Skip to Content
To date, medical providers have aimed to use AI to aid in early detection of diseases. A recent example of the AI and healthcare merger is a two-day workshop that began on May 16, 2018. Experts from the AI community and the Academy for Radiology and Biomedical Imaging Research met at the National Institute of Standards and Technology in Maryland to begin the first step in creating the equivalent of a “diagnostic cockpit” system. The system would merge a patient’s medical history, genetic information, and high-resolution imaging data with AI technology. The desired outcome would be to reduce healthcare costs and improve the quality of the diagnosis in a plethora of diseases, ranging from neurodegenerative disorders to coronary artery disease.
As part of the progression by the healthcare industry from fee-for-service to value-based care, businesses have also moved to develop healthcare products that utilize AI to assist in patient care. Therefore, it should not come as a surprise that Accenture estimates that the healthcare AI market will reach $6.6 billion dollars by 2021, and can create $150 billion in annual savings for the U.S. healthcare economy by 2026. Further, it is worth noting that the top ten AI applications that would create this $150 billion in annual savings are: robot-assisted surgery, virtual nursing assistants, administrative workflow assistance, fraud detection, dosage error reduction, connected machines, clinical trial participant identifiers, preliminary diagnosis, automated image diagnosis, and cybersecurity. Perhaps the most pragmatic example of AI playing a real-time role in providing services is Buoy Health. In November 2017, Optum, the health services business of UnitedHealth Group, invested over $200 million dollars into a startup that will advance healthcare systems. The most visible product of this investment is Buoy Health, an AI heath assistant that provides reliable information after a patient notifies it of their symptoms. It utilizes an algorithm to analyze patients' answers and then provides appropriate recommendations on what the patient should do to treat the illness.
The utilization of AI has extended to pharmaceutical companies and the drug discovery process as well. The development that AI has undergone in this sector is noteworthy because the cost to develop a new drug is currently estimated to be more than $2.5 billion dollars. On June 6, 2018, at the Spark and AI Summit in San Francisco, a Unified Analytics Platform for Genomics was announced by analytics vendor Databricks. The singular platform utilizes genomic data processing, tertiary analytics, and AI to perform drug development research. To fully appreciate why this platform is a force multiplier, it is imperative to understand a central impediment in drug discovery research. Before pursuing advanced drug research, it can take scientists several weeks to scan genomes to find a relationship between a gene and the risk of a person developing the disease they are researching (i.e., leukemia). Understanding the way a genetic variant operates is essential to creating a drug that will treat the genetic variant and, in turn, the disease. The Unified Analytics Platform accelerates a life sciences organization’s ability to generate petabytes of genomic data that will allow them to make developments in new treatments for millions of patients. Thereby, it accelerates a life sciences organization’s ability to perform critical research.
As the use of AI continues to expand into the fields of healthcare and life sciences, challenging legal issues will inevitably arise. These issues will likely include federal regulation, intellectual property law, and privacy law. As AI plays an increasing role in informing the decisions of medical professionals, the question of how to ensure patient safety through effective federal regulation will take on increasing importance. It is currently unclear whether the Food and Drug Administration (“FDA”) has statutory authority to regulate healthcare-related AI as a “medical device.” However, the federal regulatory body has taken a proactive approach to the issue, holding public workshops, publishing guidance documents, and providing premarket review and guidance for breakthrough AI technologies. For example, in April 2018, the FDA granted a “Breakthrough Device” designation to a software program called IDx-DR through its “De Novo premarket review pathway” program, which approves new medical technologies for marketing. IDx-DR uses an artificial intelligence algorithm to detect a greater than mild level of diabetic retinopathy in adults with diabetes by analyzing images of the eye taken with a retinal camera. To qualify for a “Breakthrough Device” designation, a medical device must provide for more effective treatment or diagnosis of a life-threatening or irreversibly debilitating disease or condition, and meet at least one of the following criteria: 1) it must represent a breakthrough technology; 2) there must be no approved or cleared alternatives; 3) it must offer clinically meaningful advantages over existing approved or cleared alternatives; or 4) its availability must be in the best interest of patients.
Medical AI will also present challenges related to intellectual property law as technology develops. Developing such healthcare-related AI algorithms requires large investments in research and development. However, a 2012 Supreme Court decision suggests that medical algorithms are not patentable under section 101 of the Patent Act. In Mayo Collaborative Services v. Prometheus Laboratories, Inc., the Supreme Court applied the rule that “laws of nature” cannot be patented to a diagnostic test that assessed the measurement of a metabolite level in a patient’s blood to determine whether the dosage of a drug should be adjusted. According to the Court, the “well-understood, routine, conventional activity previously engaged in by scientists who work in the field . . . is normally not sufficient to transform an unpatentable law of nature into a patent-eligible application of such a law.” Going forward, the patentability of such valuable medical AI technologies is sure to become a hot-button issue.
Finally, the merger of AI and healthcare will surely lead to issues related to patient privacy. When creating medical AI algorithms, developers utilize large amounts of healthcare data from various sources to train the algorithms and ensure their effectiveness. This data may then be shared with other organizations evaluating the algorithms. This collection and sharing of patient data clearly presents privacy issues that may not be adequately covered by the Health Insurance Portability and Accountability Act’s Privacy Rule, which applies only to individually identifiable health information and “covered entities” like healthcare providers, health insurers, health information clearinghouses, and their business associates. In response to these and other concerns, the “FUTURE of Artificial Intelligence Act” was introduced in the U.S. House of Representatives in December 2017. The Act would establish a Federal Advisory Committee that would study and assess, among other issues, “(h)ow the privacy rights of individuals are or will be affected by technological innovation relating to artificial intelligence,” “[w]hether technological advancements in artificial intelligence have or will outpace the legal and regulatory regimes implemented to protect consumers,” and “[h]ow existing laws, including those concerning data access and privacy, should be modernized to enable the potential of artificial intelligence.” Investigation into these issues will be essential to protect the privacy of millions of Americans.
It is clear that the fields of healthcare and life sciences are artificial intelligence’s next frontier. The effects of this merger of medicine and technology will assuredly be felt across all sectors of society. Stay tuned.