A sufficient supply of registered nurses (RNs) is critical for hospitals to provide high-quality patient care, and hospitals have adopted numerous strategies to respond to shortages. These strategies include instituting mandatory overtime (Hassmiller & Cozine, 2006), tolerating and adapting to chronically high vacancy rates (Rondeau, Williams, & Wagar, 2008), delivering fewer RN hours of care (Blegen, Vaughn, & Vojir, 2008), and using temporary nurses, also referred to as agency, traveling, or floating nurses (Hassmiller & Cozine, 2006; May, Bazzoli, & Gerland, 2006). Temporary nursing staff includes nurses employed both internally by the hospital (per diem and floating nurses) and externally (agency nurses) (Aiken, Xue, Clarke, & Sloane, 2007). In 2001, 56% of hospitals used temporary nurses, including agency, per diem, or traveling nurses (American Hospital Association, 2001), and Aiken et al. (2007) reported that in 2000, nearly 6% of hospital staff nurses were temporary hires. Despite their prevalence, many hospitals have tried to reduce dependence on temporary nurses because of cost and safety concerns (May et al., 2006).
Although several authors have suggested that temporary nurse staffing negatively influences safety and continuity of care, prior research has provided little empirical support for the association between use of temporary nurses and hospital outcomes (Castle, Engberg, & Men, 2008). Furthermore, a more recent study found that temporary nurses were not related to nurse or patient safety and quality problems (Aiken et al., 2007). Therefore, in this research, we examined the relationship between use of temporary nurses and nurse and patient safety outcomes at the nursing unit level.
From previous research in this area, we found opinion articles and descriptive studies of temporary nurse use but few empirical studies examining the association between use of temporary nurses and patient outcomes. Most of these empirical studies have shown a negative association between use of temporary nurses and patient outcomes. For example, the spread of nosocomial infection among patients was found to be associated with greater use of temporary nurses (Alonso-Echanove et al., 2003); needlestick injuries were more prevalent in temporary nurses (Aiken, Sloane, & Klocinski, 1997), and medication errors increased with the use of temporary nurses (Rosemen & Booker, 1995).
In contrast, using the 2000 National Sample Survey of Registered Nurses, Aiken et al. (2007) found that temporary nurses were not less qualified than permanent nurses but were more likely than permanent nurses to have baccalaureate degrees. Using data from a survey conducted in Pennsylvania, they also found that after controlling for adequacy of staffing and resources, the use of temporary nurses was not associated with nurses' and patient safety and quality problems. In some cases, having more temporary nurses was related to better patient outcomes (fewer medication errors). From these findings, Aiken et al. suggested that quality problems in hospitals using more temporary nurses might have resulted from inadequate staffing and resources rather than from hiring temporary nurses. Another current study found a nonlinear relationship between use of agency nurse aide and quality outcomes (Castle et al., 2008). Although low levels of agency nurse aid use had little impact on quality of care, high levels of use of agency nurse aid were associated with poor quality of care.
Therefore, continuing effort to understand the association between use of temporary nurses and nurse and patient safety outcomes is necessary to improve care provided in hospitals. Furthermore, this association needs to be examined at the nursing unit level instead of at the hospital level because nursing staff work in teams, and nursing units provide a proximal work context of the nursing care provided by temporary nurses (Kozlowski, Steve, & Bell, 2003). To investigate the association between use of temporary nurses and nurse and patient safety outcomes at the nursing unit level, we included both nurse safety outcomes (needlesticks and back injuries) and patient safety outcomes (patient falls and medication errors) in this study. To control for alternative explanations for nurse and patient safety outcomes, the analytic model included measures of relevant aspects of the hospital and nursing unit work environment characteristics and selected nurse and patient characteristics.
We were unable to find a conceptual model specific to understanding the relationship between temporary nurse use and outcomes in hospitals. However, Castle (2009) recently developed a conceptual model that illustrates potential reasons that might underlie such a relationship in nursing homes. His model suggests that although use of agency staff might help to increase staffing levels, it might also lead to poor patient quality of care because of influence on other staff, facility operation, and patients. Because temporary nurses and permanent nurses have to work together, use of temporary nurses may lead to decreased teamwork, increased workload, and increased supervision. The use of temporary nurses can be expensive, can increase administrative burden, and can disrupt routines. Regarding influence on patients, temporary nursing staff may interfere with continuity of care because of unfamiliarity with care practice and patients and unstable relationship with patients. Castle et al. (2008) also suggested the possibility of a nonlinear association between agency staffing levels and quality; low levels of temporary nurse use may be a proactive way of maintaining patient quality of care and may not interfere much with care delivery processes, whereas high levels of temporary nurse use may adversely influence outcomes.
