Differences in nursing practice environment among US acute care unit types: A descriptive study

G Model

NS-2368; No. of Pages 9 International Journal of Nursing Studies xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Nursing Studies journal homepage: www.elsevier.com/ijns

Differences in nursing practice environment among US acute care unit types: A descriptive study JiSun Choi a,*, Diane K. Boyle b a b

College of Nursing Science, Kyung Hee University, Seoul, South Korea Fay W. Whitney School of Nursing, University of Wyoming, Laramie, WY, United States

A R T I C L E I N F O

A B S T R A C T

Article history: Received 9 August 2013 Received in revised form 3 March 2014 Accepted 4 March 2014

Background: The hospital nursing practice environment has been found to be crucial for better nurse and patient outcomes. Yet little is known about the professional nursing practice environment at the unit level where nurses provide 24-hour bedside care to patients. Objectives: To examine differences in nursing practice environments among 11 unit types (critical care, step-down, medical, surgical, combined medical–surgical, obstetric, neonatal, pediatric, psychiatric, perioperative, and emergency) and by Magnet status overall, as well as four specific aspects of the practice environment. Design: Cross-sectional study. Settings: 5322 nursing units in 519 US acute care hospitals. Methods: The nursing practice environment was measured by the Practice Environment Scale of the Nursing Work Index. The Practice Environment Scale of the Nursing Work Index mean composite and four subscale scores were computed at the unit level. Two statistical approaches (one-way analysis of covariance and multivariate analysis of covariance analysis) were employed with a Tukey-Kramer post hoc test. Results: In general, the nursing practice environment was favorable in all unit types. There were significant differences in the nursing practice environment among the 11 unit types and by Magnet status. Pediatric units had the most favorable practice environment and medical–surgical units had the least favorable. A consistent finding across all unit types except neonatal units was that the staffing and resource adequacy subscale scored the lowest compared with all other Practice Environment Scale of the Nursing Work Index subscales (nursing foundations for quality of care, nurse manager ability, leadership, and support, and nurse–physician relations). Unit nursing practice environments were more favorable in Magnet than non-Magnet hospitals. Conclusions: Findings indicate that there are significant variations in unit nursing practice environments among 11 unit types and by hospital Magnet status. Both hospital-level and unit-specific strategies should be considered to achieve an excellent nursing practice environment in all hospital units. ß 2014 Published by Elsevier Ltd.

Keywords: Acute care units NDNQI Unit nursing practice environment Registered nurse

What is already known about the topic? * Corresponding author at: 26 Kyungheedae-ro, Dongdaemun-gu, Seoul 130-701, South Korea. Tel.: +82 2-961-0306. E-mail address: [email protected] (J. Choi).

 The nursing practice environment is an important component of better nurse outcomes, such as nurse job satisfaction and retention.

http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001 0020-7489/ß 2014 Published by Elsevier Ltd.

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 2

J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

 Researchers to date have investigated the nursing practice environment at either the individual nurse or the hospital level. What this paper adds  US acute care hospital units have favorable practice environments overall.  The nursing practice environment is the most favorable on pediatric units and the least favorable on adult combined medical–surgical units.  There are significant differences in professional nursing practice environment among 11 acute care unit types overall, as well as four core aspects of the PES-NWI subscales. 1. Background Hospitals in the United States (US) are increasingly challenged to improve patient care quality and safety while undergoing transformations in health care delivery and financing systems, changes in budgetary and regulatory constraints, and rapid increases in the aging population. In response to these major challenges, many quality improvement programs and initiatives have been developed and implemented at the local, state, and national levels. Because of these efforts, and more importantly, because nursing has been recognized as a major contributor to better quality of patient care (Institute of Medicine, 2004), increased attention has been given to creating positive work environments for registered nurses (RNs) in US acute care hospitals (Aiken et al., 2008; Lake, 2007; Schmalenberg and Kramer, 2008). Over more than three decades, the nursing practice environment has been investigated extensively and globally to address concerns about nursing workforce shortages and patient care quality and safety. Early researchers identified organizational attributes that are particularly important for nurse satisfaction and retention, including unit self-governance, professional nurse autonomy and responsibility for patient care quality, adequate nurse staffing, flexible scheduling, and visible and effective nursing leadership (Kramer and Hafner, 1989; McClure et al., 1983). Building on this research has been challenging because the nursing practice environment is complex to conceptualize and measure. Lake (2002) defined the nursing practice environment as ‘‘the organizational characteristics of a work setting that facilitate or constrain professional nursing practice.’’ She developed the Practice Environment Scale of the Nursing Work Index (PES-NWI) as a core set of five dimensions of the nursing practice environment: (1) nurse participation in hospital affairs; (2) nursing foundations for quality of care; (3) nurse manager ability, leadership, and support of nurses; (4) staffing and resource adequacy; and (5) collegial nurse–physician relations. The PES-NWI has been the only measure endorsed by National Quality Forum (NQF, 2012) as a nursing practice environment measure, and its use has grown in measuring various practice settings in multiple countries (Warshawsky and Havens, 2011).

