Nursing Turnover and Hospital Efficiency: An Organization Level Analysis

This study tests the competing arguments that organizational turnover rates are positively associated with organizational inefficiency or, alternatively, that turnover rates are positively related to organizational inefficiency only in those organizations experiencing very high or very low rates of turnover. The findings strongly support the former argument: in a national sample of 407 hospitals, turnover among registered nurses was found to be positively and linearly associated with both operating and personnel costs per adjusted admission. However, subset analyses based on hospital size, location, and teaching status, suggest that the strength of the turnover-cost relationship is contingent upon the type of institution in which turnover occurs. The majority of turnover studies, whether conducted in the health care sector or industrial sector, explicitly identify turnover as a dependent variable to be explained. Much of this research is predicated on a set of assumptions and beliefs among scholars and managers that employee turnover adversely influences effectiveness and productivity in organizations (Bluedorn 1982, Dalton and Todor 1979, Muchinsky and Turtle 1979). For example, effectiveness, or the attainment of organizational goals, may be hindered when turnover decreases job performance and familiarity with standard operating procedures. Similarly productivity, the ratio of outputs to inputs, is thought to decrease under conditions of high turnover as costs of recruiting and training new employees increase, while outputs are reduced during orientation and familiarization with the new job. Whereas most of these assumptions regarding the adverse consequences of turnover have been inviolate for years, recent writings have begun to question some of the underlying premises behind the notion that turnover is inherently bad for the organization (Dalton et al. 1979,1981, Muchinsky and Morrow 1980, Staw 1980, Staw and Oldham 1978). For example, Dalton and Todor (1979) have advanced a convincing thesis that some degree of turnover tends to be healthy for the organization. Turnover at moderate levels infuses "new blood", fresh ideas, and keeps the organization from becoming stagnant. Similarly, traditional assumptions about turnover and its relationship to cost and productivity have been questioned. Efficiencies may result from replacing full-time employees who leave the organization with part-time personnel or with entry level personnel, thereby reducing personnel and benefits costs. Although there has been recent consideration of turnover and performance at the individual level (see Dreher 1982, McEvay and Casico, Sheridan 1985), there have been almost no empirical studies of the organizational consequences of turnover that would either reinforce traditional conceptions of turnover as a negative attribute of organizational behavior, or affirm more recent notions of turnover as leading to positive consequences for the organization. In this light, the goal of the current study is to examine the relationship between voluntary turnover rates of hospital registered nurses (RNs) and hospital operating efficiency. Our focus on RN turnover stems from several considerations. First, personnel costs, on average, make up sixty percent of the operating budget of most hospitals. By far, the largest component of this cost is that associated with registered nurses. Second, unlike many occupational groups in hospitals, nurses directly participate in the core technology of hospitals — the delivery of hands-on patient care. Thus, RNs are central to the production process and are both involved and accountable for productivity as it pertains to patient care (Fottler et al. 1988). The study tests the competing arguments that 1) higher voluntary turnover rates among RNs are inversely associated with hospital efficiency or, 2) voluntary turnover rates and efficiency are related in curvilinear fashion owing to potential beneficial effects of turnover at moderate levels. The investigation will also attempt to differentiate whether efficiencies/inefficiencies incurred from turnover are associated with personnel expenditures or operational (productivity) Expenses. By distinguishing between these two domains of activity, we will be able to shed more precise light on how nursing turnover impacts organizational operations. Finally, the study will develop and test arguments related to the organizational conditions under which nursing turnover affects hospital efficiency. This phase of the investigation will examine the moderating effects of organizational size and technological complexity on the turnover-efficiency relationship. On a practical level, findings on the relationship between organizational rates of turnover and organizational efficiency contribute to management's ability to identify optimal levels of turnover rates so that better decisions about human resource utilization can be made (Dalton and Tudor 1982, Terborg and Lee 1984, Abelson and Basinger 1984). Of particular importance in such considerations are the costs associated with various turnover rates relative to those associated with different strategies for reducing turnover. Results of the current investigation should therefore provide hospital decision makers with more concrete information on the operational consequences of turnover so as to better design, fund, and implement appropriate intervention strategies to prevent RN exit from hospitals. Conversely, if RN turnover is found not to be related to hospital efficiency, or if costs of interventions exceed those of turnover, hospital administrators will have evidence to support continued use of supply side solutions to address nursing turnover in their institutions. THEORY AND HYPOTHESES Levels of Analysis Researchers have tended to focus on turnover either at the individual level, following the tradition of industrial psychology, or from the economic perspective of aggregating turnover rates