Mayo Clinic's John Halamka, MD, MS, outlines the technology infrastructure and processes that will change the way healthcare is delivered down the road.
Editor's note: This article is based on a keynote presentation at a roundtable discussion sponsored by Optum. The full report of the roundtable discussion, Strategy: The Key Factor in the Future of Healthcare Innovation, is available as a free download.
Healthcare delivery tomorrow will look much different than today for a variety of reasons. Consumer expectations, the emergence of nontraditional players, and a move to value-based care are among the driving forces. Yet nearly all advancements ride on the backbone of technology and the ability to harness a massive quantity of data now being produced.
This June, HealthLeaders convened a select group of health system executive thought leaders to discuss the topic, "Healthcare System of the Future." In his keynote address to CEOs, CFOs, CMOs, and CNOs, as well as innovation and revenue cycle executives, John Halamka, MD, MS, president of the Mayo Clinic Platform, discussed the technology stepping stones that will pave the road forward.
(Editor's note: A similar panel will convene in September to examine additional insights to develop the Healthcare System of the Future. Health system thought leaders interested in participating can contact email@example.com. Please put "Thought Leader" in the subject line.)
The ability to securely examine data in new ways, combined with tools, expertise, and partnerships, creates an ecosystem of collaboration to "cure, connect, transform, and take us to a new level of care delivery beyond bricks and mortar," Halamka says. In his presentation, he outlined six key elements that hospitals and health systems should consider as they plan for the future.
1. Start With a Technology Platform
"What tools are you going to need if you're going to build this new healthcare system of the future?" Halamka asks. "Well, certainly, you're going to need a technology platform." The process starts with data.
"If we agree that the healthcare data of the past can inform the care of patients in the future, how can you ethically use historical data?" he queries. Mayo Clinic forged into new territory to address this question.
The Rochester, Minnesota–based healthcare system has about 30 petabytes of structured data and equal amounts of unstructured data and DICOM objects (imaging data), from CT scans, MRI scans, and the like, says Halamka. In addition, there are 25 million biology samples, 30 million digital pathology slides, and active patients generating telemetry data through a variety of remote patient monitors.
"That’s a fair amount of data" to compute, secure, and share with appropriately qualified partners to deliver value to patients and providers, he says. For that reason, Mayo Clinic built an alliance with Google—"not to give data to Google, but to have a secure container for storage and compute where only Mayo Clinic has the encryption keys," he says.
Securing the data involved a complex series of processes.
"The first thing we had to do was deidentify data to an extent that’s really never been done before," he says. This involved consulting with the Department of Health and Human Services Office for Civil Rights, governments around the world, external deidentification and privacy experts in collaboration with a company called nference. "We developed a next generation deidentification algorithm that is far beyond, 'I’m going to remove the 18 HIPAA identifiers,' " Halamka says.
Additional considerations included proper consent, as well as compliance with HIPAA, the European Union's General Data Protection Regulation (GDPR), and the California Consumer Privacy Act to make the data available as an asset for a variety of collaborations.
Strong governance was another factor, involving the CEOs of Google, the Google Health team, the Google Cloud team, and Mayo Clinic's senior leadership, Halamka says. These executives worked together to outline priorities, determine who can do what with the data, and design an approval process.
As a result, Halamka says, "We have been able to achieve a level of what we’ll call a certified deidentification system that creates this cloud-hosted container … that we can then bring collaborators into. The important thing about that asset is it’s not exfiltrated; it’s not like the data is sent somewhere else. It’s a secure container where then we bring in collaborators into that container under our control with strict auditing and monitoring of all their activities and data use agreements and that kind of thing."
Mayo Clinic plans to share its model with others down the road.
2. Develop Longitudinal Patient Records
Another key component to tomorrow's healthcare delivery system is the ability to connect a variety of databases and sources to the electronic health record (EHR). Halamka says that several years ago, "Mayo created a Universal Data Platform, which was a mechanism to take data marts, link them with universal identifiers, and then have one longitudinal view over the patient’s entire experience of every element—inpatient, outpatient, ambulatory, home care, SNF care, images, and text—that had ever been gathered about them."
The ability to examine such a longitudinal patient record within a secure container could be a valuable research asset, but the process involved in deidentifying that data is complex. While the process has been challenging, Halamka says Mayo Clinic expects to add 2.5 million comprehensive unstructured records to the cloud this summer.
3. Incorporate Data From Wearable Technology
Harnessing data from wearable technology is another piece of the puzzle on the road ahead. "What are you going to do with that data?" Halamka queried. "How are you going to ingest it? How are you going to route it to algorithms and to those clinical services that would want to interpret it and provide not just information, but wisdom from the interpretation of these various signals? You need to build an orchestration layer that is capable of connecting to all these novel sources of telemetry and that provides an end-to-end workflow to connect the telemetry to people and to algorithms."
4. Employ AI and Machine Learning
Machine learning tools will play a key role in the healthcare system of the future, says Halamka. "It will help you use data, search it, and turn it into wisdom."
In concert with Google, Mayo has created an artificial intelligence (AI) factory for the organization, which examines preloaded, curated data, yet also enables an individual user to upload a local data set, Halamka explains. Then, using TensorFlow and other tools, they can create algorithms that will have measurable impact.