A large and growing number of studies using the PESNWI have found that a favorable nursing practice environment was related significantly to better nurse outcomes, including higher nurse job satisfaction, less burnout, greater nurse empowerment, and lower intent to leave (Laschinger, 2008; Li et al., 2012; Patrician et al., 2010). Further, researchers have found better patient outcomes in more favorable practice environments, including higher nurse-rated quality of care, lower mortality and failure to rescue rates, and higher patient satisfaction (Aiken et al., 2008; Kutney-Lee et al., 2009). Researchers have demonstrated that better nurse and patient outcomes exist in American Nurses Credentialing Center (ANCC) Magnet1 designated hospitals – vs. nonMagnet hospitals – because of their superior nursing practice environment to support patient care delivery (Drenkard, 2010; Kelly et al., 2011). Many US hospitals consider embarking on the Journey to Magnet ExcellenceTM as a highly effective strategy to improve hospital nurse practice environment as well as quality patient care. As of August 2013, 386 (7%) of US hospitals were ANCC Magnet designated (ANCC, 2013). To date, many of the studies on the nursing practice environment with the PES-NWI have been conducted at either the individual nurse level or the hospital level. More evidence is needed about professional nursing practice environment at the unit level, the smallest organizational unit amenable to management efforts that affect nurse and patient outcomes in acute care hospitals. Relatively few researchers have investigated the relationship of unit-level nursing practice environment with unit nurse and patient outcomes. In a study in which researchers analyzed data from 42 units in four Belgian hospitals by using a multilevel modeling approach, higher nurses’ ratings of quality of care at the unit level were associated with higher ratings of unit-level nurse management, but associated with lower ratings of hospital-level management and organizational support (Van Bogaert et al., 2010). In a recent study using data from a multi-country nurse workforce study, three aspects of nursing practice environment (promotion of care quality, doctor–nurse collegial relations, and managerial support for nursing) were examined in relation to three dimensions of nurse burnout (emotional exhaustion, depersonalization, and personal accomplishment) by testing a series of multivariate multilevel probit models (Li et al., 2012). Note that these researchers did not report the original subscale names of the PES-NWI and did not include all of the three-subscale items: nursing foundations for quality of care (promotion of care quality, 9 items), collegial nurse–physician relations (doctor–nurse collegial relations, 7 items), and nurse manager ability, leadership, and support of nurses (managerial support for nursing, 4 items). Overall, lower ratings of the nursing practice environment were significantly related to higher nurse burnout at both the unit level and the hospital level. The relationships of each of three aspects of nursing practice environment to nurse burnout were not consistent at three different levels (unit, hospital, and country). For example, all three dimensions of nurse burnout were significantly related to promotion of care quality at both levels (unit and hospital), but to

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

doctor–nurse collegial relations at the unit level only. Furthermore, the sample units in these studies were selected from only critical care, medical, and surgical units. 1.1. Unit type differences in nursing practice environment The nursing unit is the work place where nurses provide 24-hour direct care to patients and where adverse events occur. According to the conceptual theory of the clinical microsystem in health care, a nursing unit is considered a clinical system – defined as a small, functional group of staff working together to provide care to a discrete population of patients – within a hospital (Nelson et al., 2002). Although a nursing unit’s environment is affected by the hospital in which it is located, each nursing unit is likely to have different environment and nursing practice in order to meet care needs of patient population (Disch, 2006; Nelson et al., 2002). Similar to the traditionally known medical specializations (e.g., medicine, surgery, and pediatrics), the nursing units can be grouped by several dimensions such as the type of patient and the kind of service or care, resulting in different types of nursing units such as intensive, medical, and surgical units (Leatt and Schneck, 1984). In two earlier studies on some aspects of environment (e.g., nursing autonomy, medical complexity, and relationships with multidisciplinary teams) among different unit types, researchers have demonstrated that each unit type has its own unique environment (Adams and Bond, 1997; Leatt and Schneck, 1982). For example, pediatric units appeared to have the highest medical complexity, but reported lower interaction with physicians than intensive units that were highest for interaction with physicians and other groups in the hospital, whereas psychiatric units appeared to have lower medical complexity and less multidisciplinary contact than other unit types in the same hospital. Although researchers have begun to investigate the nursing practice environment at the unit level (EatonSpiva et al., 2010; Kramer et al., 2011; Li et al., 2012), it is difficult to generate comparable evidence on nursing practice environment among different unit types. In addition to few studies of the nursing practice environment at the unit level, there are some methodological issues in studies on unit type differences, including small samples of hospitals, units, or nurses; low response rate; inconsistent classification of unit types across studies; and limited number of unit types studied. The most investigated unit types in studies on unit nursing practice environment were critical care, medical and surgical units. Little is known about the nursing practice environment among other unit types such as pediatric, perioperative, and ED. Therefore, the study purpose was to describe unit nursing practice environments among 11 unit types and by Magnet status using a large, national sample of US acute care hospital units, consistent classification of unit types across hospitals (adult critical care, adult step-down, adult medical, adult surgical, adult combined medical–surgical, obstetric, neonatal, pediatric, psychiatric, perioperative, and ED), and a consistent measure of the practice environment (the PES-NWI). Because the nursing practice

3

environment is multidimensional, the specific aim was to examine differences in the unit nursing practice environment overall, as well as specific aspects of the practice environment (as conceptualized in the PES-NWI) by unit type and hospital Magnet status while controlling for other hospital characteristics (hospital size, teaching status, and geographical location). 2. Methods 2.1. Data source This study is a secondary analysis using cross-sectional data from the 2011 National Database of Nursing Quality Indicators1 (NDNQI1) RN Survey with Practice Environment Scales (PES). NDNQI, a program of the American Nurses Association, conducts an annual web-based survey to collect data on practice environments, work attitudes (e.g., job satisfaction), and demographic characteristics (e.g., age, education, and tenure) from RNs employed in NDNQI member hospitals. Each year NDNQI obtains institutional review board approval at the University of Kansas Medical Center before the Survey begins. Hospitals volunteer for participation and each appoints a site coordinator for the Survey. At the secured NDNQI website, site coordinators obtain a standardized data collection protocol and a unique hospital identification code. Also at the NDNQI website, coordinators enter the name of each participating unit and number of eligible RNs on each unit. To assure confidentiality of participation and anonymity of data, names of RNs are not collected. Site coordinators publicize the Survey internally with approved posters, offer incentives within protocol guidelines (e.g., cafeteria coupons, coffee coupons, pizza for the unit with highest response rate), and distribute two reminder postcards to eligible RNs during the designated 3-week data collection period. To be eligible for the Survey, RNs must spend at least 50% of their time providing direct patient care and must have a minimum of 3 months employment in the current unit. In 2011, 206, 085 RNs from 11,513 units in 553 NDNQI hospitals participated in the NDNQI RN Survey with PES. The overall individual RN and unit response rate were 70.4% and 73.7%, respectively. For each participating unit, the average number of responding RNs was 18, ranging from 0 to 192. 2.2. Study sample The unit of analysis for this study was the nursing unit, not individual RNs. To ensure a representative sample for data aggregation from individual RNs to the unit level, units with 5 or greater participants and at least a 50% response rate were included in the final analysis (Kramer et al., 2009; Verran et al., 1995). NDNQI units are classified first by patient population (e.g., adult, pediatric, neonatal) and then acuity level (e.g., critical care, step-down, medical, surgical) or type of care provided (e.g., psychiatric, ED). The included acute care unit types in this study were adult critical care (626 units), adult step-down (467 units), adult medical (501 units), adult surgical (394 units), adult

