Editor's note: Dr. Bindman was appointed as Director of the Agency for Healthcare Research and Quality (AHRQ) in May 2016. Prior to this, he was professor of medicine, health policy, epidemiology, and biostatistics at UCSF. He also served as Director of the University of California Medicaid Research Institute and Director of UCSF's Primary Care Research Fellowship. We spoke with him about his new role at AHRQ.

Dr. Robert M. Wachter: When you heard you were going to be taking this role, what were you most excited about?

Dr. Andrew Bindman: It's a great time to be working within the federal government. There's so much attention and energy being given to improve the health care system and tackle the biggest issues in health care coverage and delivery. AHRQ is an amazing organization in terms of its focus on doing research on the health care delivery system and trying to apply that research to make the system more effective, higher quality, safer, more accessible, more affordable, and more equitable. These themes are central to my interest in health care, and it's very exciting to be able to address them from a federal agency that takes leadership in these issues.

RW: Let's talk a little about patient safety. You did a lot of work related to patient safety, but it wasn't your central focus in your career. As you've come in, are there things that you've learned about the world of patient safety that you were less aware of when you started?

AB: You're absolutely right. AHRQ is well known for its work in patient safety, and there are a tremendous number of assets at AHRQ related to this area of focus. It is headed up by Rear Admiral Dr. Jeff Brady, who has been a fantastic asset in leading the team. I've learned from him about the robust nature of the investments that AHRQ makes in research in this area. What's been most interesting and exciting is how it's become a real model within the agency for how research is done. Research is not done for just research sake, but with an eye toward how it can be moved into the hands of clinicians and health care organizations and then scaled in order to have a major impact. There has been a tremendous building of bridges between the research and action to what matters to patients and doctors and nurses.

First, AHRQ invests in figuring out what makes a difference in, for example, things like health care–associated infections. What creates the risks? What are the strategies to try to reduce them? Then building data systems to be able to monitor those risks and ultimately to develop tools to push back out in support of frontline providers, hospitals, and health systems to reduce them. I've been so impressed by what I have learned since coming here about the impact that this work has had. That it has been a model for how research has been translated into practice and has made a difference in terms of reducing, for example, the number of catheter-associated urinary tract infections and blood-borne infections in the hospital setting. And I had not appreciated the scale of the work, the ability to move from the research into scaling this up to a national level, which has really put a dent in starting to reduce some of the adverse events in hospital settings.

RW: Do you have a sense of the trajectory of the field? Are we making significant progress, and are there areas that you think AHRQ needs to focus on to accelerate improvements?

AB: We are making a lot of progress in certain areas, particularly in the hospital setting. For example, we have a national measurement strategy around hospital-acquired conditions. We've demonstrated that between 2010 and 2014 there has been a 17% drop in those hospital-acquired conditions, which has probably prevented about 87,000 deaths over that time period and led to an estimated $20 billion in savings. So that's obviously a major set of successes. However, we still have a long way to go to move our capacity to do that kind of measurement in nursing home and ambulatory care settings, where there are likely to be similar concerns about adverse events.

RW: One dominant theme in the last 5 years is that the health care system has essentially gone from paper-based to digital, fueled in part by a large federal investment. From where you sit, how is that an enabler of what you're trying to do? How has that gotten in the way of progress in certain ways, and where do you see that going?

AB: AHRQ does have a robust research portfolio in the area of health IT. This provides opportunities for us to study the kinds of questions that you're raising so that we can understand whether the promise of health IT is actually delivering for us. AHRQ has done seminal work about barcoded medication administration and the use of computerized provider order entry, which has demonstrated value in terms of patient safety and improving quality in general. But as your question brings up, there are many more aspects of health care delivery than that, which are ripe for quality and safety improvements. For example, we need to understand how to fulfill the true promise of health IT so that it doesn't become a distraction during patient–provider interactions and doesn't undermine, but in fact supports, intelligent decision-making. These are key questions around which we are trying to develop a research portfolio. How do we make sure that the health information technology revolution is improving patient care overall? I do think that this remains in many areas an open question.

