In the first of a two-part interview, Brian Clear, MD, chief medical officer of Bicycle Health, describes an approach supported by the AMA, the American Society of Addiction Medicine, and other experts.
The cutting edge of digital health and home testing is bumping up against norms in society and medicine as the healthcare industry tackles a substance abuse epidemic. An organization that epitomizes this challenge is Bicycle Health, a Boston-based provider treating more than 8,000 patients in 29 states for opioid use disorder.
The company, which srarted as a single clinic in Redwood, City, California, uses digital health and home urine testing, alongside an approved medication for treatment called buprenorphine, the main ingredient in the commercial brand Suboxone. The success of that type of treatment plan was highlighted in a recent JAMA Psychiatry study which found that the expanded use of telehealth to treat opioid use disorder during the pandemic reduced the risk of opioid overdoses.
HealthLeaders: How did you arrive at your current role at Bicycle Health?
Brian Clear: It started 10 years ago, right through residency training. Anyone who's paying attention, especially in the hospital system, notices that substance use disorders are the underlying cause for a lot of the pathology that comes into the hospital, but they're not commonly addressed. I hope this has changed by now in my own residency program, but when I was training, I thought my attendings were excellent, and I got an excellent quality of training in all regards except substance use disorders, where it was pushed off, like it's not a medical problem.
Brian Clear, MD, chief medical officer for Bicycle Health. Photo courtesy Bicycle Health.
This is a social problem. This is a criminal justice system problem. Patients who are referred to abstinence-based recovery programs, often costing tens of thousands of dollars, get a month of time, essentially in a hotel, and then they go back to where they were, and return to their old habits, with nothing fixed.
The funny thing is there's a compelling, highly effective solution to it, that just because of an abundance of regulation, a lot of stigmas, and general medical provider reluctance to engage in the care of patients with substance use disorders, it's not used.
How often do you find an opportunity like that in healthcare, where you have a national health emergency and you have a solution for it that is ready-made and available but being underutilized? It's just a huge opportunity to improve public health.
I ended up working in an opioid treatment program in Tallahassee, Florida, after residency, did my buprenorphine and opioid use disorder training program through them, and started my own small kind of micro practice in the neighborhood. Then I got an opportunity to move out to San Francisco and work in a large opioid treatment program with integrated primary care and behavioral health. That got me plugged into public health systems in San Francisco, a lot of incredible doctors doing great work in the care of patients with opioid use disorder. And then I got introduced to Ankit Gupta, founder of Bicycle Health, through the California hub and spoke program.
I stepped into this role, and it's grown way beyond what I thought it was going to be, where I was expecting to direct a program in California, and then we expanded to Florida, which made me a little bit nervous having two licenses on opposite ends of the country. But then you learn how much need there is outside of California. And once you get used to the diligence process, you can safely operate in other states. And lo and behold, two and a half years later, we're in 29 states and continuing to grow and seeing amazing outcomes for our patients, which is pretty cool.
HL: What are the components of this program?
Clear: There are three medications that are shown to reduce death and disability from opioid use disorder. The one that can be prescribed and taken daily as a medication in a patient's home safely and effectively is buprenorphine. Methadone is another effective medication, which is less safe, and it's given through a really tightly regulated opioid treatment program structure. The other medication that's effective, but much less effective than either buprenorphine or methadone, is naltrexone, which goes by the brand name Vivitrol. We focus on buprenorphine, because that's the tool that's available as a first-line treatment. For patients with opioid use disorder, it has an equivalent rate of reducing death and disability compared to methadone, but a much better safety profile.
HL: Why is it so underutilized?
Clear: Stigma is one element. Historically, the medical system has not looked at substance use disorders as a chronic disease or as a medical problem. This dates back all the way to the post-World War I era, when the American Medical Association essentially took the stance that substance use disorders were not within the umbrella of medical practice, that it should be regulated by the Department of Justice rather than physicians. It's taken a very long time to reverse that stance and convince doctors that there is a highly effective medical treatment for opioid use disorder specifically.
Also, it's very tightly regulated. There's a lot of fear of enforcement action by the DEA [US Drug Enforcement Administration], fear of medical boards taking action against licenses. Buprenorphine is an opioid, a medication that at a very low level stimulates and blocks opioid receptors. But it's not separated out in the Controlled Substances Act as being different from other opioids, even though it is. Providers are afraid that as the DEA expanded enforcement action against overprescribing of oxycodone and hydrocodone, it's going to be a red flag practice to be prescribing buprenorphine, which is ironic and kind of horrible, because that is precisely the solution to the problem that was caused by overprescribing of oxycodone and hydrocodone, and more recently, the fentanyl epidemic.