We hypothesized a positive association between high levels of temporary nurse use and adverse nurse and patient outcomes. Figure 1 presents the model proposed in this study on the basis of Castle's (2009) model. We tested the direct relationships between use of temporary nurses and patient safety outcomes (illustrated in the solid boxes and solid arrows in Figure 1). The underlying mechanisms of these direct relationships (dashed boxes and dashed arrows in Figure 1), including the impact on other staff (decreased teamwork, increased workload, and increased supervision), facility operations (i.e., increased expense, increased administrative burden, disrupted routines), and patients (i.e., reduced continuity of care), were not tested in this study.
Nurse Safety Outcomes
Needlesticks and nurse back injuries are a relatively common injury with potentially serious outcomes (Aiken et al., 1997; American Nurses Association, 2008; Vieira, Kumar, Coury, & Narayan, 2005). Currently, safety equipment and lifting devices are available to lower risk of percutaneous needlesticks and low back injuries, and some researchers have found that injuries have been reduced with the adoption of these devices (Clarke, Rockett, Sloane, & Aiken, 2002; Tan, Hawk, & Sterling, 2001; Vieira et al., 2005). However, some handling of exposed needles is likely to be a permanent feature of nursing practice. In addition, it is unlikely that technologies or procedural guidelines can entirely eliminate needlesticks resulting from unpredictable movements of disoriented patients (Clarke et al., 2002). Temporary nurses frequently need increased supervision (Deitzer, Wessell, Myles, & Trimble, 1992) and may disrupt routines (Merolle, 1988). Thus, temporary nurses can be distracting to other staff who may be carrying out sensitive nursing functions in which they are handling sharps or trying to lift patients. Thus, permanent nurses as well as temporary nurses may be at risk for these injuries.
Furthermore, permanent staff working on units that depend heavily on temporary nurses may also experience frequent patient reassignment (Castle et al., 2008), and temporary staff may not be able to perform all of the required tasks and may not be familiar with equipment and patients (Manias, Aitken, Peerson, Parker, & Wong, 2003). Therefore, the presence of temporary staff may increase the workload for permanent staff (Amenta, 1977) and the frequent reassignment of staff can increase physical job demands, which in turn lead to increased nurse back injuries. In addition, temporary nurses may not be familiar with the culture of help seeking in their new workplace. Therefore, they may work alone and take unnecessary risks in trying to lift patients. This practice can also increase risks of back injuries of temporary nurses.
Hypothesis 1a: Greater use of temporary nurses will be associated with high levels of nurse needlesticks.
Hypothesis 1b: High levels of temporary nurse use will be associated with high levels of nurse back injuries.
Patient Safety Outcomes
Temporary nurses might provide more labor that allows for more complete falls risk assessments. However, beyond such standardized risk assessments, "knowing the patient," which is defined as knowledge specific to the patient's fall risk and need for help, is considered as a core aspect of fall prevention (Rush et al., 2009). This strategy for fall prevention includes effort not only by individual nurses but also by the entire nursing unit, which is an aggregate of individuals on the unit (Rush et al., 2009). Recent research has shown higher nurse staffing levels to be associated with lower fall rates (Dunton, Gajewski, Taunton, & Moore, 2004; Potter, Barr, McSweeney, & Sledge, 2003; Sovie & Jawad, 2001). However, temporary nurses may not be familiar with practices and patients (Deitzer et al., 1992) and may be less engaged in team work (Bloom, Alexander, & Nuchols, 1997). As a result, nursing units with a high proportion of temporary nurses may be less effective in preventing falls than nursing units with more permanent nurses.
Similarly, temporary nurses' lack of familiarity with the specific types of medications used on a nursing unit and lack of knowledge about unit-specific medication administration policies may pose a risk for medication errors. Roseman and Booker (1995) found that medication errors increased when the number of patient days worked by temporary nurses increased but decreased as overtime worked by permanent nursing staff members increased. On the contrary, Aiken et al. (2007) found that the proportion of temporary nurses was negatively associated with nurse-reported frequency of medication errors in which patients received the wrong medication or dose. This relationship was significant after controlling nurse characteristics and staffing-resource adequacy. Because medication errors result from multiple factors in complex health care systems (Reed, Blegen, & Goode, 1998), researchers have emphasized the importance of technology, policies, and practices related to correct administration of medications as well as interaction among health care providers involved in medication administration (Carlton & Blegen, 2006; Kohn, Corrigan, & Donaldson, 2000). However, when nursing units rely on temporary nurses, permanent staff and mangers may experience frequent nursing staff changes, which may lead to inadequate and insufficient interaction among staff nurses. Furthermore, temporary nurses' lack of knowledge about policies and procedures and lack of sufficient information about a patient's condition may also lead to insufficient interaction among staff members. In such nursing units, medication errors may be more likely to occur.
Hypothesis 2a: High levels of temporary nurse use will be associated with high levels of patient falls.
Hypothesis 2b: High levels of temporary nurse use will be related to high levels of medication errors.
