International Benchmarks for the Practice Environment Scale of the Nursing Work Index: A Meta‐Analysis

ABSTRACT

The Practice Environment Scale of the Nursing Work Index has been used worldwide to measure nurse work environments. International benchmark values for this scale can assist managers in assessing their work environment. The objective was to conduct a meta‐analysis of this instrument's composite and subscale values across continents, nursing unit types, and time. Studies published up until September 30, 2023 were identified in CINAHL, MEDLINE, and Embase. One‐hundred and sixty publications representing 38 countries were included. Most studies were rated as high certainty and low to moderate risk of bias. The pooled point estimate (2.70) indicated that hospital work environments were modestly positive. The weakest work environment domain was Staffing and Resource Adequacy (2.47). Europe had significantly weaker work environments than Asia and North America. Africa and South America had few studies. Better work environments were reported in neonatal intensive care, as compared to medical surgical and critical care units. A small, positive slope over time, which was detected in the three continents, was significant in North America. To promote evidence‐based management globally, benchmark values are now available by setting and continent.

Summary.

• This systematic review and meta‐analysis of mean values for the Practice Environment Scale of the Nursing Work Index reveal modestly positive hospital work environments worldwide.

• Among the five domains comprising this work environment measure, Staffing and Resource Adequacy were consistently rated significantly weaker than the remaining domains.

• To promote evidence‐based management globally, benchmark values are now available by setting and continent.

1. Introduction

The nurse work environment, conceptualized as organizational features that influence a nurse's professional autonomy (Lake 2002; Lake 2007), has been a research focus for several decades. Research has demonstrated that variation in the work environment is significantly related to patient health outcomes and nurse job outcomes.

The global nursing crisis demands the improvement of nurses' work environments internationally. This crisis comprises demand for health services outstripping current supply, an unevenly distributed, under skilled in the context of emerging technologies, aging, and mobile workforce experiencing unprecedented burnout following the pandemic (World Health Organization 2023). Improving working conditions is noted as key to attracting and retaining healthcare workers.

A widely used instrument, the Practice Environment Scale of the Nursing Work Index (PES‐NWI) (Lake 2002), was developed to measure nurses' work environments. This instrument could be utilized to address shortcomings in work environments, but the lack of international benchmarks on the instrument's scores limits managers' capacity to focus improvement efforts. That is, their capacity would increase by having benchmarks for “peer” contexts, e.g., particular continents or nursing unit types, to compare with their unit's subscale and composite values. Thus, a literature gap exists on benchmarks synthesized from extant research.

The PES‐NWI is endorsed as a national quality indicator in the United States (National Quality Forum 2022) and by Health Promotion Switzerland (Switzerland 2024). The instrument statements describe organizational traits, e.g., “Opportunity for staff nurses to participate in policy decisions,” in five subscales: Nurse Participation in Hospital Affairs; Nursing Foundations for Quality of Care; Nurse Manager Ability, Leadership, and Support of Nurses; Staffing and Resource Adequacy; and Collegial Nurse–Physician Relations. Nurses are asked to what extent they agree that the trait is present in their current job. The instrument takes 7 to 10 min to complete. Typically, nurses in the same organization are surveyed annually or biennially.

Research use has focused on linking the work environment to health and job outcomes. Two systematic reviews and one meta‐analysis have summarized this evidence (Lake et al. 2019; Lee and Scott 2018; Warshawsky and Havens 2010). These concluded that most studies identified statistically significant associations between the composite or subscale scores of the PES‐NWI and health and job outcomes as would be theorized. The cross‐sectional evidence has been augmented by two longitudinal studies documenting that work environments change over time and when they do, safety and quality in a hospital change also (Lake et al. 2019; Sloane et al. 2018). In a reliability generalization meta‐analysis (Zangaro and Jones 2019) comprising 51 studies, 19 countries and 80 563 nurses the authors concluded that the instrument is reliable in samples within and outside of the United States.

