OR WAIT null SECS
The concepts that power cognitive computing and AI have grown to include substantial healthcare advancements.
For many people, their first exposure to artificial intelligence (AI) was seeing IBM’s Watson supercomputer beat Ken Jennings on Jeopardy in 2011. Since then, the concepts that power cognitive computing and AI have grown to include substantial healthcare advancements.
Cognitive computing is technology that uses artificial intelligence to make predictions based on data. Artificial intelligence is technology that uses reasoning, natural language processing, machine learning, and even human interaction through speech and vision interactions. What makes cognitive computing systems so smart is that they evaluate patterns from data and get better with each use. The goal is for systems to be able to anticipate problems and behavior, offering suggestions and solutions before an issue occurs.
“Cognitive computing is basically predictive analytics. It analyzes patient data and indicates results and a suggested plan that a doctor can use to decide a course of action,” says Jaspinder Grewal, CEO of CareSkore, a population health management technology company that utilizes artificial intelligence. “Artificial intelligence takes that a step further by leveraging adaptive computing algorithms to formulate the best course of action and, within constraints imposed by the practitioner, implements that action. Cognitive computing acts as an advisor. Artificial intelligence takes that a step further by acting as an assistant.”
Big names in technology including IBM, Microsoft, and Google have made significant investments in finding healthcare uses for cognitive computing, and providing a platform for various startup firms that create specialized applications.
The healthcare cognitive computing market will be more than $5 billion by 2022, according to a 2016 study by Grand View Research, Inc. Fifty-two percent of the current market resides in North America, with increased demand for personalized medical care, pressure to reduce healthcare costs, and the rapid growth of technology in healthcare being drivers to innovation.
With an increased focus on finding solutions for people with chronic conditions, coupled with policy decisions that aim to decrease healthcare costs, it is imperative that health systems understand and begin to utilize cognitive computing systems to meet the demands of the changing healthcare landscape.
First, a brief history lesson to understand the evolution of AI in healthcare. After Watson’s appearance on Jeopardy in 2011, IBM began working with health systems in 2012 to enhance oncology treatments and with medical colleges to expand learning.
In 2015, IBM Watson Health was established, and the company began partnering with Johnson & Johnson, Apple, and Medtronic, and acquiring other technology companies. Apple’s HealthKit and ResearchKit both use IBM Watson technology.
In February 2017, Microsoft announced the Healthcare NExT initiative based on collaborations that will leverage cloud technology to generate insights about patient health and adherence, make quicker samples to answer processes for genome analysis, and enhance conversational health tools. Other Microsoft AI technology aims to use predictive analysis, rules management, and other best practices to lower healthcare fraud and waste.
Next: AI tech startups
In the meantime, several AI tech startups have developed, either using platforms developed by larger technology companies or using their own platforms to address healthcare issues.
Anil Jain, MD, vice president and chief health informatics officer for IBM Watson Health, says that the current capabilities of cognitive computing stem from having a history dealing with healthcare providers and systems.
“We see a lot of people doing artificial intelligence in the market. We like to call our operations augmented intelligence. We are not trying to replace traditional insights. We have a proven place in the market as offering cognitive solutions in the marketplace. The longer this technology is in place, the smarter it becomes,” Jain says, adding that the combination of using data analytics, traditional and cognitive insights in a population health setting allows for the best results. “It’s not just technology, our expertise is surrounded with data technicians, scientists. It’s an ecosystem, not just pure technology.”
Grewal says that health organizations looking to make an investment in cognitive computing and AI should think about patients’ needs first before getting caught up in all the diverse capabilities.
“Operationalizing analytics, which is what cognitive computing and AI do, can be used across the spectrum of care. But the simplest step to take first is in patient engagement,” Grewal says. “Cognitive computing provides insights that help guide a case manager to decide how to interact with a patient.” Insights include best practices for care plans, optimized length of hospital stay, personalized care plans and identification of gaps in care planning.
Using AI to manage non-face-to-face interactions, including pain assessments and appointment and medication reminders, can help offload these activities from providers, allowing them to interact and engage more with patients. For example, AI might be used to identify and auto-enroll patients with chronic diseases into care plans to reduce readmission, and engage those same patients at a higher rate. By identifying those patients, the AI technology also automatically schedules transportation for patients at high risk of no shows, and can send medication reminders.
“Relying on AI to engage with a patient having a heart attack is reckless today, but probably not in the future. There are too many variables that cannot yet be coded into a computing model,” Grewal says. “But standardized, more mundane interactions that do not require a face-to-face engagement are prime candidates for the automation AI can provide.”
Jain agrees, adding that the decision-making capabilities of AI technology assist various stakeholders with data and cognitive intelligence.