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

4

Table 1 Registered nurse (RN) sample characteristics by unit type. Unit type Critical care Step-down Medical Surgical Med–Surg Obstetric Neonatal Pediatrics Psychiatric Perioperative ED Overall

N 17,516 11,527 11,369 8740 15,988 12,968 7059 7136 3058 16,975 11,248 123,584

N. Respondents per unit, Mean (SD)

Age, Mean (SD)

Unit tenure, Mean (SD)

BSN or higher, %

Women, %

Non-Hispanic White, %

Full-time, %

30.27 26.75 24.43 23.95 24.12 31.49 36.79 24.75 13.07 19.01 34.57

38.42 37.61 38.27 38.81 38.72 41.47 41.44 36.29 48.07 46.23 39.07

6.56 4.88 5.41 6.20 5.50 9.05 10.17 7.12 7.20 8.49 5.59

(6.93) (5.31) (5.81) (6.68) (6.00) (8.09) (8.90) (7.35) (7.26) (7.81) (5.87)

68.04 56.29 55.32 55.10 53.60 55.18 64.58 68.86 50.38 52.91 57.84

86.00 90.07 91.77 92.57 91.89 99.56 98.68 96.02 84.00 92.18 83.62

75.18 66.84 64.21 69.01 63.77 78.43 79.11 80.90 69.69 80.40 77.38

84.39 83.68 83.21 81.84 81.90 69.74 71.94 75.10 68.82 72.57 79.05

6.83 (7.08)

58.05

91.46

73.16

78.44

(15.50) (12.88) (10.38) (11.49) (11.04) (19.21) (29.02) (16.57) (06.47) (12.00) (22.35)

25.45 (15.99)

(10.75) (10.76) (11.20) (11.68) (11.13) (11.08) (11.27) (10.55) (11.77) (10.68) (10.45)

40.12 (11.38)

Note: Med–Surg = combined medical–surgical; ED = emergency department; BSN = bachelor of science in nursing.

combined medical–surgical (715 units), obstetric (450 units), neonatal (207 units), pediatric (309 units), psychiatric (262 units), perioperative (1020 units), and ED (371 units). The final analytical sample included 5322 units in 519 US acute care hospitals. For the 519 hospitals represented in the sample, 28% had 300 staffed beds or more, 48% were teaching hospitals, 84% were not-for-profit hospitals, and 31% were Magnetdesignated hospitals. These distributions of the sample hospital characteristics were similar with those of all NDNQI member hospitals in 2011, but higher in our sample compared with those in all US hospitals reporting to the American Hospital Association (AHA) Annual Survey in 2011: hospitals with 300 beds or more (15%), not-for-profit hospitals (26%), and Magnet-designated hospitals (6%). The geographical locations of the sample hospitals, as classified by four census regions, were as follows: West (17%), Midwest (25%), South (40%), and Northeast (18%). The sample and all AHA hospitals had similar geographic distributions. Our sample units included 123,584 RN survey participants with the average number of respondents of 25 per unit. The average age and tenure on the current unit were 40 and 7, respectively. The majority were female (92%), non-Hispanic White (73%), and full-time employees (78%), and 58% had a bachelor’s or higher degree in nursing. Descriptive characteristics of the sample RN respondents are presented by unit type in Table 1. RNs in pediatric units were the youngest (M = 36), whereas RNs in psychiatric units were the oldest (M = 48), followed by RNs in perioperative units (M = 46). RNs in step-down units had worked the shortest years on the current unit (M = 5), whereas RNs in obstetric and neonatal had worked the longest years on the current unit (M = 9 and 10, respectively). RNs in critical care and pediatric units had the largest percentage of a bachelor’s or higher degree in nursing (M = 68 and 69, respectively). 2.3. Measures The PES-NWI (Lake, 2002) is included as part of the NDNQI RN Survey to measure the nursing practice environment at the nursing unit level. The PES-NWI

consists of 31 items in five subscales: nurse participation in hospital affairs (hospital affairs, 9 items); nursing foundations for quality of care (nursing foundations for quality, 10 items); nurse manager ability, leadership, and support of nurses (nurse manager, 5 items); staffing and resource adequacy (staffing and resource adequacy, 4 items); and collegial nurse–physician relations (nurse– physician relations, 3 items). RNs are asked to rate their level of agreement that each of 31 organizational traits is present in their current job on a 4-point Likert-type scale, ranging from 1 (strongly disagree) to 4 (strongly agree). The possible mean score of each of the five subscales ranges from 1 to 4; higher scores indicate the presence of a more favorable nursing practice environment. For this study, one of the five subscales (hospital affairs) was not included in our analyses because items in this subscale reflect nursing practice environment at the hospital level (Lake, 2002) and it was correlated highly with another subscale (nursing foundations for quality, r = .86) in the current sample. The mean scores on each of the four subscales were calculated by averaging individual RN responses to items on each subscale. These RN scores then were averaged for each unit to obtain a unit-level score for each of the four included subscales of the PESNWI. For each unit, a composite score of the PES-NWI also was calculated by averaging the unit-level mean scores for four included subscales to obtain a score of overall unit nursing practice environment. Reliability and validity of the PES-NWI has been established at the different levels (e.g., individual nurse, unit, and hospital), as well as in various clinical settings and countries (Lake, 2007; Warshawsky and Havens, 2011). In the current sample, the individual RN-level internal consistency reliability was high for the composite score (a = .94), as it was for each of the four subscales, with alphas ranging from .87 to .90. The unit-level reliability was assessed by estimating the intraclass correlation coefficient [ICC (2)] from a one-way analysis of variance (ANOVA) model. The ICC (2) indicates the degree of homogeneity of responses from individuals in the same group. Thus, the ICC (2) can be used to assess within-group agreement for the mean as a group measure and is computed with the following formula (Bliese, 2000):