over groups of organizations or industries (Johnson and Vaughn 1982, Long 1951, Mirvis 1977, Cawsey and Wedley 1979). The few studies that have examined nursing turnover and its consequences have had a common focus on the cost associated with the departure and replacement of the individual nurse. Dane (1972), in a comprehensive study, found that RN replacement costs were between 438 and 830 dollars for orientation and training. These costs were derived from breaking down the average cost related to the steps necessary to obtain a ^replacement. These included initiation of need, interviewing of applicants, physical examinations) orientation and training, the non-productivity of new employees during the orientation and training period, paperwork processing, and overtime required from other employees to take up slack caused by turnover. In a less extensive study, Tuchi and Carr (1971) estimated that replacing either an LPN or LVN would cost $1,335 and $1,133 respectively. More recent writings on the subject have been based largely on anecdotal or case study information and suggest that costs of recruiting and orienting a professional nurse to a hospital may range from $3,000 to $5,000 (Donovan 1980; Weisman et al. 1979,1980). Despite increasing emphasis in the theoretical literature, there has been little empirical consideration of the relationship of turnover rates at the organizational level and organizational operating efficiency in the aggregate (Baysinger and Mobley 1983, Bluedorn 1982, Mobley 1982). Conceptualizing turnover and its consequences at the organization level places the emphasis not on individual employee decisions, but on turnover and turnover effects as properties of the organization itself. From a theoretical perspective, the issue of level of analysis is central because functional relationships involving turnover may differ between individual and organizational levels (Terborg and Lee 1984, Wallace 1983). For example, if individual turnover is associated with increased cost of recruitment, training, and reduced productivity, it does not necessarily follow that turnover rates at the organization level will be positively correlated with personnel or production costs for the organization as a whole. Indeed, progress in personnel and industrial relations will be limited until researchers are able to establish vertical synthesis across levels of analysis (Heneman 1969, Terborg and Lee 1984). From a managerial perspective, the analysis of turnover as an organizational attribute has a number of advantages. Conceptualizing and measuring turnover at the organization level shifts the explanation of the consequences of turnover from the individual to the organization and thus opens the way for administrative intervention through changes in organizational design and staffing arrangements (Pfeffer and O'Reilly 1987, Scott and Shortell 1988). It is also one aspect of human resources management that is commonly monitored for both intervention and for making personnel policy projections. Finally, the informatiaa.ner.essary.ta study turnover and its •-• consequences at the organization level of analysis does not require a measurement of perceptions, thoughts, and feelings of people which may be less reliable than recording the organizational and environmental context in which employee decision making occurs (Pfeffer and O'Reilly 1987). Turnover Rates and Organizational Performance If we accept the theoretical and applied utility of examining turnover and its consequences at the organization level, what processes might account for such relationships? Arguments regarding the relationship of organizational turnover rates and organizational performance are based frequently on the premise that turnover disrupts the input/throughput/output cycle of organizational production and thereby reduces efficiency (Gouldner 1957, Grusky 1963, Staw 1980). Under the open systems perspective, organizations obtain energy from the environment to transform inputs into outputs. These outputs are introduced into the environment thereby stimulating further energy exchange. A primary organizational function is to perpetuate this process or, alternatively, to prevent entropy or dissipation of energy exchange. Countering the entropic process in social systems is achieved not only by sustained acquisition of energy from the external environment, but also from the ability to maintain the structures necessary to effectively stabilize the exchange process. Social systems are structured so as to create a "unity in their completion or closure" and a balance between energy intake and organizational activities (Katz and Kahn 1966). The ability of these social system structures to maintain stability or constancy in energy exchange is central to the notion of organizational homeostasis or equilibrium. Maintenance structures that focus on ensuring stability or predictability in exchange relationships are built on a common set of norms. New entrants, therefore, must be socialized into the norms of the organization. High turnover rates cause organizations to potentially expend more energy in maintaining the input-throughput-output process than they take in from the environment. Turnover is likely to prevent new structures from emerging that have any degree of permanence and at very high rates, it becomes increasingly difficult to counteract such disruption through the organization's ma iniensiiee mechanisms (Katz and Xahn 1966, Staw 1980). • •A second but related theoretical argument regarding the relationship of turnover rates and performance is based on the notion that turnover is part of a more generic problem of organizational control (Price 1977). To be effective, organizations must maintain members willing and able to perform the work necessary to produce the output of the organization. High rates of turnover affect the basis of organizational control by eliminating the normative foundation on which much control is exercised. "Norms specifying work to be done cannot be obeyed unless an