In its recent report, “Artificial Intelligence: Healthcare’s New Nervous System,” Accenture predicts that growth in the AI health market will reach $6.6 billion by 2021. That’s a compound annual growth rate of 40 percent.
That impressive growth rate is feasible because AI has wide-reaching applications across the spectrum of healthcare. Think cybersecurity, robot-assisted surgery, fraud detection, virtual nursing assistants and more. Accenture estimates that, combined, these applications represent a savings potential of $150 billion.
Two areas I’ve had my eye on in recent months are radiology and administrative workflow assistance. Global research and consulting firm Frost & Sullivan predicts that in coming years, “Clinical support from AI will strengthen medical imaging diagnosis processes. In addition, the use of AI solutions for hospital workflows will enhance care delivery.”
As a venture capitalist, I’m already seeing the potential of AI becoming reality in the form of startup companies leveraging this technology to turn possibility into delivered solutions.
Definition of terms
The potential and implications for AI vary widely by industry, so let’s begin with a common understanding. I like Amazon’s definition because it conveys both the depth and simplicity of AI, which it calls “the field of computer science dedicated to solving cognitive problems commonly associated with human intelligence, such as learning, problem solving, and pattern recognition.”
Medical Futurist calls AI “the next successful area of cooperation between humans and machines.” The general idea is that artificial intelligence allows computers to learn and improve over time. AI is especially adept at quickly processing large amounts of information with high accuracy—or, to put it differently, performing tasks that can overwhelm or fatigue humans. Odds are high if you work in healthcare that you’ve encountered overwhelmed or fatigued colleagues. AI can help. Let’s look at how.
Automated image diagnosis: A $3 billion savings potential
Just under a third of all cancer in women is breast cancer. Early detection can make all the difference in a patient’s outcome. Cancer Research UK found that 90 percent of women diagnosed within the earliest stage of cancer survive their disease for at least five years beyond their diagnosis date.
IBM Watson is developing a solution called Cognitive Radiology Assistant which analyzes medical images and combines that data with information from a patient’s medical records to facilitate accurate early detection and diagnoses. The desired end result is the answer to a binary question: Does this woman have potential breast cancer or not? If it’s a potential yes, the patient would come in for a secondary or 3D scan. But if it’s a no, the AI has successfully ruled out the cancer diagnosis with no human component required.
Because of the promise of this application, some fear that AI will completely replace radiologists in the not-too-distant future. But I tend to agree with a more conservative take, as described by The Medical Futurist: It’s not a matter of AI replacing radiologists; it’s a matter of radiologists who use AI replacing those who don’t. We will continue to call on radiologists’ expertise to parse complex issues and supervise diagnoses—they’ll just be empowered by AI technology for handling simpler cases and repetitive tasks.
Administrative workflow assistance: An $18 billion savings potential
Another company I’ve got my eye on is PriorAuthNow, which is revolutionizing the prior authorization process within the revenue cycle. Research published by Health Affairs estimates that the annual cost of prior authorization compliance is between $23 billion and $31 billion each year. Currently, this process is painfully manual. It’s handled by a member of the physician office staff or a hospital representative who manually collects the pertinent patient information from the medical record and proceeds to do one of the following submission methods: 1. Perform duplicate data entry into one of the many direct payer portals; 2. manually transfer the pertinent information onto paper and then fax it to the payer; or 3. submit the authorization by phone which often results in long hold times and several follow-up phone calls. To make matters worse, many types of procedure prior authorizations force staff to do a combination of all three methods. Because the prior authorization process is still manual for providers, this often has a negative downstream effect on payers. If a provider calls in a prior authorization, then the payer has to have someone on their end to receive and do the manual intake. Payer portals aren’t the answer either, as many times it is unclear and confusing as to which payer portal the staff should use. Therefore, they often revert to flipping the patients’ insurance card over and calling the number on the back. Not to mention, almost every payer these days has a portal, making the maintenance of submitting and following up on each case an unbearable task. Even if the prior authorization gets submitted successfully via a portal, oftentimes the payer’s medical review process still falls into a manual queue for a nurse to review. To put it plainly, this process is inefficient, time-consuming, human resource-dependent and wasteful.
PriorAuthNow, on the other hand, is setting out on course to automate the entire process. Its solution identifies the critically necessary pieces of information from a patient’s medical record based on the patient’s specific payer and plan and then electronically sends those critical data elements over to the payer for review. The technology auto-fills out the administrative portion of the authorization as well as integrates with the payer’s different clinical systems and partners, making the authorization more streamlined and less of an administrative burden. After a decision has been rendered by the payer, some in real time, PriorAuthNow can push the pertinent details of the decision back into the respective host hospital and physician office, eliminating enormous amounts of manual work on all sides. Automating the authorization process drastically reduces the manual workload of employees at the payer, health system, physician's office and/or revenue cycle management company, freeing them to perform higher value, non-automated patient-centric tasks.
The interoperability question
It’s no secret that most clinical settings struggle to integrate the technology systems already at their disposal. So how can AI solutions like IBM Watson and PriorAuthNow possibly gain traction in the daily running of hospitals and physician offices?
Luckily, many electronic health record companies are adopting a new platform for the electronic exchange of healthcare information called Fast Healthcare Interoperability Resources (FHIR). Created by the non-profit organization Health Level Seven International and accredited by the American National Standards Institute, FHIR has achieved buy-in from major (and historically competitive) EHRs including Epic, Cerner and AthenaHealth. FHIR promises to significantly improve the interoperability gap by serving as a common mechanism for information integration across multiple platforms and devices. This common language could eventually allow legacy systems to communicate with one another across health systems and providers.
Looking ahead
It’s easy to get caught up in the excitement surrounding AI’s potential impact on healthcare. The industry is in need of disruption and I believe an orders-of-magnitude change is just on the horizon. Integrating AI into applications like image scanning and administrative workflows will not only help health providers operate more effectively—it will ensure providers are able to offer quality care as the needs of the population continue to rise. Although change is always scary, I believe the future of AI in healthcare is a very bright one.
# # #
About the Author
Sarath Degala is VP at venture capital firm BIP Capital, where he leads its healthcare investment activities. He offers 16 years of healthcare administration experience achieving noteworthy operational and financial results within a variety of healthcare settings. Connect with Sarath at sdegala@bipcapital.com or on Twitter at @BIPCapital.