issue Summer 2021

Search for Health

By Judy Masterson
David T. Feinberg, MD ’89, MBA
Photo by Michael R. Schmidt

David T. Feinberg, MD ’89, MBA, spoke with medical students on April 9 as part of the annual Aaron Brown Memorial Lecture, sponsored by the Chicago Medical School chapter of Phi Delta Epsilon. Here, Dr. Feinberg, the former CEO of two of the nation’s most innovative hospital systems — UCLA and Geisinger in Pennsylvania — shares his thoughts on the digital transformation of health care from his perspective as VP of Google Health. In August, the Missouri-based health information technology supplier Cerner announced that Dr. Feinberg will take the reins as its CEO and president in October. Excerpts from the conversation have been edited for brevity and clarity.

HELIX: What attracted you to Google?

DR. FEINBERG: I wanted to learn if tech can actually help in health care. Having been here for almost two years, I am positive with absolute certitude that this technology can save lives at scales that none could imagine. But what I have learned is we won’t be able to do it if we don’t do two things: 1. Earn people’s trust. If people don’t trust us, then lifesaving technology can go sit on a shelf. 2. Google is going to help save lives through others. We are going to do this through doctors and nurses and pharmacists and social workers. We are going to do this through partnerships. Our tools, as amazing as they are, are really going to be a way for families and communities to take better care of themselves, and for doctors and nurses to spend more time at patient bedsides and less time in front of computers. Our tools are going to let care managers predict who is going to get sick. But the care, fundamentally, is not technology. Health care, fundamentally, is people caring for people. That sometimes means a family caring for a family member.

For us to be successful as a tech company in health care, trust and partnership are going to be probably more important than the coolest machine-learning AI I have ever seen in my life.

Sometimes it means a professional caregiver. Sometimes — and I think we’re living it with COVID — it means a community caring and in some cases not caring for itself. For us to be successful as a tech company in health care, trust and partnership are going to be probably more important than the coolest machine-learning AI I have ever seen in my life.

What do you think is the single most important piece of technology for improved healthcare delivery?

Google Search! So 30 years ago, when I graduated, jokingly, I thought I knew everything, or at least the senior doctors did and the patients didn’t know anything. They didn’t go to the library and look up “melana.” There was a real imbalance. Twenty-one years ago, Google Search came out, and for the first 10 years, it was pretty bad. Patients came in and they had printed out the internet. And you were like, “Come on now — you’ve gone down a rabbit hole.” The last few years, we’ve really worked to get that information authoritative, useful, dignified, culturally sensitive, accessible. And it has fundamentally changed the relationship from a paternalistic one between your doctor and you into one that is actually now a partnership. And you can actually find pretty good information. A recent Harvard study concluded that doctors should recommend patients go to Dr. Google. That’s a big flip over 20 years’ time. It took a while. But that technology of organizing information, in this case health information, and making it accessible, I think has been the biggest tech success that I have seen in my career.

How is Google working to automate or improve data infrastructure for electronic health records, which have been a source of burnout for so many health professionals?

We’re working exactly on that in partnership with Ascension. What we’ve done is go with Ascension and help the doctors normalize (organize) an individual patient’s record. And patient records — they may be with Athena as an outpatient, or they may use Epic or Cerner (healthcare software) and they may have gone to CVS. That information is spread across multiple systems. So this project — new software we’re calling Care Studio — has been in the works for a number of years. It’s now live. So we’ve normalized the data and the screen for the doctor is what looks like a Google Search homepage. It’s just a Google Search bar. The software offers a single, centralized view that automatically brings forward a patient’s important information. If you get called to see a patient who has melana, you can type “melana” or “anemia” or “GI bleeding” — that same query expansion that we all use in the regular part of our world with Google happens for that patient’s record. Within a microsecond, everything comes up that you need to know. Anemia. What kind of medications they’re on. It’s looked at all of their hemoglobin levels from multiple hospitalizations normalized in a really easy-to-use interface — actually, this patient always runs a low hemoglobin, so they don’t need to be transfused — and then we help you write the note. And writing the note is the same way we help you complete your Gmail. The tool actually allows us to do predictions around who’s going to get sick or what’s going to happen. It allows us also to think of data insights from a population health standpoint. The thing that it really does that blows me away — with enough training, it can actually write the person’s medical record for the next week or two or a year. We can tell you what’s coming next. And when you have that, now you have the ability to say, “Well now, let’s tweak it. If we gave this medicine, what would happen?” We are just organizing the information and making it accessible. I don’t want us to build an EHR. I want us to build this on top of the EHR ability. Ultimately, we want to take that same data on the individual and combine it with what we do on the consumer side.

In light of widespread vaccine hesitancy and skepticism, how can trust in medicine and science be improved, especially in terms of data collection?

Trust is absolutely crucial. Getting information out is really, really important. But people make decisions based on human connection. So your doctor or your cousin or your friend or your baseball friends are really influential on how people make decisions. So we’re looking right now at vaccine issues. We were really focused originally on information about COVID and then information about COVID testing, and now “vaccine near me” has gone up considerably on Google Search. We want to get people the right information, and we can see by search queries and populations that there are certain groups that aren’t doing the query enough. We can tell where there is vaccine hesitancy. Sometimes it’s an access issue, sometimes it’s a trust issue and sometimes it’s just a blatant misinformation issue. So we’re working super-hard on that. For example, the only subject researchers from Vanderbilt University were unable to get a conspiracy theory to recommend a further conspiracy theory or video was on vaccines. That was our team. So if you type in “vaccines are bad,” the next video you get is authoritative information on vaccines. If you have bias, you can find it on the internet, but we don’t have to put it on the first page. Then there are sometimes gaps, places where the authoritative folks haven’t spoken up. We then engage with them to get that content so that we have pages and pages of good stuff before you get to the bad stuff.

Watch Dr. Feinberg’s full talk with CMS students at