This has really been a bipartisan effort. A number of provisions on interoperability were helped along in the complex governmental processes by my predecessors at ONC, including the 21st Century Cures Act, which was signed into law in December of 2016.

The law states that there should be application programming interfaces without special effort; there shall not be information blocking; and so on. But in terms of what that means, we had hundreds of stakeholder meetings and public presentations. With all of that input — including input from athenahealth — what we accomplished is putting a robust pathway in so that patients are going to be able to get their data and have it under their control.

The Cures Act lays in place the foundation for patients to access their data on their smartphone — and operationally, they can start becoming in charge of their healthcare. It's one thing, realistically, to be able to access your data on a portal. It's a whole other thing to be able to have your data on your smartphone to do with as you wish. I believe that is going to be transformative in healthcare. It's going to make care much better, much richer. And it obviously puts a vast light of transparency into what we can know and the way we've practiced medicine

I’ve been a practicing doctor since the early ’80s. I think this [legislation] will bring medicine and the practice of medicine into some of the same dynamics that we have in the rest of our consumer lives. That’s a big thing. Obviously other big things are the supporting elements — the provision as we go to standard, readily usable application programming interfaces, and the rules on information blocking — we need to make that a workable thing.

We’ve really not had much automation in healthcare, so I think that has potential to transform the American healthcare system over the next five to eight years. When providers or payers have access to clean data, they can use modern machine learning and big-data techniques.

When I went into medicine, I did my residency and then decided to learn computer science. I did that because I was tired of the waste-of-time and the clerical, no-value-add activities that I had to do throughout my residency that were expensive, brittle, unsafe, and slow. And it was clear that automation would be a big help. The EHR and its first instance has become a building tool…You can’t find information in a chart because they are full of templated text that has little or no value. In the last administration we’ve done some major things to try to reduce the burden of chart pollution [including revising evaluation and management codes]. We've also started a number of things on integrating clinical financial data to get rid of some of the other burden generators, this sort of goofy prior authorization system that we have.

We haven't had to focus on operational efficiency. That’s intrinsic to computers. And we haven’t had to focus on the increasing consumerization of healthcare, higher deductibles. It is well within our grasp to eliminate almost all of the payment-related burden in healthcare using the technologies that already exist today.

Most of the burdens really come because we don't have a great way of paying for healthcare.… We have all of these proxies for value and they’ve gone sideways. If you look at things like quality measures as done by government agencies or private payers, it’s really a classification problem. In computing, you know who are the “best” doctors, the “best” hospitals. That classification is easy to do with the dozens and dozens of things that big data provides that’s no burden on providers.

While we've had some health information exchanges on a national level and in a number of states, I think we've underused those resources — both in the diagnoses of the pandemic in lab tests and how it correlates to race, ethnicity, comorbidities, medications. A number of the HIEs have some pretty stunning data, especially the state and local ones that capture all of the patients in a district.

Public health authorities have historically loved mandated reports. Well, look — as clinicians we know that is salaciously burdensome and answers yesterday's question. We’ve had more mandates frankly than I would have liked to see in reporting of test results. That information, because it’s typically deidentified, means we’re leaving most of the clinical information unknown.

And the whole immunization information system has been sort of a weak sister in healthcare. It’s like fighting a war with a rifle as opposed to a cannon. Our current Vaccination Adverse Event Reporting System was designed to mitigate vaccine manufacturer liability. That isn't going to get population-level data on the impact of vaccination.

The IISs (Immunization Information Systems) aren’t rich systems like EHRs. We need to rethink the entire stack of public health data.

In my ideal world, public health agencies, for the most part, wouldn't be collecting data. It's duplicative. It's after-the-fact. These out-of-process reports are expensive. They should be using privacy protections — some of which are extraordinary these days — and data captured in the course of care from health information exchange. We need to move the entire public health scrum from reporting to exchange and have our world be based on that rich-knowledge machine discovery.