Nurse turnover is a complex phenomenon that is of particular concern for healthcare administrators during times of shortage. Shortages limit the pool of nurses available to fill vacancies in healthcare organizations (HCOs), but at the same time offer nurses opportunities to change jobs in search of better working conditions or increased compensation. Intuitively, one would expect nurse turnover to be costly; however, because so many of the costs related to turnover are hidden or difficult to estimate, there tends to be uncertainty about how much turnover actually costs or how important it really is. Little recent research on nurse turnover costs is available, and much of the research on nurse turnover has focused on the causes, not the costs, of nurse turnover.
This is the second part of a 2-part series examining the costs of nurse turnover in HCOs. Part 1 discussed nurse turnover within the context of human capital theory and, using human resource accounting measures, illustrated the replacement costs associated with turnover. The Nursing Turnover Cost Calculation Methodology (NTCCM)1 was described, including recent efforts to evaluate its continuing relevance, identify and validate new and emerging nurse turnover cost categories, and update the costs and calculation methods.
The Original Study
The NTCCM and the methods upon which this study is based were previously developed and tested in a retrospective, descriptive study to estimate the costs of nurse turnover at 4 acute care hospitals.1,2 In the original study, hospital-level data were gathered from chief nurse executives and other designated hospital representatives by using the NTCCM to determine registered nurse (RN) turnover rates and associated costs incurred during the 1988 fiscal year (FY).
The FY 1988 study found that the 4 hospitals tracked RN turnover information using data on hand, such as the number of nurses who had left and the total number of nurses employed, to calculate RN turnover rates.2 Even then, calculation of turnover rates varied between departments within the same organization and between organizations. Employing a common method for calculating RN turnover rates across the 4 hospitals revealed an average annual RN turnover rate of almost 27% for the hospital sample. The overall cost of RN turnover averaged almost $1 million per hospital for FY 1988, and the mean annual cost per RN turnover was estimated to be $10,198. Interestingly, even though none of the hospitals studied actually had quantified or estimated the costs of RN turnover, all of the nurse administrators confirmed the importance of this information for decision making.
Changes in the healthcare environment and in nursing since this original work suggest that methods used in the past to quantify the costs of RN turnover may need to be modified for use today. Further, the application of human capital theory in the current study will advance our understanding of nurse turnover.
Methods and Sample for the Current Study
A retrospective, descriptive design was used to gather FY 2002 nurse turnover cost data at an acute care hospital with more than 600 beds. Data were gathered at 2 levels: service-level data were gathered from 3 clinical nursing directors; hospital-level data were gathered from the associate chief nurse. Gathering data at these 2 levels was deemed an advantage over the original study methods because the associate chief nurse had access to most of the departmental-wide data needed to estimate turnover costs, and service line administrators had knowledge of nursing employment and budget data closer to the point of service. Data were also gathered from other departments and individuals within the organization, as recommended by the 4 nurse administrators.
Nurse turnover in this study represented those RNs who terminated employment—or turned over externally—at the study hospital during the study period. The rationale for including only nurses who turned over externally is that any nurse who turned over internally had to be replaced, he or she likely would have filled an existing vacancy, for which turnover costs were being incurred. Also, internal turnover data are difficult to obtain because some hospitals and HCOs do not adequately capture these data. Both voluntary and involuntary nurse turnovers were included in study calculations.
The service lines represented by the clinical nursing directors participating in this study—surgery, women's, and children's—provided care for 377 inpatient beds (approximately 55% of the hospital's inpatient beds) on 16 inpatient care units. These 3 service lines employed 493 nurses in approximately 419 full-time equivalent (FTE) positions, which represented about 45% of the hospital RN FTEs. In these areas, 96 nurses turned over (∼47% of hospital nurse turnovers), and 178 nurses were hired (∼43% of all hospital nursing hires) during the FY. These 3 areas also had an average of 51 vacant RN positions (approximately 21% of hospital nursing vacancies) open during the FY; these vacancies included nurse turnover–related vacancies plus RN vacancies that existed at the beginning of the study period, and new RN positions created during the FY.
To calculate nurse turnover rates, data were gathered on the number of RNs who turned over during the FY and the number of RNs employed in the service lines at the beginning and end of FY 2002. Turnover rates were then calculated as,1
Equation (Uncited)Image Tools
Using this method, the mean turnover rate for the 3 services lines studied was found to be 19.4%, compared to a turnover rate of approximately 18.5% hospital-wide.