HL: Where are the American Medical Association and the Food and Drug Administration today on their position on this?
Clear: We've got more evidence for the efficacy of buprenorphine than we have for any other medication in the history of medicine. There are multiple Cochrane Reviews establishing the efficacy of buprenorphine at this point. No reputable society can ignore that. Buprenorphine use for opioid use disorder is universally embraced. It just takes a very long time to change attitudes of individual physicians and expand that practice to a whole network of medical practices nationwide. But the societies are all on board, including the AMA and the ASAM [American Society of Addiction Medicine].
But there's a real lack of understanding amongst regulators, in some states, especially of the key difference between problematic opioids and buprenorphine. A waiver system was created where providers have to get special training, and do attestations every year to maintain that. It's a lot to navigate.
HL: How does Bicycle Health use telehealth to make this therapy available and effective?
Clear: As of 2018, a full 40% of counties in the United States didn't have a single provider who is able to treat opioid use disorder with buprenorphine. In my experience working in Florida and also San Francisco, patients often have to commute insane distances. In my Florida practice, the average patient commute to get to the program each day was almost an hour. We had patients coming over 100 miles every single day to come to the program, just because care is very sparse. So that selects for patients who [live] in dense urban areas close to programs, also patients of means, patients who have access to transportation. In rural areas, especially patients who are of lower means often have no ability to access care whatsoever, because they can't travel that far. If they're working, they can't take that time off work every morning to come into the program.
The most obvious advantage of telehealth is that it transcends that geographic barrier if they have internet access.
Editor's note: Part 2 of this HealthLeaders interview with Bicycle Health Chief Medical Officer Brian Clear, MD, will be posted on Thursday, October 13.
Adaptive techniques could improve diagnostic effectiveness in five key disease areas, a Government Accountability Office report states, but only if the data is high quality.
Low-quality data is hampering artificial intelligence (AI) and machine learning (ML) from making more inroads in healthcare diagnostics, according to a new report from the US Government Accountability Office (GAO).
In addition, the report found, these technologies are yet to fully demonstrate real-world performance in diverse clinical settings.
"Our policy options--like improving data access and collaboration--may help address the challenges," the report stated.
Potential benefits of machine learning in the diagnostic process include earlier detection of diseases, more consistent analysis of medical data, and increased access to care, particularly among underserved populations, the report said.
The GAO identified a variety of ML-based technologies for five selected diseases: certain cancers, diabetic retinopathy, Alzheimer's disease, heart disease, and COVID-19. Most rely on imaging data such as x-rays or magnetic resonance imaging (MRI), but the report noted that these technologies have yet to be widely adopted.
Three broader approaches could assist these diagnoses: autonomous, adaptive, and consumer-oriented ML diagnostics.
According to the GAO, relying upon information supplied by the US Food and Drug Administration (FDA), which oversees use of these algorithms in diagnoses, incorporating additional data during the machine learning process (the adaptive approach) may improve accuracy, but only if the data being automatically updated is of high quality. Barring that, these processes could cause algorithms to perform poorly or inconsistently.
Diagnostic errors affect more than 12 million Americans each year, with aggregate costs likely in excess of $100 billion, the GAO said, citing a report by the Society to Improve Diagnosis in Medicine.
The report recommends that policymakers promote collaboration among technology developers, providers, and regulators when developing or adopting machine learning diagnostic technologies. This collaboration could expedite the creation of ML-ready data, according to officials at the National Institutes of Health interviewed by the GAO.
Providers should consider setting aside time for their employees to engage in these innovation activities, the GAO said.
Research by the Children's Hospital of Philadelphia saw good results in a telemedicine platform used by providers to manage care at home for children.
Researchers are encouraged by early efforts to read pediatric epilepsy patients via telemedicine, but say more work is necessary to reach wider populations.
The Epilepsy Neurogenetics Initiative (ENGIN) at Children’s Hospital of Philadelphia (CHOP) reported that across nearly 50,000 visits, patients continued to use telemedicine effectively, even when outpatient clinics reopened a year after the onset of the COVID-19 pandemic.
Still underrepresented, though are socially vulnerable families and racial and ethnic minorities, the hospital reported.
These findings represent the largest study of telemedicine in child neurology to date, and were recently published in the journal Developmental Medicine & Child Neurology.
