need to make it as slim and adaptable as possible. In the startup and systems communities, we need to take notes from the e-Health and e-records challenges that are still being faced more than a decade after that “solution” was promised to healthcare teams and patients. Teams developing AI solutions must be willing to work around existing systems and processes – as well as be mindful of all the existing regulations as well as some to come – to truly provide impact. We’ve learned this lesson ourselves as we ended up spending more time than expected learning about various team practices and processes in reading radiology scans. We then spent even more time developing the simplest, lightweight widget that sits atop any system currently in place.
Our Hope in Healthcare
We’ve been working daily with so many practitioners and academics in the healthcare space over the last few years. We’ve shared stories and had impassioned conversations about our shared vision of being able to help more people get and stay healthy around the globe. The hope that we can help make healthcare more affordable and accessible across geographies and socioeconomic levels is very personal to us, as we’re sure it is to many of you. That’s what motivated our approach to providing our technology at a low per scan rate. Unfortunately, we can’t stop people from getting sick or injured, but hopefully we can make it easier for anyone to get diagnosed and treated. Our hope, at Zebra Medical Vision, is that other existing or developing companies see this as a way to shift their own models. We hope that as people develop new AI technology for doctors and patients, that they see they can be a successful business by providing their technology and services broadly. AI and other new technologies do cost a lot of time and resources to develop, but many of these endeavors will be funded by private and investor or venture capital funds. Given those funding sources, we hope that companies like us will think about how to extend those initial funds to help the greatest amount of people over the long term.
The healthcare industry is ripe for all kinds of innovation and development on local and global levels. To allow for the best possible integrations of these developments, like AI, we will have to be patient. These changes will take decades to develop and fully realize. And we will have to think about new ways to provide these technologies broadly, so they can lift the standard of care any patient can benefit from around the world, regardless of condition or means.