Prevalence of turnover intention among emergency nurses worldwide: a meta-analysis

The PubMed, Web of Science, Embase, CINAHL, and Cochrane Library databases were searched for related articles from the inception of these databases to February 1, 2023. To conduct a comprehensive review, a search strategy was implemented that combined three keyword groups, specifically focusing on (a) turnover intention, (b) nurses, the study population, and (c) the emergency department occupational environment. The search terms included nurses [MeSH term]) OR nurse * OR nursing care OR nursing staff [Title/Abstract]) AND (emergency services, hospital [MeSH term]) OR (emergency department OR emergency unit OR emergency [Title/Abstract]) AND (Personnel Turnover [MeSH Terms] OR (Intention to leave OR Turnover intention OR Intention to quit OR Turnover [Title/Abstract])). Appendix A provides the search strategies used for each database. Furthermore, the comprehensiveness of the search was ensured by meticulously reviewing the reference lists of related articles and all included studies.

Inclusion criteria: According to the PRISMA statement, the following inclusion criteria were established utilizing the PICOS framework: Participants (P): nurses in emergency settings; Intervention (I): not applicable; Comparison (C): not applicable; Outcomes (O): providing the overall turnover intention prevalence rate or containing sufficient original data to calculate the prevalence rate of turnover intention. Study design (S): cross-sectional study and longitudinal study. If there was a cohort of nurses with overlapping studies, the study that included the most significant number of nurses was chosen.

The exclusion criteria were as follows: (1) no peer review or original articles, including reviews, editorials, notes, letters, case reports, meeting minutes, books, news, unpublished papers, or dissertations; (2) unable to obtain full-text research; (3) research that cannot obtain complete data; and (4) articles published in languages other than English.

The outcomes of the systematic search were incorporated into the reference manager (EndNote X9). After eliminating duplicate studies, XYC and XWH, two authors, independently assessed eligible publications by screening titles and abstracts following the inclusion and exclusion criteria. These articles were included in the full-text evaluation when at least one author deemed the abstract eligible. The articles’ full texts were assessed independently by two authors for final inclusion, and any disagreements between them were reconciled by the third reviewer’s (RH) opinions and the three authors’ consensus.

The two reviewers independently extracted the following data via standardized data tables: author, publication year, survey time, country/region, research type, sampling method, survey method, sample size, prevalence of turnover intention, data source, and participant characteristics (average age). The extracted data were validated by a third reviewer (RH).

The risk of bias assessment tool, developed by Hoy and colleagues to determine the internal and external validity of prevalence studies, was used by two investigators to assess the quality of each study reviewed [17]. The tool comprises 10 items and is segregated into two components: the external validity subscale (comprising 4 items) and the internal validity subscale (comprising 6 items). Each item was assigned a score of 1 (‘yes’ for ‘high quality’) or 0 (‘no’ for ‘low quality’). The total score for each study was determined by adding the scores of all 10 items, which were ranked from 1 to 10. The studies were classified into three groups on the basis of the total score: high quality (≥ 9), medium quality (6–8), and low quality (0–5).

The primary finding of this systematic review was the prevalence of turnover intentions among emergency nurses, with a confidence interval of 95%. All the statistical analyses were performed via Stata (version 17.0). Heterogeneity between studies was assessed via I² statistics, with I² values of 25%, 50%, and 75% considered low, moderate, and high heterogeneity, respectively [18]. If any heterogeneity was present, a random effects model was employed to compute the pooled prevalence and 95% confidence interval of the turnover intentions of emergency nurses. Otherwise, a fixed effects model was employed. The research results and characteristics are illustrated as a forest map. The publication bias was examined by applying a funnel plot, Egger test, and Begg test [19, 20].

To investigate possible sources of heterogeneity, a subgroup analysis was conducted in this study to explore the prevalence of turnover intentions among emergency nurses. The data were classified into subgroups on the basis of various factors, including sample size, age, region, survey time, publication time, and evaluation tools. Pooled estimates of the prevalence of turnover intention were subsequently calculated with 95% confidence intervals. According to the Cochrane Handbook, every subgroup should contain a minimum of four studies [21]. Hence, an analysis will only be conducted on a subset if it comprises more than four studies.

This study identified 744 literature records from five databases and eliminated 201 duplicates. After the title abstracts were read, 241 out of the 543 studies were eligible for inclusion in the full-text evaluation. Among the 47 studies, the full texts could not be obtained, whereas 87 were excluded because they did not conform to the study types, 69 studies could not extract data, and 26 articles were of nonconforming language types. Finally, this study included 12 studies [9, 22‐32]. The details of the screening process are shown in Fig. 1.

The basic information of the 12 studies included in this study is shown in Table 1. A total of 4400 participants from nine countries were investigated, and the sample sizes of the 12 studies ranged from 102–976. Most of the included subjects were female nurses, 3575 out of 4290 (84%), except for one document that did not specify the sex ratio. All studies included in the analysis were cross-sectional, with no other types of studies meeting the eligibility criteria for inclusion. The included literature was published between 2012 and 2022. Among them, six (50.00%) were conducted in Western Pacific countries (China, South Korea, Australia), two (16.66%) were conducted in American countries (the United States, Canada), and three (25.00%) were conducted in European countries (Belgium, the Netherlands, Turkey). One study was conducted in Ethiopia, an African country. The prevalence of turnover intention was calculated via the following equation:

The bias risk assessment tool developed by Hoy and colleagues was used to assess the methodological quality of the 12 studies included in this study. The details of the assessment process are displayed in the appendix. One study was rated as high quality (9 points, 8.3%), whereas the other 11 were evaluated as moderate quality, with scores ranging from 6–8 points (91.7%).

