So where are these images coming from? Do patients have to give permission to have their images be part of this greater diagnostic atmosphere? Are there HIPAA challenges to consider here? Where are you in terms of fine tuning these types of diagnostic functions? And how far away are we from seeing this type of technology used by physicians nationwide, do you think?
HIPAA is the Health Insurance Portability and Accountability Act. It’s important to note that the ‘P’ in HIPAA stands for portability. HIPAA is designed to allow medical data to be shared where that sharing can improve patient outcomes. Clearly, sharing image data to allow our algorithm to be built, which can detect disease more accurately and earlier, is a great use of medical data. The only real HIPAA challenges are to ensure that we are careful to treat that patient data with great respect and great care, which is as it should be.
The data comes from hospitals and radiology service providers. By providing the data, the organizations can leverage the technology to improve their own patients’ care, while also contributing to better patient care more widely. So it’s a great deal for patients and hospitals alike.
Our partner, Capitol Health, is in Australia, so I hope that we’ll see patients benefiting from this technology in Australia within 12 months. The U.S. may take longer – the U.S. is the most complex place for health-care innovation I know of in the world. Unfortunately, HIPAA is very vague, and so it makes a lot of people uncomfortable knowing exactly in which situations they can share patient data – the greyness of it is confusing and uncomfortable.
Furthermore, the alignment of financial outcomes with patient outcomes is on the whole very poor. Our technology allows patients to be treated better and cheaper, which, weirdly enough, is not always the best financial outcome for an American hospital. Whether or not we’ll see patients benefit from this as quickly in the U.S. as we’re seeing in Australia will really depend on whether there are forward-thinking medical centers in the U.S. who care enough about patient outcomes to invest in this technology.
Patients don’t have to give permission to have their images be part of this. The only thing they can be used for is to help ensure that other patients who have a similar affliction in the future can have their affliction identified more quickly and more accurately. In some cases, a mistake may even be found in the original diagnosis, allowing the patient to receive the treatment they need (or avoid an unnecessary test).
What’s the long-term value proposition of such technology for providers in the long run? What about patients?
For providers, doctors that are using this technology will make far fewer mistakes and will get through their work a lot faster, allowing them to see more patients. So, clearly, from a financial point of view, they will be more profitable. From a patient-care point of view, they’ll have the satisfaction of knowing they’re doing a better job for their patients. And, also importantly, the work which is more menial is the work being automated, so they can spend time doing things that are more interesting, and in the areas where they can really add value as a doctor.
The patient will get answers much more quickly. Rather than having to wait sometimes days or weeks, they’ll pretty much get answers straight away. They’ll be able to trust in those answers. They’ll be more accurate. They’ll also know how exactly accurate they are, because the level of accuracy will be provided. So, if it’s somewhat uncertain, they’ll know.
They’ll also have more information that they’ll be able to see to understand the context in which they can make their own medical decisions. The software provides a lot of information about that particular patient’s medical situation. And of course, they’ll also see the cost of medical care significantly decreased with this technology.
Most excitingly, in parts of the developing world where previously they had no access to modern medical diagnostics, they’ll be able to get a very high level of care for the first time.
Are there any roadblocks technical or otherwise that will make adopting this technology difficult?
The biggest roadblock is the silo-ization of data, particularly in the U.S. For example, for a rare children’s disease, getting enough data together to be able to recognize that reliably, to diagnose kids with that disease in the future, often requires combining data from multiple hospitals. Generally speaking, those hospitals’ data sets are totally separate. So this is a major challenge. We know it’s costing many lives today.
Is Enlitic’s sole focus on image searching for diagnosis of diseases and cancers, at this point? What’s next?
Our current focus is radiology. Next we’ll be looking at pathology, genomics, and analyzing natural language, to look at electronic medical records. Eventually we want to combine all types of medical data so that we can help provide insights to physicians based on a full understanding of the patient’s situation.