Nursing is generally considered a “cost” rather than revenue in a hospital context, which makes nursing a constant target for cost reductions. Because institutions are not directly compensated for providing nursing care, unlike physician services, there is little motivation for providing the right “dose” of nursing to meet patients’ varying needs. Some argue that direct reimbursement for nursing in hospitals would have a positive impact on quality of care by better aligning intensity of nursing care with hospital reimbursement levels and thereby reducing incentives to inappropriately ration nursing care (Welton, Fischer, DeGrace, & Zone-Smith, 2006
). Indeed, one of the leading nursing research journals recently published a report on a tool for measuring the rationing of nursing care (Schubert, Glass, Clarke, Schaffert-Witvliet, & DeGeest, 2007
). Given the general trend in health care payment toward bundling services together for reimbursement, direct reimbursement for nursing may not prove to be feasible, and thus other strategies need to be considered that might achieve the same ends—more appropriate levels of nursing care.
The systematic review and meta-analysis of the association of RN staffing levels and patient outcomes commissioned by the Agency for Health care Research and Quality concluded that there is substantial evidence that increased RN staffing is associated with better patient outcomes (Kane et al, 2007
). This comprehensive review helps refocus attention on the strongest component of a business case for nursing, that is, that more nursing may pay for itself by preventing costly adverse patient outcomes. In the health services research literature, cost off-sets refer to spending in one resource category to achieve an equal or greater savings in another. Given the fragmented system of financing of health care in the United States, cost offsets often do not benefit the same entity making the initial investment, thus undermining the incentive. For example, better nurse staffing in nursing homes results in reduced use of hospital emergency room visits and inpatient stays for nursing home patients (Aiken, Mezey, Lynaugh, & Buck, 1985
). However, better nurse staffing in nursing homes would cost Medicaid money, whereas the beneficiary of reduced hospital expenditures would be the Medicare program. In the case of hospital nurse staffing and better patient outcomes, the same financial entity—the hospital—stands to benefit if better nursing prevents the need for costly patient care expenditures.
Aiken and colleagues, in a 20-hospital study of inpatient AIDS care, found that Magnet hospitals achieved better risk-adjusted patient outcomes without higher overall expenditures. Although Magnet hospitals had better nurse staffing than other hospitals in comparison, they also had lower risk-adjusted mortality, used fewer days in the intensive care unit (ICU), had a shorter average LOS, and had lower average ancillary costs for tests and drugs (Aiken, Sloane, Lake, Sochalski, & Weber, 1999
). The Kane et al. (2007)
meta-analysis estimated that if the association between nurse staffing and improved patient outcomes found in so many studies was causal (something difficult to confirm without unlikely clinical trials), an additional RN per patient day would avoid 7 cases of infected wounds and 4 cases of nosocomial sepsis per 1,000 hospitalized surgical patients, in addition to mortality prevention; in ICUs, an increase by 1 RN-FTE per patient day would avoid 7 cases of hospital-acquired pneumonia, 7 cases of respiratory failure, 6 cases of unplanned extubation, and 2 cases of cardiac arrest per 1,000 patients. Rothberg, Abraham, Lindenaur, and Rose (2005)
demonstrated that the cost of saving a life through investments in nurse staffing was in line with the costs of saving a life through commonly accepted medical care practices such as thrombolytic therapy for acute myocardial infarction and routine cervical cancer screening.
Hospitals that have more favorable nurse staffing and work environments have not only better clinical outcomes that should offset the costs of increased staffing but also better nurse retention, another offsetting cost savings (Aiken, Clarke, Sloane, Lake, & Cheney, 2008
). Reduced nurse turnover produces substantial savings to hospitals when all the costs of replacing nurses are considered, including recruitment, overtime, and use of supplement agency nurses (Jones, 2004
). Higher job dissatisfaction and nurse burnout, both precursors to voluntary turnover, are associated with less favorable nurse staffing and poor work environments (Aiken et al., 2008
). Presumably one of the reasons why a record number of hospitals have pursued Magnet designation is the evidence that nurse recruitment and retention is better in Magnet hospitals (Aiken, 2002
A less studied but vitally important dimension of cost offsets associated with investments in nursing are the interrelationships between types, costs, and outcomes of nursing investments. Hospital care is labor intensive and thus there is long-standing interest in the potential for substituting less expensive labor for more expensive personnel, specifically substituting less expensive personnel for professional nurses. In the health services research literature there are many studies on nursing skill mix. Most of these suggest that better outcomes are associated with a higher proportion of RNs (Estabrooks et al., 2005
; Landon et al, 2006
; McCloskey & Diers, 2005
; Person, Allison, Keife, Weaver, & Williams, 2004
). Needleman and colleagues used a particularly compelling cost offset approach to evaluate the use of LPNs in hospitals. They concluded that contrary to popular opinion, both lives and money could be saved by replacing LPNs in hospitals with RNs (Needleman, Buerhaus, Stewart, Zelevinsky, and Mattke, 2006
Another strategy for determining nursing cost offsets has been pursued by Aiken and associates to evaluate the relative impact of different types of investments in RNs. They have evaluated sequentially the association of nurse staffing, nurses’ education, and the nurse work environment on patient outcomes, considering each individually and collectively. Their research first showed that each additional patient added to the workload of a hospital staff nurse was associated with a 7% increase in both mortality and failure to rescue following common surgical procedures (Aiken, Clarke, Sloane, Sochalski, et al., 2002