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

(MSB  MSW)/MSB, where MSB is the between-group mean square and MSW is the within-group mean square. The ICC (2) above .6 is considered to be substantially reliable (Shrout, 1998). The estimated ICC (2) for the composite score was 0.84 and ranged from 0.78 to 0.87 for each of the four subscales of the PES-NWI, indicating high reliability of the unit-level measure of the PES-NWI. Additionally, we computed rWG(J), the most commonly used measure of within-group agreement for multiple-item scales, using the following formula (James et al., 1984): J[1  (sxj2/ sEU2)]/J[1  (sxj2/sEU2)] + (sxj2/sEU2), where J is the number of items, sxj2 is the mean of the observed variances on the J items, and sEU2 is the expected random variances. The rWG(J) for the composite score (22-item PES used in this study) was 0.97 and ranged from 0.84 to 0.95 for each of the four subscales of the PES-NWI, indicating high within-group agreement. 2.4. Data analysis To examine differences in the unit nursing practice environment overall as well as in four specific aspects of the practice environment among the 11 unit types, several statistical methods were applied in two stages. At the first stage of data analysis with unit-level composite scores of the PES-NWI, we first examined the relationships of hospital characteristics (hospital size, teaching status, Magnet status, and geographical location) to the overall unit nursing practice environment using a two-level linear regression analysis. In this analysis, a random hospital intercept was included to account for the units clustered by hospitals. Unit type also was included to account for the unit differences in patient characteristics such as patient acuity level and patient population. Then, to examine differences in overall unit nursing practice environment across 11 unit types, we used a one-way analysis of covariance (ANCOVA) with a Tukey-Kramer (TK) post hoc test. The TK test is a single-step multiple comparison method based on a studentized range distribution to determine the critical differences in all possible pairs of means, and is a preferred method when the sample sizes are unequal (Hayter, 1984). The included covariates were the same hospital characteristics tested above in the linear regression analysis. At the second stage of data analysis with unit-level scores on four subscales of the PES-NWI, we examined differences in specific aspects of the nursing practice environment by the 11 unit types and hospital Magnet status while controlling for other hospital characteristics (hospital size, teaching status, and geographical location). A two-way multivariate analysis of covariance analysis (MANCOVA) was used because it allowed us to examine the effects of two independent variables on multiple dependent variables simultaneously. It is also a suitable approach to minimize potential Type 1 error rate due to the use of multiple univariate tests (Tabachnick and Fidell, 2007). In addition to two independent variables in the model (unit type and Magnet status), the interaction between unit type and hospital Magnet status (unit type  Magnet status) also was included to see whether differences in four specific aspects of the unit nursing

5

practice environment among the 11 unit types differ between Magnet and non-Magnet hospitals. Pillai’s trace was used to evaluate the significance of the multivariate effects as it is the most reliable statistic for unequal group sizes (Tabachnick and Fidell, 2007). If significant multivariate effects were found, further evaluation was performed using multiple univariate tests to assess on which specific aspects of the unit nursing practice environment the differences existed, and then using TK post hoc tests to assess where differences existed by the 11 unit types on each PES-NWI subscale. All data analyses were conducted using Stata version 11. 3. Results Table 2 presents the relationships of hospital characteristics to the overall unit nursing practice environment. Except for hospital size, all tested hospital characteristics were associated significantly with unit nursing practice environment. More favorable unit nursing practice environments were found in Magnet than nonMagnet hospitals, and in non-teaching than teaching hospitals. Among geographical locations of hospitals, unit nursing practice environments were significantly more favorable in West region hospitals compared with those in the Northeast region. The effect sizes of these three hospital characteristics were similar. For units in a Magnet, a non-teaching, or a Northeast region hospital, the estimated unit-level composite scores of the PES for units were about a 0.06 higher, about a quarter of a standard deviation, compared to those in a non-Magnet, a teaching, or a West region hospital. Table 3 presents unit-level PES-NWI composite and four subscale mean scores by unit type. The mean composite score of the PES-NWI was 2.96 for all sample units, ranging from 2.91 for combined medical–surgical units to 3.08 for pediatric units. All mean composite scores of the PES-NWI were above the midpoint of 2.5 on the 4point response scale, indicating an overall favorable unit nursing practice environment. Results from the ANCOVA with TK post hoc test indicated that there was a significant difference in overall unit nursing practice environment across the 11 unit types (F = 24.04, p < .001). The TK post hoc test showed that unit-level composite scores of the PES-NWI were significantly higher in neonatal units and pediatric units than all other unit types. Likewise, all unit-level mean scores on four subscales of the PES-NWI were above the midpoint of 2.5, indicating Table 2 Hospital characteristics associated with unit-level nursing practice environment (N = 5322). Characteristics

Coefficient (SE)

p

300 beds (vs. <300 beds) Teaching hospital (vs. non-teaching) Magnet hospital (vs. non-Magnet) Geographical location (vs. West) Midwest South Northeast

.014(.013) .056(.013) .061(.013)

.115 <.001 <.001

.009(.019) .021(.017) .059(.020)

.646 .201 .003

Note: Unit type was included to adjust.