Benchmark values from international studies may be useful for researchers to compare with their sample values and for managers to compare with their facility values. These values may reflect cultural differences across continents, and care delivery differences across hospital nursing units. Patterns across the subscales may reveal universally stronger or weaker domains of the work environment. Stronger domains may reflect organizational traits that are easier to achieve. Weaker domains may warrant policy, research, or managerial attention. Two decades of evidence may reveal broad changes in work environments over time. Presently, such benchmark values, subscale patterns, and trends are unavailable, demonstrating the gap in understanding about nurse work environments this study addresses.

The purpose of this study was to provide international benchmarks for a widely used nurse work environment instrument, the PES‐NWI. We aimed to compare patterns in the subscale values, values across continents, and nursing unit types and to analyze trends in composite scores.

2. Design

The design was a meta‐analysis, i.e., a technique that statistically combines the results of quantitative studies to provide a more precise effect of the results (Grant and Booth 2009). This design requires an exhaustive search, a certainty appraisal, a risk of bias assessment, and a numerical synthesis. The review was registered in PROSPERO CRD42023399357 in February 2023.

3. Materials & Methods

3.1. Eligibility Criteria

Inclusion criteria were descriptive statistics of the mean and standard deviation (SD) for the composite or a subscale from hospital nurse survey responses on the PES‐NWI and publication in a peer‐reviewed journal. We required samples to have at least 90% registered nurses (RNs). Excluded were data from nursing students or advanced practice nurses. A paper had to report at least one mean value from the composite or a subscale to be included. Papers without a reported SD were included. These criteria allowed the maximum number of studies to be included. The number of observations (papers) for each analysis varied according to extracted reported scores. Papers reporting secondary data analysis were excluded if their sample overlapped with a previous publication reporting descriptive statistics. Papers or subsamples from Magnet or military hospitals were excluded due to their unique characteristics. Magnet hospitals are designated by the American Nurses Credentialing Center for achieving rigorous standards of nursing excellence (American Nurses Credentialing Center 2024). Multiple values were extracted from the same sample if they represented various nursing unit types. In such papers, medical‐surgical unit values were chosen for estimating overall hospital values, because these are the most plentiful unit types.

3.2. Information Sources and Search Strategy

In October 2023, with a university librarian's guidance, MEDLINE, CINAHL, and Embase were searched for the term “Practice Environment Scale of the Nursing Work Index” or the abbreviation, “PES‐NWI.” Relevant studies without date limits were identified.

3.3. Selection Process

Results were processed using COVIDENCE software. After combining search results and excluding duplicates, two co‐authors independently completed the article evaluation for inclusion and exclusion criteria. A third co‐author resolved disagreements.

3.4. Data Collection

Multiple co‐authors extracted the following data, which were reviewed by all authors in team meetings: continent, nursing unit type, year of data collection, number of nurse respondents, response rate (if reported), mean and SD for the composite and any subscale(s), and whether it was primary or secondary data. For studies without year of data collection reported, we contacted the corresponding author to acquire it.

3.5. Risk of Bias Assessment

To appraise risk of bias, three criteria on sources of potential bias were utilized (see Table 1): sample representativeness, response rate, and instrument reliability. Risk of bias ranged from 0 to 6; higher scores indicated lower risk. We classified studies with scores of 5 or 6 as low risk, 2 to 4 as moderate, and 0 to 1 as high. Two co‐authors rated these criteria. Results were reviewed and confirmed by the team.

TABLE 1.

Risk of bias and certainty of evidence rating scales.

