AI in Healthcare: How It’s Changing the Industry

Healthcare has always involved the intersection of human judgement and scientific data. Advancements in artificial intelligence (AI) are bringing those two elements closer than ever—and the industry is feeling the impact.

Defined as “computer systems able to perform tasks that usually require human intelligence,” data-based artificial intelligence analyzes large amounts of data using algorithms to learn how to do tasks without being explicitly programmed. That capability is creating waves of change as AI in healthcare proves to be a critical component in diagnosis, treatment, care delivery, outcomes and cost.

From big data to policy, artificial intelligence is significantly changing the healthcare industry.

1. Clinical Decision Support

Across the healthcare industry, artificial intelligence is changing the way clinical providers make decisions. More than ever, it’s playing a key role in clinical decision support as it delivers data to providers to aid in diagnosing, treatment planning and population health management.

Consider all the vast amounts of data that AI has the potential to harness—from genomic, biomarker and phenotype data to health records and delivery systems. The technology is already being used to support decisions made in data-intensive specialties like radiology, pathology and ophthalmology. In the future, it may even be possible to perform certain tasks autonomously using this technology.

The key to safe and effective integration of AI in healthcare is rigorous and ongoing evaluation. Systems will continue to improve as AI-enabled decision-making is applied practically, in hospitals and doctor’s offices around the world.

Artificial intelligence even has the potential to decrease the administrative burden on clinicians by improving clinical decision software. With natural language processing, the technology can help translate clinical notes in EHRs. That means a clinician only needs to enter data once. AI-enabled software can also provide access to data from multiple sources—including medical images, EHR data and even consumer devices such as activity trackers, smartphones and connected medical devices. This expands the diagnostic and treatment options clinicians can propose—and has the power to transform health outcomes and create more personalized care delivery.

With the rise of big data from multiple sources and in vast amounts, it’s been difficult to aggregate and analyze that data. But we are now seeing predictive analytics falling under the broader umbrella of artificial intelligence. It is allowing clinicians to discover patterns in multiple sources of data that can lead to better decision-making. For example, it can help nurses determine the appropriate number of days a patient should stay in the hospital—and that can enhance care planning and management to prevent complications, improve patient satisfaction and reduce costly readmissions.

2. Cost

As the industry shifts, there is great opportunity to use AI in healthcare to help drive cost savings.

Leonard D’Avolio, PhD, founder and CEO of Cyft—a company helping organizations use technology to adjust their workflow and systems to reduce cost and improve outcomes—sees artificial intelligence not as a solution but as a capability, allowing us to learn from data in ways we never could before. And that takes an investment. During a HIMSS TV interview, Dr. D’Avolio acknowledged that change is hard in healthcare, but that change is also a prerequisite for improvement.

His advice for those beginning AI implementation projects: “You need executive level sponsorship, and you need to pick a problem with a really solid five-to-one return on investment. Fall in love with the problem, not necessarily the solution.” He encourages healthcare organizations to form a working group with data scientists working arm-in-arm with project managers and the IT team to effectively implement the technology. And then measure the impact.

“You don’t want to wait a year and then look at admission rates or total medical expense. You need to measure baseline, activities, outcomes, every step of the way. If you want people to get behind AI, you need to be able to show them it’s worth it.”

Jonathan Bush, executive chairman of Firefly Health, sees artificial intelligence as a way to clean up the estimated $91 billion in wasted healthcare spending that comes from inefficient administration. “Most healthcare executives are still unsure of their AI strategy. They sense that AI will be a game changer, but they’re not sure how. I love that healthcare has heroic ambitions for a promising new technology, even after years of high-tech disappointment. But while we shoot for the moon, let’s clean up the muck that’s bogging us down today, unleashing our potential to transform healthcare.”

Watch Dr. D’Avolio talk with HIMSS TV on operationalizing artificial intelligence for better care at a lower cost.

3. Healthcare Policy

With this technology evolving rapidly, it only makes sense that policy is racing to keep up. But developing and deploying AI requires a regulatory approach that fosters innovation, growth and engenders trust, while also avoiding regulatory and non-regulatory actions that hamper its expansion.

By developing oversight mechanisms for the technology—that apply both in the U.S. and abroad—regulatory friction could be reduced, allowing technologies and advances to be transferable across national borders, helping developers and the entire innovation field thrive.

To make this happen, transparency is needed to enable regulators to review the process used to achieve an AI-based result or recommendation. They believe regulatory oversight, should focus on trying to get greater transparency without infringing on intellectual property by encouraging the private sector to be able to answer the following types of questions:

• What is the purpose of the AI algorithm?
• What is it designed to do?
• What challenge is it trying to solve?
• How was the data collected to power your algorithm?
• What are the origins of the data?

To safely and effectively integrate AI in healthcare, it will be a slow and careful process, requiring policymakers and stakeholders to find balance between keeping patients safe and secure, and providing innovators with the tools and space needed to make products that improve public health.

4. Patient Data

As the technology is integrated into healthcare, it will become easier to find meaning in the massive mountains of patient data. Lily Peng, MD, PhD, product manager in the Google Brain AI Research Group, explained that while human intelligence is best suited for integrating small numbers of very large effect factors, AI is particularly adept at combing through and identifying patterns in vast numbers and obscure factors.

