Healthcare may lag behind other industries in adopting artificial intelligence (AI), but it could be the ultimate proving ground where AI can truly demonstrate its value on many dimensions – by bringing greater productivity and responsiveness to a sector that is thriving financial bottlenecks, paperwork and regulations are burdened, while improving outcomes in life-or-death decisions.
From a financial standpoint, wider adoption of AI could result in savings of 5% to 10% in U.S. healthcare spending — about $200 billion to $360 billion a year in 2019 — within the next five years, according to a recently published report Nationally published paper Bureau of Economic Research. In addition to cost control, AI can also offer non-financial benefits, “such as improved quality of healthcare, better access, better patient experience, and greater clinician satisfaction.”
AI has the potential to deliver startling improvements in health outcomes, the NBER paper’s authors, led by McKinsey’s Nikhil Sahni, point out. “Medical knowledge is growing so rapidly that only 6 percent of what the average new doctor learns in medical school today will be relevant ten years from now. Technologies like AI could provide the doctor with valuable clinical data at the time of diagnosis.” This can also help reduce the wasted hours and degraded patient experience seen in today’s healthcare settings, they add.
Industry observers and solution providers agree that AI can give a boost to healthcare as we know it. The timing couldn’t be better as healthcare systems are in crisis. “We’ve seen hospitals close or stop providing essential services like obstetrics and emergency rooms,” says Mudit Garg, CEO of Qventus. “AI-powered care automation helps healthcare systems increase surgical revenue by maximizing operating room utilization through improved planning. By automating elements of the discharge process, hospitals can also reduce the average patient stay. At the same time, hospitals are struggling with staff shortages exacerbated by the Covid-19 pandemic. They no longer have the bodies to deal with procedural tasks and they have to free up their contractors to do the most important work. AI-supported care automation makes this possible.”
The healthcare industry is “rapidly recognizing the inherent benefits of using AI to study health trends and provide patients with better, value-based care,” said David Friede, vice president of strategic partnerships at DrOwl. “For example, AI can be used to create a digital twin of a patient that you can apply a variety of different treatments to and use this technology to assess possible outcomes, leading to a better patient journey and overall experience.”
Again, healthcare could be the ultimate testing ground for AI, as it’s “an enormously complex sector and possibly the most heavily regulated business in the country,” says Garg. “Also, it’s very resource-intensive, and much of day-to-day care consists of many small but critical tasks performed by doctors and staff. Hospitals and healthcare systems simply don’t have the resources to do all of these tasks manually anymore. AI can simplify many of these processes.”
AI has the power to “improve care, reduce administrative burdens, and deliver better patient outcomes,” agrees Sachin Patel, CEO of Apixio. “By integrating AI into clinical and administrative workflows, complete and accurate decisions can be made at scale. For example, we have seen a 20% increase in fully and accurately capturing a population’s burden of disease, along with helping clinical staff generate patient records five times faster than manual review.”
AI is helping “automate mundane tasks and give employees more time to do work that technology can’t replace,” Patel continues. “An example of this in healthcare is AI combing through mountains of data in the electronic medical record to identify all existing and suspected patient conditions so the doctor can spend more time practicing the art of medicine.”
AI can not only prove itself in hospitals and medical practices, operations in the life sciences and pharmaceutical industries can also be revolutionized by AI. “The time and accuracy that AI brings to drug development and monitoring is simply breathtaking,” said Elizabeth Smalley, Director of Product Management at ArisGlobal. “Think back to just a few years when much of this work was paper-based and cumbersome. It’s easy to conjure up the image of a doctor’s office cluttered with Manila binders overflowing with patient files and records. Then there was the human component of remembering that a patient had an adverse event to a drug, pulling out their chart, reading through the notes, comparing their medical history to another patient with similar reactions, and looking for similarities or clues . This did not even include the reporting phase, where reactions were manually reported one at a time and examined over a period of time to look for patterns.”
AI can change the face — and efficiency — of research and development, says Smalley. “Did you know that studies show that the clinical testing phase of new drugs using conventional methods takes nine years and costs an average of $1.3 billion per drug? Many of these studies still use outdated methods such as manually logging patient diary entries, faxing medical records, and snail-mailing results to regulators, among other processes. Each of these areas and more can be done faster, cheaper, and more accurately with AI.”
However, although the potential of AI is enormous, there are obstacles that need to be overcome to fully realize it in healthcare. For example, accuracy must be 100% at all times – there is little to no room for machine-generated error. “The biggest obstacle preventing AI adoption in highly regulated industries like healthcare is the highest level of accuracy required to achieve adoption,” says Patel. “For a streaming service, for example: ‘If their algorithms are inaccurate, the worst-case scenario is that a customer doesn’t like the recommended movie or series. Although this can be uncomfortable or even annoying, it is not life threatening. However, when healthcare AI is not accurate enough, it can result in an ineffective treatment plan that puts the patient’s health at risk.”
The transition to AI in healthcare “requires more time for mainstream adoption,” says Patel. “The technology needs to be developed consistently with diverse and rich datasets, along with the right telemetry to enable a clear understanding of how AI recommendations are made. Ultimately, the goal is for clinicians to trust the recommendations of the AI.”
However, “It can also be difficult to trust a machine to make a prediction – this is often an issue in the healthcare industry,” says Friede, “Doctors are used to relying on human interactions and specifics to determine the best possible course of action, which include the assessment of a patient’s mental status, expectations, medical history and more.”
It’s important to note that “while AI and machine learning will help us predict the best course of action, they will not replace a doctor’s treatment decision, but complement it to provide additional opportunities,” Friede continues. “The increasing use of AI in healthcare has changed the game. It empowers healthcare professionals to analyze big data to make faster, more informed healthcare decisions, ultimately leading to a better patient experience, reduced costs and more efficient business processes.”
Despite its size and complexity, “we think of healthcare as human-to-human interactions,” Garg agrees. “We want care decisions to be made by doctors, nurses and technicians, not software. Some people mistakenly believe that AI will remove human judgment and compassion from the formula. In fact, it allows providers to spend more time working directly with patients by automating the tasks that take up so much time.”
Follow me on Twitter.
I am an author, independent researcher and speaker covering information technology innovations, trends and markets. I co-chaired the AI Summit in 2021 and 2022 and have also attended the IEEE International Conference on Edge Computing and the International SOA and Cloud Symposium Series. I am also co-author of the SOA Manifesto, which outlines the values and guiding principles of service orientation in business and IT. I also regularly contribute to Harvard Business Review and CNET on topics shaping careers in business and technology.
Much of my research is affiliated with Forbes Insights and Unisphere Research/ Information Today, Inc. and covers topics such as artificial intelligence, cloud computing, digital transformation, and big data analytics.
In a previous life, I was the communications and research manager for the Administrative Management Society (AMS), an international professional association dedicated to advancing knowledge in the fields of IT and business management. I am a graduate of Temple University.
Read moreRead less