Editor's note: Jack Needleman, PhD, is a Professor in the Department of Health Policy and Management at UCLA School of Public Health and Associate Director of the UCLA Patient Safety Institute. He has performed some of the key studies on the impact of the nursing workforce on health outcomes, including seminal New England Journal of Medicine articles on this topic published in 2002 and 2011.

Dr. Robert Wachter, Editor, AHRQ WebM&M: What got you interested in nurse staffing?

Dr. Jack Needleman: Two things about nurse staffing issues intrigue me. One is that nurses tend to be viewed as a cost center, rather than a core service line of the hospital whose work is important for the quality of care, the safety of patients, and the efficiency with which the hospital operates. The tension between those two views is interesting and important in understanding the behavior of hospitals as organizations and their ability to improve quality and efficiency.

It's also clear that patients being discharged from hospitals, nursing homes, and rehabilitation units are sicker than they have been. Major changes in technology are underway in terms of computerization, electronic health records, and the ability to use technology in monitoring and delivering care. Given the ongoing shortage of nurses, nursing is going to have to change the way it does its work. Because the work is so complex—nurses are constantly juggling different demands for different patients—the redesign of the work must involve the frontline staff. How the work is changed to make patients safer and deliver care more efficiently, both the processes for making change and the changes themselves, is both fascinating and important to study.

RW: Does new technology make nurses' lives easier or harder, and does it make us need more or fewer nurses?

JN: As hospitals try to shorten length of stay by discharging patients who don't need to be in the hospital, the remaining patients have more need for nursing care. However, nurses practicing in rehabilitation hospitals, skilled nursing facilities, and home health agencies are also dealing with patients who are sicker. Throughout the health system, all nursing care is fundamentally dealing with patients who have been sicker than they ever have been. In the normal course of things, that would imply we need more nurses, not fewer nurses. But, there is a shortage of nurses, which may be eased but is unlikely to disappear, and therefore we must figure out how to reorganize the care. Little research has examined how to get maximum value out of the nurses or how to allow them to shed the work that doesn't need to be done by them, and perhaps doesn't need to be done at all. That I think is part of the future agenda. Making more effective and creative use of technology is part of that agenda.

RW: From having watched nursing care, not only through the research you've done but also your evaluation of the Robert Wood Johnson "Transforming Care at the Bedside" program, how much of the nurses' workday do you think is wasted?

JN: I would be very hard pressed to document this in any formal sense, but my suspicion is somewhere between 10% to 25% of a typical hospital nurse's day is spent in work that makes little contribution to the effective delivery of care.

RW: Is part of the problem with nurse staffing ratios and the nature of their work that hospitals have not been under pressure to achieve value—that essentially the business case to think about the right size of the nursing workforce, the right training, and the right work hasn't been there for the entities that employed them?

JN: Hospitals have been under a substantial pressure to contain costs. That pressure is going to continue, if anything it's going to get worse. Labor is the largest component of hospital costs and nursing the largest component of labor. There's strong pressure on hospitals to follow the Willie Sutton theory of cost containment, which is to go where the money is, and the money has been in nursing. What we saw through the 1990s was length of stay going down, patient acuity going up, and nursing not increasing to reflect the higher demands associated with more rapid turnover and sicker patients. That was one thing that contributed to the calls in the mid to late 1990s for more study of patient safety in hospitals because of potential low staffing of nursing. The calls were for efforts to document that patients were at risk when nurse staffing is low.

RW: In the future, if hospitals and other health care organizations operate in an increasingly transparent environment around value and maybe even one in which pay is predicated on measureable value, do different things happen around the organization of nursing care and nurse staffing? One could at least muse about a future state where organizational incentives on value are so intense that hospitals automatically try to figure out the right number of nurses in ways that might be very different than they have in the past.

JN: I think that has potential and is one of the reasons why nursing groups need to embrace the payment-for-value models. They need to embrace systems in which reimbursement is going to be tied to reduction of never events, hospital-acquired infections, and other preventable adverse events. There's some concern in the nursing community that if things do not improve nurses will be blamed. I truly believe that nursing should be embracing those options because it changes the calculation about what's important and when it's important to employ nurses. I did work with Peter Buerhaus and colleagues looking at the business case for nursing. Under existing reimbursement systems, the business case for maintaining your staffing at levels that keep patients safe, reduce adverse events, and enable nursing teams to operate efficiently to reduce length of stay was not strong. It cost more than was saved in costs, and given the current reimbursement systems it probably cost a hospital more than it would recover in payment. But, as the payment systems change the business case has to be reevaluated, and with more public reporting increasing hospitals' concerns about their reputations, I think the equation will change and we'll see hospitals paying increasing attention to assuring they deliver effective nursing care.

RW: As you thought through the business case, did you factor in issues of nurse retention?