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

6

Table 3 Unit-level mean scores of the Practice Environment Scale of the Nursing Work Index (PES-NWI) by unit type. Unit type

N

Composite Mean (SD)

QOC Mean (SD)

NM Mean (SD)

Critical care Step-down Medical Surgical Med–Surg Obstetric Neonatal Pediatric Psychiatric Peri-op ED

626 467 501 394 715 450 207 309 262 1020 371

2.95 2.93 2.94 2.95 2.91 2.97 3.05 3.08 2.96 2.97 2.96

3.07 3.13 3.13 3.14 3.11 3.12 3.13 3.16 3.05 3.07 2.99

2.92 2.98 3.02 3.01 2.99 2.93 2.94 3.03 2.91 2.92 2.95

Overall

5322

2.96 (.25)

(.24) (.23) (.24) (.24) (.24) (.24) (.21)* (.21)* (.30) (.27) (.26)

(.20) (.18) (.20) (.18) (.20) (.19) (.20) (.18) (.26)yy (.22) (.22)**

3.10 (.21)

(.35)yy (.31) (.33) (.30) (.31) (.34) (.34) (.29) (.43)** (.38)yy (.34)

2.96 (.35)

S&RA Mean (SD)

N-P Mean (SD)

2.74 2.63 2.62 2.64 2.58 2.78 2.97 2.92 2.78 2.85 2.65

3.05 2.99 2.99 3.02 2.96 3.07 3.16 3.19 3.11 3.03 3.24

(.34) (.36) (.35) (.35) (.35) (.34) (.27)* (.30)* (.40) (.36)y (.39)

2.73 (.37)

(.25) (.23) (.25) (.26) (.24) (.27) (.27) (.27)y (.31) (.27) (.26)*

3.05 (.27)

Note: QOC = nursing foundations for quality of care; NM = nurse manager ability, leadership, and support of nurses; S&RA = staffing and resource adequacy; N-P = collegial nurse/physician relations; Med–Surg = combined medical–surgical; Peri-op = perioperative; ED = emergency department. All analyses are adjusted for hospital characteristics (hospital size, teaching status, Magnet status, and geographical location). * Significantly higher than all others at p < .05, not including y. ** Significantly lower than all others at p < .05, not including yy.

that all four aspects of the unit nursing practice environment were favorable across the 11 unit types. The highest mean score in the full sample of units was reported for the nursing foundations for quality subscale (M = 3.10), followed by the nurse–physician relations (M = 3.05), nurse manager (M = 2.96), and staffing and resource adequacy (M = 2.73). For specific unit types, however, the highest mean scores reported in neonatal, pediatric, psychiatric, and ED units were for the nurse–physician relations subscale. For each of four subscales of the PESNWI, the highest mean scores were reported in pediatric units (for nursing foundations for quality and nurse manager), neonatal units (for staffing and resource adequacy), and EDs (for nurse–physician relations). The lowest mean score of the nursing foundations for quality subscale was reported in EDs. For the staffing and resource adequacy and nurse–physician relations subscales, the lowest scores were reported in combined medical–surgical units.

The unit-level four-subscale mean scores of the PESNWI on 11 unit types by hospital Magnet status are presented in Table 4. Results from the two-way MANCOVA indicated that the multivariate effect on four subscales of the PES-NWI was statistically significant by unit type (Pillai trace = .41 F = 60.59, p < .001) and Magnet status (Pillai trace = .03 F = 45.28, p < .001). The multivariate interaction effect of unit type and Magnet status was not statistically significant; the difference on four specific aspects of the unit nursing practice environment across 11 unit types did not differ by Magnet status. Follow-up univariate tests showed that mean scores for all four subscales were significantly higher in Magnet than nonMagnet hospitals (M range: 2.77–3.14 vs. 2.71–3.07, all p < .001). Also, there were significant differences across 11 unit types on each of four subscales of the PES-NWI (F range = 7.77–54.83, all p < .001). Compared with all other unit types, EDs had a significantly lower mean score on the nursing foundations for quality subscale except for

Table 4 Unit-level mean scores on four subscales of the Practice Environment Scale of the Nursing Work Index (PES-NWI) among unit types by hospital magnet status. Unit type Critical care

Step-down

Medical

Surgical

Med–Surg

Obstetric

Neonate

Pediatric

Psychiatric

Peri-op

ED

Overall

3.17 3.10

3.17 3.10

3.17 3.08

3.15 3.10

3.15 3.11

3.20 3.12

3.10 3.02

3.10 3.05

3.03 2.97

3.14* 3.07

Nurse manager ability, leadership, and support Magnet 2.96 2.99 3.06 Non-Magnet 2.89 2.97 2.98

3.03 2.99

3.02 3.00

2.93 2.92

2.96 2.93

3.05 3.01

2.96 2.87

2.93 2.92

2.94 2.94

2.98* 2.94

Staffing and resource adequacy Magnet 2.78 2.65 Non-Magnet 2.72 2.62

2.67 2.59

2.67 2.61

2.65 2.55

2.78 2.77

2.99 2.96

2.94 2.91

2.86 2.73

2.86 2.85

2.69 2.63

2.77* 2.71

Collegial nurse–physician relations Magnet 3.07 3.02 Non-Magnet 3.03 2.97

3.05 2.95

3.06 2.99

3.01 2.93

3.09 3.05

3.11 3.18

3.21 3.17

3.16 3.08

3.04 3.02

3.26 3.22

3.08* 3.03

Nursing foundation for quality of care Magnet 3.11 3.15 Non-Magnet 3.04 3.11

Note: Med–Surg = combined medical–surgical; Peri-op = perioperative; ED = emergency department. * Significantly higher at p < .01.

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

psychiatric units but a significantly higher score on the nurse–physician subscale except pediatric units (see Table 3). In psychiatric units, a significantly lower mean score was found on the nurse manager subscale than those in all other unit types except critical care and perioperative units. Neonatal and pediatric units had significantly higher mean scores for the staffing and resource adequacy subscale than those in all other unit types except perioperative units.