Telemedicine as a novel method of care in child neurology has not been systematically explored until now, CHOP officials said. A prior study in Neurology in 2020 reported that patients and clinicians were highly satisfied with telemedicine, and were looking forward to utilizing telemedicine for future visits.
The latest observational study was based on a cohort of 34,837 in-person visits and 14,820 telemedicine outpatient visits between October 2019 and April 2021, spanning a total of 26,399 child neurology patients.
“In 2020, the COVID-19 pandemic necessitated the use of telemedicine visits, but now that telemedicine visits have been established as part of the care we are able to deliver, we had the opportunity to compare them more thoroughly to in-person visits,” the study’s primary author, Michael Kaufman, MS, a data scientist with ENGIN at CHOP, said in a press release. “With data on nearly 15,000 telemedicine visits, we were able to identify trends in how telemedicine was being used by individuals of different demographic backgrounds, neurological conditions and other variables.”
Certain patients used telemedicine more often than in-person visits--in particular, patients with epilepsy and attention-deficit hyperactivity disorder. Other patients, including those with certain neuromuscular and movement disorders, younger patients, and those needing specific procedures, were less likely to receive telemedicine care.
Self-reported racial and ethnic minority populations in the study as well as those with the highest social vulnerability--a measure of community resilience to stressors on human health--participated at lower rates in these telemedicine visits.
Compared to less vulnerable individuals, some the most vulnerable individuals studied were less likely to utilize online patient portals, and were at greater risk to receive delays in care, the study found.
The Detroit-based federally qualified health center is using digital health tools to improve care outcomes.
A Detroit-based federally qualified health center (FQHC) is using digital health technology integrated with its electronic health record platform to improve care management for patients living with high blood pressure.
The MicroMD electronic health record software in use in three clinics at the Institute for Population Health (IPH) was integrated in less than four weeks with Carium, a leading virtual care platform, in a partnership with Smartlink Health Solutions.
A private nonprofit, IPH took over Detroit's public health functions in 2012 as the city was approaching bankruptcy. Although most such services transferred back to Detroit in 2015 after the city emerged from bankruptcy, IPH continued to offer clinical outpatient services at three locations in the city.
The Health and Human Services Department's Health Resources and Services Administration (HRSA) has continued to fund IPH operations to the tune of $650,000 per year in 2020 and 2021. IPH provides general primary care and health screenings including mental health, prenatal care, telehealth, and adult dental services.
Prior to Carium's integration, the IPH team was manually notifying clinicians of patients with abnormal vital readings captured through remote monitoring. Now, Carium-supplied blood pressure cuffs report data directly into IPH's EHR.
The integration creates a more complete patient record, incorporating data from within the patient's daily routine, such as blood pressure or weight.
With this data, the IPH care team can better monitor, assess, alert, intervene, educate, coach, and communicate with patients.
The integration between MicroMD and medical devices deployed and supported by Carium allows physicians at IPH to click-to-order virtual services like remote patient monitoring directly through their MicroMD EHR, seamlessly enrolling patients to Carium.
IPH went live with this integration on June 10.
"The integration between Carium and our EHR allows us to be there for our patients every day," says IPH Chief Operating Officer Anthony Harris, MSW. "Our patients know we care and are keeping an eye on them, and it's also a great tool for accountability. Before, we didn't know if they took their medication until they come in the next time for a follow-up. Now we take their blood pressure twice a day, so we have an accurate account of what they're doing."
Anthony Harris, MSW, chief operating officer for the Institute of Population Health. Photo courtesy IPH.
"The integration process was easy, straightforward, and the project team was very responsive," adds Mark Lynn, IPH's chief information officer for technology services. "By incorporating the timely data from patients' daily life into our clinicians' workflows, they're able to communicate more efficiently, and make recommendations or adjustments in real-time when necessary."
In addition, discrete data automatically flows from Carium-managed devices to the connected EHR, enabling the IPH team to easily run required quality reports.
Harris has served health departments during a 25-year career, starting out managing HIV and STD patients, then moving on to IPH when it formed in 2012.
"People still kept coming here, so we stayed open," Harris says.
In his role as COO, "I have 10 hats here," he says.
So far, IPH has enrolled 30 patients in the remote monitoring program, but the funding provides for up to 80 patients, so recruitment continues to fill the remaining spots.
The Carium platform also allows for tracking more conditions over time, and Harris says the organization will expand its base to monitor chronic conditions like diabetes.
The EHR used by IPH did not have a complete set of application program interfaces (APIs) required to fully integrate the Carium platform, which is where the partnership with Smartlink Health Solutions enters the picture.