Hospital, Nursing Unit, Nurse, and Patient Characteristics
The importance of the work environment in nurse and patient safety outcomes has been emphasized in literature pertaining to health care safety. In a previous study of temporary nurses (e.g., Aiken et al., 2007), work environment characteristics played an important role in explaining the association between use of temporary nurses and outcomes. Therefore, in this study, hospital and nursing unit characteristics were included to adjust for factors that have been shown to affect nurse and patient safety outcomes (Clarke, 2007).
Hospital characteristics include hospital size, technological sophistication, teaching status, and Magnet certification. Larger hospitals may have better support systems for patient care (Daft, 1992) and thus fewer adverse events. Nurses in teaching hospitals are likely to experience higher patient acuity and greater work complexity (Iezzoni et al., 1990), both of which may contribute to a higher likelihood of adverse events for both patients and nurses. Advanced technological services have been linked to better quality of care (Kuhn, Hartz, Gottlieb, & Rimm, 1991). In addition, previous studies have found that nurses in Magnet-certified hospitals reported fewer adverse events (Aiken et al., 1997).
At the nursing unit level, work complexity, unit size, and availability of support services are considered as work environment conditions that may help explain nurse and patient safety outcomes (Mark et al., 2007). In a nursing unit with greater work complexity and more beds, nurses may more frequently encounter time-sensitive situations. This might influence adherence of nurse staff to safe practices, which in turn may affect work-related nurse and patient safety outcomes. The availability of support services may help to reduce nurse workload, which may contribute to the reduction of nurse injuries and patient safety (Mark et al., 2007).
Certain nurse characteristics may also play a role in nurse injuries. For example, older nurses may be at higher risk for back injuries (Sherehiy, Karwowski, & Marek, 2004). Nurse tenure and education levels and the proportion of total nursing care hours provided by both permanent and temporary RNs have been found to be associated with patient outcomes (Clarke, 2007; Berney, Needleman, & Kovner, 2005; Kane, Shamliyan, Mueller, Duval, & Wilt, 2007).
In terms of patient characteristics, studies indicate that medication errors and falls are more common among elderly patients (Thomas & Brennan, 2000). In addition, those with lower perceived health status are likely to be more acutely ill (Mommersteeg, Denollet, Spertus, & Pedersen, 2009) and therefore more at risk for falls and medication errors.
The current study is a secondary analysis of data from the Outcomes Research in Nursing Administration Project II (ORNA II), a nonexperimental, longitudinal causal modeling study (Mark et al., 2007) that used the nursing unit as the unit of analysis. The ORNA II study was undertaken to investigate the relationship between RN staffing adequacy, work environments, and organizational and patient outcomes. The hospital sample was selected through a SAS computer-generated random selection procedure using the American Hospital Association Survey of Hospitals. Using the inclusion criteria of hospitals (nonfederal, nonpsychiatric, not-for-profit, JCAHO-accredited acute care hospitals with more than 99 beds), 1,401 hospitals were entered into the selection pool. Inclusion criteria for the two nursing units in each hospital were "general" medical-surgical units and medical-surgical specialty units. The final ORNA II sample consisted of 286 medical, surgical, and general medical-surgical nursing units from 143 hospitals throughout the United States. All RNs working on each nursing unit who had worked on the unit more than 3 months were invited to participate in the study, and staff nurses responded to survey questionnaires measuring work complexity and availability of support services in the first of three rounds of data collection. Ten patients who were 18 years or older, able to speak English, and hospitalized on the unit for at least 48 hours were randomly selected from each participating unit to complete a patient survey that contained information on their age and health status. A total of 2,720 patients responded (response rate of 91%). ORNA II data collection began in 2003 and ended in 2004. Because of missing values for the selected study variables, the final dataset for the current study consisted of 277 nursing units from 142 hospitals. The staff nurse response rates were 75% (4,911) at the first round, 58% (3,689) at the second round, and 54% (3,272) at the third round.
Data from each nursing unit were obtained during three rounds of data collection conducted over six consecutive months. Nurse managers in each nursing unit reported the care hours provided by temporary nurses, the incidence of staff nurses' needlesticks and back injuries, and the number of patient falls and medication errors during the 6-month study period. Staff nurses completed a survey to measure work complexity and availability of support services in their work unit at the first month of the data collection period (first round). Hospital, nursing unit, and nurse characteristics were collected at the first round. In the last month of data collection, patients provided information about their age and how they perceived their health.