Indicators	Response options
Risk of bias
Sample representativeness	0 = sampling approach not stated 1 = convenience 2 = probability or census
Response rate	0 = not reported or low (0 to 32%) 1 = medium (33 to 66%) 2 = high (67 to 100%)
Reliability of the instrument	0 = not reported or low (< 0.70) 1 = medium (0.70 to 0.79) 1.5 = Medium to high (scale range includes values between 0.70 to 0.79 and 0.80 or higher) 2 = high (0.80 or higher)
Certainty of evidence
Research question(s)	0 = not stated, difficult to identify, or minimally stated 1 = clearly stated
Subjects in the sample	0 = not described or minimally described (three or fewer characteristics) 1 = well described (four or more characteristics)
Setting	0 = not described 1 = institution or source of participants described

3.6. Descriptive Statistics (Effect) Measures

The mean and SD for the composite and any subscale(s) were the “effect” measure for the synthesis.

3.7. Synthesis Methods

Studies were eligible for synthesis overall, by continent, and nursing unit type. Missing SDs were imputed from the remaining observations (Copas et al. 2014).

We meta‐analyzed six continuous outcomes: the composite and five subscales. Given the intrinsic heterogeneity in these outcomes across studies, we used univariate random‐effects meta‐analyses to obtain the pooled point estimates and 95% confidence intervals (CIs). Heterogeneity was quantified by the Higgins' and Thompson's heterogeneity measure $I^2$ (Higgins et al. 2003). An $I^2$ value of 0%–25% represents insignificant heterogeneity, 26%–50% represents low, 51%–75% represents moderate, and > 75% represents high. The evidence of statistically significant differences between subgroups was investigated using the Cochran's Q statistics (Cochran 1954). All analyses were performed using the R package “meta” in R, Version 4.2.1 (Schwarzer 2022). We conducted a sensitivity analysis in the studies with complete SD data to assess the results' robustness.

We conducted subgroup analyses for continents and unit types for all outcomes. We classified statistics into medical surgical, adult critical care, neonatal intensive care, and other nursing unit types. “Other” unit type statistics were not meta‐analyzed.

To evaluate work environment changes over time, a meta‐regression on the continuous data collection year for all studies in which the composite hospital value and year were reported or year was acquired (n = 134 or 83% of papers with composite values) was conducted. We analyzed data within continents, to maintain the cultural context of nursing practice and the evaluation of the work environment, for continents with sufficient observations (Asia, North America, Europe). These analyses were conducted for the composite score only, which had the most observations.

3.8. Reporting Bias Assessment

Potential publication bias was not a concern because our focus was descriptive statistics rather than treatment effect sizes in randomized controlled trials.

3.9. Certainty Assessment

To appraise certainty, we utilized three criteria: clear statements of the research question, the subject characteristics, and the source of subjects (Table 1). The certainty ranged from 0 to 3; lower scores indicated lower certainty. We classified studies with scores of 3 as high certainty, 2 as moderate, and 1 as low.

4. Results

4.1. Study Selection

Search results are displayed in the PRISMA diagram (Figure 1). There were 562 unique references. The final analytic sample comprised 166 samples across 160 studies. Table S1 presents the table of evidence displaying overall extracted data (Part 1) and PES‐NWI composite and subscale values (Part 2).

FIGURE 1.

PRISMA diagram.

4.2. Study Characteristics

The included studies, representing 38 countries, were predominantly from Asia (n = 72; 45%), North America (n = 49; 31%), and Europe (n = 30; 19%); few were from Australia (n = 6), Africa (n = 1) or South America (n = 2). All references for the included studies can be found in Table S1.

4.3. Appraised Risk of Bias in Studies

The studies exhibited moderate‐to‐low risk of bias, with average ratings of 1.25 for sample representativeness, 1.26 for response rate, and 1.31 for reliability (of a maximum of 2.00, reflecting least bias). The sample was rated as mostly moderate risk of bias (54%), then low (40%), and high (6%). These ratings likely overestimate potential bias because any study that did not explicitly report a criterion was rated a zero, but the actual representativeness, response rate, and/or reliability value may have been satisfactory.