“Our engagement manager helps providers … understand what processes are and are not working. It helps them determine things like possibly using a text message to communicate with patients instead of email,” Jain says. “We can take those who are in the most need and reach out to them. By applying cognitive thinking, advanced machine learning and cognitive insights, you can only reach out to the folks that you can make the biggest impact with. We can scale that type of expertise to all corners of healthcare.”
Next: The market for collaboration
One of the most important aspects of cognitive computing is multiple stakeholders working together to solve some of healthcare’s biggest problems. Health systems, universities, and technology companies large and small are combining their areas of expertise to give a thorough look at complex health issues, and help patients better understand solutions.
“We rely on collaborations to help patients interact with cognitive systems. We support other organizations with cognitive capabilities. We use our capabilities and ecosystems and others can build on top of that platform for patients,” Jain says.
For example, Microsoft’s Healthcare NExT initiative partnered with the University of Pittsburgh Medical Center to improve healthcare delivery through a series of projects.
IBM Watson has leveraged multiple collaborations, including Sugar.IQ by Medtronic.
IBM Watson technology works with the continuous glucose monitoring application to find patterns in users’ behavior that may lead to spikes or changes in glucose levels.
Also, IBM Watson Genomics from Quest Diagnostics, creates an end-to-end solution for profiling tumors to identify potential targeted therapies and clinical trials. The service, launched in October 2016, involves laboratory sequencing and analysis of a tumor’s genomic makeup to help reveal mutations that can be associated with targeted therapies and clinical trials.
Watson then compares those mutations against relevant medical literature, clinical studies, pharmacopeia, and annotated rules created by leading oncologists, including those from Memorial Sloan Kettering Cancer Center in New Jersey. Watson for Genomics ingests approximately 10,000 scientific articles and 100 new clinical trials every month.
Since Quest Diagnostics already serves half of physicians and hospitals in the U.S., partnering with IBM allowed the company to leverage their insights to make the most impact, says Jay G. Wohlgemuth, MD, senior vice president of medical research and development and chief medical officer of Quest Diagnostics.
“Working together, we have the potential to deliver actionable information to clinicians in a community setting who may not otherwise have access to these insights,” Wohlgemuth says.
Quest and IBM worked with Memorial Sloan Kettering to increase the precision oncology knowledge involved with the genomic platform. “It isn't enough to have information-the data has to be insightful and actionable,” Wohlgemuth says. “It has to be clinically appropriate, and it has to be provided in a timely manner, in the place where it can inform decisions. Quest's relationship with IBM Watson is about taking data and, to put it simply, making it useful. That's how healthcare quality and efficiency improves.”
As AI becomes more commonplace, Jain says health organizations can bring their area of specialty to work with technology specialists to get better a understanding of health problems. “Healthcare organizations looking to utilize cognitive computing need to be willing to collaborate. The results will be more gratifying,” Jain says
Next: The future
One of the biggest misconceptions about AI is that the technology will eventually put clinicians out of work. But Jain says it is a mistake for a healthcare organization to invest in cognitive computing thinking it will replace the expertise of humans.
“One mistake is trying to replace people instead of trying to make people more effective. Our systems have to be taught and trained,” Jain says.
Though nearly 40% of jobs in the U.S. will be automated or affected by AI by 2030, the healthcare industry has the lowest risk of people losing jobs because of this, according to an analysis by PricewaterhouseCoopers released in April 2017.
In fact, data science with a focus in IT and healthcare is one of the most in-demand career fields, according to a February 2017 report by CareerCast.com. As more healthcare organizations seek to invest in cognitive computing, the need to hire people who can interpret data and use machine learning applications will increase 16% by 2024, according to the job openings website.
Jain stresses the important of having a diverse team of specialists-including physicians, data scientists, analysts, and those with business expertise-to help interpret and put into action the results from AI analysis.
“Many think that using this type of technology is like buying a DVD player-use it one time and it should work perfectly. Augmented intelligence is a personal journey. You can’t try to jump in and fix everything,” Jain says.
In the next 10 years, Jain says that cognitive computing will be a more mainstream part of the healthcare landscape, and AI capabilities will be sought by patients who want a personalized healthcare experience. He says that clinicians who can work with cognitive computing systems may be seen by health plans as more efficient, and therefore receive higher ratings when it comes to patient care.
“In 10 years, we may take for granted how providers and physicians will always be using an AI platform. Consumers will expect us to have cognitive assistants with all providers,” Jain says, adding that as a physician he understands the benefits of cognitive computing. “I look forward to practicing alongside Watson so that I don’t miss anything. Patients will ask for it as they search for physicians. Health plans will say, perhaps a doctor who uses the latest technology is more efficient. It’s just another way to assist them in making the most appropriate decisions.”
Donna Marbury is a writer in Columbus, Ohio.