Some believe artificial intelligence has the potential to transform health care from fighting diseases to preventing them in the first place.
Health care data is one of the fastest-growing segments of the digital universe. If the 153 exabytes (or 153 billion gigabytes) of health data that existed in 2013 were downloaded to computer tablets and stacked, they would stretch 5,000 miles high. By 2020 the tablets would stretch 82,000 miles high.
This vastness of digitized health information makes it impossible to manually mine for the insights needed to improve care, according to an industry brief by Dell EMC, an information technology subsidiary of Dell Technologies. Health data grows 48 percent annually, the firm says, gathered daily from doctor office visits, electronic medical records and research to prescriptions, health-tracking wearables and more. And while health data is multiplying faster than Tribbles on the starship Enterprise, there’s a silver lining to the challenge this data growth poses to health care.
Artificial intelligence — computers coded to think like humans—thrives on big data. It can quickly parse bytes of information — petabytes, exabytes and zettabytes — to find ways to improve care. In just the last decade, AI-driven technologies have introduced innovations such as 3D X-rays, predictive health analytics, new drugs with startling curative powers and automated disease management.
As a payer, we see the holy grail as providing consumers with access to … care designed specifically to their needs.
Privacy-protected health data is stripped of individual identifiers (e.g. names, Social Security numbers) before it is fed into AI-coded computers for insight. In some cases, researchers add third-party data, such as demographics and buying patterns, to get a more holistic understanding of particular health issues.
Experts hope artificial intelligence can significantly advance the health industry’s push toward value-based care – that is, paying for care in ways that reward quality and cost-effectiveness. True success will come when AI consistently improves patient outcomes, detects and prevents fraud and waste, and reduces costs.
Ultimately, innovators hope to see artificial intelligence transform health care from simply writing prescriptions for issues that already exist to preventing the maladies in the first place.
“Artificial intelligence has the opportunity to fundamentally change health care,” says Himanshu Arora, executive director of analytics for Blue Cross and Blue Shield Plans in Illinois, Montana, New Mexico, Oklahoma and Texas. “As a payer, we see the holy grail as providing consumers with access to health care that engages them personally with care designed specifically to their needs. AI is getting us closer than we’ve ever been before.”
Ash Damle, CEO and founder of a startup predictive analytics company called Lumiata, likens the use of artificial intelligence in health care to the 21st Century microscope. Just as the first compound microscope 400 years ago revolutionized medicine with cellular exploration, AI deep dives into data to uncover patterns, in real time, that could help identify health risks.
AI’s helping hand comes at an opportune moment for health care. The World Health Organization reported that the current shortage of more than 7 million physicians, nurses and other health care workers worldwide is expected to nearly double by 2035.
Compounding this shortage, doctors are seeing a growing prevalence of chronic disease in the general population. Chronic disease — such as heart disease, stroke, cancer, type-2 diabetes, obesity and arthritis — ranks as one of the most common, costly and preventable health issues we face. Heart disease and cancer, alone, accounted for 46 percent of all deaths in 2014, according to the Centers for Disease Control and Prevention.
Many AI endeavors in health care today are focused on diagnosing, managing and preventing disease. IBM’s cognitive computer Watson, for example, is being used at a handful of cancer hospitals. The company says Watson can speed DNA analysis and help physicians make diagnoses. Medical startup Enlitic uses a deep learning algorithm to improve the accuracy and speed of radiologist’s interpretation of X-rays.
Similar AI-powered technologies are exploring genomics, cell biology, organ therapy — all of which scientists are working to harness to fight disease. From that exploration, for example, came advances such as 3D organ printing, artificial eyes, eye drops that dissolve cataracts and an artificial pancreas that senses blood glucose levels and delivers insulin.
Think of AI as Russian dolls nested within each other. Housed within the definition of artificial intelligence (mimicking human behavior) is machine learning (coding machines to learn on the go), deep learning (analyzing epic proportions of data), neural networks (replicating the way the brain processes information), natural language processing (recognizing and responding to human speech) and more.
Applications of machine learning in the broader marketplace include the self-driving car, ride-sharing apps such as Uber and Lyft, email spam filters, internet search functions, facial recognition and online chatbots.
The Food and Drug Administration recently created a Digital Health Unit to keep pace with rapidly expanding digital technologies. The unit will overhaul the way the administration regulates the digital health industry in an effort to get safe products to market faster.
Data-mining company CBInsights identified 106 companies using machine learning and predictive analytics in everything from drug discovery and virtual assistance to imaging analysis.
Pharmaceutical companies and medical scientists are trying to apply artificial intelligence in various realms, such as disease identification, personalized treatment, new drug discovery and clinical research. It’s too early yet to determine the impact of this work, but expectations are high.
Drugmaker AstraZeneca recently entered a research deal with Berg, a company that specializes in using AI in drug development, to search for new therapies for Parkinson’s disease and other neurological disorders.
MIT’s Clinical Machine Learning Group is developing algorithms, among other endeavors, to better understand disease process and design for effective treatment of diseases like Type 2 diabetes. Microsoft’s Project Hanover has teamed with the Knight Cancer Institute to use machine learning to personalize drug combinations for acute myeloid leukemia.
Lumiata has developed a predictive analytics tool designed to make predictions related to symptoms, diagnoses, procedures and medications for individual patients or patient groups.
Zebra Medical Vision, another startup, uses AI to read and diagnose medical images. Intermountain Healthcare, the University of Virginia and others are using the company’s analytics engine to identify lung, liver, bone, cardiovascular and neurological findings on CT, mammography and radiography scans.
Other companies are attempting to use AI to help people maintain their health and manage chronic conditions.
AiCure uses facial recognition via smartphones or tablets to visually confirm if patients are taking their medication. A study by the National Institutes of Health published in the journal Stroke showed a 50 percent improvement in adherence in a small clinical trial.
Next IT uses conversational AI for customer engagement and workforce support. It is developing a virtual customer service representative to prompt patients about medications adherence, ask about their symptoms and inform patients’ doctors of any issues.
AI is already transforming the retail industry with voice-controlled assistants like Amazon’s Alexa and sophisticated search engines that anticipate a consumer’s wants and needs. But experts say the technology is capable of far more than selling wares — it could help save and enrich lives.