"Smartlink has the ability to automate some user interface procedures, where we can then interface with Smartlink's APIs to drive the pushing of the data in the EMR using the user interface and a set of automation technology that really does mouse clicks and button pushes within the EMR," says Scott Pradels, Carium's chief operating officer and co-founder.
Carium provides a range of connected remote monitoring devices from various manufacturers, using both cellular and Bluetooth connectivity. The platform is also able to support connectable devices supplied by patients.
Carium charges a small monthly fee for each connected patient, Pradels says.
The federal oversight agency recommends CMS adopt new coding procedures to compare care quality to in-person visits.
With pandemic-fueled temporary waivers on telehealth leading to a surge in telehealth visits in 2020, especially on audio-only platforms, the practice is overdue for its own exam for effectiveness and privacy, according to a new Government Accounting Office (GAO) report.
The use of telehealth services topped 53 million visits in the period between April and December 2020. During the same period in 2019, only 5 million such visits occurred. Many of those were conducted by phone or non-video telehealth, which was rarely allowed prior to the pandemic.
The Centers for Medicare & Medicaid Services has monitored some risks to program integrity related to these telehealth waivers, but the GAO report found that CMS "lacks complete data on the use of audio-only technology and telehealth visits furnished in beneficiaries' homes," in part because no billing mechanism exists to identify all these telehealth visits.
"Providers are not required to use available codes to identify all instances of audio-only visits," the GAO reported. "Moreover, providers are not required to use available codes to identify visits furnished in beneficiaries' homes."
The GAO said this coding is important to monitor the quality of these telehealth services as compared to equivalent in-person services.
"CMS has not comprehensively assessed the quality of telehealth services delivered under the waivers and has no plans to do so, which is inconsistent with CMS' quality strategy," the GAO said. "Without an assessment of the quality of telehealth services, CMS may not be able to fully ensure that services lead to improved health outcomes."
The GAO offered three recommendations for CMS going forward:
Develop a new billing modifier or make clearer how to bill audio-only office visits for better tracking;
Require providers to use existing site of service codes when beneficiaries receive Medicare telehealth services at home; and
Assess the quality of telehealth services delivered during the public health emergency.
Finally, the GAO urged the Health and Human Services Department's Office of Civil Rights to offer additional education, outreach, and other resources to providers to help them explain risks to privacy and security that patients may face during telehealth visits.
In the second of a two-part interview, Dennis Chornenky, Optum's senior vice president and chief AI officer, looks for inspiration from finance and institutional review boards to steer AI toward maturity.
In early 2022, Dennis Chornenky, MPH, became chief AI officer and senior vice president at Optum Health, a subsidiary of the UnitedHealth Group health plan. In part 1 of his conversation with HealthLeaders, posted on Wednesday, he addressed the role of AI in creating data-driven insights to prevent disease and personalize care; how to apply governance frameworks to AI, and his prior role serving Presidents Trump and Biden in crafting national strategies on AI and telehealth.
HeathLeaders: Where are we on the maturity cycle of AI? Where are the limits of AI, particularly with the eye toward healthcare?
Chornenky: Maturity in the AI space is certainly an evolving concept. The further out something is, the less defined it becomes, and the more variability you're going to run into. My view is that we're early in the maturity stages, in terms of development potential.
Dennis Chornenky, MPH, senior vice president and chief AI officer at Optum Health. Photo courtesy Optum Health.
Some of the work that I’m privileged to be advancing, that creates governance and a portfolio management model for our organization, is quite innovative and I would say ahead of the curve. I'm doing that by leveraging a combination of experiences and emerging frameworks.
With regard to managing a portfolio of AI investments, we can pull in some insights from the financial sector. We can look at institutional investment policy statements, for example, and approaches to outlining which risks you may be likely to encounter as a portfolio manager. What are you doing to mitigate those risks? What's the strategic goal of the portfolio? What level of risk are you willing to accept to achieve your target return? These kinds of established approaches can be helpful to keep in mind.
I also draw on my epidemiology experience with institutional review boards, which are designed to ensure ethics and safety in clinical trials. There is an emerging recognition in healthcare that we can do something very similar for the AI space, so we're seeing the emergence of “AI review boards,” loosely modeled on institutional review boards, that can screen AI projects, models, and applications for various types of risks in operational and clinical environments. We want to make sure that we have consistent and reliable processes that help ensure our models are safe, ethical, and as fair as they can be for any particular use case.