4.4. Results of Syntheses

Figure 2 and Table S2 display the number of studies, pooled point estimates, and 95% CIs. The overarching pooled point estimate for the composite in 152 studies was 2.70 (95% CI 2.66; 2.75) indicating that hospital nurses worldwide rate their environments as modestly positive. This interpretation is based on 2.70 relative to the scale midpoint (2.50) and theoretical range (1.00 to 4.00), reflecting the responses strongly disagree = 1 through strongly agree = 4. Higher scores reflect greater agreement that valued organizational traits are present in the nurse's job. If all nurses in an institution strongly disagree that all valued traits are present, the composite value would be equal to 1.00. Accordingly, a mean value of 2.70 exceeds the neutral midpoint but falls short of the value of 3.00 (“agree”).

FIGURE 2.

Point estimates from meta‐analysis of practice environment scale of the nursing work index composite and subscales in hospital nurses.

Fewer studies reported subscale values (n = 130 to 134). The most highly rated domains were Nursing Foundations for Quality of Care (mean of 2.84) and Collegial Nurse‐Physician Relations (2.83). Nurse Manager Ability, Leadership and Support (2.76) also exceeded the composite mean. Two subscales fell below the composite: Nurse Participation in Hospital Affairs (2.57) and Staffing and Resource Adequacy (2.47). Notably, this last subscale fell in the unfavorable range, below the neutral midpoint, 2.50. Results of the sensitivity analysis in the subgroup with complete SD data were essentially identical.

Continent results are shown in Figures 3 and 4 and Table S3. The highest mean scores were for North America (2.76) and Asia (2.75). The lowest were for Europe (2.54), significantly lower than both continents (both p values < 0.001). In all continents, the lowest rated subscale was Staffing and Resource Adequacy and the highest was Nursing Foundations (Figure 2). The second highest subscale in Asia, Australia, and North America was Nurse‐Physician Collegial Relations. In Europe, the second highest was Nurse Manager Traits. There were insufficient studies in Africa and South America for valid comparisons.

FIGURE 3.

Point estimates and confidence intervals from meta‐analysis of practice environment scale of the nursing work index composite in hospitals by continent.

FIGURE 4.

Bubble plots of composite score (y axis) by data collection year (x axis); slope from meta‐regression.

Table 2 presents results by inpatient unit type: medical surgical (n = 14), neonatal intensive care (n = 4), and adult critical care (n = 14). Here, the composite value was highest in neonatal intensive care (2.93), middling in medical surgical (2.74), and lowest in critical care (2.69). These differences were nonsignificant. In all unit types, the two lowest rated subscales were Staffing and Resource Adequacy and Nurse Participation in Hospital Affairs. The Staffing and Resource Adequacy value was lowest for medical‐surgical units (2.44), then adult critical care (2.58), and then neonatal intensive care (2.77) (p < 0.05). In medical‐surgical units, the highest rated subscale was for Nursing Foundations for Quality of Care. In both critical care settings, the highest rated subscale was Nurse‐Physician Collegial Relations.

TABLE 2.

Point estimates from meta‐analysis of the practice environment scale of the nursing work index composite and subscales by nursing unit type.

	Medical‐surgical	Neonatal Intensive Care Unit (NICU)	Adult Intensive Care Unit (ICU)
	Point estimate [95% CI] Number of studies
Composite	2.74 [2.61; 2.86] 14	2.93 [2.77; 3.08] 4	2.69 [2.54; 2.84] 14
Nurse Participation in Hospital Affairs	2.59 [2.43; 2.75] 10	2.67 [2.13; 3.22] 2	2.53 [2.33; 2.74] 11
Nursing Foundations for Quality of Care	2.95 [2.82; 3.08] 12	3.06 [2.81; 3.30] 3	2.79 [2.59; 2.99] 11
Nurse Manager Ability, Leadership, and Support of Nurses	2.84 [2.73; 2.95] 12	2.85 [2.73; 2.97] 4	2.69 [2.48; 2.90] 12
Staffing and Resource Adequacy	2.44 [2.31; 2.57] 12	2.77 [2.54; 3.01] 4	2.58 [2.41; 2.75] 12
Collegial Nurse‐Physician Relations	2.84 [2.70; 2.97] 11	3.09 [2.91; 3.26] 4	2.91 [2.74; 3.08] 13