For example, data from a wearable is delivered via telemetry into the database. It then goes to an algorithm from the AI warehouse, which could deliver a care plan or treatment advice. The AI factory has already produced success stories, Halamka says:
- By examining ECG data from an Apple Watch or other wearables, "We can predict, with a pretty high degree of accuracy," he says, "your ejection fraction and your likelihood of developing atrial fibrillation in the future."
- Other devices can ingest neurological data and predict seizures.
- Algorithms have been developed to predict depression and anxiety.
- Algorithms have been created to examine DICOM images of CTs or MRIs and create a radiation oncology treatment plan that can be delivered through a linear accelerator.
"Imagine the healthcare system of the future is a series of experts creating a series of value-added algorithms that are able to connect to an ecosystem and then turn data received into high-quality, low-cost care," Halamka says.
For instance, the radiation oncology auto contouring treatment plan mentioned above tends to deliver lower doses with fewer side effects than most typical programs, he says. In the future, a provider might be able to send an image of a patient's tumor to Mayo Clinic, which will then deliver back a linear accelerator treatment plan "that is going to have few side effects and presumably higher quality and better outcomes," Halamka says. "I imagine healthcare systems across the world will offer these sorts of things. It will be a much more digitally enabled ecosystem."
5. Address Health Disparities and Consider the Human Element
Addressing socioeconomic disparities related to reduced access to technology among some populations will also play a role in the future of healthcare, Halamka says, who cites a story to illustrate his point. "About a year ago, a large tech company sent a number of engineers to Boston and wanted to visit a Medicaid clinic to understand the considerations of a person who may have issues of technology access, familiarity, literacy, or even the ability to have a cellular plan or a reasonably modern phone."
He continues, "I brought these [roughly] 28-year-olds, all wonderful people, to this clinic and they walk up to a homeless gentleman and the first question they ask is, 'What’s your favorite wearable?' He looked at them and said, 'Socks.' "
Halamka says that those using technology to enhance healthcare delivery need to understand that "we’re not [always] engineering [solutions] for an iPhone 11 and [placing] five different devices on your body. How do you create something that works on the simplest feature phone? How do you do it via SMS text? How do you meet the patient at their level of technological comfort?"
One way to address disparities is through human intervention," he suggests. "We may need a new position in healthcare called a care traffic controller—somebody who, on your behalf, is actually doing some of this digital work and is your interpreter. My mom is almost 80 and when I tell her, 'Don’t you realize the ONC information blocking rule enables you to use a FHIR API to gather your longitudinal patient record on your phone?' She says, 'I don’t know what any of that meant, but why would I even want it?' I explained to her why having data stewardship is good, and she said, 'Great, you're my data steward. Go do it for me.' We absolutely must think about language and literacy, technological comfort, and ability to afford and engineer for all the different kind of delivery mechanisms we’ll need, including a human delivery mechanism."
6. Seek an Innovation Niche
While smaller hospitals and health systems may not have the same resources as Mayo Clinic and other high-profile academic health systems, they still can play a role in transforming healthcare delivery and innovating the future of healthcare, Halamka says. As virtual care gains ground, building alliances and partnerships with health systems such as Mayo will enhance the level of care offered in all communities. Urgent care clinics, emergency departments, and ICUs, as well as critical access hospitals and community hospitals, "will be connected to other organizations offering services they may not be able to invest in themselves," he says.
Yet it is still possible to innovate at a local level. "It depends upon the visionary leadership of the organization," Halamka says. "Even a small organization can do amazing things. Pick a niche and do it well. There are lots of niche possibilities if you have a willingness to take a risk."
Innovation Roundtable Report Available
To obtain a copy of Strategy: The Key Factor in the Future of Healthcare Innovation, click here.
Thought leaders participating in the innovation roundtable discussion included:
- Michelle Conger, chief strategy officer, OSF HealthCare, and CEO, OSF Saint Gabriel Digital Health, Peoria, Illinois
- S. Nicholas Desai, DPM, MBA, enterprise chief medical information officer, Houston Methodist, Houston, Texas
- Bill Gillis, MS, chief information officer & vice president, Beth Israel Lahey Health Performance Network, Boston, Massachusetts
- Dan Liljenquist, JD, senior vice president and chief strategy officer, Intermountain Healthcare, Salt Lake City, Utah
- Steve Miff, CEO & president, Parkland Center for Clinical Innovation, Dallas, Texas
- Mitch Morris, MD, president, Optum Advisory Services, Austin, Texas
- Karen Murphy, PhD, RN, executive vice president and chief innovation officer, Geisinger, and founding director, Steele Institute of Health Innovation, Danville, Pennsylvania
- Rob Steele, MD, MBA, executive vice president/chief strategy and innovation officer, Children’s Mercy Kansas City, Missouri
Editor's note: This article was updated with the link to the roundtable report on August 14, 2020.
“Imagine the healthcare system of the future is a series of experts creating a series of value-added algorithms that are able to connect to an ecosystem and then turn data received into high-quality, low-cost care.”
John Halamka, MD, MS, president, Mayo Clinic Platform
Mandy Roth is the innovations editor at HealthLeaders.
The ability to deidentify data and securely share it through a secure platform is an essential building block to the healthcare system of the future.
The ability to develop AI-driven algorithms will enable providers to turn "data into wisdom" and may position large healthcare systems with the ability to master these innovations into resources to develop treatment plans for smaller systems.
"Care traffic controllers" may offer a human element to address disparities in access to technology.
Smaller health systems can play a role in the future of healthcare by focusing on an innovation niche.