As for the limits of AI, we can frame this from the perspective of the trade-offs between performance when doing precise, clearly defined tasks against the ability to perform a broad range of tasks and solve a broad range of cognitive problems creatively. Machines excel at doing repetitive tasks with precision and humans excel at navigating dynamic environments and solving new problems. What business leaders should really be thinking about in this context is how to drive the kind of digital transformation in their enterprise that optimizes the collaboration of humans and machines in a way the amplifies the strengths of both and minimizes the limitations of both.
HL: The fact that you have this title of chief AI officer at Optum suggests that the day is not far off where AI becomes part of the standard of care. Is that something you ever hear discussed?
Chornenky: It's an interesting question. For me, this conversation started in telehealth, where the question was, could telehealth be the standard of care in the sense that there were concerns many years ago around the risks of doing telehealth? A health system that's providing care delivery through telehealth may be exposing itself to legal action if some harm comes from offering virtual care. The other side of that argument was that there may come a day where health systems that don't provide telehealth as a modality may be the ones exposing themselves to lawsuits, because they’re the ones providing less access to care, or restricting care, because they’re forcing people to only get care in person, potentially disadvantaging certain populations.
It's the same question with AI. Does using AI introduce legal risks, or will there come a day when it's expected and not using AI is what introduces legal risks? Because if you're not using more advanced technologies, you might be relying on imperfect and more variable human decision making in diagnosis and care delivery, without the use of more precise machine recommendations. We're pretty far away from something like that, maybe decades, because of the very limited maturity of the use of these technologies and the lack of scale. And in some ways, we may never entirely get there.
There's always going to be some sort of downstream human decision making. It's unlikely that we're going to get to a point where machines will completely control all aspects of the diagnostic or care delivery process. There's always a requirement, a natural need, to have humans make the broader decisions around whatever machine insights may be derived from data, and the kinds of actions we may want to take as a consequence. That said, over time we will certainly see a growing automation of tasks and more care decisions being informed by data-driven insights.
HL: Before the pandemic, Vinod Khosla was going on and on about how we don't need doctors. The pandemic changed that dialogue.
Chornenky: The pandemic forced a recognition of the importance of investing in technology. But I don't think It did much to support Khosla's case, which, if it can be described as doctors inevitably being replaced by computers one day, is a poor formulation. There are many tasks currently performed by physicians and medical technicians that will indeed be replaced with AI, i.e., computational methods and machines that simulate or surpass human cognitive capabilities. But the overall role of the physician is much too broad and requires the kind of big picture thinking, creativity, and bedside manner that machines are unlikely to achieve in the coming centuries, if ever.
CHIME and others say hospitals and other providers are not ready and need more help from HHS.
Despite assertions from the Office of the National Coordinator for Health IT (ONC) that it has gone the extra mile to help health systems abide by new information blocking rules that begin to take effect on October 6, several are lobbying for a significant extension of the deadline.
In a statement, CHIME officials asserted the organization has been an ardent supporter of information sharing and continues to advocate for patients’ ability to access their healthcare information in a digital format. But members have been hampered by scarce resources in trying to meet the deadline, the organization said.
"CHIME members remain steadfast in their dedication to be a trusted partner for patients and safeguard their ability to access their healthcare records, but it’s clear that more time is needed to ensure that providers have a thorough understanding of these important policies," said Russ Branzell, CHIME's president and CEO. "There has not been enough guidance on best practices and potential enforcement."
Citing what it said was overwhelming feedback from CHIME members representing a broad range of providers from across the healthcare continuum, including electronic health record (EHR) vendors, CHIME leadership maintained that these stakeholders are not fully prepared for the deadline.
With clinician burnout rates increasing, CHIME officials said HHS' unclear and inconsistent information sharing about data requirements could further strain healthcare providers and their support staff and may inadvertently undermine HHS’ goals to reduce provider burden, improve interoperability, and empower patients with their information.
Branzell said the “need for clear guidance is so important, and we need to make sure that all healthcare providers fully understand the many nuances of these complex policies.”
Small and rural providers, who more heavily rely on their EHR vendors for regulatory compliance support, are particularly unready, CHIME officials said.
Postponing compliance was not CHIME's only request. The organization also requested that HHS use corrective action warning communication to providers-- especially before they impose any financial penalties or begin formal investigations.
CHIME urged the HHS to ensure that providers and clinicians have the guidance, education, and technology to support these new policies before full implementation and enforcement of information sharing regulations.
