A 2016 Massachusetts Institute of Technology (MIT) report “Convergence: The Future of Health,” suggested the most pressing issues facing healthcare today can only be solved if the American scientific community embraces the concept of convergence and applies it directly to research on the improvement of human health. The report defines Convergence in Healthcare as the coming together of professionals from communities such as mathematics, physical science, life science, engineering and information technology in an effort to create new, dynamic approaches to researching solutions to health issues that continue to adversely affect the wellbeing of humanity in the age of advanced medicine. Through convergence, the medical research community has the best opportunity to change medicine—how it’s diagnosed, treated and prevented—in ways never before possible.
Big Data and Health IT is one of four approaches suggested in “Convergence: The Future of Health” which may hold significant potential for spurring convergence in the healthcare sector. Below are three ways this particular approach is aiding the acceleration of the Convergence in Healthcare revolution.
Along with the Internet of Things (IoT) and artificial intelligence (AI), machine learning is among the topics most widely discussed by tech industry pundits in the last several years. Its potential is apparent in the impact it has had on some of the most revolutionary recent technologies—it is a key component of what makes services such as Netflix, Google, Siri for iPhone and Amazon Echo function. In this sense, it is unsurprising that machine learning has a place in the Convergence in Healthcare revolution and is already changing the way that physicians perform life-saving work.
For example, research shows machine learning may be an invaluable tool for doctors seeking early diagnoses of certain diseases and syndromes, such as the prevention of sudden cardiac arrest in patients with heart disease. Heart disease kills more than half a million Americans each year, yet could be prevented when diagnosed and treated in a timely manner. Machine learning could help doctors identify potential cardiac arrest candidates early, leading to intervention in patients at an elevated risk for cardiac arrest before they experience a critical episode.
Health IT devices for patients
The more information a physician has about a patient, the more comprehensive the picture of patient health and the ability to more accurately make diagnoses to save a life. Health IT is the convergent way to provide a vast pool of structured data, since consumer-focused health IT devices allow doctors to gather specific and passive patient health data to create a full portrait of patient health. This approach is visible in the invention of items such as FitBit and their accompanying apps. Convergence has also encouraged the creation of a variety of smartphone apps dedicated to the collection and transmission of health data and records using HIPPA-compliant processes.
The health IT movement continues to impact the convergence revolution through the creation of new sensors to monitor biological systems and collect passive patient data using methods of advanced signal processing, ultrasound and bio-instrumentation. Additionally, health IT is spurring a growth in convergent software programs to scan physical characteristics or movements to identify metrics such as pain levels for patients who are hesitant to vocalize their distress. It is also helping earlier diagnose of certain illnesses by using IT-based screening applications.
High-throughput molecular profiling
Health IT has a less direct but more impactful potential in areas such as high-throughput molecular profiling. In this instance, machine learning is combined with additional efforts in biology, analytical chemistry, next-generation sequencing, in situ gene expression profiling by imaging and computation to create a technology capable of a specific single-cell analysis that allows researchers to identify the antigen candidates most likely to enable a cancer patient’s T-cells to identify and attack antigen cells present in protein fragments from cancerous mutations. This could effectively lead to the development of personalized cancer vaccines.
This form of high-throughput molecular profiling may pave the way for a more broadly deployable cancer vaccine using data from fewer patient trials. Machine learning could synthesize vast amounts of data from a small collection of patients who show strong positive reactions to distinct forms of immunotherapy. When combined with other machine learning-based programs, such as software used to identify tissue’s cellular patterns, the technology could help medical researchers create a widely applicable cancer vaccine.