NTCCM Turnover Cost Calculations
After gathering FY 2002 nurse turnover cost data via the data collection tool and interviews with respective nurse administrators, data were entered into the spreadsheet model. This spreadsheet contains several linked worksheets and numerous calculations that are required to determine the total and per RN turnover costs. Costs were separated into 2 main categories, prehire and posthire, using the updated NTCCM described in Part 1. Details on how turnover costs were calculated are contained in Table 1.
Some nurse turnover cost categories were adjusted for non–turnover-related vacancies. This was done to account for (1) RN vacancies that existed prior to the beginning of the study period and (2) RN vacancies that arose from newly created positions during the study period. These cost adjustments were necessary because hospitals incur these costs to fill all RN vacancies, and not just those attributable to turnover during the study period. The cost categories that were adjusted to account for these non–turnover-related vacancies were advertising and recruiting; vacancy; hiring; orientation and training; and newly hired RN productivity. It was not necessary to adjust the remaining cost categories (preturnover productivity and termination) because the NTCCM calculates these costs directly on a per RN turnover basis.
The per RN and total turnover costs for each of the 7 NTCCM cost categories are presented in Tables 2 and 3, respectively. As shown here, per RN turnover cost estimated in this study ranged from approximately $62,100 to $67,100, and the total nurse turnover cost ranged from approximately $5.9 million to $6.4 million for the 3 service lines studied. Relative proportions for each of the NTCCM cost categories are included in Table 2, and indicate that, on average, 4 cost categories—vacancy, orientation and training, newly hired RN productivity, and advertising and recruiting costs—accounted for more than 90% of the total and per RN costs of turnover. Table 4 shows this was also the case with the original NTCCM, when it was applied at different hospitals 14 years ago.1 In both studies, hiring and termination costs represent a relatively small proportion (<5%) of overall nurse turnover costs.
Vacancy costs, or the costs incurred while attempting to replace nurses who leave, by far, represent the single largest category of nurse turnover costs. The costs in this category included those incurred for hiring temporary (and traveling) RN staff, paying overtime to employed RN staff, closing beds, patient deferrals, programs developed for addressing vacancies, and productivity losses of permanent staff, all in an effort to fill RN turnover vacancies in the short run. This category, previously labeled “unfilled” positions' in the original study, also represented the single largest category of nurse turnover costs in the original application of the NTCCM.1 However, the proportion represented by vacancy costs increased from approximately 35% in the original study to 75% reported here. Several reasons may explain this increase. First, there was greater reliance on the use of temporary nurses to fill nurse turnover vacancies in the hospital in 2002 than at the hospitals in the original study. Second, the revised NTCCM included additional costs that were believed to be influenced by RN vacancies—patient deferral costs, productivity of permanent staff working overtime or with temporary nurses, the losses from simply having vacant positions, and the adoption of new staffing plans to fill vacancies—that were not included in the original study. Third, data for the 2 studies were gathered from 2 completely different hospital samples, in different geographic locations, focused on different levels of analyses, and likely reflect differences in organizational missions, models of care, and other important aspects of nursing care delivery.
The orientation and training cost category was the second highest cost category in both 2002 and 1988. After removing the largest cost category, vacancy costs, orientation and training costs accounted for approximately one third of the remaining RN turnover costs in both 1988 and 2002. In FY 2002, preceptors and dedicated training staff accounted for more than 90% of orientation and training costs at the study site, while equipment and space accounted for the balance of these costs.
Newly hired RN productivity costs ranked third in the present study and fourth in the 1988 study. In the original study, these productivity costs were determined for all newly hired nurses in the aggregate. In the current study, however, newly hired RN productivity costs were determined separately for new RNs (ie, recent graduates and/or those with little or no nursing experience) and experienced RNs. To determine these costs, learning curves were estimated on the basis of the time it takes newly hired RNs—both new RNs and experienced RNs—to reach a productivity level equivalent to 90% of incumbent staff members; this period of time was then divided into thirds and productivity percentages were estimated for each third.