Note: ${\mathrm{I}}^2$ values within each subgroup of the meta‐analyses on all outcomes are > 80.2%, the p of Cochran's Q test: Comp: 0.0699; Participation: 0.8545; Found: 0.2260; Manager: 0.4154; Staffing: 0.0412; Nurse: 0.0880.

Figure 4 displays a meta‐regression on data collection year for samples with this variable (n = 129; 90% of papers with composite values) in Asia (n = 59), North America (n = 41), and Europe (n = 29). The x axis spans the years 1985 to 2020. These plots show a slight positive slope (β = 0.01) for all continents, which was statistically significant only in North America.

4.5. Certainty of Evidence

The included studies exhibited high certainty, reflected by average ratings of 0.99, 0.95, and 1.00 (of a maximum of 1.00) for research question, sample description, and setting.

5. Discussion

We were motivated to generate international benchmark values on the most widely used instrument to measure nurses' work environments. We identified 160 studies, predominately from Asia, North America, and Europe. Using this scale, work environments were rated as modestly positive overall. The value of 2.70 demonstrates suboptimal performance worldwide. The results shed light on the instrument's capacity to discern performance variation across and within facilities to support quality through nursing.

Two prior systematic reviews on this instrument, including 37 (Warshawsky and Havens 2010) and 46 (Swiger et al. 2017) studies, reported descriptive statistics in ranges, making comparison to our point estimates difficult. Within our study, the lowest rated subscale was Staffing and Resource Adequacy (2.47), below the neutral midpoint, reflecting disagreement that the elements in this subscale are satisfactory. These elements comprise enough staff to get the work done, RNs to provide quality patient care, support services that allow me to spend time with my patients, and time and opportunity to discuss patient care problems with other nurses. Our results are consistent with both the above reviews and a scoping review of 20 Latin American studies (de Oliveira Riboldi et al. 2021) [13 Portuguese and 7 English], including five utilizing the PES‐NWI, reporting that the Staffing and Resource Adequacy subscale was most often the lowest‐scored subscale. Both also found the Nursing Foundations for Quality and Nurse‐Physician Collegial Relations were the highest rated, consistent with our results from 130 or more studies.

Swiger et al. (2017) differentiated scores in non‐Magnet, emerging/aspiring Magnet, and Magnet subgroups from three studies (Kutney‐Lee et al. 2015; Lake and Friese 2006; Ma and Park 2015). Comparing their ranges with our point estimates, we observed that our estimates align with values for non‐Magnet hospitals for all subscales/composite. Only our point estimate for the Nurse Manager subscale crossed into the values from the Magnet hospitals. We infer that hospitals worldwide reflect work environment conditions/statuses like U.S. non‐Magnets; they do not resemble Magnet hospitals.

Our comparative results across continents indicate that the best hospital work environments are in North America and Asia. Those in Europe are decidedly poorer. There were too few studies in Africa and South America to draw conclusions. Across time within‐continent, slight, significant improvements in North America, but non‐significant in Europe and Asia over the two to three decades of evidence, were noted. We posit that forces in the United States, that is, the Magnet Hospital Program, and potentially the Pathway to Excellence Program, which is researched less frequently, account for U.S. work environment improvements. A European multi‐country research consortium linking work environments to outcomes may partly explain improvements noted in Europe (Sermeus et al. 2011). Engagement in such a consortium motivates improvements in work environments to improve quality and outcomes.

