In 2007, the annual meeting of the Society of Hospital Medicine concluded, as usual, with a speech by Robert Wachter, MD, FACP, professor and chair of the department of medicine at the University of California, San Francisco, and a father of the hospitalist field.
The title was “Hospital Medicine in 2017: A Note from the Future.” Since that distant future has now arrived, ACP Hospitalist recently spoke with Dr. Wachter to assess his predictions.
Q: First, you very accurately predicted there would be 50,000 hospitalists in 2017. Based on that, do you want to make a prediction for 2027?
A: Did I say that? Boy, that was smart. I'm guessing it will be about 60,000. I don't think the growth curve will be as rapid as we've seen in the past 10 years, because I think the number of hospitals and hospital beds will shrink, and we've come close to saturating most of the mid-size to large hospitals.
In some ways, it depends on the definition. If you consider physicians and other clinicians working full time in postacute care and skilled nursing facilities as hospitalists, then it might be a little bit larger than that, because that's increasingly going to be where the action is. If it's purely hospitals as we think of them currently, there will be continued growth, but it won't be quite as fast as we've seen.
Q: A less accurate prediction was that you said you'd be shocked if resident duty hours weren't reduced below 80 per week, the limit that was put in place in 2003. What do you think happened there?
A: That clearly was wrong, but what did happen was further tightening, just not the total number of hours. They kept it at 80, but what they did [in 2011] was to tighten the rules in other ways—so, for example, interns can no longer stay overnight, which they could in 2007.
My sense at the time was that the pressure to change resident work conditions and rethink the role of residents in academic health centers was going to continue to grow. I didn't anticipate that the bulk of the research would show that duty-hour limits didn't work very well. I believe that if the research had shown that if you shrink duty hours, not only do you have happier and better-rested residents but you also have important impacts on patient safety, then the pressure to reduce them further would have been even greater.
What we've seen is that there are tradeoffs here and not everybody is wildly happy about the duty-hour limits. We reached a state of equilibrium at 80 that I didn't expect.
Q: Looking at quality measurement, you predicted it would have “matured” by now, with outcome measures replacing process and structure measures and pay for performance having “regained its legs after false starts.” How do you feel about the current status of those issues?
A: I think directionally that was right. There's clearly been a backlash against process measures. There has been more emphasis on outcome measures than there was previously, and if you look at the measures that people obsess over these days, it's less about “did you give this patient a beta-blocker” and more about your readmission rate or your central-line-associated bloodstream infection rate.
It hasn't gone as fast as I thought, and I think the limitation there has been the data. Because of that, I suspect that it will speed up over the next 10 years. Part of why outcome measures were hard to both get and adjust based on severity was, in 2007, we were still mostly on paper. Now we're mostly digital, and as we achieve interoperability, the ability to track outcome measures will be better.
In terms of pay for performance, it's a mixed bag. Maybe I was a little bit too fast, because at the level of the individual physician, there still is not all that much pay for performance embedded in day-to-day practice. Clearly the decision by CMS and others to push toward value-based payments comes with a much stronger emphasis on changing the payment system to hinge on performance. It took a little longer than I thought it would take—rolling out in 2018, 2019—but by the time the Medicare Access & CHIP Reauthorization Act (MACRA) fully rolls out, I think it will be accurate to say that there's a lot more emphasis being placed on performance.
Q: On the topic of digitization, you accurately predicted that electronic health records would be a “must-have” for hospitals, but it seems like interoperability moved a little bit more slowly than you expected. Your predictions included a Google Universal Record that would have all of patients' health data, self-populating with their physiologic measurements and calculating expected outcomes. What happened there?
A: Wow, and that was pre-Fitbit. I was influenced by being on Google Health's advisory board at the time, and that was what Google was trying to do. When you're sitting in the Googleplex talking about the future of the world, you can easily believe everything you're hearing. A year or two later, Google Health had collapsed and I would have been much smarter then.
