AAIH Positions Itself As ‘Trusted Voice’ For AI Use in Healthcare

2135
The AAIH's role and mission are is to advocate on behalf of the broader AI healthcare community. (GETTY)

By Deborah Borfitz

Close to 30 companies have joined a fledgling advocacy organization with big plans for aligning the healthcare sector around a shared vocabulary, standards of excellence and realworld potential of artificial intelligence (AI). The Alliance for Artificial Intelligence in Healthcare (AAIH) launched in January and literally hit the ground running, according to Annastasiah M. Mhaka, AAIH president and a senior advisor with Adjuvant Partners. 

The U.S. Food and Drug Administration (FDA) heard from the AAIH in May with a formal response to the agency’s planned approach to oversight of Software as a Medical Device (SaMD) products. Most notably, Mhaka says, the AAIH suggested that the proposed framework “work to address and accommodate the realities of regulatory oversight during the iterative process of AI product development.” 

The AAIH is also about to weigh in on the portion of President Trump’s AI Executive Order dealing with the use of federal data for research and development purposes. “It’s important for us to provide some thoughts around data issues and challenges specifically related to healthcare,” Mhaka says.

While our role and mission are is to advocate on behalf of the broader AI healthcare community, at the moment we want to serve as a resource so legislative bodies and policymakers will consider healthcare when they think about AI,” says Mhaka. During a recent set of meetings on Capitol Hill, AAIH representatives met 14 key members of Congress in the House and Senate to drive awareness, discuss the challenges of AI and build familiarity with the organization.

Earlier this year, the AAIH also began a dialogue with stakeholders in Canada and the Asia-Pacific region to promote cross-border harmonization, Mhaka continues. Chief technology officers and data scientists from various life science companies have separately been working with the National Institute of Standards and Technology on standards development for AI in the U.S., issuing detailed commentary on high-priority topics from a healthcare perspective.

With the help of plain-language translators on the communications team, the Technology & Standards Development Committee also authored a white paperAn AI in Healthcare Primer,” which includes definitions of commonly misused or misunderstood terms and a sampling of use cases in healthcare, she notes. The paper is expected to be released for public review in early August. The same committee is involved in early planning for a “company of the month” feature on the AAIH website where companies will be able to regularly showcase concrete ways their AI solutions are being used in real-world healthcare environments.

Education will continue to be a huge focus throughout the second half of 2019, Mhaka says, when the AAIH will be addressing workforce development through cross-training and reskilling of adults as well as creating an internship program for computer science majors. Academia and industry will be working together on pilots to scale the concept.

The AAIH also plans to partner with multiple market research and data companies to start tracking data and performance metrics for the AI-in-healthcare industry and “becoming more of a trusted voice in the sector, she adds. A test case is expected by the end of the year.

During the recent AI World Government conference in Washington, D.C., AAIH hosted a workshop on the vision for AI in healthcare from the vantage point of the U.S. Department of Energy, the U.S. Department of Defense, the National Institutes of Health, FDA and academia. Mhaka recites her five key takeaways as follows: 

  • Value of awareness and understanding of AI terms and relatable use cases 
  • Significance of underlying compute power and architecture with future needs for aligning emerging chip design with healthcare needs 
  • Data stewardship, including guidelines for data gathering, avoiding unintended bias and model standards 
  • Importance of real-time data for decision-making and real-time intervention in the clinic 
  • Emphasis on value and need for public-private partnership, including cross-industry and cross-federal engagement

Seeing the Possibilities

chemistry grad who did her Ph.D. in medicine at Johns Hopkins University, Mhaka brings a well-rounded background in biomedical discovery and drug development, diagnostics and business consulting to her presidential role with the AAIH. As director of business development & strategic alliances at Johns Hopkins Medicine, she worked on a regenerative medicine strategy that introduced her to the Alliance for Regenerative Medicine (ARM)—what would later become her model for successfully scaling the AAIH. Prior to that, she worked in biotech with focus on new company formation and early-stage technology development.

Her lifelong passion has been finding a place for new products and technologies in the healthcare sector where they will be viable economically, commercially and clinically, she says. It wasn’t until her stint at Hopkins Medicine that Mhaka saw the delivery side of healthcare firsthand, including keeping people healthy and avoiding unnecessary costs.

But it was the second time she saw the potential of AI—then talked about in terms of interoperability and prospective decision-making—to solve a host of real-world healthcare challenges. It resurfaced yet again during Mhaka’s brief stint with a contract research organization, when she came to appreciate that clinical trials are a makeorbreak proposition for any clinical advancement.

Building Critical Mass

Parallels between the challenges around bringing AI and regenerative medicine to the market were informal topics of conversation Mhaka had been having with her colleagues at Adjuvant and a few AI companies for several years, she says. A decade ago, regenerative medicine was overhyped and misunderstood with no sustainable way forward. Today, thanks to the ARM, it has its own FDA designation and international advocacy community with over 300 members, and more than 1,000 clinical trials are ongoing worldwide. ARM also took on standards, manufacturing issues and capital formation to effectively coalesce the sector.

“Obviously because AI is new technology with new challenges and requirements, we have to be thoughtful about what the barriers are and how we pace ourselves in trying to address them,” says Mhaka. Responding to requests for input on deadline-sensitive government initiatives has been the one notable exception. The goal is critical mass across the healthcare continuum—from biomedical discovery to clinical research and development, medical diagnostics and devices, and precision medicine approaches—to foster delivery of better medicine to patients and populations.

Founding members include Amazon Web Services, Bayer, Beyond Limits, BlackThorn Therapeutics, The Buck Institute for Research on Aging, CyclicaEnvisagenics, GE Healthcare, Genialis, GSK, Insilico Medicine, Janssen, Minds.ai, NetriasNuMedii, Numerate, Nuritas, OWKIN, Progenics Pharmaceuticals, Recursion, SimplicityBio (now QuartzBio, part of Precision for Medicine) and the University of Pittsburgh.

Recruitment efforts began with AI biomedical platform and pharmaceutical companies, making the membership temporarily skewed toward drug discovery, she says. Companies of “all different sizes and flavors” are joining the organization and will ultimately balance AAIH. “It would be a huge mistake to not think holistically about healthcare… we’d end up being siloed again.”

AAIH will need a combination of AI technology developers and end users, including not-for-profit groups, as well a diversity of infrastructure and architecture providerssays Mhaka. The clinical delivery side of the membership will get a boost from greater involvement of universities with affiliated medical systems.

With AI, a win for one is a win for all, Mhaka says. One sinking boat can likewise take down everyone. The “value of 30 opinions in one go” also gives individual voices more power and influence on matters of standards, regulation and policy.

AAIH will be dealing with global issues and cross-border harmonization projects, says Mhaka. Local chapters will be created once a particular jurisdiction has enough representation to focus on issues specific to that country. Members currently hail largely from the United States, the European Union and Canada, including a fair number of multinational corporations and younger international companies.

Happily, the prevailing mood among existing and potential AAIH members is comradery, even among companies that may otherwise be competitors, she continues. Bayer has chosen to be involved in almost every activity of the AAIH, but companies can do as much or as little as they wish and pick their areas of interest.

The rest of 2019 will be a balance of recruitment activities and growing the resources and infrastructure for the planned partnerships and pilots. With a few “exemplary projects under our belt,” she adds, “we can go out and demonstrate real value early on.”