By nursing unit type, sufficient numbers of studies were available on medical‐surgical and adult critical care units. Whereas SRA was lowest for medical‐surgical units (2.48) [below the neutral midpoint, indicating unsatisfactory], that of adult critical care was relatively higher (2.59). These results are consistent with a comparative descriptive study comparing workloads on medical‐surgical and critical care units, finding that medical surgical but not critical care were understaffed with RNs (Hughes et al. 2015). Similarly, Nurse‐Physician Collegial Relations was slightly lower in medical‐surgical (2.88), than in critical care (2.90). Conversely, the remaining subscales were higher on medical surgical but lower in critical care. By contrast, the work environments in neonatal intensive care are uniformly superior, that is, the composite and all subscale values exceeded the other unit types. These differences reveal that the work environment domains differ across nursing unit types and can inform managers in these settings about which areas to focus on improvement efforts.

Our analyses spanning two decades revealed slight upward slopes in all three continents with sufficient observations to analyze. Formal trends would be revealed in longitudinal panel studies, which are uncommon due to logistical challenges. Two studies support the North American trend identified here that work environments have improved over these decades. Kutney‐Lee et al. (2013) found in a sample of 137 Pennsylvania hospitals, from 1999 to 2006, that most hospitals (39%) had an improved work environment, as compared to those unchanged (28%) or worsened (33%). Lake, Riman, and Sloane (2020) analyzed data from nurses in 458 hospitals in four U.S. states in 2006 and 2016. They found that the average hospital's work environment composite score increased. These studies' sequential time periods (1999 to 2006; 2006 to 2016) encompass most years analyzed herein, confirming the positive trend in our results.

Over the three decades of data reflected herein, the hospital setting has increased dramatically in complexity. This includes increases in patient acuity, extensive quality improvement initiatives, patient safety and satisfaction monitoring, electronic health record implementation, and nurse staffing and overtime standards. The underlying work environment may support or hinder managers to address these challenges successfully.

Regarding study certainty and risk of bias, nearly all included studies were graded high certainty and moderate or low risk of bias. These ratings are nearly identical to a meta‐analysis of reliability coefficients from 51 reports using this instrument (Zangaro and Jones 2019).

This meta‐analysis did not account for potential confounding variables that could influence the work environment scores, such as cultural differences in nursing practices, assessment of the work environment, or healthcare policies across countries. The estimated values likely reflect these distinctions. Continent‐specific benchmarks are important to hold constant the cultural context. Decades of evidence globally show consistent relationships between the work environment, utilizing the PES‐NWI, and patient and nurse outcomes (Lake et al. 2019; Nascimento and Jesus 2020; Wei et al. 2018; Zhao et al. 2019), indicating that these phenomena transcend cultural norms and policy variation. Even if the observed PES‐NWI composite value reflects these differences, the instrument properly captures the construct of interest.

5.1. Implications for Management

Given a worldwide modestly positive rating, the message to policymakers, hospital administrators, and managers is that attention to work environments is needed and overdue. What actions can policymakers take to foster work environment improvements? An example in Switzerland is the Health Promotion Switzerland initiative, a national non‐profit organization that provides employers with surveys, including the Staffing and Resource Adequacy subscale of the PES‐NWI, to monitor workplace health management to improve employees' mental health and reduce stress (Switzerland 2024). A U.S. example is the Leapfrog Group, a non‐profit organization developed in 2000 by employers and purchasers to obtain transparency about hospitals' quality and safety practices (Milstein et al. 2000). Leapfrog publicly reports hospital performance from annual surveys across four levels, from “limited achievement” to “achieved the standard.” This group plans to begin surveying members utilizing the PES‐NWI in 2025 to expand their nursing workforce reporting. Similar entities in other countries could report on their nurses' work environments. Currently, in the state of Colorado, all large hospitals are required to collect data using the PES‐NWI annually (Colorado Hospital Quality Report by Measure 2022). Therefore, state or country hospital associations could require similar surveys and publicly report the results.