In a statement, CHIME officials said, "We believe strongly in information sharing and want to see these policies succeed. A successful implementation of information sharing necessitates that all stakeholders have the critical tools, knowledge, guidance, and systems required to comply. This is simply not the case for most of the providers represented by CHIME."
An ONC spokesperson noted that the agency has provided a wealth of information about information blocking, including fact sheets, FAQs, blogs, and webinar recordings.
The spokesperson said these resources have been continually updated since April 5, 2021, the date when the regulation became official, and pointed to a fact sheet posted at that time.
Starting October 6, "actors will be expected to avoid interfering with access, exchange, or use of the full scope of EHI, except when applicable law mandates the interference or an information blocking exception is met," the spokesman said.
The ONC also said it would post a related blog post ahead of October 6 on its Buzz Blog.
In 2016, the 21st Century Cures Act made sharing electronic health information the expected norm in healthcare by authorizing the HHS Secretary to identify "reasonable and necessary activities that do not constitute information blocking." The ONC's 2020 Cures Act Final Rule established information blocking exceptions to implement the law.
The CHIME letter was also signed by the American Hospital Association, American Medical Association, Federation of American Hospitals, and Medical Group Management Association.
In the first of a two-part interview, Dennis Chornenky, Optum's senior vice president and chief AI officer, describes how he is driving AI tech for 15 million members of the United Health subsidiary.
At the start of 2022, Dennis Chornenky, MPH, became chief artificial intelligence officer and senior vice president at UnitedHealth Group health plan subsidiary Optum Health, after having served as a senior advisor and presidential innovation fellow in the White House in both the Trump and Biden Administrations.
The chief AI officer is one of the newest titles in the C-suite. Only a handful exist, in places such as the US Department of Health and Human Services and at technology companies like IBM, Elevance Health, and eBay.
In this two-part HealthLeaders interview, Chornenky describes just what a chief AI officer does, how it dovetails with pressing needs in Optum and all of healthcare, and what AI means for the future of healthcare.
HealthLeaders: Optum has pushed for value-based healthcare. What role is AI playing in driving that?
Dennis Chornenky: At Optum and UnitedHealth Group we’re driving healthcare transformation toward comprehensive value-based care, and AI is playing a big role in that. It’s a key focus of our growth strategy, helping more patients and care providers transition from fee-for-service to value-based approaches. We’re applying advanced technologies to drive better and more consistent care outcomes at lower overall cost.
Dennis Chornenky, MPH, senior vice president and chief AI officer at Optum Health. Photo courtesy Optum Health.
We have around 15 million members participating in value-based arrangements with over 1,000 hospitals and over 100,000 providers. Through our OptumCare delivery organizations we’re leading the industry in terms of the proportion of the patients we serve participating in value-based arrangements. I think we're expanding that at the highest rate out of any other care delivery organization in the US as well.
The way AI can help us accelerate this expansion into value-based care is by leveraging data to identify patients and members best fit for value-based care models and the clinical innovations and operational efficiencies that are most important in driving that transformation. There is a spectrum of data-driven insights that help us to better understand which patients may benefit most from which types of interventions and which types of care plans. That ends up getting broken down into a whole lot of different things, whether we're looking at disease prevention or surveillance, or integrating telehealth and virtual encounters into care modalities.
We are applying supervised machine learning techniques to improve our ability to predict disease progression and enable earlier interventions and unsupervised techniques like clustering to help us better understand the natural cohorts in our patient and member populations to advance more personalized care models. Overall, we are looking at anything that can help us improve patient outcomes, advance clinical innovation, and reduce costs.
HL: How can healthcare audit the AI it's starting to consume and use that to drive improvement?
Chornenky: You're right to make the connection that the way we approach the risks involved in deploying AI applications can be an important opportunity to drive improvement. AI governance is an emerging field that can leverage industry frameworks like Responsible AI to facilitate auditability and mitigate regulatory and reputational risks. A more technical framework referred to as ML Ops can help mitigate technical and model lifecycle risks. When done correctly, AI governance in healthcare helps to improve access to care and advance health equity. Without it, AI applications can run the risk of actually amplifying existing healthcare disparities.
I’m really encouraged that healthcare leaders are starting to understand that AI governance is an important area of investment and that it can help enterprises identify and mitigate the technical, regulatory, and financial risks posed by AI.
It’s also fascinating how rapidly innovation has been evolving in this field, with more and more startups and AI enterprise companies launching new offerings for Responsible AI, ML Ops, and bias and fairness assessments. The more forward-thinking health systems are also making investments in internal processes. Mayo Clinic, for example, has recently stood up a governance model they refer to as the “AI Translation Assessment” process, led by a distinguished group of experts.
