Making AI Real In Healthcare Requires A Chief Human To Lead The Way
By David Dimond, Chief Innovation Officer of Global Healthcare, Dell EMC
Whether you call it cognitive computing, machine learning, deep learning or artificial intelligence (AI), the era of collaborative human-machine intelligence has begun, and the implications for healthcare are enormous.
Without the leverage of AI, there’s just simply no other way to turn the massive volumes of data coming from diverse and rapidly growing sources into the meaningful insights so critically needed to move into the new age of precision medicine and rise of healthcare consumerism. In fact, according to a PwC report, 54 percent of healthcare consumers worldwide are already open to receiving AI-enabled healthcare.
Humans are critical to this next wave. The humans of healthcare—physicians, caregivers, researchers, administrators, policy makers—will increasingly rely on thinking machines to uncover patterns and inform decisions that benefit patients, populations, health systems and society at large.
Figure 1. A good example of why AI is needed in healthcare — consider the number of registered studies that continue to accelerate year over year. Clinicians will likely need AI just to keep up with research in their own area of specialization. (Data sourced from ClinicalTrials.gov on February 19, 2018.)
#1 Investment
Recently, Dell Technologies commissioned a follow up survey of 3,800 business leaders for their take on the Dell Technologies and IFTF Realizing 2030 report. Across industries, AI was selected as the “top technology investment to become a digital business by 2030.”
Among healthcare respondents, 81 percent reported that they expected humans and machines to be working “as an integrated team” in their organization within the next five years, and 27 percent say this is already the case in their organization. Examples of AI at work in healthcare today run the gamut from mining EHR data to identifying at-risk patients to segmenting images to helping design treatment plans and managing patient appointments.
Moving From Fear Of Disruption To Embracing New Opportunities
Although healthcare faces unique challenges integrating AI into what will always be a high-touch and highly regulated patient-centric industry, I believe that the industry will “leapfrog” and become a leader in what promises to be the most disruptive wave of digital transformation yet.
First, healthcare is indeed “ripe for disruption”, as Fortune magazine puts it, opening the opportunity for creative and pervasive use of AI in everything from predictive diagnostics to claims and billing to facilities and operations management.
Second, the real value of AI will come, not from proprietary data or IP, but from sharing data and intelligence to gain actionable insights—something that many in healthcare understand and have been working to do for years. Current examples include sharing EHRs across accountable care organizations, developing million cohort databases like NIH’s All of Us and crowdsourcing databases like DNA Land, and big data platforms like DataRobot which enable healthcare data science to scale as a service.
Third, the transformative technology needed to move to the next phase of data innovation is falling into place. The required compute power is already available through high performance computing. Analytics engines required to support AI exist today and will only become more powerful. And data sharing is becoming easier and more secure with the development of multi-cloud environments.
The Next Phase In Human-Machine Partnerships Is Distinctly Human
Progress aside, enabling humans to freely and securely share data and AI-enabled services across the healthcare ecosystem takes more than technology and data to achieve meaningful results. It also takes vision, strategy, and people across the organization who buy into both.
Data from the 2030 study support this notion with 60 percent of healthcare respondents identifying lack of vision and strategy, and workforce readiness as barriers.
Sometimes, we need to look at past historical trend points to help us focus on what may be the next logical step in overcoming barriers in the future. In looking at the past progress in our industry, a recognition of the need for a cohesive enterprise strategy has driven healthcare organizations to appoint a CIO, CTO, Chief Data Officer, and then a Chief Security Officer to develop and realize the organization-specific vision and strategy for health IT (HIT), technology, data, and security, respectively (Figure 2).
Figure 2. The progression of Chief “fill in the blank” Officers in healthcare over the past two decades
Now is the time for healthcare organizations to create the role of Chief AI Officer to lead the effort to explore potential opportunities, develop a cogent AI strategy, and harness the necessary funding, professionals, technology, and organizational resources to implement them.
The Right Human For The Job
In fact, according to the 2030 study, 74 percent of the healthcare leaders surveyed recommended assigning a Chief AI Officer.
In filling that position, healthcare must remember its “lessons learned” with other C-execs. The right human for the Chief AI Officer job needs qualifications that go far beyond keeping up with the latest technical advances in deep learning. They must fully understand the clinical workflow—the front-line workforce and the culture that drives care delivery. They must understand patients, markets, and the strengths and weaknesses of the organization itself.
Who do they work for? Who cares, let’s work that out later. It’s a malleable and agile new position which needs to be embedded into the organization where the rate of business opportunity is high, and the technology adoption barriers are low.
Indeed, the best candidate for making AI real in a healthcare organization may already be working there. Those that take notice and capitalize on that fact will see themselves leading this next wave of human-machine partnerships as proposed in the 2030 report.