organization can maintain its members. The greater the expenditures of scarce resources to maintain its membership, the more an organization must neglect the work necessary to produce its basic output" (Price and Mueller 1981). Inability to maintain norms is a control problem facing all organizations. However, high rates of turnover exacerbate the problem and result in diversion of resources from basic production into controlling membership, a process counterproductive to organizational effectiveness (Hage and Aiken 1974). The logic of these arguments suggest that there is a positive, linear relationship between turnover rates and disruption and control problems in organizations: the greater the turnover, the greater the disruption and control difficulties. Because of the negative relationship between disruption, control problems and efficiency, the implication is that turnover has an indirect, linear and negative impact on efficiency: the greater the turnover, the lower the efficiency since disruption and control problems should decrease efficiency. Applying these theoretical arguments to the hospital, we would expect general organizational inefficiency to increase with increasing rates of turnover among hospital RN staff. Thus, Hypothesis 1: RN turnover rates in hospitals are positively associated with hospital operating inefficiencies. A competing argument suggests that the relationship between turnover and organizational productivity is not positive and linear, but curvilinear. Such claims are based on the premise that at certain levels, turnover in organizations has positive consequences for organizational performance. Serious discussions of positive outcomes of turnover have only recently begun to , , appear in the literature (Price 1977; Muchinsky and Tuttie 1979; Staw. 1980;-BIuedom 1980). One of the positive organizational consequences of turnover relates to its potential impact on personnel and fringe benefit costs. Higher levels of turnover in organizations may produce lower payroll and fringe benefit costs because the rates of pay for new hires are often substantially lower than rates for experienced employees (Jeswald 1974). As employees exit the organization in large numbers, management may replace them with new hires at entry level wages. Similarly, eligibility for some fringe benefits does not occur until seniority is established. Thus, replacing leavers with entry level personnel affords organizations savings on insurance premiums, vacation pay, sick pay and other related benefits. With respect to productivity, the long-held assumption that turnover and organizational productivity are negatively related is now being questioned by a number of writers (Dalton and Todor 1979, Staw, 1980). If organizations replace departing employees with new arrivals who are more highly motivated and who possess better job skills than their longer-tenured, exiting counterparts, productivity may, in fact, increase. Further, in certain types of organizations where physical or psychological demands are high on employees (e.g., hospital nursing), low levels of turnover and long lengths of service can be detrimental to organizational productivity as employees become more dysfunctional over time (Muchinsky and Morrow, 1980). In the hospital setting, specifically, high nursing turnover may make it easier for the hospital administration to introduce cost effective changes since traditional or long standing operating procedures are eroded by the movement of employees into and out of hospitals. Such arguments gain additional credibility in environments characterized by rapid change and uncertainty, which require responsive adaptation on the part of hospitals (Rakich et al. 1977, Mobley 1982, Seybolt 1984). Although the few empirical studies conducted on turnover and its consequences for organizational performance are far from conclusive (Grusky 1963, Wells and Pelz 1966, Eitzen and Yetman 1972, Leviatan 1978, Allen et al. 1979), the relationship most consistently supported by these and related studies is an inverted U-shaped one in which organizational performance is highest at intermediate levels of turnover. Extending these theoretical and empirical studies to hospitals, we hypothesize that: Hypothesis 2: Relative to hospitals with high or low turnover rates among staff registered nurses, hospitals with moderate rates of turnover will experience greater operating efficiencies. To this point our discussion of turnover rates and organizational efficiency has been general. However, organizations are not homogeneous and may vary both in their susceptibility to turnover and in the manner in which turnover impacts efficiency (Bluedorn 1982). Specifically, we expect the relationship of organizational efficiency and turnover rates to be moderated by certain contextual attributes of organizations. We anticipate that the effects of turnover rate on organizational efficiency will be more pronounced in those institutions where organizational structure and technologies are complex. Bluedorn (1982), for example, has argued that organizations operating with non-routine technologies require more decisions based on experience, judgement and intuition, all of which require lengthy learning periods. Routine technologies, on the other hand, are characterized by few exceptions in the work process as well as by tasks that are highly analyzable. These depend to a lesser extent on organizationor task-specific learning. Thus, relative to those organizations possessing more routine technologies, turnover is likely to be more disruptive in organizations possessing non-routine technologies since the learning periods associated with effectively operating in such contexts are longer. In complex, technologically advanced hospitals, RN turnover will result in higher costs since it becomes relatively more difficult to integrate new nurses into technically demanding jobs and to familiarize them with the involved operational and administrative protocols characteristic of complex institutions. The