I was right about digital records being essential. However, I would have gotten that wrong had it not been for the U.S. economy imploding in 2008 and the stimulus package creating $30 billion of federal incentives. Had that not been the case, we probably would not be at 95% of hospitals digitized; we'd be at more like 50%. In retrospect, it was clear there needed to be a federal push to get digitization done, and there also needs to be more federal push to get interoperability done. I would be shocked if it's not completely done by 2027. It feels like it's gathering momentum. There is less pushback on it than there was, and even the entities that have not been helpful—particularly the big IT vendors, but also large health care systems—now know it's inevitable, know they have to get on board, know that Congress is going to push them. So it's just a matter of time.
Q: Your imagined hospital of the future also had CT scanning in the entrance to the ED as well as a scan given with breakfast for every inpatient. What are your current thoughts in that area?
A: It sounds like I was almost being facetious. However, pretty much every big ER now has a CT scan. You don't roll through it on your way in, but it's almost hard to get out of there without getting one.
In 2007, we were just beginning to think about overuse of radiation as an issue, and not even quite beginning to think about the issues of overuse and the Choosing Wisely sensibility. Thankfully, those have both become important forces in the last 10 years. But the thinking captured in my prediction is probably not wrong; as we have technologies to scan or sense that are cheaper and safer and easier and more ubiquitous, the temptation to go ahead and just do it is awfully hard to resist.
One of the things we've learned in the last 10 years is those things are not costless. It may seem costless to monitor pain as a vital sign or get an MRI on somebody when you're not all that suspicious of something terrible. But then, after you treat pain as a fifth vital sign, all these people are now getting opiates that maybe shouldn't. We're doing all these scans and we're finding stuff and some of them are false positives. This may well be the dominant issue for the next decade. For health care systems and doctors who are now being measured and paid based on value, just because you can do a test, should you? To me, the answer is no, but that's a hard change for American health care culture.
Q: Finally, on the subject of telemedicine, the last slide of your predictions showed a hospitalist seeing a patient, with the patient's daughter pulled up on one screen, and another screen where the hospitalist was choosing between overseas nephrologists who could consult on the patient. Do you think that vision is still in our future?
A: Yeah. We had to get the technology right, and if you look at the explosion in telemedicine the last 5 years, I think we finally have. It's now becoming mainstream and lots of patients are getting their care through telemedicine.
The obstacles are workflow, workforce, and payment. Those are nontrivial, but when I talk about all this now, I say we're in the middle of twin revolutions in health care. One is intense pressure to improve value and the second is that we have finally become a digital health care system. It's the marriage of those two things that create innovation. My mistake in 2007 was to say that if the technology can do those things, they will happen. They don't, unless the organization is feeling intense pressure to deliver higher-value care and then it begins to ask hard questions about which doctor is going to do this or that or how we get family members engaged.
It takes 5 or 10 years during which people begin to say, “Oh, there's a completely new different way of doing this thing that involves new people and new relationships and new flows of money.” As I think back on that vision, it still feels right to me. I can't tell you the number of very sick people I see in the hospital where I really need to know what the nephrologist thinks I should do. Right now, I call a consult and the specialist comes and sees the patient 12 hours later, and then leaves a note, and I read the note the next day. If we could just figure out a way of having a conversation—get past the logistics and the billing piece—we could solve the problem in 10 minutes. Now that the technology can enable that, I think it's just a question of the health care delivery system feeling enough pressure to make those kinds of changes.
Q: Any closing thoughts on your predictions?
A: In some ways it sounds like nothing went faster than I thought and a few things went slower. If so, that violates the usual principle that we overestimate how much change we can see in a year and underestimate what we can see in a decade. But what I think is going to be different about this next decade is digital. By markedly speeding up the flow of data and allowing for new entrants into our field, people and companies that think about the work in a different way, I think digital will facilitate a more rapid pace of change.