Another approach in the United States would be to add requirements to the Medicare Conditions of Participation, that is, a regulatory approach (Centers for Medicare and Medicaid Services 2022a). Hospitals must fulfill these to receive payment for the hospital care of older Americans (Medicare beneficiaries). The 2019 Conditions of Participation Nursing Services standards require an RN to direct nursing services and RNs to have valid licenses, to supervise and evaluate each patient's nursing care and to assign patients to nurses based on patient need, staff qualifications, and competence (Centers for Medicare and Medicaid Services 2022b).

These Conditions of Participation could be expanded as recommended by a recent think tank addressing pandemic challenges in nursing care delivery, which recommended Conditions of Participation changes for organizations to regularly assess the health of the work environment and demonstrate evidence of continual improvement (Partners for Nurse Staffing Think Tank 2022). This think tank represents national nursing, financial management, and healthcare improvement organizations. Ministries of health in countries with public hospital systems could require reporting of nurse work environments.

Hospital administrators can join a voluntary hospital network to monitor nursing performance, including the work environment. The U.S. National Database of Nursing Quality Indicators includes data from RNs surveyed every year or two in participating hospitals (Press Ganey 2022). Although about one‐third of U.S. hospitals (i.e., about 2000) participate, most measure human resource indicators such as RN hours per patient day, or clinical indicators such as patient falls per 1000 patient days. Relatively few (~600) conduct RN surveys including the PES‐NWI. Hospitals pay tens of thousands of dollars to participate in this network. Joining such a network provides a mechanism for administrators to optimize outcomes through nursing quality metrics. U.S. programs that foster work environment improvements, such as the Magnet Program and Pathway to Excellence, deserve consideration in other continents. A promising approach is the European Union‐funded study, Magnet4Europe, an intervention that twins European hospitals with U.S. Magnet hospitals (Sermeus et al. 2022).

Nurse managers are pivotal to the creation and maintenance of favorable work environments. Wei et al. (2018) in a systematic review of 54 work environments studies in the United States published between 2005 and 2017 identified nurse leadership as the foundation of a healthy work environment. If managers are in the networks mentioned previously, they will receive reports benchmarking their work environment scores against similar nursing units in peer hospitals. Most nurse managers do not have such data available. Nurse managers could request their supervisors' support to survey their nurses periodically and utilize the values published in this paper for benchmarking. They could interpret the results in the context of this paper's findings and then make changes where needed. A creative use of the Nurse Manager subscale by nursing administrators in a health system was to identify the most highly rated managers in their system and conduct focus groups with them to identify, then share, key elements in their managerial success (Anderson et al. 2010). Values of respect and empathy and nursing leadership characteristics of visibility and communication were identified by the participants. Because effective strategies in managerial success are likely influenced by a health system's culture, this approach may be replicated and applied in other health systems to improve the effectiveness of their nurse managers.

The newly published, validated five‐item version of the PES‐NWI (the PES‐5) (Lake et al. 2024) offers a much less burdensome survey but without the capacity to measure distinct subscales. For some of the implications noted above, it would be satisfactory, but for others, for example, addressing specific domains, the full instrument or subscale is required.

5.2. Implications for Research

What next research steps are needed to advance work environments worldwide? This meta‐analysis revealed geographic gaps. Researchers and research sponsors should support evaluation of work environments in Africa and South America. Regarding variation across unit type within hospitals, the available evidence supported benchmarks in medical‐surgical, neonatal intensive care and adult critical care units. Evidence is lacking from other common unit types including labor and delivery, emergency, and surgery/post‐anesthesia.

5.3. Limitations and Strengths

The exclusion of publications in non‐English languages was a limitation. The trend analysis excluded studies (17%) without year of data collection, a potential bias. Studies with low survey response rates may have yielded biased values. Some studies (~17%) did not report response rates. The large I ² values reflect large heterogeneity in the empirical literature, potentially limiting the ability to generalize estimates across settings and continents for practice and policy purposes.

