For decades, we have discussed, debated and anticipated ‘The Future of Healthcare.’ Every 50 years, healthcare takes a great leap forward into what has now been dubbed The Healthcare Revolution. As the healthcare industry joins the digital age, data-driven discoveries and technologies are brought to the forefront, and we arrive at the cusp of unprecedented advancement, patient empowerment and evidence-based medicine.
Tomorrow came yesterday and the future is now – welcome to the golden era of innovation in healthcare.
But let’s not get ahead of ourselves. While artificial intelligence could provide novel insight into disease or assist in developing personalized treatment plans, the role of healthcare professionals still remains as vital as before. “Artificial intelligence” could really be looked at as “augmented intelligence,” with AI functioning as an aid or assistant to healthcare specialists.
AI has been a long-standing hot topic in the world of tech, but only now has AI become a visible reality in healthcare (and in some cases, an FDA-approved one).
However, there is still a considerable amount of debate on the topic, including the hurdles that must be overcome prior to its adoption by the healthcare industry. Here are some of the opportunities and considerations for AI in healthcare.
– Discovery: AI will make a large impact on drug discovery by identifying patterns within billions of aggregated data points.
Technology: Deep Genomics, an AI-driven genetic medicine company, has developed a platform designed to unlock the best drug candidates for patients. The technology’s knowledge of cell biology and a cell’s ‘molecular machinery’ has aided Deep Genomics in the discovery of new classes of AI-identified gene therapies and advanced them for clinical evaluation.
– Diagnostics: AI has received significant attention in diagnostics and biomarker detection. AI-powered image pattern recognition has been found to be effective in early detection of certain cancers and diseases. Additionally, AI is increasing in its application of treatment response monitoring and early cancer detection via blood tests.
Technology: Arterys, an AI-centered medical imaging company has created a web-based imaging and analytics platform that will function as an ‘assistant’ to radiologists by evaluating medical images and shortening the process from 45 minutes to 15 seconds. The platform also flags specific images that require a specialist’s attention and further evaluation, allowing radiologists to spend their time more efficiently, while helping to achieve better patient outcomes. As the first FDA-cleared broad oncology imaging suite powered by deep learning, the company’s platform abilities include cloud supercomputing and patient data storage, as well as AI solutions in the cardio, lung, and liver imaging space.
– Data Parsing: Various AI platforms have the capacity to increase the “intelligence threshold,” or the amount of data that can be analyzed to extract meaning. This could allow AI to identify connections that were previously unknown or overlooked as unrelated contexts.
Technology: Bigfoot Biomedical is seeking to transform the standard of care for people with insulin-requiring diabetes with AI-driven systems designed to more precisely and automatically administer insulin doses. The company’s virtual clinic offered critical foresight into the efficacy and performance of their automated insulin delivery system during trial, allowing them to simulate clinical trials in a matter of minutes for a fraction of the cost. Bigfoot’s solution is founded in machine learning algorithms, while leveraging existing smartphone and cloud-connectivity technologies to create a smart, connected ecosystem for people with insulin-requiring diabetes
– Clinical: Starting this year, we will begin to see AI implemented in the clinic. Clinical optimization includes engagement and risk factor prediction, streamlining workflows, virtual medical assistants, diagnostic prediction via data extracted from wearables, intelligent clinical decision support at point of care, assigning patients into specific categories of care management intervention – the list goes on.
Technology: Spry, a digitally driven healthcare company, has created a wearable health monitoring and insight technology, or “Loop.” Aimed towards patients with chronic health conditions, Loop’s capabilities will include continuous vital sign data monitoring, cloud-based analysis, and provision of actionable insights to deliver ‘the right care at the right time’ when vital signs indicate an acute health event may occur.
– Health IT: One of the concerns raised by the implementation of AI in healthcare is the potential exposure of personal health data. One issue being where machine learning physically occurs, whether that be on a consumer device, or a private vs. public cloud. Various platforms are positioned to mitigate that risk by securing data through Blockchain interfaces. Patient data privacy and security have always been a priority within healthcare. As the industry shifts, it becomes even more imperative to understand the extent of privacy and possible IT risks that may arise.
– Regulatory Roadblocks: The root of this issue begins with a question: how does one regulate a self-improving and evolving algorithm? While there is no cut and dry answer, we know that there must be an evolution of standards for AI in medical application. The FDA’s Digital Health Program strives to deliver clarity through practical approaches that weigh the benefits and risks within the digital health field.
– Value: Herein lies the assessment of which AI-powered applications use intelligence to resolve unmet healthcare needs versus those that exercise the hype train of AI for its marketing value. AI has the ability to improve and streamline efficiency, as well as open the door to new discovery. But in order to provide true value to the healthcare industry, AI should be used to advance patient care by helping improve outcomes or providing further accessibility and affordability for more people, rather than a mere means of promotion.
– Workflow Integration: One of the most substantial hurdles for AI lies in implementation and integration into the healthcare system – be it your doctor’s office or the operating room. The digitalization of health records is exemplary of this issue as the process cannot happen overnight, nor should it. AI’s use within discovery and treatment, as well as diagnostics, imaging, and treatment monitoring may be met with by stakeholders in order to evaluate value, transparency, and compliance structure that would ensure a smooth adoption into the clinic. Additionally, AI implementation must demonstrate its ability to streamline clinical workflow and offer interoperability rather than encumber clinicians.
In the last year we’ve witnessed increasing excitement and interest in AI, but in the last few months we’ve seen some skepticism grow in parallel. While we cannot predict the future (yet?), we can educate ourselves on the current state of the healthcare industry and where it’s headed – and something tells us we’ve got big things to look forward to.
– Lara Lingenbrink