To measure use of temporary nurses, we used three variables (care hours provided by internal temporary nurses [float and per diem], care hours provided by external temporary nurses [agency RNs], and total temporary care hours provided by internal and external temporary nurses). In previous studies, researchers examined only external temporary nurses such as agency staff (Castle, 2009) or percentages of supplemental/float nurses (Aiken et al., 2007). This study used an advanced approach to measure use of temporary nursing staff by examining the relationship separately between internal temporary nurses/outcomes, external temporary nurses/outcomes, and total temporary nurses/outcomes. This provided us with a unique opportunity to examine whether there might be any differences in the relationships between temporary staff and outcomes on the basis of the specific type of temporary nurses. The care hours provided by temporary nurses were calculated by the total hours of care delivered by temporary RNs divided by the total nursing care hours delivered by RNs, yielding a percentage. The distribution of temporary RN hours was right skewed; therefore, they were grouped into four categories, and the cutoff points from the study of Aiken et al. (2007) were used: zero (reference group), low (i.e., greater than 0% to less than 5%), moderate (5-65%), and high (greater than 15%).
Nurse safety outcomes were measured as the number of incidents (needlesticks and back injuries) reported during the 6 months of data collection. Needlesticks were defined as any break of the skin with a needle or sharp instrument that was used on a patient; back injuries were defined as any musculoskeletal disorder of the back caused or made worse by the physical demands of the work of caring for patients (Mark et al., 2007). Because the number of those incidents per RN within a nursing unit during the 6-month study period was so small, the analysis used the number of needlesticks or back injuries reported per 10 RNs on a unit.
Patient falls, defined as an unplanned descent to the floor, were measured as the rate of patient falls per 1,000 patient days. Patient falls included any kind of patient falls. Medication errors were defined as the wrong dose, the wrong patient, the wrong time, the wrong drug, the wrong route, or an error of omission; these were also measured as the number of incidents per 1,000 patient days. To reduce the problem of underreporting of medication errors, this study used a measure of medication errors that resulted in increased nursing observation, technical monitoring, laboratory and radiographic testing, medical intervention or treatment, or transfer of the patient to another unit (Hofmann & Mark, 2006).
Control variable definitions are displayed in Table 1. Because of limited space, we present here selected control variables requiring additional explanation. Work complexity was measured using a 7-item Likert-type scale from the Perceived Environmental Uncertainty Scale (Salyer, 1996). Cronbach's alpha for the scale in the current study was .85. These items asked about the extent to which (a) nurses perceive a need for more information about their patients, (b) physicians change orders frequently, and (c) frequency of admissions, transfers, and discharges makes it difficult to get the work done. Higher scores indicate a higher level of work complexity. The availability of support services was measured by a check list of 21 items, completed by staff nurses who indicated the availability (not available, inconsistently available, and consistently available) of selected support services (Mark et al., 2007). Examples of support services included computerized order entry, automated medication administration system, and intravenous team service.
Because the nursing unit was the unit of analysis in this study, selected study variables (work complexity and availability of support services) measured at the individual level were aggregated to the nursing unit level. To justify the aggregation of lower level data to higher level units of analysis, we used measures of intraclass correlation coefficient (ICC) and interrater agreement (rwg). ICC(1) assessed how within-group variance contrasted with between-group variance, and rwg was used in the event that observed group variances differed from some theoretically expected random variance. The common threshold for such justification is an rwg value equal to or greater than .70, and a larger ICC(1) is accepted as a greater similarity within group.
The current analysis with a sample size of 277 units had sufficient power to test the proposed model (Cohen, 1988). When the dependent variable is the count of some outcome and its distribution is skewed right, count models are used (Hutchinson & Hotman, 2005). In this study, a Poisson regression model with an adjustment for overdispersion was used for this reason.
The means, standard deviations, and, where appropriate, ICC(1) and rwg for the study variables are presented in Table 1. Results of a Poisson regression with control variables are presented in Tables 2 and 3.
Across the 277 nursing units, 57 (20.58%) nursing units did not use any temporary RNs, and 57 (20.58%), 78 (28.16%), and 85 (30.69%) nursing units used low (greater than 0% to less than 5%), moderate (5-15%), and high (greater than 15%) levels of temporary RNs, respectively. One hundred thirty-nine (50.18%) nursing units did not use external temporary nurses, and 69 (24.91%), 44 (15.88%), and 25 (9.03%) nursing units fell in the low, moderate, and high levels of external temporary nurses, respectively. Seventy-five (27.08%) nursing units did not use internal temporary nurses, and 76 (27.44%), 78 (28.16%), and 48 (17.33%) nursing units used low, moderate, and high levels of internal temporary RNs, respectively. During the study period, on average, nursing units reported 0.29 needlesticks and 0.67 back injuries per 10 nurses. On average, there were 4.05 patient falls and 0.79 medication errors per 1,000 patient days.
The association of nurse safety outcomes with temporary nurse care hours when hospital and nursing unit work conditions and nurse and patient characteristics were controlled is presented in Table 2. Incident rates of more than one indicate a higher likelihood of nurse safety outcomes. As evidenced by Table 2, the greater use of external or internal temporary nurses had no significant association with the reporting of nurse needlesticks.