JN: In the 2006 Health Affairs paper, we did not consider turnover as an avoidable cost of low staffing. Several years later, I analyzed the effect of turnover, drawing upon some of the research on staffing and intent to leave. It reduced the gap in the difference between cost offsets or savings, but didn't completely close it. But as the reimbursement system changes, I think we're going to see the business case look different in the future. As we move back into an environment in which nurses are in tight supply, issues like turnover and training are going to be increasingly important.

RW: From your work in evaluating "Transforming Care at the Bedside," what did you learn about the bedside work of nurses that surprised you?

JN: What I knew from looking at the literature and what I think the public understood was that nursing is physically and emotionally demanding work. What I think the public didn't recognize, and what I came to appreciate as I talked to people during the early analysis, is that it's intellectually demanding work. This whole issue of being able to observe your patients, assess them, anticipate how they're doing and what could be going wrong, and take appropriate intervention steps requires a high level of training.

I have also come to appreciate that nursing, even for the frontline staff, is managerially complex work. The staff nurse that comes onto a med–surg unit has 4 to 8 patients assigned, each of whom has a to-do list at the start of the shift, and has to figure out how to integrate those to-do lists, how to keep revising their daily to-do list as patient conditions change, as new things get ordered, as things that should have happened get missed, and as new patients come onto the unit. Patricia Ebright refers to this as "managing the stack." Think about your to-do list as a stack of things to get done. That job of anticipating what the stack needs to look like, dynamically changing it over the course of the day as things change from what you anticipated, is not only an intellectually demanding task but also a managerially demanding task in a way that's very complex.

Addressing this managerial complexity is one reason why the work of nurses needs to be redesigned, and its complexity requires that nurses be an integral part of work redesign. When thinking about improving care or implementing evidence-based practice, the tendency is to look at specific problems one-by-one. Patients have skin problems. They're getting decubitus ulcers. The "solution" is to put in an assessment form and an action plan associated with the assessment. Patients are at risk of falls. The "solution" is an assessment form and an action plan related to the risk of falls. They're at risk of delirium. For each individual problem, the natural inclination is to simply add another assessment form and action plan and add to the stack of work. Well, you can't do that. Nurses only have a limited amount of time in the day. We need to figure out how to better integrate that work, and the frontline staff knows more about the integration activities that need to be done than anybody else.

RW: You mentioned that it's emotionally, physically, intellectually, and managerially complex. What have you come to understand about its sociological complexity and particularly the relationships between nurses and doctors?

JN: One of the critical issues for nurse satisfaction is effective communication among the team and in particular effective communication between physicians and nurses, since nurses are often the focal point for the negotiation with all the other members of the care team. When that communication is not effective, there's a lot less satisfaction with the work environment and the care environment for the patients is also not as good.

RW: What was your hypothesis when you started studying actual staffing compared to staffing targets?

JN: We studied patient outcomes when actual staffing was substantially below hospital-set targeted or recommended staffing. Almost all prior work, including work my colleagues and I did, was cross-sectional, comparing high-staffed and low-staffed hospitals. When I looked at the variation in staffing across hospitals, I saw many hospitals whose reported staffing was so low that I would not want to be a patient there. Minimum staffing requirements [such as exist in certain states like California] probably keep the patients in those institutions safer than they otherwise would be. But fixed ratios do not solve the problem of places that have high acuity patients that require substantial nursing. The medical center we studied for our 2011 New England Journal of Medicine article was a magnet institution, and it had high staffing levels because it treated very sick patients. Its staffing targets were high and they hit their targets 84% of the time—that's the best case. Its staffing was well above the targets generally put forward in legislation, and it both looks like and is a well-staffed hospital. What we found is even in that environment, patients exposed to shifts where the actual staffing was substantially below what the hospital's own system defined as the appropriate level of staffing were at increased risk of mortality, and I suspect at increased risk of other problems as well. The real issue is figuring out for an individual institution what level of staffing is needed to keep their patients safe, to deliver care reliably, to deliver care efficiently, and how to ensure that it can hit those targets most of the time. The first message from this more recent work is that patients are at increased risk when staffing is not sufficient to meet patient needs, but that may differ from a standardized nurse-to-patient ratio. The second important message from that work is hospitals that have high average staffing don't get a free pass. To the extent their high average staffing reflects the fact that their patients are high nursing acuity, they need to make sure that they're able to meet the nursing needs of those patients every day of their shift or come as close to that as they can.

RW: Speaking as a researcher, is there evidence that a regulatory approach, such as a state-mandated ratio, actually improves safety?

JN: I've been looking at the literature on the impact of the California ratios, including some papers that are currently in process. For hospitals that have very low staffing, improving the nursing ratio should, all other things being equal, make patients safer in those institutions. There's some evidence that patients have been made safer, but all other things are not equal. Without payment that supports the improved staffing, hospitals have to look at other ways to meet the staffing needs within the constraints of current revenues. It's not clear what hospitals are doing. To the extent that they undertake efforts to control costs by investing less or cutting back on other things that keep patients safe, the net effect of the ratios will be reduced. So it's a complex environment. We need to build payment systems that create incentives to maintain safe staffing levels.