At the federal level, the FDA is developing regulatory guidelines for Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices. NIST is developing a new AI Risk Management Framework. There is a growing body of emerging legislation across the US and in the EU that will have a significant impact on how we develop and deploy AI. I participate in several industry groups focused on developing best practices and standards for AI governance in healthcare and strategies for regulatory engagement.
HL: What learnings did you take from your time at the White House?
Chornenky: Serving in a non-political role across Republican and Democrat administrations, especially during the pandemic, gave me a lot of perspective on how our federal government works and how to successfully formulate and advance national policy, particularly in the healthcare and technology sectors.
As a senior advisor and a Presidential Innovation Fellow I was initially focused on advising our US chief technology officer on national AI strategy and our federal chief information officer on federal AI strategy. National AI strategy is how we think about cultivating growth and innovation in the private sector and the markets regarding AI/ML technologies, scaling up investment in R&D, academic partnerships, and also building trust in these technologies among the American people. This is also where we start getting into AI ethics and Responsible AI, or trustworthy AI.
Federal AI strategy is how do we think about standing up better data science capabilities across the federal government. This is everything from vanilla IT cloud migration, to upskilling the existing workforce with data literacy and analytics curricula, to thinking about data science as a career path, creating new job codes with OPM, and new processes and programs for engaging data science talent, recruiting, and retention. I had a portfolio of agencies I worked with to advance innovation, AI governance models, and AI capability maturity roadmaps.
As part of this work, we launched a new federal AI community of practice that was meant to bring together leaders and practitioners from across federal agencies to share best practices and do the type of collaborative work that they might not always be able to do within their usual, perhaps more constrained, agency environments. I also helped manage a government coordination committee that produced the executive order promoting the use of trustworthy AI in the federal government. This was a very important initiative that was bipartisan in nature, and the government is implementing the provisions of that executive order today.
When the pandemic hit, I was able to apply my training as an epidemiologist to help coordinate response efforts across federal agencies and our private sector partners, including technology companies, health systems, and payers.
What turned out to be more consequential for me, however, was my background in telehealth. I previously had an AI-driven telehealth and smart-scheduling company out of Palo Alto. Through that work I got to know everybody in the industry, the CEOs of the larger telehealth companies, the different industry associations and who led them, and top law firms working on telehealth regulatory issues around the country. So I really ended up in a unique position to pull all of that together and very quickly formulate and advance a national strategy on telehealth and how we were going to work across federal agencies with our private sector partners to make telehealth accessible to as many Americans as quickly as possible.
I think probably the biggest silver lining, if you will, of the pandemic, was that it accelerated telehealth adoption and access to virtual care for Americans across the board. It wasn't only for mitigating the risk of the spread of infectious disease, but also helped to ensure continuity of care for non-COVID related cases.
There was a tremendous amount of work done from an administrative and a policy perspective. Within just a few weeks, we put out over 50 waivers to enable telehealth, a couple dozen new billing codes, and a new modern website, telehealth.hhs.gov, to help patients and providers adopt telehealth in safe ways. We also convened a telehealth innovation summit, which was a great way to celebrate a lot of the work that had been done, particularly out of the deputy secretary's office at HHS, and with our private sector partners, but more importantly, to align on what the next steps should be to continue to advance adoption of telehealth and expanding access to care for all Americans.
Editor's note: Part 2 of this HealthLeaders interview with Optum SVP and chief AI officer Dennis Chornenky will be posted on Thursday, Sept. 29.
Starting October 6, healthcare organizations will be expected to avoid interfering with access, exchange, or use of the full scope of electronic health information (EHI), except when applicable law mandates the interference or an information blocking exception is met. Between April 5, 2021, and October 6, 2022, the scope for information blocking was USCDI v1.
While additional resources may be forthcoming from HHS, the Sequoia Project resources may help fill some of those immediate needs. They include:
Good practices for information-sharing and information-blocking compliance;
A further exploration of the expanded definition of electronic health information (EHI) and related considerations;
An infographic depicting the web of information systems included in the expanded definition of EHI;
Operational implications of the move to an expanded definition; and
"The Sequoia Project is grateful to the public for taking the time to provide feedback on these documents, as well as to the members of the IBWG [Information Blocking Compliance Workgroup] and our subject matter experts whose collaboration and dedication was integral to the development of these resources," Mariann Yeager, chief executive officer of The Sequoia Project, said in a press release. "We look forward to continuing to work to advance interoperability and build a community of practice around this critical topic."