In terms of back injuries, we found that nurses working on units with 15% or more of total nursing hours provided by external temporary RNs were 1.730 times more likely to have back injuries than nurses on units without using any temporary RNs. In other words, as nurses working on nursing units with no agency RNs reported 0.56 back injuries per 10 nurses on average, those working on nursing units with 15% or more agency RN hours reported 0.97 (1.730 × 0.56) back injuries per 10 nurses during a 6-month study period. That was equal to a 73% increase in nurse back injuries, which was statistically significant (p < .05). Similarly, nurses working on nursing units with 15% or more internal temporary nurses were likely to have 1.548 times (p = .05) as many back injuries as those on nursing units that did not use temporary RNs. However, the relationship between total temporary nurse care hours and back injuries was not statistically significant (incident rate = 1.533, p = .056). In addition, we found that older nurses reported more back injuries (4% increase, p < .05). One interesting finding was that nursing units with a high proportion of RNs with bachelor's degrees or higher reported more incidents of back injury than nurses employed on units with a lower proportion of nurses with bachelor's degrees (142.6% increase, p < .05).
Table 3 presents the association of patient safety outcomes (patient falls and medication errors) with temporary nurse care hours after controlling for hospital and nursing unit work characteristics and nurse and patient characteristics. Nurses in nursing units providing 15% or more direct care hours by temporary RNs reported an 18.8% increase in patient falls compared with those nursing units that did not use temporary RNs; the difference was statistically significant. As expected, the incidence of patient falls was negatively related to average patient health status on the unit. Patients on nursing units in hospitals that had greater involvement in teaching were less likely to have patient falls than patients in hospitals that were less involved in teaching (32.9% decrease, p < .05).
Nurses working on nursing units with 5% to 15% external temporary nurse care hours reported fewer medication errors than those nursing units that did not use external temporary RNs (incident rate = 0.564, p < .05). This suggested that when nursing units with no agency RN hours reported 0.91 medication errors during the study period on average, those with 5% to 15% reported 0.51 (0.564 × 0.91) medication errors, which was a 43.6% decrease in medication errors.
Another important finding was a negative relationship between nurse education level and medication errors. Nursing units with high proportions of bachelors or higher degree-prepared nurses reported fewer medication errors than units with lower proportions of nurses possessing a bachelor's degree or higher (87.5% decrease, p < .05). Although unit tenure (0.5% increase) and RN hours (1.4% increase) had statistically significant associations with medication errors, the magnitude of their impact was small. To sum up, the results suggested that greater use of temporary RNs was associated with high levels of nurse back injuries and patient falls. Moderate levels of temporary nurses were also related to lower levels of medication errors.
The use of temporary nurses has been used frequently as a short-term strategy for managing nurse shortages (May et al., 2006). In this study, the use of temporary nurses in hospitals was prevalent. We found that 75% of nursing units used either external or internal temporary nurses. These findings showed an increase in use of temporary nurses in hospitals compared with previous studies (American Hospital Association, 2001). Also, we found that some nursing units depended heavily on temporary nurses to provide nursing care. At the most, nursing units used external temporary nurses to deliver 32.7% of nursing care hours. In the case of internal temporary nurses, nursing units used them to make up 41.46% of nursing care hours. When considering the number of nursing care hours provided by both external and internal temporary nurses, nursing units used temporary nurses to deliver patient care up to 62.37% of nurse care hours.
The most important findings of this study were the positive associations between use of temporary RN hours and nurse back injuries and patient falls. Nurses working on nursing units with 15% or more external temporary RN hours or internal temporary RN hours were more likely to report back injuries than nurses working on nursing units that did not use temporary RNs, after controlling for hospital and nursing unit characteristics and nurse and patient characteristics. When we used all temporary RN hours, this result was not significant (p = .056). The increase in back injuries might result from frequent reassignment of staff (Castle et al., 2008), physical job demands of permanent nurses, and temporary nurses' unnecessary risk taking to lift patients. As stated earlier, nurses working on units with 15% or more external temporary RN hours reported a 73% increase in nurse back injuries than those working on units with no external temporary RNs. The incidence was 0.67 back injuries per 10 nurses during a 6-month study period, on average. In other words, a nurse working on a unit can expect one back injury about every 8 years (during a year 0.134 [0.67 × 2 / 10] back injuries per a nurse = every 8 years 1 back injury per a nurse). Although the incidence of back injuries is very small, the increase associated with heavy use of temporary nurses is of concern.
Compared with nurses on units that did not use temporary RNs, nurses on nursing units providing 15% or more direct care hours by temporary RNs were 18.8% more likely to report patient falls. These findings were consistent with previous studies that found a negative relationship between use of temporary nurses and patient safety (infections and medication errors; Aiken et al., 1997; Alonso-Echanove et al., 2003). This implies that although at certain levels of use temporary nurses may provide necessary labor to prevent the risk of patient falls, the high levels of dependency on temporary nurses seemed to increase patient fall incidents. As Rush et al. (2009) pointed out, it is important for individual nurses and the entire nursing team to know the patient to help prevent patient falls. To do that, a stable work relationship and teamwork is a core element. Nursing units heavily depending on the temporary nurses cannot maintain that stable work relationship.