This study's many strengths included the databases searched, efforts to maximize the inclusion of studies, and the comparative descriptive design. Three large databases relevant to nursing were searched. Efforts to include studies otherwise excluded comprised acquiring year of data collection for trend analysis and imputing unreported SDs. Multiple comparisons (continent, nursing unit type) make the results informative to various stakeholders. We assert that these strengths greatly outweigh the noted limitations.

6. Conclusion

Hospital nurse work environments globally are rated by nurses as modestly positive. Staffing and Resource Adequacy is consistently rated lowest. The researchers found that mean values for the PES‐NWI composite or subscales vary statistically significantly across continents and nursing unit types. Small improvements in the work environment are evident globally over the past two decades, significantly in North America. The cumulative evidence points to the capacity of the instrument to differentiate across geographic areas. Due to data gaps, evaluation of work environments in Africa and South America is needed.

The overarching conclusion is that work environments warrant improvement internationally to maximize the productivity of the nursing workforce. Staffing and Resource Adequacy require the most attention. To promote evidence‐based management globally, benchmark values are now available by setting and continent. As the post‐pandemic era presents challenges and opportunities for workforce retention and stabilization, the instrument can provide managers with needed feedback. In the coming years, attention should shift to evidence‐based approaches to work environment improvement.

6.1. Relevance for Clinical Practice

Our findings that work environments are only modestly positive worldwide have similar implications for practicing and student nurses. The work environment warrants attention and support from their managers. Although practicing nurses may recognize that their work conditions are poor, they may be unaware of the “work environment” concept. If nurses understand that the work environment is the foundation for high quality nursing practice and patient outcomes, work environments may improve. This could occur if nurses are motivated to complete relevant surveys and participate in improvement efforts. Likewise, if they are empowered to expect a better work environment and to address it with their manager there could be subsequent changes in outcomes.

Understanding these concepts is important for student nurses and should be incorporated into nursing school curricula. This may include content on the nurse work environment presented through simulations or vignettes, which could exhibit poor versus favorable environments and their related domains, such as effective nurse managers and non‐collegial relationships. Care philosophies that reflect a biomedical versus nursing perspective could also be incorporated into these learning situations.

Author Contributions

Eileen T. Lake: conceptualization, data curation, writing – review and editing, writing – original draft, funding acquisition, investigation, methodology, project administration. Bingyu Zhang: conceptualization, methodology, investigation, formal analysis. Jiayi Tong: conceptualization, investigation, methodology, formal analysis, writing – original draft, writing – review and editing. Gloria Mpundu: conceptualization, data curation, investigation, validation, writing – original draft, writing – review and editing. Kelsey N. Gross: conceptualization, data curation, investigation, writing – original draft, writing – review and editing. Malini H. Correa: conceptualization, data curation, investigation, validation, writing – original draft, writing – review and editing. Lauren B. Wethman: conceptualization, data curation, investigation, validation, writing – original draft, writing – review and editing. Lynne Moronski: conceptualization, data curation, investigation, validation, writing – original draft, writing – review and editing. Domenique Villani: conceptualization, writing – review and editing, methodology. Yong Chen: investigation, formal analysis, methodology, writing – original draft, writing – review and editing.

Conflicts of Interest

Dr. Lake is the developer of the Practice Environment Scale of the Nursing Work Index. The remaining authors have no potential conflicts to declare.

Supporting information

Supplementary Table S1. Table of Evidence—Part 1: Extracted Overall Data.

Supplementary Table S2. Point Estimates from Meta‐analysis of Practice Environment Scale of the Nursing Work Index Composite and Subscales in Hospital Nurses.

Supplementary Table S3. Point Estimates from Meta‐analysis of Practice Environment Scale of the Nursing Work Index Composite and Subscales in Hospital Nurses by Continent.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.