Sandy Hathaway of Exit3x expands on how to use AI to your advantage in the healthcare tech revolution.
While medicine is often thought of as a hotbed of innovation and scientific discovery, the healthcare industry as a whole has historically resisted the widespread implementation of new technologies for a variety of reasons.
But change is coming whether providers and patients are ready or not.
In particular, continuous advancement in the field of artificial intelligence is opening up new possibilities for disruption across nearly every industry, and healthcare is no exception.
That’s a good thing because some of the industry’s most pressing problems are problems that can be addressed by AI. Among them is a massive shortage of physicians, which the Association of American Medical Colleges predicts is only going to worsen. In fact, a new report by the AAMC forecasts a shortage of anywhere from 40,800 to 104,900 doctors by 2030, suggesting that American healthcare providers might soon be overwhelmed by the needs of an aging population.
Another problem is that healthcare providers are swamped with data, and when there's so much information at your fingertips, it's hard to keep up on the latest recommendations, guidelines and innovations. AI can translate all that data into key insights that can be applied to a patient.
Fortunately, AI can alleviate some of the burdens that providers face now and will face in the future. The purpose of AI is to help — not replace — humans. So, in fact, AI could rather stand for augmented intelligence.
Great expectations
For starters, AI technology can make physicians more accessible. At present, electronic health records represent a valuable store of data for reference and use, but they also create more work for physicians, who need to enter patient data in order for it to be viable. Machine learning technology — the precursor to true AI — can now receive and store healthcare data automatically and intelligently, allowing physicians to spend more quality time with patients.
Moreover, machine learning technology is becoming a powerful foundation for diagnostic tools. Millions of pages of medical records can be quickly scanned and analyzed to provide effective diagnoses and treatment options in a matter of seconds. For example, in one study, IBM Watson was able to detect brain cancer in a patient and suggest a treatment plan in just 10 minutes — it took human doctors 160 hours to do the same. Watson can also help doctors discover patterns in cancer patients, allowing them to match patients with the best treatment paths and most accurate survival rates.
Machine learning has also led to advances in triage. The technology can take into account current best practices and evidence-based guidelines, as well as study the nuances of each individual case — much like an experienced doctor.
Automated diagnostics could potentially lead to better patient outcomes while simultaneously saving providers a lot of money. In fact, according to McKinsey & Company, AI-enabled healthcare initiatives could save the U.S. $300 billion annually, and full AI adoption could make registered nurses up to 50 percent more productive.
But for all its promise, AI is far from ubiquitous in the healthcare industry, and a number of hurdles stand in the way of widespread adoption.
Reality check
Perhaps the biggest obstacle preventing the technology from taking hold is access to data. Data is the fuel that powers artificial intelligence tools, and in the healthcare industry, it’s especially abundant.
The hard part is reaching it.
Healthcare data largely exists in private, tightly controlled EHR systems, laboratory and imaging repositories, physician notes and health insurance claims. In order for AI to be viable in the future, all of this information must be merged in colossal integrated databases so machines can capture and analyze it. Such a shift will inevitably take time.
Patient perspectives on AI might also have to shift. According to McKinsey, it’s unknown how much trust patients would place in AI-powered diagnostics and treatment plans. Even if, in theory, machines were more capable than doctors in the identification of health problems, the lack of understanding from patients as to how machines reach a particular diagnosis could prove difficult to overcome.
Likewise, regulators may be hard-pressed to endorse AI without understanding the operations that lead to a diagnosis. Acquiring that knowledge wouldn’t be easy, as black box algorithms process billions of data points that are far too expansive for the human brain to comprehend.
Similarly, concerns exist about AI’s potential to minimize or remove the human-to-human interactions that are the cornerstone of the provider-patient relationship. That said, we are still a long way away from a world of robot doctors, and despite AI’s vast potential, the relative lack of digital adoption in the healthcare space up until now means the industry will likely lag behind others in harnessing that potential.
Even so, there are a number of companies in the space that are already implementing AI technology in innovative ways, and their early success offers an exciting glimpse of what the future might hold.
1. Ada Health
Ada Health is an AI-powered app that's often referred to as the "Alexa for health," and it's one of the fastest-growing medical apps of 2017. The app uses a chat interface to interact with patients, help diagnose their ailments, and then connect them with real doctors.
Although it's been in development for six years, it didn't launch until late 2016 — and since then, more than 1.5 million people have used it. The Berlin-based startup also recently raised $47 million in funding to open an office in the U.S., hire more employees, and continue improving the product.
Ada's goal is to make healthcare more accessible to people around the world, combining AI technology with medical expertise to create a patient-centric experience. Because the app uses AI to start the diagnosis process until a doctor takes over, this saves patients a trip to the doctor's office in cases of minor ailments and saves doctors time by already having a prediagnosis by the time they examine a patient.
2. Cardio Cube
CardioCube, an early-stage startup and participant in the current cohort of Startupbootcamp Digital Health Berlin’s 2017 program, is tackling the problem of unplanned hospital readmissions due to cardiovascular disease. Using voice assistants such as Amazon Alexa, Google Home and its own hardware, Cardio Cube is able to initiate an ongoing dialog with at-risk patients.
The company aims to merge its growing collection of home patient data with electronic health records and then power the analysis of that data with cloud-based deep learning and AI algorithms to determine whether a patient needs to seek hospital care based on American Heart Association clinical guidelines
Using the power of natural language processing, AI can perform complex analyses that help patients make faster and more effective decisions about their health.
3. CareSkore
CareSkore is a startup that uses machine learning technology to aggregate and normalize billions of data points and millions of patient records to provide end-to-end care management. In fact, according to CEO Jas Grewal, the company has “23 data partnerships [and] data from 15 different hospitals on around 39 million patients.”
Using all that data, the company is able to offer providers a 360-degree view of their patients’ health while reducing the administrative burden for clinicians and allowing them to spend more time caring for patients.
AI-powered innovation in the healthcare space is still in its nascent stage, but with companies like Google, Amazon, IBM, Microsoft and Facebook using the power of their data to develop AI systems aimed at healthcare applications, traditional players in the space need to move fast or move over. Healthcare incumbents need to be engaging with startups that are pioneering these new approaches or watch them be snatched up by the major tech players that want to take over.
The entire ecosystem is changing. Organizations that recognize this now and form collaborative partnerships with others leading that change will ultimately prevail. Those that don’t will fall by the wayside.
Sandy Hathaway is a founding partner of Exit3x and also co-founded technology startups RetentionGrid, which focuses on big data and predictive analytics, and AVARI, involving machine learning and personalization. For the past 20 years, Sandy has worked in strategy, business innovation, go-to-market and growth roles in the medical device industry, including with U.S. market leader Medtronic and German remote monitoring pioneer BIOTRONIK, and she specializes in working with companies in heavily regulated and deep tech industries.