We found an interesting relationship between use of external temporary nurses and medication errors. Nursing units with 5% to 15% external temporary RN hours reported fewer medication errors than those nursing units that did not use temporary RNs. Aiken et al. (2007) also found a similar result after accounting for the difference in the adequacy of staffing and resources. Nurses in hospitals with high levels of nonpermanent staff were less likely to report high levels of dispensing the wrong medication. Our findings showed that the reduction in medication errors occurred only at moderate levels (5-15%) of external temporary RN use. This may imply that in nursing shortfalls, using temporary nurses may be a proactive way to maintain patient quality of care.
Among control variables, we found nurses working on nursing units with a high proportion of RNs with bachelor's degrees or higher reported more incidents of back injury than nurses employed on units with fewer bachelors prepared nurses. One possible explanation might be that the educational programs of nurses with BS degrees emphasize the importance of reporting these events. In contrast, education level was associated with fewer reported medication errors. Because our definition of medication errors focused on those that resulted in increased nursing observation, technical monitoring, laboratory and radiographic testing, medical intervention or treatment, or transfer of the patient to another unit, this result is not likely due to underreporting. The relationship between education level and fewer medication errors thus may be explained by better preparation in the educational program regarding the importance of adhering to safe medication practices.
Limitations and Suggestions for Further Research
The first limitation was that the study did not account for the characteristics of the temporary nurses that were used. In this study, the use of temporary nurses was evaluated by using only patient care hours provided by temporary nurses. However, to more accurately assess the association between use of temporary nurses and nurse and patient safety outcomes, the temporary nurse characteristics should also be assessed. These characteristics, such as education level, tenure at the workplace, experience working on certain types of nursing units, can affect the ability of temporary nurses to deliver patient care and to make adjustments to unit policy and practice related to patient care. For future research, these characteristics of temporary nurses should be considered.
This study attempted to control hospital (size, technological sophistication, teaching status, and Magnet certification), nursing unit (work complexity, size, and availability of support service), nurse (age, education, unit tenure, and staffing), and patient (age, health status) characteristics to eliminate alternative explanations for nurse and patient safety outcomes. However, it may not have been possible to completely exclude alternative explanations for safety outcomes such as resource adequacy and orientation for temporary nurses and the unit and hospital's ability to recruit and retain permanent nurses (Aiken, 2007; Aiken et al., 2007). For future studies, such variables need to be accounted for in the model to obtain the true relationship between temporary nurses and nurse and patient safety outcomes.
This study found different impacts on falls and medication errors of different kinds of temporary staff. In terms of patient falls, only total temporary nurse use more than 15% of RN hours contributed to more falls. Thus, the type of temporary staff, be it internal or external, may not be the critical factor related to patient falls; rather, it may be the total hours of care provided by temporary staff. In contrast, total temporary hours had no relationship to medication errors. What did make a difference was the type of temporary staff. In this case, only the use of external temporary nurses for between 5% and 15% of hours was significantly related to fewer medication errors; there was no relationship between internal temporary nurse use and medication errors. One possible explanation might be that temporary nurses from outside the hospital may recognize and respect their lack of familiarity with the medication administration systems, are therefore more careful, and make fewer reported errors. Another possible explanation might be that external temporary nurses may have a long-term contract on a unit and may be familiar with medication processes. To make a conclusion regarding impacts on falls and medication errors, future studies need to include the temporary nurse characteristics, such as working experience, contract period, and education. In addition, studies that examine specifically the hypothesized mechanisms underlying these relationships in terms of increase workload, decreased teamwork, and continuity of care (Castle et al., 2008) would help to elucidate how these changes depend on the magnitude and type of temporary nurse utilization.
Permanent nurse' overtime practice can be used to control nurse shortfall instead of hiring temporary nurses. However, we did not have information what percentage of RN hours were provided by permanent nurse overtime hours. Because of this limited information, we were unable to investigate whether nurse mangers used nurse overtime hours as a substitute for temporary nurse care hours. Thus, future studies need to collect information regarding overtime hours worked by permanent nurses to examine the relationship between use of temporary RNs and patient outcomes.
These study findings have several implications for practice. The current nursing shortage contributes to inadequate hospital nurse staffing. Hospitals use temporary nurses, either external (agency RNs) or internal (float/per diem RNs), to manage nurse staffing shortfalls as a short-term strategy, which seems to be a prevalent practice (American Hospital Association, 2001; May et al., 2006). Using the data collected from medical-surgical nursing units, this study found that the use of temporary RNs for greater than 15% of total RN hours was positively related to both the number of back injuries nurses reported and the patient falls. Health care institutions, especially hospitals, need to monitor the levels of temporary nurse usage and maintain this level below 15% of RN hours to ensure both nurse and patient safety. At the same time, nurse managers need to consider how use of temporary nurses influences the care process, such as a disrupted continuity of care and an increase in the administrative workload of permanent nurses who attempt to supervise and orient temporary nurses. To reduce such adverse influences, nurse managers need to assess more carefully the capability of temporary nurses and the fitness of these nurses to carry out nursing tasks. Also, an orientation program for temporary nurses conducted before temporary nurses come into nursing units may be an effective way to reduce the burden of permanent staff (Aiken, 2007).