The 3 clinical nursing directors were asked to estimate the length of time required for newly hired RNs to reach 90% productivity. They reported that it took an average of 14 weeks for new RNs—considered by these clinical directors as nurses with less than 1 year of experience—to reach 90% productivity. These 14 weeks were divided into thirds (approximately 4.7 weeks in each third), and at the end of each third, new RNs were estimated to be, on average, 30%, 50%, and 90% effective. On the other hand, newly hired but experienced nurses required an average of only 6 weeks to reach 90% productivity, and it was estimated that at the end of each of the three 2-week periods experienced RNs were 75%, 85%, and 90% productive. The resulting learning curves estimated for both new and experienced RNs are shown in Figure 1. Salary and productivity estimates in each period were then used to estimate a cost for the newly hired RN learning period. In the present study, the newly hired RN productivity costs for new RNs were 6 times that for experienced RNs. Combined, the orientation and training, and newly hired RNs productivity cost categories represented over one half of the non–vacancy costs in both studies.
Advertising and recruiting costs—more so than any other category—represent obvious costs of nurse turnover for which data tend to be readily available. This cost category was ranked third in the original NTCCM application, and fourth in the current study. Even though, advertising and recruiting costs were still considerable, they appeared to be of less concern than previously determined, and certainly less costly than many other costs of nurse turnover. In FY 2002, media purchases, personnel, and student nurse programs accounted for more than 90% of advertising and recruiting costs.
Hiring costs were ranked fifth in both studies, and represented a relatively small fraction of overall turnover costs in FY 2002. Termination costs were ranked last in both studies, and represented less than 1% of RN turnover costs in FY 2002. Taken together, these 2 cost categories represented approximately one tenth of turnover costs after vacancy costs were removed.
A new cost category, decreased preturnover productivity, was included in the updated NTCCM and was conceptualized to reflect staff productivity changes in the 3 months prior to the occurrence of turnover. Not surprisingly, these and other productivity costs were very difficult to capture. While the costs reported here represent a very small proportion of overall turnover costs, these costs may actually be higher than those reported because of the difficulty in estimating the costs, and given our current state of knowledge and lack of understanding about this preturnover period.
Comparison of Current and Earlier Studies
Nurse turnover category costs in 2002 were generally 1 to 3 times those costs in 1988, with the exception of vacancy costs, which were 12 times as large in 2002. The differences between nurse turnover cost estimates derived from the original NTCCM and those derived from the updated NTCCM may be attributed to several factors. First, data were gathered at different hospital samples for the 2 data-collection periods, separated by 14 years—FY 1988 and FY 2002. Second, inflation occurred in the 14-year period between studies. Indeed, a review of Consumer Price Indices for medical care services and hospital and related services indicates that the cost of medical care services in 2002 was 2.1 times that in 1988 and the cost of hospital and related services in 2002 was 2.5 times that in 1988.
Third, the inclusion of additional costs and cost categories in 2002 and revisions to the NTCCM explain some of the differences observed in costs. For example, the 2002 NTCCM includes costs not included in the 1988 study1 (departing nurse preturnover productivity costs; vacancy costs, such as patient deferrals, implementation of new staffing plans in response to vacancies, and decreased productivity of staff working with temporary nurses and at times of RN vacancies). Approximately 29% of RN turnover costs in 2002 were attributable to cost categories and components that were not included in the original NTCCM. Fourth, there was a remarkable increase in vacancy costs between the FY 1988 data1 and the study reported here, even after adjusting for new vacancy cost components.
To compare FY 19881,2 and FY 2002 results, a common basis was desired: the preturnover productivity costs were removed from the 2002 calculations (there was no similar category in 1988), and vacancy costs were removed from both 1988 and 2002 results because the over 10-fold increase was not observed in any other cost category. Doing so resulted in nonvacancy, per RN turnover costs of approximately $6,100 in 1988 versus an average of $14,900 in 2002. When the 1988 costs were adjusted for inflation (ie, the 1988 costs were multiplied by the inflation factors of 2.1 and 2.5 noted above), the equivalent 2002 costs would ranged from $12,800 to $15,200, which was in line with the 2002 NTCCM calculations. Thus, aside from vacancy costs, these other costs (in aggregate) have increased approximately at the same rate as medical care services and hospital and related services costs. However, vacancy costs—the largest cost category—have increased at 5 to 6 times that indicated by inflation. One reason for this increase may have been the heavy reliance on temporary nurses to fill vacancies and keep beds in service at the 2002 study hospital: temporary nurse costs accounted for almost all of the inflation-adjusted vacancy cost increase between the FY 1988 and FY 2002 data-collection periods. What is unknown, however, is the organizational costs that would have been incurred at the FY 2002 study hospital had beds been closed or alternate staffing strategies been employed.