The Sequoia Project is the recognized coordinating entity for the Trusted Exchange Framework and Common Agreement (TEFCA), an initiative of HHS' Office of the National Coordinator for Health IT, designed to facilitate interoperability between health systems and other entities it regulates.
The health system's value-based care model is built around home-grown population health technology.
Crossover Health, a primary care provider catering to self-insured employers, has built its own population health tool to improve patient and clinician experience alike.
The San Clemente, California-based healthcare organization, which counts Amazon among its customers, launched its tool to show clinicians how they were improving the health of their patient populations, says Stephen Ezeji-Okoye, its chief medical officer.
Crossover Health has 42 clinics throughout the United States, a mixture of on-site and near-site clinics, as well as a national virtual medical practice. The organization built its population health tool to overcome what Ezeji-Okoye calls the "tyranny of the visit" that dominates traditional electronic health record software.
"We don't work in a fee-for-service arrangement," he says. "What we're focused on is how do we improve the health and well-being of the population because that's how we show our value."
Such considerations impact Crossover Health's design of its clinics and its exam rooms, he says, so it made sense to extend design considerations to the population health tools utilized by its physicians.
Crossover Health employs traditional primary care providers as well as specialists offering mental health, physical therapy, and chiropractic services -- all augmented by nurse practitioners, health coaches, and care navigators.
"A lot of programs in advanced primary care don't actually include all of the elements I've just described," he says. "We have a very holistic approach. We believe that you've got to go beyond the biomedical model of just treating disease. You've got to get into the things that really affect health."
Ezeji-Okoye says 80% of the factors affecting health are not medical interventions. They might include the behaviors of patients and their living environments.
Stephen Ezeji-Okoye, chief medical officer of Crossover Health. Photo courtesy Crossover Health.
"We brought all of this together in a measurement system that allows us to have a good view of what's going on in the practice and allows providers to understand the impact they're having on the patients they see," Ezeji-Okoye says
An example of that impact: One of Crossover Health's clinics uses the population health dashboard to identify gaps in care among patients undergoing colonoscopies to screen for colorectal cancer.
"The population was one where people were primarily hourly workers," Ezeji-Okoye says. "They needed to be able to be at work as much as possible. Taking a day off for a colonoscopy was actually a huge barrier to care."
To address that gap, the clinic began providing those workers with at-home colonoscopy kits.
"One of the benefits of the tool is it makes these care gaps visible and allows us to work in a proactive fashion to be able to improve the quality of care of the population," Ezeji-Okoye says.
The platform enables clinicians to factor social determinants of health into care management, he says, identifying instances where the root cause of a health concern might not be clinical.
Crossover Health collects data from within its Elation electronic health record system to populate its population health dashboard, he says.
"Then we also use other measurement tools to capture things such as the therapeutic index and the severity index, to help measure how we're doing in those domains," he adds.
For now, Crossover Health's system does not send alerts back to the EHR.
"Certainly, alert fatigue is a real thing," Ezeji-Okoye says. "You'd need to limit the alerts to the ones that are most important. We don't do that at this time, but we're looking at that."
The system helped Crossover Health deal with the pandemic as well.
"Pre-pandemic, we were probably 3% virtual," Ezeji-Okoye says. "During the pandemic, we were probably 90% virtual. Now we're about 70% in person and 30% virtual. Regardless of how members choose to use us, they can expect the same quality of care, and we're going to have the same proactive approach to care."
Over time, he says, Crossover Health will continue to incorporate more data from its patients' visits to other health systems, including claims data.
"We need to make sure that we provide the human intelligence behind the data feeds, to make sure we've got the right treatment plan for each member," Ezeji-Okoye says.
Another feature of the tool allows Crossover Health to measure how quickly patients can get appointments. This drives the company's hiring decisions and helps to avoid overburdening current providers.
When the primary care team must initiate referrals, Crossover Health's care navigators curate lists of medical specialists outside the company who most identify with the company's proactive health values, Ezeji-Okoye says. Health records from such specialist visits then come back into Crossover Health's population health system, to drive further quality of care.
In time, Crossover Health may expand its offerings beyond self-insured employers. It has a commercial offering on a subscription basis, but so far that is a small portion of the company's business.
"Our goal is to be able to make Crossover accessible to anyone who's interested in having a better healthcare experience," Ezeji-Okoye says.