We are also very interested in not just how health IT supports patient care at the time when the patient is sitting in front of you, but also the capacity of the data entered into electronic health records to be mined and used to generate knowledge and to support decision making for other patients with similar problems. We are never going to be able to do randomized trials on every clinical problem that comes up. The electronic health record—with appropriate standardization, with appropriate ways to protect information—can be mined in a way that could support new and more rapid knowledge generation and result in improvements in decision making when this information, combined with outcomes, is shared with patients.

RW: Another theme over the last several years has been growing rates of clinician burnout, and there are many reasons for that. One seems to be the electronic health record and the data entry burden, but there are others. One thing that you hear from people working in the improvement space is that they don't have anymore capacity, and that just admonishing people to try harder feels like it's reaching the end of its rope. What can be done to deal with issues of burnout not only of practicing doctors and nurses, but also people in the performance improvement world?

AB: It is a very complex and important issue. Certainly it has gotten attention because of the issues you're raising about data entry and the feeling of it being work that could detract from the actual interaction of talking with a patient around emotionally sensitive issues like their diagnosis and treatment. But I do think it's more than that. It's also how our payment incentives sometimes create that sense of just racing on the hamster wheel.

In the area that I know best, primary care, substantial questions must be asked about whether we have adequate staffing to perform the responsibilities that are being asked of us now. The job has become much more complex given the aging of the population, the growing burden of chronic disease, and the expectations to not only diagnose and prescribe medications but to also coordinate care within health care and across the community. So we have to think about whether our staffing is adequate in certain areas, and whether primary care has the capacity to address these challenges.

We need to have payment models in place that encourage new models of care, such as greater use of telehealth, which can leverage our available resources. AHRQ has been engaged in studying these new models. For example, one project AHRQ supported many years ago that has really generated a lot of interest is called Project ECHO. Project ECHO started in New Mexico as a telehealth hub. It first focused on hepatitis C and trying to support rural primary care providers who did not have adequate access to a liver specialist in their communities. It used the hub of the University of New Mexico to support primary care providers out in the community. What was remarkable about this program, and AHRQ was an early supporter of it and evaluated the work that they were able to do, is that they were actually able to achieve quality outcomes for patients with hepatitis C receiving care through primary care that compared favorably to those of patients going to the specialty clinic at the University of New Mexico. This has led to a rapid spread of this model and high satisfaction. Primary care providers have learned a lot from interacting with the specialists around how to provide this care.

But it also has brought up another issue. When this model was started, support was given to the hub to allow them to do teaching with the primary care providers. However, the primary care providers had to donate their time to be able to provide the cases and talk through the cases online. We're coming to an era in which primary care providers must have the knowledge when they're delivering service. We're probably going to have to pay for that in some way or build that in. We cannot just count on a provider going home at night, reading a bunch of journals, and getting up to speed. This relates to a different kind of payment model and expectations of what the work life will look like. We are having ongoing discussions at AHRQ about how we might explore some of these workforce and payment challenges that will hopefully lead to solutions that will reduce some of the stress that's leading to burnout. This is an incredibly important issue, and a big part of it is the enormous amount of information that is now coming at providers. It is too much to be assembled by a physician just reading it on a Saturday evening to be caught up. It now needs to be part of the workflow, and our health care system has not figured out quite yet how to do that.

RW: Now throw in genomics and patients having apps and sensors that are giving off data that theoretically the primary care doc now has to manage—it seems undoable.

AB: It's a lot more information to manage. It's too much to try to do by yourself as a physician. It clearly has to become a team sport. But while we're saying it's a team sport, we haven't quite figured out yet who does what and the communication pattern among those roles. We see clearly key elements of it. Information technology is probably going to be a centralized information hub, but it's not quite at that point yet. We're building in the alarms faster than we're building in some basic functions that we need to make sure that we can all do our job. We don't quite know yet what capacity is reasonable given the health risk of a population for someone to manage. A lot of research still needs to go on in the delivery system to inform realistic expectations. I think a lot of the frustration that you're referencing is that we haven't really set realistic expectations for providers about what the amount of work is that they should get done and at what level their knowledge supports should be to be able to accomplish that. We've thrown these tools in that are partially built and said, "Oh, it should all work great now." Research still needs to make sure that matches up correctly to what the job really looks like.

