Does the ethical development of AI rely on the “algorithmically” underserved? CHAI’s mission

For AI to flourish in healthcare, the industry must focus on the “algorithmically underserved,” said John D. Halamka, MD, MS, president of Mayo Clinic Platform, at the HLTH conference. 2022 this month in Las Vegas. Giving visibility to the algorithmically underserved – people who are not generating enough data/represented enough in health datasets for AI to make a decision – is only a requirement to overcome the prospect of AI bias in healthcare. And identifying and correcting sources of bias in AI must be a priority area for an industry striving to develop ethical and fair AI, Halamka shared.

For example, what if there was a national registry housing all the metadata needed to power the responsible development of algorithms for use in healthcare? Building this kind of standardization in the relatively dark nature of AI development is among the priorities of the Coalition for Health AI (CHAI), which was launched earlier this year. In addition to his leadership of the Mayo Clinic Platform, Dr. Halamka is co-founder of CHAI, alongside Brian Anderson, MD, Chief Medical Officer of Digital Health at MITRE.

What is CHAI?

CHAI’s mission, according to the organization, is to provide guidance for the ever-changing landscape of healthcare AI tools to ensure high-quality care, increase credibility with users, and meet to health care needs. The coalition serves to identify where standards, best practices and guidance need to be developed for AI-related research, technology and policy.

CHAI has been trained by Change Healthcare, Duke AI Health, Google, Johns Hopkins University, Mayo Clinic, Microsoft, MITER, Stanford Medicine, UC Berkeley, UC San Francisco and others, and is observed by the US Food and Drug Administration States and the National Institutes of Health and the Office of the National Health IT Coordinator.

One of CHAI’s primary goals is to help health informatics buyers make informed decisions about the AI ​​solutions they choose, based on academic research and using vetted guidelines. The coalition’s toolsets and guidelines are also designed to ensure that underrepresented populations are not affected by algorithmic bias. The “Guidelines for Responsible Use of AI in Healthcare” being developed by CHAI will intentionally promote resilient AI assurance, safety and security, the organization says.

Below, CHAI co-founders Dr. Halamka and Dr. Anderson, and CHAI member Suchi Saria of Bayesian Health, discuss the importance and timeliness of CHAI’s mission, and how the organization plans to prioritize patient safety, reliability, fairness, transparency, and trust in the process of developing healthcare AI.

Q: What are the goals of the Coalition of Health AI, and why is its existence so important?

Dr Halamka:

“The goal of the Coalition for Health AI (CHAI) is to develop voluntary “guidelines and safeguards” for industry to drive the adoption of credible, fair and effective health AI systems. transparent.

The power of machine learning to unlock better and more efficient healthcare delivery at scale is undisputed. We have long moved from discussing “should we use AI in healthcare” to asking “how can we design a framework for the responsible use of AI in healthcare, guided by the principle of health equity?

To answer this last question, we brought together a community of stakeholders – including academic healthcare systems and AI experts – to discuss critical concerns such as patient safety and algorithmic bias in workshops.

These in-depth discussions are taking place in conjunction with federal observers such as the U.S. Food and Drug Administration (FDA), Office of the National Health Information Technology Coordinator (ONC), National Institutes of Health ( NIH) and the White House Office of Science. and technology policy.

What makes CHAI’s work important and extremely relevant – beyond its mission – is timing. We are at an inflection point where AI is poised to take off exponentially if we can come together to harmonize standards and reporting for AI in healthcare and educate end users on how to assess these technologies to encourage their adoption. We have little time to develop and recommend shared standards and industry practices that ensure all communities will benefit in the future.

Dr AS Saria

“AI as a field is changing very rapidly. As a result, there is varying expertise among groups on how to apply it correctly and assess whether what they have implemented is working. There is an opportunity important to accelerate the adoption of AI by sharing best practices and developing safeguards that the wider community (government, payers and provider groups) can benefit from.

Q: What are the biggest challenges to applying AI in healthcare today, and how will the Coalition’s work help overcome some of them?

Dr. Anderson:

“Understanding and trust are definitely two of the biggest issues today from a patient perspective.

For example, as a physician, I have to ask myself, “Does my patient understand the role machine learning plays in their healthcare?” And, “should my patient believe that an algorithm was trained on data that reflects their demographics?”

Ultimately, I can’t expect my patients to have confidence in AI if I myself don’t have confidence in how the AI ​​systems and models that help define the policies and interventions are constructed and evaluated. And right now, it’s the lack of consistent standards, guidelines, and transparency that poses a potential threat to trust and adoption.

Too often, frameworks are discussed or deployed after widespread adoption of a technology. CHAI is trying to get a head start while Health AI is still in its infancy, to help determine the rules of the road to ensure the most positive impact for as many patients as possible.

Dr AS Saria

“CHAI brings together a diverse group of individuals with deep expertise in AI and AI for health, its translation into different use cases, regulatory and reimbursement knowledge, and diverse policy levers to accelerate safe adoption.”

Q: What does the Coalition hope to accomplish over the next 2-3 years? What would a “successful” outcome look like for the coalition?

Dr. Anderson:

“CHAI is evolving rapidly. After a summer and fall of meetings – both virtual and in-person – we are rapidly finalizing our framework and recommendations to be shared publicly by the end of the year. Although the short-term goal is to release a framework, consider it a 1.0 offering.

Given the speed at which AI is advancing, we expect that the initial framework and recommendations will need to be adjusted and calibrated as we study the impact of care and develop actionable data.

However, one thing that will not change is the prioritization of CHAI principles around patient safety, reliability, fairness, transparency and trust.