4. Discussion Using a large, national sample of acute care units with a consistent definition of unit type across hospitals and the NQF-endorsed PES-NWI, our study is among the first to explore variations in the nursing practice environment at the unit level. The practice environment was found to be generally favorable across all 11 unit types included in our study. In previous studies of the professional nursing practice environment using the PES-NWI, relatively few researchers reported average composite scores of the PEW-NWI. Further, most of these scores were computed at the individual nurse level and the hospital level, rather than the unit level. In previous studies, the reported average composite scores on the PES-NWI in US acute care settings ranged from 2.48 for Pennsylvania hospitals to 2.99 for the first seven ANCC designated Magnet hospitals (Warshawsky and Havens, 2011). Although the mean composite scores of the PES-NWI reported here may not be comparable to those from previous studies, it appears that overall nursing practice environment in our sample units is as favorable as in some of the ANCC designated Magnet hospitals in the late 1990s. Still, the level of agreement on a set of core organizational traits that support the professional nursing practice environment in all sample units of the current study is not at the maximum degree of theoretical agreement (2.96 on a 1–4 scale). In our study, significant differences in RN perceptions of the unit nursing practice environment were found among the 11 unit types while controlling for potential confounding hospital characteristics (e.g., teaching status, Magnet status, and geographical location). In comparison to other unit types, pediatric and neonatal units have the most favorable practice environments. The highest scores were reported for nurse–physician relations in these two unit types among the four PES-NWI subscales. Further, pediatric units had the highest scores on two of the four PES-NWI subscales, nurse manager and nursing foundations for quality. Although the staffing and resource adequacy score was relatively lower than other subscale scores in pediatric units, the staffing and resource adequacy score was significantly higher than the staffing and resource adequacy score in all other unit types, except neonatal units. In our sample, pediatric and neonatal workgroups had among the highest mean tenure on the unit and the highest percentage of BSN or higher prepared nurses. The more stable and higher educated pediatric workforce likely contributes to an improved practice environment. Further study of pediatric and neonatal unit practice environment, particularly nurse–physician

7

relations, may provide insight into improving the practice environment on other unit types. Obstetric and psychiatric unit types demonstrate moderate scores on the composite of the PES-NWI, as well as staffing and resource adequacy and nurse– physician relations subscales. Of interest is that the two unit types had among the lowest scores of nurse manager support. Moreover, psychiatric units had the highest mean tenure on the unit and the lowest rate of BSN and higher nurses. In these unit types, it may be critical to invest in leadership training and provide resources for BSN completion. Compared with all other unit types, EDs had a significantly lower score for the nursing foundations for quality subscale and among the lower scores on nurse manager and staff and resource adequacy subscales. Further, ED nursing workgroups were among the lowest in mean years of tenure on the unit. EDs are characterized by high patient turnover, lack of established patient relationships, and use of many standardized protocols (Boyle et al., 2006), all of which may contribute to lower tenure on the unit and weaken autonomous nursing practice. However, EDs had a significantly higher score on the nurse–physician relations subscale than all other units. This finding may not be surprising in the context of ED environments because nurses work constantly beside ED physicians. While adult critical care, step-down, medical, and surgical units share similar practice environment score patterns such as lower composite scores, higher scores on the nursing foundations for quality subscale, and lower scores on the staffing and resource adequacy subscale, combined medical–surgical units had the least favorable nursing practice environment of all unit types and the lowest scores on two of the PES-NWI subscales (staffing and resource adequacy and nurse–physician relations). Among these unit types, we found that combined medical– surgical unit nursing workgroups in our sample also had the lowest mean tenure on the unit and the lowest percentage of BSN and higher prepared nurses. These combined medical–surgical workforce characteristics may contribute to less favorable professional nursing practice environments. A potential strategy to improve practice environment in these unit types might be to include more availability of advanced practice nurses for consultation (Boyle et al., 2006) and more support for attainting BSN education. Findings indicate that there is much room for improvement in specific areas of the practice environment in all unit types. Among the four aspects of the practice environment examined in our study, the lowest scores across all unit types were for the staffing and resource adequacy and nurse manager subscales. These two practice environment aspects may require the most attention when developing strategies for improvement. Across all unit types except neonatal units, scores on staffing and resource adequacy ranged from 2.58 for medical–surgical units to 2.97 for neonatal units. This is consistent with findings from a literature review of studies using the PES-NWI that the lowest score of the PES-NWI was most often the staffing and resource adequacy

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 8

J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

subscale (Warshawsky and Havens, 2011). Previous research indicated that nurses’ perceptions of staffing adequacy were significantly influenced not only by the total number of nurses on a unit but also by several other factors, including the number of beds, patient acuity, and inadequate number of assistive personnel (Kalisch et al., 2011; Mark, 2002). Given that there is a large body of research demonstrating the relationship between better nurse staffing and better patient outcomes (e.g., lower mortality rates, fewer falls, shorter length of stay) (Lake and Cheung, 2006; Shekelle, 2013; Unruh, 2008), strategies to improve this aspect of practice environment are paramount. Nursing administrators and managers should consider both quantity and quality of nurse staffing to provide care for patients in each specific unit when they develop strategic plans for staffing. Our findings suggest that combined medical–surgical, medical, surgical, and step-down units could benefit most from improvements in the practice environment aspect of staffing and resources. There was a significant difference in scores of the nurse manager subscale among the 11 unit types. However, all scores across unit types were highly similar and relatively lower than those on the nursing foundations for quality and nurse–physician relations subscales. Researchers have found that nursing unit managers play an important role in developing positive work environments and thereby improving nurse job satisfaction and retention (Duffield et al., 2010). Our findings underscore the need to develop better front-line nursing leadership particularly in psychiatric, critical care, and perioperative units. It is important that all unit nurse managers make themselves aware of the impact of their leadership on both nurse and patient outcomes. Hospital-level organizational support also is needed to assist front-line nurse managers in developing the leadership skills required for their positions. Kramer et al. (2012) stated that ‘‘another common way of defining and measuring work environments is by equating them with nurse job satisfaction’’ (p. 149). It is interesting to note that our findings echo the levels of job satisfaction reported in an earlier study of RN workgroups in various US acute care units (Boyle et al., 2006). Using unit-level data from the 2004 NDNQI RN survey, Boyle et al. (2006) examined differences in RN workgroup job satisfaction among 10 unit types. RN workgroups in pediatric units were found to be the most satisfied, whereas those in EDs were least satisfied. In addition, RN workgroups in critical care, step-down, and medical– surgical units reported moderate job satisfaction overall, but they reported the lowest scores on job enjoyment scales compared with all other unit types (e.g., pediatric, maternal-newborn, and rehabilitation) except EDs. Similarly, in another recent study of nurse turnover among three unit types by population age (neonatal, pediatric, and adult) using unit-level NDNQI data, nurse turnover (total nurse and RN turnover) was found to be lowest on neonatal units and highest on adult units (Staggs and Dunton, 2012). A positive relationship between work environment and nurse outcomes has been well documented in the literature. Although a direct relationship between nursing practice environment and nurse outcomes such as job