). The relationship between workload and mortality was also tested in the International Hospital Outcomes Study led by the University of Pennsylvania. The association between staffing and mortality was found to be similar in English hospitals (Rafferty et al. 2007
), and a nursing skill mix association was found in Canada (Estabrooks et al., 2005
) and in New Zealand (McCloskey & Diers, 2005
). Then the Aiken group established that, independently of nurse workloads, each 10% increase in the proportion of hospital staff nurses with a baccalaureate or higher degree was associated with a 5% decline in mortality (Aiken et al., 2003
). A similar finding was documented in Canada by Estabrooks et al. (2005)
and Tourangeau (2007)
A 2003 study by Aiken and colleagues also reported on the joint effects of nurse staffing and education on mortality, revealing an interesting example of an offset that is illustrated in . The lowest mortality was found in hospitals where nurses on average cared for 4 patients each and 60% had a baccalaureate degree. The highest mortality was found in hospitals where nurses cared for an average of 8 patients each but only 20% had baccalaureate degrees. The offset is suggested by the two middle categories where the mortality rate is the same, although not the best, in two widely different staffing scenarios. The mortality is the same in hospitals where nurses care for an average of 8 patients each but at least 60% have a baccalaureate degree and hospitals where nurses care for 4 patients each but only 20% or less have a baccalaureate degree.
Deaths per 1,000 General Surgery Patients in Hospitals With Differing Workloads and Percentages of Nurses With Bachelor of Science in Nursing Degrees (BSNs)
This finding has been difficult for many staff nurses to envision because of a tendency to think of education as an individual characteristic of a nurse. A hospital staff nurse with a baccalaureate degree usually has difficulty imagining caring effectively for 8 patients. However, the findings reported above consider nurses’ education as a characteristic of a hospital; education is measured some-what like nursing skill mix at the institutional level. It is presumed that management decisions can impact the proportion of nurses in a hospital with baccalaureate education through selective recruitment, retention, and support for existing staff to obtain higher education. Care in an environment where the overall educational levels of nurses is higher may afford patients with advantages that would require more staff to produce in institutions where the nurse workforce is less educated. Although the causal links need further study, the findings raise the potential that a smaller RN workforce might be possible in the future if the overall educational levels were higher.
Currently only 43% of staff nurses have baccalaureate degrees, which means there are not enough nurses for all hospitals to have at least 60% with at least a baccalaureate degree. More than 60% of new nurses now graduate from associate degree programs and relatively small proportions appear to be going on to obtain their bachelor of science in nursing degrees (Bevill, Cleary, Lacey, & Nooney, 2007
). If subsequent research confirms the findings of an offset of higher nurse education on hospital nurse staffing requirements, efforts to rebalance the educational mix of the nurse workforce could potentially have a significant impact on the outcomes and cost of the hospital workforce.
A similar kind of offset can be hypothesized about the nurse work environment and nurse staffing requirements. The literature is replete with examples of operational failures in the hospital work environment that undermine the efficiency and safety of nursing care (Page, 2004
; Tucker & Spears, 2006
). Recent research at the University of Pennsylvania shows that the nurse practice environment has a significant effect on patient mortality and failure to rescue, independently of nurse staffing and nurse education (Aiken et al., in press; Friese, Lake, Aiken, Silber, & Sochalski, 2008
). The approach by Aiken’s group, like other investigators, has been to test additive models examining the level of incremental improvement in patient outcomes that can be achieved by additional investments in nursing. Results of this body of literature suggest that investments in improved nurse staffing, nurse education, and work environments all contribute individually and collectively to improved patient outcomes. Further exploration of these findings suggests that there might be an interaction between the effects of nurses’ work environments and nurse staffing on patient mortality as illustrated in .
Interaction Between Hospital Nurse Staffing and Work Environment on Surgical Mortality: Exploratory Findings
The results presented in are exploratory and require further study. They use data from the 1999 Pennsylvania hospital study (Aiken, Clarke, Sloane, Sochalski, et al., 2002
) and models that hypothesize an interaction effect of staffing and the practice environment. We find in this exploratory work that mortality rates after common surgical procedures are the same in hospitals with a range of nurse staffing levels between 4 patients per nurse to 8 patients per nurse as long as the quality of the practice environment is in the best category. In contrast, there are wide differences in mortality by different nurse staffing ratios in hospitals with poor nurse practice environments. Nurse education in models tested here is assumed, for illustrative purposes, to have an additive effect.
The point of showing these exploratory findings (which will be followed up in larger studies) is to illustrate that there is some beginning empirical evidence that there are offsets between various forms of investments in nursing. Some types of investments are likely to result in better value than others. We need the science base not only to make the case that investments in nursing yield better outcomes for patients and cost savings to institutions but also to guide decisions about the relative value of different kinds of investments in nursing.