Dr. Hayworth: Errol made an excellent point about AI. There is a product that’s being used in Europe and in some other parts of the world that can provide 85% of care through an algorithm without a patient even having to speak to a nurse or doctor. The company that offers the product claims a high level of patient satisfaction and a very low error rate.
We are a long way from the point at which—and I don’t anticipate that we’ll ever get there—AI fully replaces human providers, but there’s enormous and growing potential for data aggregation and machine learning to enhance, exponentially, the capabilities and capacity of care teams.
The most immediate applications for AI in the United States are in diagnostics, pathology, and the mapping of protocols for patients with cancer who will benefit from access to investigational interventions and clinical trials. As we gain experience in those areas, acceptance and confidence will lead steadily to broader deployment of AI, enhancing the quality of care and the efficiency of delivery and saving costs.
Dr. Norwitz: AI is a tool to assist providers. It is not going to replace us, which is the fear.
Ms. Pearson: From the consumer perspective, again, there is concern that if not enough data are available from Black, indigenous, and people of color, the levels won’t start out in a good place. The criticism over mammography randomized controlled trials (RCTs) has existed for a long time. The big trials that got all the way out to mortality did not include enough women of color; and so women of color rightly say, “Why should we believe these guidelines developed on results of the RCTs?” My point is that because of historical inequity, logical solutions such as algorithms do not always work for communities that were previously excluded or mistreated.
Dr. Levy: Your point is incredibly well taken. That means that those of us researching and working with AI need to ensure that the data going in are representative, that we are not embedding implicit biases into the AI algorithms, which clearly has sometimes already happened. We have to be careful to embrace input from multiple sources that we have not thought of before.
As we look at an algorithm for managing a postpartum patient or a postoperative patient, have we thought about how she’s managing her children at home after she goes home? What else is happening in her life? How can we impact her recovery in a positive way? We need to hear the voices of the people that we are trying to serve as we develop those algorithms.
Perspectives on future health care delivery
Dr. Levy: To summarize so far, we are thinking about a Woman’s Medical Home, a capitated model of comprehensive care for women that includes mental health, social determinants, and home care. There are different models, but a payment structure where we would have the capital to invest in community services and in things that we think may make a difference.
Dr. Norwitz: I think the health care system of the future is not going to be based in large academic medical centers. It’s going to be in community hospitals close to home. It’s going to be in the home. And it will be provided by different types of practitioners, whose performances are tracked using more appropriate outcome metrics.
Dr. Levy: I also think we will have community health workers. While we haven’t talked about rural health and access to care, there are some structural things we can do to reach rural communities with really excellent care, such as training community health workers and using telemedicine. It does require thinking through a different payment structure, though, because there really isn’t money in the system to do that currently, at least to my knowledge.
Janice, do we have enough motivation to take care of women? Women are so underrepresented when we look at care models.
Dr. Huckaby: I do think there is hope, but it will truly take a village. While CMS (Centers for Medicare and Medicaid Services) has its innovation center in the Medicaid space, it’s almost like we have to have the payers, the government, the specialty societies, and so on say that we need to do something better. I mention the government because it is not only a payer but also a regulator. They can help create some of these things.
There are opportunities with payers to say, “Let’s move to this kind of model for that.” But still, we are implementing change but on a fairly minor scale.
We could have the people who care about issues, help deliver the care, pay the bills, and so on say, “This is what we want to do,” and then we could pilot them. It may be one type of pilot in a rural area and one type of pilot in an urban area, because they are going to differ, and do it that way and then scale it.
Telemedicine, or telehealth, is part of creating access. Even some nontraditional settings, such as retail store clinics, may work.
Continue to: Dr. Levy...