satisfaction and turnover was not tested in the current study, our findings provide insight into the practice environment that is related to these outcomes. We also found that our sample units in Magnet designated hospitals had slightly more favorable practice environments than those in non-Magnet hospitals, while the overall difference in nursing practice environment across 11 unit types did not differ by hospital Magnet status. It should be noted that effects of hospital Magnet status on unit nursing practice environment were small, with statistical significance being likely in part to the large sample size. Despite these small differences between Magnet and non-magnet hospital units, the mean scores for all four subscales were constantly higher in Magnet than non-Magnet hospitals, except two unit types (EDs for the nurse manager subscale and neonatal units for the nurse–physician relations subscale). Given that Magnet hospitals have to maintain their already superior nursing practice environment over time for redesignation, it is assumed that the practice environment in all hospital units within Magnet designated hospitals would be more favorable than that in non-Magnet hospitals. Also, nonMagnet hospital units in our sample were likely to be at hospitals that were either applicants or preparing for initial application for Magnet designation because the NDNQI’s reports of unit-level data on nursing performance could be used by hospitals to submit data to the ANCC for Magnet appraisal process. Hospital-level initiatives to meet Magnet designation criteria may contribute to facilitate unit-based efforts to improve nursing practice environment. Our findings provide support that the ANCC Magnet Recognition Program1 may be an effective mechanism for achieving positive practice environments on all hospital units. Although we used unit-level data on the nursing practice environment in a large, national sample of hospitals from the NDNQI database, our findings cannot be generalized to all US acute care hospital units. The sample hospitals voluntarily participate in NDNQI, resulting in a tendency for larger and teaching hospitals to be over-represented compared with all US hospitals. Also, the sample sizes of unit type were unequal, and several unit types, particularly neonatal, pediatric, and psychiatric units, comprised a relatively small proportion of the study units. Finally, some unit types such as rehabilitation, outpatient clinics, and labs were omitted in our study, while nursing practice environments were examined on a wide range of units in acute care hospitals. 5. Conclusions This study provides comprehensive, descriptive, and comparable information on nursing practice environment among 11 unit types as well as hospital Magnet status. Study findings indicate that there are significant variations in unit nursing practice environments among 11 unit types and by hospital Magnet status, as well as where changes may be needed to improve the nursing practice environment in the 11 different unit types. These findings provide empirical evidence that both hospital-wide and unitspecific strategies should be considered to improve the

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001

G Model

NS-2368; No. of Pages 9 J.S. Choi, D.K. Boyle / International Journal of Nursing Studies xxx (2014) xxx–xxx

professional nursing practice environment, which can affect nurse and patient outcomes. It is crucial to identify best strategies tailored to the nature of specific units. Further research is needed to develop unit-specific interventions in targeted areas of the nursing practice environment along with other unit structural elements such as nurse staffing levels, RN education levels, and unit RN tenure. Moreover, it is needed to explore what specific strategies can be implemented to develop and maintain a favorable practice environment in any particular unit type. Conflicts of interest None declared. Funding This study was conducted under a contract with the American Nurses Association Ethical approval This study was approved by the University of Kansas Medical Center Human Subjects Committee. References Adams, A., Bond, S., 1997. Clinical specialty and organizational features of acute hospital wards. J. Adv. Nurs. 26 (6) 1158–1167. Aiken, L.H., Clarke, S.P., Sloane, D.M., Lake, E.T., Cheney, T., 2008. Effects of hospital care environment on patient mortality and nurse outcomes. J. Nurs. Adm. 38, 223–229. Amercian Nurses Credentialing Center, 2013. Find a Magent Hospital. http://www.nursecredentialing.org/FindaMagnetHospital.aspx (accessed 08.06.13). Bliese, P.D., 2000. Within-group agreement, non-independence, and reliability: implications for data aggregation and analysis. In: Klein, K.J., Kozlowski, S.W.J. (Eds.), Multilevel Theory, Research, and Methods in Organizations: Foundations, Extensions, and New Directions. JosseyBass, San Francisco, CA, pp. 349–381. Boyle, D.K., Miller, P.A., Gajewski, B.J., Hart, S.E., Dunton, N., 2006. Unit type differences in RN workgroup job satisfaction. West. J. Nurs. Res. 28, 622–640. Disch, J., 2006. Clinical microsystems: the building blocks of patient safety. Creat. Nurs. 12 (3) 13–14. Drenkard, K., 2010. The business case for Magnet1. J. Nurs. Adm. 40, 263–271. Duffield, C.M., Roche, M.A., Blay, N., Stasa, H., 2010. Nursing unit managers, staff retention and the work environment. J. Clin. Nurs. 20, 23–33. Eaton-Spiva, L.A., Buitrago, P., Trotter, L., Macy, A., Lariscy, M., Johnson, D., 2010. Assessing and redesigning the nursing practice environment. J. Nurs. Adm. 40 (1) 36–42. Hayter, A.J., 1984. A proof of the conjecture that the Tukey-Kramer multiple comparisons procedure is conservative. Ann. Stat. 61–75. Institute of Medicine, 2004. Keeping Patients Safe: Transforming the Work Environment of Nurses. National Academies Press, Washington, DC. James, L.R., Demaree, R.G., Wolf, G., 1984. Estimating within-group interrater reliability with and without response bias. J. Appl. Psychol. 69 (1) 85–98. Kalisch, B., Friese, C.R., Choi, S.H., Rochman, M., 2011. Hospital nurse staffing: choice of measure matters. Med. Care 49 (8) 775–779. Kelly, L.A., McHugh, M.D., Aiken, L.H., 2011. Nurse outcomes in Magnet1 and Non-Magnet hospitals. J. Nurs. Adm. 41, 428–433.