Another practice implication is related to the positive association between moderate levels of employment of temporary nurses and fewer medication errors. This study found that when nursing units use temporary nurses at a level between 5% and 15% of RN hours, these nursing units reported fewer medication errors when compared with those that did not employ temporary nurses, suggesting that a moderate level of temporary nurse use may be appropriate. At the same time, additional evidence regarding the association between use of temporary nurses and outcomes is needed before this becomes a standard practice.
Temporary nurses are commonly used in hospitals to fulfill nurse shortfalls (Aiken et al., 2007; American Hospital Association, 2001). It is believed that use of temporary nurses is associated with poor quality (e.g., Deitzer et al., 1992). The results of this study showed that high levels of use of temporary RN staff (i.e., 15% or more temporary RN hours) were associated with high levels of back injuries and patient falls. We also found that moderate utilization (i.e., 5-15% temporary nurse care hours) were related to fewer medication errors. Although we controlled possible variables related to safety outcomes, omitted variables may still explain our findings. Future studies should investigate not only the possibility of inadequate resources, poor orientation for temporary nurses, and deficiencies in recruiting and retaining nurses, these studies should also attempt to measure the underlying causal processes we outlined in our conceptual framework. In this way, we can develop more definitive conclusions about the association between use of temporary nurses and nurse and patient safety outcomes.
This work was supported by grant no. 5R01NR003149 from the National Institute of Nursing Research and had an exemption of the University at Buffalo institutional review board approval because of using a de-identified secondary data.
American Hospital Association. (2001). The healthcare workforce shortage and its implications for American's hospitals
. Retrieved December 15, 2008, from http://www.aha.org/aha/content/2001/pdf/FcgWorkforceReport.pdf
American Nurses Association. (2008). 2008 study of nurses' view on workplace safety and needlestick injuries
. Retrieved September 8, 2009, from http://www.nursingworld.org/MainMenuCategories/OccupationalandEnvironmental/occupationalhealth/SafeNeedles/2008InviroStudy.aspx
Aiken, L. H. (2007). Study: Temporary nurses
not a threat to quality. Healthcare Benchmarks and Quality Improvement
Aiken, L. H., Sloane, D. M., & Klocinski, J. L. (1997). Hospital nurses' occupational exposure to blood: Prospective, retrospective, and institutional reports. American Journal of Public Health
Aiken, L. H., Xue, Y. D., Clarke, S. P., & Sloane, D. M. (2007). Supplemental nurse staffing in hospitals and quality of care. Journal of Nursing Administration
Alonso-Echanove, J., Edwards, J. R., Richards, M. J., Brennan, P., Venezia, R. A., Keen, J., et al. (2003). Effect of nurse staffing and antimicrobial-impregnated central venous catheters on the risk for bloodstream infections in intensive care units. Infection Control and Hospital Epidemiology
Amenta, M. M. (1977). Staffing through temporary help agencies. Supervisor Nurse
(12), 19-20, 23, 25-26.
Anderson, F. D., Maloney, J. P., Knight, C. D., & Jennings, B. M. (1996). Utilization of supplemental agency nurses in an army medical center. Military Medicine
Berney, B., Needleman, J., & Kovner, C. (2005). Factors influencing the use of registered nurse overtime in hospitals, 1995-2000. Journal of Nursing Scholarship
Blegen, M. A., Vaughn, T., & Vojir, C. P. (2008). Nurse staffing levels: Impact of organizational characteristics and registered nurse supply. Health Services Research
(1 Part I), 154.
Bloom, J. R., Alexander, J. A., & Nuchols, B. A. (1997). Nurse staffing patterns and hospital efficiency in the United States. Social Science & Medicine
Carlton, G., & Blegen, M. A. (2006). Medication-related errors: A literature review of incidence and antecedents. Annual Review of Nursing Research
Castle, N. G. (2009). Perceived advantages and disadvantages of using agency staff related to care in nursing home: A conceptual model. Journal of Gerontology Nursing
Castle, N. G., Engberg, J., & Men, A. (2008). Nurse aid agency staffing and quality of care in nursing homes. Medical Care Research and Review
Clarke, S. P. (2007). Hospital work environments, nurse characteristics, and sharps injuries. American Journal of Infection Control
Clarke, S. P., Rockett, J. L., Sloane, D. M., & Aiken, L. H. (2002). Organizational climate, staffing, and safety equipment as predictors of needlestick injuries and near-misses in hospital nurses. American Journal of Infection Control
Cohen, J. (1988). Statistical power analysis for the behavioral sciences
. Hillsdale, NJ: Erlbaum.