The nurse turnover costs derived from the NTCCM can also be compared with other nurse turnover cost estimates.3 As would be expected, the costs reported here are much higher than estimates from the 1980s and early 1990s. The per RN costs derived from the updated NTCCM are higher but relatively consistent with 1999 estimates made by the Advisory Board Company.4 It should be noted that the RN turnover costs reported here, unlike in the Advisory Board Company study, were not differentiated by specialty versus general patient care area because both specialty and general patient care areas were represented in all of the service lines participating in this study. Also, nurse executives reported similar difficulties in recruiting new RN hires for both general and specialty units, and in many cases, it was more difficult—and perhaps more costly—to recruit nurses for general versus specialty areas.
The FY 2002 nurse turnover costs reported here are 1.2 to 1.3 times the average salaries for incumbent nurses (including those who turned over). This finding puts RN turnover costs in the middle of the 0.75 to 2.0 rule of thumb approach to estimating turnover costs,5 but greater than the 1.0 times salary6 rule of thumb for turnover cost estimation. In some cases, these rule of thumb estimates may be appropriate; however, additional research is needed to value those human capital and other attributes of departing nurses (eg, education, experience, and type of position held) that might value them at different points inside or outside the range of 0.75 to 2.0 times salary.
The per RN turnover costs determined in this study are also greater than the most recent estimates of nurse turnover costs, most likely due to the inclusion of detailed vacancy costs in this study. For example, the per RN turnover cost reported here is more than double the per RN turnover cost reported by Waldman et al,7 and approximately triple the 2003 costs reported by Stone et al.8
The NTCCM provides a systematic way to better obtain the “true” costs of nurse turnover, and aids in determining baseline nurse turnover costs so that more costly categories can be tracked on an ongoing basis. The NTCCM allows comparisons to be made across cost categories, facilitates the determination of the costly aspects of replacing nurses who turn over, and helps highlight where interventions to retain nurses—and thereby reduce costs—should be targeted. In this study, hidden costs captured by the NTCCM, such as those included in the vacancy, newly hired RN productivity, and preturnover productivity cost categories, accounted for over 25% of turnover costs; thus, these costs are very important to consider when calculating nurse turnover costs.
The NTCCM captured all known significant costs of nurse turnover to the extent that data were available. However, there were some difficulties encountered during data collection that are worth noting. First, the NTCCM requires extensive data collection that crosses multiple departments in a HCO. While much of the data needed for nurse turnover cost calculations were available, completing the NTCCM required (1) gathering some primary data retrospectively for the study period, (2) reformatting some existing data for use in the NTCCM, and, (3) in some cases, estimating certain aspects of turnover costs. For example, because data on productivity changes were not routinely tracked or quantified, the nurse executives and directors participating in data collection estimated productivity changes based on their knowledge of the clinical area and nursing practice within those areas.
In general, the hospital and service lines had much of the information needed to calculate nurse turnover costs, but had not done so. The nurse administrators recognized the value of this information, however, and the act of compiling these existing data brought about an increased interest in nurse turnover costs. It was determined during data collection that some nurse turnover and related cost data were maintained at the hospital level, some were maintained at the nursing division level, and some were maintained at the service line level. Because of the difficulties in attributing specific nurse turnover costs to service lines, the costs reported herein represent combined totals for the service lines participating in this study.
Different departments within the hospital also maintained different categories of data, and some duplication and overlap of data existed between the departments. This duplication, while perhaps inefficient for the organization, was used as a way of verifying nurse turnover and cost data collected. Some data contradictions occurred between departments or units, but this was not surprising given the size of the organization and the fact that different departments may have made different assumptions about nurse turnover and related costs. Because statistics were maintained and analyzed differently across some organizational units, reporting was inconsistent and required reaching consensus on certain data elements to be used in calculations. Overall, these discrepancies were resolved with minimal effort.
This application of the NTCCM and determination of nurse turnover costs is limited in generalizability, given that it represents only 3 service lines within one large, acute care hospital. Also, as is always the case with studies involving data collection in hospitals, the application was limited by the availability and accuracy of data used in nurse turnover calculations. Thus, research that uses the NTCCM in a larger, more broadly representative sample, and takes a prospective and/or more longitudinal view of nurse turnover costs to determine the stability of certain costs over time and account for temporal changes in the nursing workforce pool is needed.