RW: Why do you think there is so much more research on nurse staffing levels and the work of nurses than comparable research on physicians? There's a moderate amount of literature on the right amount of staffing, its impact on outcomes, and the right qualifications for nurses. I've tried to find the parallel literature about the right number of doctors, how hard and how many hours doctors should work, and there's virtually nothing.

JN: A couple of things are going on. First, we see a fair number of simple descriptive studies about the hours that physicians work, how many patients they're seeing, length of visits for outpatient care, some research on specialty versus generalist, and specialty certification on patient outcomes. Thinking about the way the physician side of care is delivered and organized has probably been understudied. As policy and practice require us to think about the issue of primary versus specialty care, the impact of electronic health records on both efficiency and quality, the implementation of accountable care and patient-centered homes, and expansion of telemedicine in primary and specialty areas, we're going to see increased attention to issues of physician organization and productivity. One of the reasons why we see less research on physicians is the data are harder to come by. Information on ambulatory practice, outpatient practice, even physicians working in hospitals is hard to come by and hard to link to individual patients. In contrast, the overwhelming majority of nurses work in hospitals, so research about nursing care is fundamentally research about hospital care, and the databases have been better, with the state hospital discharge dataset, which includes information on hospitals' financial performance that often include information on staffing.

The second reason is that nurses were quite appropriately expressing concern all through the 1990s about increasing workload, slow growth in the number of nurses, and that nursing staffing was not adequate to keep patients safe. When the Institute of Medicine (IOM) studied this and issued a report in 1996, the committee basically said that it believed nurses made a difference in patient safety but the research base was thin. Part of the growth in the late 1990s and through the last decade in articles and studies specifically looking at nursing care has been in response to IOM's call to build an evidence base about this very important topic. In contrast, there's been an assumption that physicians may be overworked, but by and large they've been doing a good job in the outpatient area. Now we're increasingly finding inconsistencies in the quality of care delivered by physicians in the outpatient area and efforts to understand what accounts for that. So those inconsistencies are driving a research agenda that may answer the questions.

RW: Nurses generally are employed by organizations and therefore there's a natural tension around whether the organization hired enough of them. Physicians have traditionally worked for themselves, and so they would have to blow the whistle on themselves if they say, "my ratio is not right" or "I'm seeing too many patients." That's changing very rapidly, and I wonder if we will see a similar dynamic in the future with more and more employed physicians.

JN: Yes, and there was that article that appeared in the New York Times recently about how the physician work force is changing as they become employees rather than individual entrepreneurs. Don Berwick used to talk about the "bad apple" theory of quality, that the problem is bad apples and my apples are fine, leave me alone. This is in sharp contrast to adopting a systems view, which is that systems can make it hard or easy to deliver high quality care. I think that it is easy to let the bad apple theory dominate your thinking when you're in a small practice with you and a couple of colleagues and you believe that you're doing a good job—certainly doing it as well as you know how to.

RW: Yeah, and you worked 100 hours a week in your training too. So it becomes inconceivable that a 70-hour workweek would be too much.

JN: Kübler-Ross talks about the three stages of grief, and there are stages I would characterize as "quality denial." Very predictably, as people are presented with quality information, the first reaction is "your data are wrong." The second stage is: "my patients are sicker and you haven't taken that into account." Then you get to the third stage, which is: "I'm practicing as well as I know how to already, there's nothing more I can do." We're learning to get people through those first two stages. That third one is challenging. It's challenging in part because people are working 50, 60, 70, 80 hours a week where the amount of time that's available to reflect on practice is very attenuated. Our models for continuing education have not been terribly good about getting physicians to reflect on how to improve their routine performance in practice as opposed to learning new things. These have been the kinds of major challenges that we have to get the delivery side and the outpatient side performing as well as we would like it to.

RW: Anything else that you'd like to talk about that we haven't covered?

JN: We've been talking about safety, reliability, and quality. But the agendas for the next decade are: bending the cost curve, increasing the efficiency with which we deliver care, and moving from what I would characterize as a middling level of quality and safety to high levels of quality and safety. We need to do those simultaneously. The days when CEOs, COOs, nurses, and physicians could say we're worrying about quality on Mondays and Tuesdays and we're worrying about efficiency, Wednesday, Thursday, Friday, and most of the weekend, and we treat those as two separate agendas, those days must end. That's not a productive way to do it. Just as I talked about the need for nurses to think about how to integrate new practice and new technology into the way they work day to day, the cost containment agenda, the efficiency agenda, and the quality agenda need to be integrated into the thinking of everybody involved in improving care.