9

Kramer, M., Hafner, L.P., 1989. Shared values: impact on staff nurse job satisfaction and perceived productivity. Nurs. Res. 38 (3) 172–177. Kramer, M., Halfer, D., Maguire, P., Schmalenberg, C., 2012. Impact of healthy work environments and multistage nurse residency programs on retention of newly licensed RNs. J. Nurs. Adm. 42 (3) 148. Kramer, M., Maguire, P., Brewer, B.B., 2011. Clinical nurses in Magnet hospitals confirm productive, healthy unit work environments. J. Nurs. Manage. 19 (1) 5–17. Kramer, M., Schmalenberg, C., Brewer, B.B., Verran, J.A., Keller-Unger, J., 2009. Accurate assessment of clinical nurses’ work environments: response rate needed. Res. Nurs. Health 32 (2) 229–240. Kutney-Lee, A., McHugh, M.D., Sloane, D.M., Cimiotti, J.P., Flynn, L., Neff, D.F., Aiken, L.H., 2009. Nursing: a key to patient satisfaction. Health Aff. (Millwood) 28 (4) w669–w677. Lake, E.T., 2002. Development of the practice environment scale of the Nursing Work Index. Res. Nurs. Health 25 (3) 176–188. Lake, E.T., 2007. The nursing practice environment. Med. Care Res. Rev. 64 (2 Suppl.) 104S–122S. Lake, E.T., Cheung, R.B., 2006. Are patient falls and pressure ulcers sensitive to nurse staffing? West. J. Nurs. Res. 28, 654–677. Laschinger, H.K.S., 2008. Effect of empowerment on professional practice environments, work satisfaction, and patient care quality: further testing the nursing worklife model. J. Nurs. Care Qual. 23 (4) 322– 330. Leatt, P., Schneck, R., 1984. Criteria for grouping nursing subunits in hospitals. Acad. Manage. J. 27 (1) 150–165. Leatt, P., Schneck, R., 1982. Work environments of different types of nursing subunits. J. Adv. Nurs. 7 (6) 581–594. Li, B., Bruyneel, L., Sermeus, W., Van den Heede, K., Matawie, K., Aiken, L., Lesaffre, E., 2012. Group-level impact of work environment dimensions on burnout experiences among nurses: a multivariate multilevel probit model. Int. J. Nurs. Stud. 50 (2) 281–291. Mark, B.A., 2002. What explains nurses’ perceptions of staffing adequacy? J. Nurs. Adm. 32 (5) 234–242. McClure, M.L., Poulin, M.A., Sovie, M.D., Wandelt, M.A., 1983. Magnet Hospitals: Attraction and Retention of Professional Nurses. Amerian Nurses’ Association, Kansas City, MO. National Quality Forum, 2012. NQF-Endorsed Standards. http:// www.qualityforum.org/QPS/QPSTool.aspx (accessed 08.05.13). Nelson, E.C., Batalden, P.B., Mohr, J.J., Godfrey, M.M., Headrick, L.A., Wasson, J.H., 2002. Microsystems in health care: Part 1. Learning from highperforming front-line clinical units. Joint Comm. J. Qual. Patient Saf. 28 (9) 472–493. Patrician, P.A., Shang, J., Lake, E.T., 2010. Organizational determinants of work outcomes and quality care ratings among Army Medical Department registered nurses. Res. Nurs. Health 33, 99–110. Schmalenberg, C., Kramer, M., 2008. Essentials of a productive nurse work environment. Nurs. Res. 57 (1) 2–13. Shekelle, P.G., 2013. Nurse–patient ratios as a patient safety strategy: a systematic review. Ann. Intern. Med. 158, 404–409. Shrout, P.E., 1998. Measurement reliability and agreement in psychiatry. Stat. Methods Med. Res. 7, 301–317. Staggs, V.S., Dunton, N., 2012. Hospital and unit characteristics associated with nursing turnover include skill mix but not staffing level: an observational cross-sectional study. Int. J. Nurs. Stud. 49, 1138–1145. Tabachnick, B.G., Fidell, L.S., 2007. Using Mltivariate Statistics. Allyn & Bacon, Boston. Unruh, L., 2008. Nurse staffing and patient, nurse, and financial outcomes. Am. J. Nurs. 108 (1) 62–71. Van Bogaert, P., Clarke, S., Roelant, E., Meulemans, H., Van de Heyning, P., 2010. Impacts of unit-level nurse practice environment and burnout on nurse-reported outcomes: a multilevel modelling approach. J. Clin. Nurs. 19 (11–12) 1664–1674. Verran, J.A., Gerber, R.M., Milton, D.A., 1995. Data aggregation: criteria for psychometric evaluation. Res. Nurs. Health 18, 77–80. Warshawsky, N.E., Havens, D.S., 2011. Global use of the practice environment scale of the Nursing Work Index. Nurs. Res. 60 (1) 17–31.

Please cite this article in press as: Choi, J.S., Boyle, D.K., Differences in nursing practice environment among US acute care unit types: A descriptive study. Int. J. Nurs. Stud. (2014), http://dx.doi.org/10.1016/j.ijnurstu.2014.03.001