Because little is known about how nurse turnover affects quality of care, patient satisfaction and safety, staff satisfaction and safety, and productivity and organizational performance,9,10 an important line of research is to examine the consequences of nurse turnover for patient, provider, and systems outcomes. Doing so necessitates more accurate estimates of the associated costs and benefits of nurse turnover. For example, nurse turnover may have detrimental effects on quality of care and patient safety in HCOs when nurses and other staff are overworked, when there are vacant positions and a revolving door of staff, when nursing tasks are left undone, or when patient symptoms or care needs are missed. Moreover, when new or temporary staff come into HCOs and are unfamiliar with organizational processes, patient safety may be jeopardized. These detrimental effects may ultimately translate into costs of nurse turnover that were omitted in this version of the NTCCM, but should be included in the future so that the ability to quantify such costs improves.
Using Turnover Costs to Build a Business Case for Nurse Retention
Knowledge of nurse turnover and its costs and/or changes in spending relative to turnover can serve as an indicator for impending changes in an organization's workforce, and as an impetus for taking action to mitigate the potentially negative consequences of turnover. Knowledge of nurse turnover costs can also aid in developing organizational practices aimed at improving working conditions and quality of care, and retaining nurses. Based on the assumption that turnover consumes resources that could be directed elsewhere—ie, toward care delivery and staff retention—nurse turnover cost information can be used by HCOs to make more appropriate resource allocation decisions, set organizational turnover and retention policy, and develop a business case for nurse retention. Building a business case for any organizational problem requires the ability to calculate rates of return on investments, break-even points, costs and benefits of both turnover and retention, and other financial indices.
Calculating these indices, however, requires data on costs and benefits associated with the issue and its resolution. The costs and benefits of nurse turnover and retention are interrelated; that is, the opportunity cost of nurse turnover is retention, and vice versa. Thus, when nurse turnover costs are quantified, the costs can be viewed as a potential cost savings of retention efforts, and can be assumed to be a benefit of nurse retention. In this way, the costs and benefits of turnover and retention are complementary and both are necessary to build a business case for nurse retention.
Turnover is costly.4 However, during times of nursing shortage or under constrained labor market conditions, nursing resources are at a premium. If the costs of turnover are known, administrators in HCOs may be willing to invest in nurse retention activities, which in turn may contribute to increases in nurses' human capital and an organization's social capital. For example, one strategy commonly used to address nurse turnover is increasing wages and/or providing bonuses. While this may or may not be the best long-term retention strategy, it does provide an easy way to think about how turnover costs might be reallocated. If the $6.4 million cost of nurse turnover had been paid out as a one-time retention bonus to the 493 nurses employed in these 3 service lines, each nurse would have received approximately $13,000. What is unknown is whether this bonus alone would have increased nurse retention sufficiently over the long run to justify the expenditure. It is likely that investing in multiple retention efforts—remuneration and otherwise—would yield better returns over the long run.
Turnover and its costs—both to patients and organizations—are critical concerns facing administrators in HCOs today. What is known about nurse turnover is considerable, but there is a lot yet to learn. For example, there is minimal knowledge about the costs and benefits of nurse retention; the effects of turnover and retention on patient satisfaction and safety, patient loyalty and defection, staff satisfaction and safety, and organizational productivity and performance; and the human capital aspects of nurse turnover. As suggested by the Human Capital Theory, increases in human and social capital can be expected to increase individual and organizational productivity and returns on investment. Beyond tangible returns on investment, investments in RN retention may give HCOs a competitive advantage by having a stable, productive, and satisfied nursing workforce, and, in turn, improve consumers' perceptions of workforce quality and increase demand for HCO services.
As administrators in HCOs strive to operate more efficiently and effectively, broader knowledge of factors that affect expenses, revenues, and productivity is needed. This study addressed this need by defining important aspects of nurse turnover costs and providing an updated application of the NTCCM to calculate the costs of nurse turnover. Results of the study provide insight into the true costs of nurse turnover (1.2 to 1.3 times RN salary at the study hospital) and facilitate the evaluation of policies designed to address nurse turnover and its consequences.
A Faculty Research Opportunity Grant from the School of Nursing, UNC – Chapel Hill, provided partial support for this study. The author also gratefully acknowledges the thoughtful comments of colleagues Dr Barbara Mark and Dr Mary Lynn on an earlier version of the article.
© 2005 Lippincott Williams & Wilkins, Inc.