departure of experienced nurses from such organizations will have a relatively larger impact on productivity since they are less "substitutable" than RNs working in less complex hospitals. That is, it becomes more expensive for hospitals to replace a highly trained and experienced nurse in more technologically advanced hospitals. For purposes of this analysis we make the assumption that teaching hospitals have more complex patient care technologies than non-teaching hospitals. "This assumption is based on the premise that teaching hospitals treat a sicker group of patients that demand more intensive, complex medical interventions and more intensive nursing care. Teaching hospitals are also more likely to engage in untried or "experimental" medical care procedures that require much more discretionary judgement on the part of health care professionals working in these institutions (Daft and Becker 1978). Hence, Hypothesis 3: The association between nursing turnover rates and hospital operating efficiency will be stronger in teaching hospitals than in non-teaching hospitals. A second context variable that may potentially moderate the turnover-organizational efficiency relationship is size. Students of organizational adaptation suggest that smaller organizations possess less slack and are thus less capable of absorbing the negative effects of external or internal disruptions such as turnover on their operations (Bluedorn 1982). Similarly, smaller sized organizations are likely to feel a greater impact of turnover because the effects of exits are magnified by the reduced scale of the organization. That is, turnover is likely to be more salient (disruptive or positive) when played out against a smaller organizational context than a larger one because staff members in smaller organizations make a disproportionately higher contribution to the production function of the organization. Conversely, as organizations grow and become more differentiated, they are better able to absorb the impact of turnover since individual staff members are relatively less important in terms of their contribution to overall organizational productivity. Thus, we expect smaller hospitals to feel the impact of turnover to a greater degree than larger hospitals. Hypothesis 4: The association between nursing turnover rates and hospital operating efficiency will be stronger in smaller hospitals than in larger hospitals. Clearly, organizational efficiency is affected by a number of variables besides turnover. Research on hospital efficiency, specifically, has been extensive and grounded primarily in the discipline of health economics (Davis 1974, Sloan and Steinwald 1980), This research has focused on modeling hospital cost functions, including the inputs, outputs and contextual factors that influence efficiency and costs. These studies have emphasized the precarious nature of hospital operations resulting particularly from exogenous factors related to multiple environments (markets). Less emphasis in these cost function models has been placed on the internal organizational characteristics of the hospital, particularly human resource characteristics such as employee turnover. To isolate the relationship between organizational efficiency and voluntary turnover rates, characteristics of the organization and its environment are included in the model as control variables. By also including in the model other potentially confounding factors such as wage rates, level of competition, product mix and organizational technology, we can estimate the effect of turnover rates on organizational costs "holding" constant these other factors. For example, hospitals treating a "sicker" mix of patients are likely to incur greater costs and thus may appear less efficient. To the extent that casemix severity is also related to RN turnover (e.g. nurses treating sicker patients "burn out" more easily), it may account for all or part of any empirical association between voluntary turnover rate and efficiency. In the regression model, the association of turnover and operating efficiency will net out any effect of casemix. Clearly, any number of variables may potentially affect hospital efficiency. To guide our selection, we used two criteria: 1) support in the health economics literature for an association with hospital cost and 2) potential as a competing or alternative explanation of the RN turnoverhospital cost relationship. For example, geographic region has been demonstrated to have an association with costs (e.g. hospitals in the Northeast experience higher cost than those in the South) and with nursing turnover (e.g. higher in the West). It is important to emphasize that our model is not intended to represent a hospital cost function-that is to include all demand and supply factor variables related to cost. Our purpose is restricted to examining the relationship of RN turnover rates and hospital costs per unit of output, while controlling for those hospital and environmental variables that represent potential alternative explanations of this relationship. Our control variables were divided into two categories: i) organizational characteristics, and 2) environmental characteristics. The organizational control variables included organizational size, ownership/control (e.g. for-profit, government), teaching status, operating capacity, and input complexity/uncertainty. The environmental controls included geographic region, urban/rural status, regulatory intensity by state, local economic climate, area wage rates, competition (organizational), and competition (labor supply). These variables are discussed in detail under measurement. METHOD Data Sources Multiple sources of data were used in this study. The primary data set, which defined the study sample of American hospitals was the Nurse Personnel Survey, conducted by the American Hospital Association (AHA) in 1981. This survey gathered aggregate (hospital level) information about vacancies and turnover among hospital nursing personnel. The Nursing Personnel Survey