RW: We've both seen 20 or 30 years of primary care being touted as part of the cure for the system, yet the required focus and resources never quite comes to fruition. When a medical student asks what you think about a career in primary care, what do you tell them?

AB: Primary care can be a sensational career. But it depends a little bit on the setting. There's something incredibly fulfilling about having continuity of relationship with a patient. But if you have to do that in a context in which you are not really given sufficient time to develop that relationship and you do not have the support to allow you to be there at those critical moments that you want to be there for the patient, then it can be a frustrating experience. Even within an environment like UCSF, we could see that under different kinds of payment incentives and work expectations, some primary care providers could feel much more satisfied. Many who work in safety-net settings can feel quite satisfied because of the mission and feeling connected to that. But even then, the expectations for workload have gone up without the infrastructure to support that greater workload. That can be quite frustrating. We need to invest in primary care in a way that can allow what is currently a very mission-driven specialty to actually fulfill the expectations in a way that will be more satisfying over a career.

At AHRQ, primary care is one of our central focus areas, and we are putting a lot of attention into how we can better understand the realistic expectations for work and the tools that can allow a primary care practice to meet the opportunity in a satisfying way. I am optimistic that we can make progress, but I recognize that there is still frustration because many places don't have the infrastructure in place to allow people who want to go into primary care to do it in the way that they think is best.

RW: The President and others have talked about the ACA working incredibly well, yet there are things that need to be improved and tweaked, and that's been difficult in the political environment. What do you think AHRQ's role is in evaluating the ACA and informing decisions about its future?

AB: AHRQ has a lot of resources that can be helpful for researchers, both those at AHRQ and more importantly the broader community of health services researchers, to continue to assess health care overall and the health care delivery system. For example, we have a resource called the Medical Expenditure Panel Survey (MEPS), which is one of the most in-depth national surveys that gives us insight into the experience of individual patients in the health care system from the point of view of their health needs, their health care costs, and so forth. We have data from 2014 and 2015 that has allowed us to see transitions of individuals who were previously uninsured coming into coverage, how their health care utilization is changing, and how their health is changing. The ACA is very complex; it's very large. So it's unlikely that there will ever been any kind of single evaluation of it. It will take putting together multiple perspectives on different aspects of the law to understand it and to give insights about ways that it has been successful and how things might need to be improved over time.

Another great resource we have is something called the HCUP database, which is virtually the entire country's hospital claims and most of the country's emergency department visit claims. It is a valuable way of getting insights into the kinds of problems that emergency departments and hospitals deal with. These are important resources to assess whether the expansion of coverage is changing the pattern of care away from the hospital and emergency department when many health problems could be managed in primary care or other outpatient settings. I do think we are well positioned in terms of data resources to answer this and other important questions.

We're in the midst of beginning another very important data source to support this kind of analysis. It's called the Comparative Health System Performance Initiative, which will characterize all of the health systems in the United States. We are anticipating being able to update that over time so that we can monitor consolidation, for example, that might be going on in the health care marketplace. We can also determine different characteristics of the health systems in the United States and look at how those different characteristics ultimately impact quality of care, cost of care, and value. We are interested in improving our understanding of the role health systems play in generating evidence, adopting evidence, and applying evidence in a systematic way at the frontlines of care. We are particularly interested in patient-centered outcomes research, or PCOR, the sort of evidence that comes from comparative effectiveness studies. AHRQ is thinking a lot about what data will be most useful to support health services researchers to evaluate any health policies that come along—in essence, to ensure that we have an ability to do ongoing surveillance of the health care system.

RW: Obviously AHRQ operates in an incredibly complex political environment and there periodically are threats to the agency's existence. How does that influence the way you do your job?

AB: The reach of AHRQ is tremendous. It's a moderate-sized agency but has had an enormous impact. When we think about some of the most important health care problems in the United States, including the health consequences of adverse medical events, our investment in AHRQ pales in comparison to the investments we make in cancer and heart disease, perhaps the only two conditions that have a death rate higher than what we believe is associated with adverse events. So in reality, the investment is, if anything, too small. I try to at least make the case for that so people understand the choices they are making when they raise questions about whether AHRQ deserves the budget that it has. I think that what AHRQ does with its limited resources is extraordinary.