Chris DeRienzo, MD, MPP, FAAP, senior vice president, chief quality and medical staff officer, WakeMed Health and Hospitals, sees the potential for artificial intelligence to “bend the laws of physics” by allowing one experienced doctor to treat many more patients. “We live in an era of augmented intelligence, where our smartphones use algorithms to guess where we want to go or what we want to say next. The key benefits of artificial intelligence are grounded in this ability to continuously comb the EMR, training on past patients’ trajectories in the same way clinicians are trained and hone the lenses of their own prescription glasses, patient by patient by the millions.”

Mark Weber, senior vice president, Infor Health Solutions, also sees the potential in applying machine learning and AI in healthcare similar to how it’s applied to consumer marketing. “For example, if I buy hiking boots online, suggested items pop up that I can use as well, like bug spray or sunscreen. The data analytics behind those recommendations includes a wealth of information about me—my demographics, such as age, gender, education and income level, as well as where I live and other factors that influence my buying decisions. My prediction is that we will be able to apply the same principles to healthcare data—to improve patient outcomes, costs and efficiencies.”

5. Precision Medicine

Precision medicine, per the U.S. National Institutes of Health, is “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” Artificial intelligence algorithms can take precision medicine to the next level by increasing the accuracy and prediction of outcomes through mining large quantities of genetic, clinical, social, lifestyle and preference data across broad, heterogenous populations, shared Dixie B. Baker, PhD, FHIMSS, senior partner, Martin, Blanck and Associates.

A report from Chilmark Research goes so far as to state that precision medicine must be accompanied by machine learning and artificial intelligence to reap its full rewards—due to the ability to analyze large data sets faster than clinicians and medical researchers.

AI delivers major benefits to advancing precision medicine, not only in predicting outcomes of current patients, but even in predicting the probability of future patients’ having a disease. With this level of insight, providers can know the best care plan not only for individuals, but for entire populations.

6. Provider Satisfaction

In an era of physician burnout, technology like artificial intelligence has the potential to overwhelm providers and contribute to the problem. But, with a solid strategy and effective workflows, artificial intelligence can empower clinicians, rather than exhaust them.

“A human being can take into account about seven different factors in making a decision. With the volume of data and the complexity of data in a clinical situation now, there’s actually too much data now for a human being to make a decision,” shared Dimitri Fane, director of product management of TrakCare at InterSystems, during a HIMSS TV interview. “That’s where we’re going to start seeing AI really getting value out of the data. Not replacing the human making the decision but to support and guide that decision.”

Dr. DeRienzo also sees opportunities to use artificial intelligence as a tool to return humanity to healthcare. “I truly believe we can in two fundamental ways. First, by serving up connections among data points either too numerous or too vast for people to have previously been able to access—allowing humans to do new things,” said Dr. DeRienzo. “Second, by freeing up people to spend more time connecting with and serving other people and less time on data entry, data discovery and simple sorting tasks—allowing people to do more of the right things.”

7. Security

The healthcare industry must be vigilant in securing and protecting technologies with vulnerabilities that can be exploited like artificial intelligence and machine learning. “AI is a dual-use technology that can be deployed defensively or offensively,” said Lee Kim, JD, CISSP, CIPP/US, FHIMSS, director of privacy and security at HIMSS. “The malicious use of AI will impact how we construct and manage our digital infrastructure as well as how we design and distribute AI systems, and will likely require policy and other institutional responses.”

In the cybersecurity realm, AI in healthcare can be used to automate phishing, the initial point of compromise in most cyberattacks. Spear-phishing, often tailored to the recipient using intelligence gathered about the recipient is even more damaging, with fully automated attacks potentially disruptive for many organizations. Artificial intelligence systems may also be used for other nefarious purposes, such as automated cyber-attacks on hospital networks, potentially putting lives at risk—especially those patients that depend upon life-saving or life-sustaining devices that have network connectivity.

But artificial intelligence can also be used to protect organizations by detecting and otherwise learning about potentially malicious activity through means such as heuristic learning and adaptation. Such technology can help prevent and thwart cyberattacks.

After patient records were targeted by a hacktivist group, Boston Children’s Hospital began using the technology as a defense tool to strengthen existing security structures and protocols. “By using AI, we can do a better job at being more prospective and staying one step ahead and starting to be able to detect that anomalous behavior or activity as it’s happening,” says Dr. Daniel Nigrin, the hospital’s senior vice president and chief information officer.

Clearly, AI in healthcare is changing the industry across the spectrum. But as humans, we don’t just want to know how. We want to know why—because we are capable of asking the big questions and looking at the ramifications of technology and infrastructure.

Dr. DeRienzo said it well:

“Our central purpose in designing, developing and deploying artificial-intelligence-based patient care technology solutions must be grounded in helping clinicians return to their patients’ bedsides, be they at home or in hospital, and serve them with more connected information, more upstream interventions and more time to spend with the patients who most need their human bond… Once our purpose is clear, we must incorporate human factors from the outset of design, and continually ask and answer the question, throughout development, how will this tool help humans better serve other humans?”