Daft, R. L. (1992). Organization theory and design
(4th ed.). St. Paul, MN: West Publishing.
Deitzer, D., Wessell, J., Myles, K., & Trimble, P. (1992). Agency nurses: The right solution to staffing problems? Journal of Long Term Care Administration
Dunton, N., Gajewski, B., Taunton, R. L., & Moore, J. (2004). Nurse staffing and patient falls
on acute care hospital units. Nursing Outlook
Hassmiller, S. B., & Cozine, M. (2006). Addressing the nurse shortage to improve the quality of patient care. Health Affairs
Hofmann, D. A., & Mark, B. (2006). An investigation of the relationship between safety climate and medication errors
as well as other nurse and patient outcomes. Personnel Psychology
Hutchinson, M. K., & Holtman, M. C. (2005). Analysis of count data using Poisson regression. Research in Nursing & Health
Iezzoni, L. I., Shwartz, M., Moskowitz, M. A., Ash, A. S., Sawitz, E., & Burnside, S. (1990). Illness severity and costs of admissions at teaching and nonteaching hospitals. JAMA
Kane, R. L., Shamliyan, T. A., Mueller, C., Duval, S., & Wilt, T. J. (2007). The association of registered nurse staffing levels and patient outcomes: Systematic review and meta-analysis. Medical Care
Kohn, L. T., Corrigan, J., & Donaldson, M. S. (2000). To err is human: Building a safer health system
. Washington, DC: National Academy Press.
Kozlowski, S. W. J., Steve, W. J., & Bell, B. S. (2003). Work groups and teams in organization. In W. C. Borman, D. R. Ilgen & R. J. Klimoski (Eds.), Comprehensive handbook of psychology: Industrial and organizational psychology
(Vol. 12, pp. 333-375). New York: Wiley.
Kuhn, E. M., Hartz, A. J., Gottlieb, M. S., & Rimm, A. A. (1991). The relationship of hospital characteristics and the results of peer review in six large states. Medical Care
Manias, E. B., Aitken, R. B., Peerson, A., Parker, J., & Wong, K. (2003). Agency nursing work in acute care settings: Perceptions of hospital nursing managers and agency nurse providers. Journal of Clinical Nursing
Mark, B. A., Hughes, L. C., Belyea, M., Chang, Y., Hofmann, D., Jones, C. B., et al. (2007). Does safety climate moderate the influence of staffing adequacy and work conditions on nurse injuries? Journal of Safety Research
May, J. H., Bazzoli, G. J., & Gerland, A. M. (2006). Hospitals' responses to nurse staffing shortages. Health Affairs
Merolle, D. L. (1988). The nuts and bolts of agency nursing. RN
Mommersteeg, P. M., Denollet, J., Spertus, J. A., & Pedersen, S. S. (2009). Health status as a risk factor in cardiovascular disease: A systematic review of current evidence. American Heart Journal
Potter, P., Barr, N., McSweeney, M., & Sledge, J. (2003). Identifying nurse staffing and patient outcome relationships: A guide for change in care delivery. Nursing Economics
Reed, L., Blegen, M. A., & Goode, C. S. (1998). Adverse patient occurrences as a measure of nursing care quality. Journal of Nursing Administration
Rondeau, K. V., Williams, E. S., & Wagar, T. H. (2008). Turnover and vacancy rates for registered nurses: Do local labor market factors matter? Health Care Management Review
Roseman, C., & Booker, J. M. (1995). Workload and environmental factors in hospital medication errors
. Nursing Research
Rush, K. L., Robey-Williams, C., Patton, L. M., Chamberlain, D., Bendyk, H., & Sparks T. (2009). Patient falls
: Acute care nurses' experiences. Journal of Clinical Nursing
Salyer, J. (1996). Development and psychometric evaluation of an instrument to measure staff nurses' perception of uncertainty in the hospital environment. Journal of Nursing Measurement
Sherehiy, B., Karwowski, W., & Marek, T. (2004). Risk factors for work-related musculoskeletal decoders in nursing profession: A systematic review. Ergonomia
Sovie, M. D., & Jawad, A. F. (2001). Hospital restructuring and its impact on outcomes: Nursing staff regulations are premature. Journal of Nursing Administration
Tan, L., Hawk, J. C., & Sterling, M. L. (2001). Report of the Council on Scientific Affairs: Preventing needlestick injuries in health care settings. Archives of Internal Medicine
Thomas, E. J., & Brennan, T. A. (2000). Incidence and types of preventable adverse events in elderly patients: Population based review of medical records. BMJ
Vieira, E. R., Kumar, S., Coury, H. J., & Narayan, Y. (2006). Low back problems and possible improvements in nursing jobs. Journal of Advanced Nursing