For the meta-analysis, 12 studies were included in total. There are numerous methods used to evaluate turnover intention, with the turnover intention questionnaire (TIQ) (16.7%) [22, 28] and Leiden quality of work‒life questionnaire for nurses (LQWQ-N) (16.7%) [33, 34] being the most frequently utilized tools. On the basis of the references and the classification criteria for turnover intention, two researchers, a third researcher and two nursing management experts, distinguished individuals with turnover intention into ‘high’ and ‘extremely high’ groups [35]. One study objectively evaluated turnover intention and reported that 90.2% of emergency nurses, with an average age of 29.49 years, had turnover intention [22]. It represents the highest prevalence of turnover intention among the included studies. When publicly available tools such as the TIQ [35], Price and Mueller’s [36], MDS [37], LQWQ-N [38], TIS [39], and Scale Assessing the Intention to Resign or Remain at the Hospital [40] are used to assess turnover intention objectively, the range of turnover intention varies from 24.3% to 90.2%. In contrast, the prevalence of turnover intention in studies utilizing homemade questionnaires ranges from 23.0% to 77.5%. A meta-analysis of the prevalence of turnover intentions among emergency department nurses can be conducted using data from 12 studies and 4400 participants.

The prevalence of turnover intentions among emergency nurses in 12 studies varied from 23.0% to 90.2%. A random effects model was employed because of the substantial heterogeneity (I² = 98.89%, P < 0.001). The pooled prevalence of turnover intention among emergency nurses was 45% (95% CI: 0.32, 0.58). Figure 2 shows the forest plots obtained from the meta-analysis. The funnel plot (Fig. 3) appeared symmetrical during visual inspection, and the results of the Egger test (P = 0.2541) and Begg test (P = 0.1926) indicated the absence of publication bias.

A sensitivity analysis was performed on the 12 included studies, as depicted in Fig. 4, and Stata 17 software was used to evaluate the robustness of the meta-analysis outcomes. By sequentially excluding each study, we observed no significant variation in the pooled effect size relative to the overall combined estimate. This consistency suggests that the findings of the present study exhibit a high degree of reliability and stability.

Subgroup analyses were conducted to investigate the origins of heterogeneity on the basis of factors such as age, sample size, publication time, survey time, region, and assessment instrument (Table 2).

The subgroup analysis of the sample size results indicated that the prevalence rate of turnover intentions among emergency nurses was 43% (95% CI: 24–63%) in studies with sample sizes greater than 277. In contrast, it was 46% (95% CI: 28–64%) in studies with a sample size of less than 277 (Fig. 5). The results of the subgroup analysis on the investigation time indicated that the prevalence rate of turnover intention from 2016–2022 was 55% (95% CI: 37–72%), whereas from 2009–2016, it was 37% (95% CI: 17–58%) (Fig. 6).

An analysis of subgroups based on region revealed that the prevalence of turnover intention was 54% (95% CI: 28–81%) in Asia, whereas in other areas, it was 38% (95% CI: 27–48%) (Fig. 7). The time frame of the publication revealed that the prevalence of turnover intentions among emergency nurses was 53% (95% CI: 35–71%) between 2020 and 2022 and 33% (95% CI: 22–45%) from 2012–2020 (Fig. 8). The subgroup analysis of assessment instrument revealed that the prevalence of turnover intention was 41% (95% CI: 24–58%) in studies utilizing published scales, whereas it was 52% (95% CI: 28–76%) in studies employing self-developed scales (Fig. 9).

Furthermore, subgroup meta-analysis revealed notable disparities in the age-based turnover intentions of emergency nurses. The prevalence of turnover intention reported by studies with an average age of 35.6 years or older (30%; 95% CI: 26–33%) was often lower than that reported by nurses with an average age of less than 35.6 years (61%; 95% CI: 31–92%), and there was a significant difference (P < 0.05) (Fig. 10).

Because of the significant heterogeneity observed among the studies included, subgroup analysis was employed in this meta-analysis to investigate the potential sources of differences. Subgroup analysis revealed that Asian nurses had the highest combined prevalence of turnover intention (54%), whereas other regions (Europe, North America, Australia, and Africa) presented considerable heterogeneity, with only 38% prevalence. Nurses in the Americas demonstrate a comparatively lower propensity for turnover intention, which may be attributed to the robust economic conditions and the high priority placed on nurse retention management in developed countries, such as the United States and Canada [47]. Notably, only one study was conducted in Africa, revealing relatively high turnover intention (77%). Moreover, only a single study was carried out in Oceania, indicating significant turnover intention (48%), and no studies conducted in South America were identified in the search. This suggests that the demand for research into the prevalence of nurse turnover intentions in this region remains unmet. To assess the global prevalence of turnover intentions among emergency nurses more accurately, further exploration informed by existing research will be necessary for future studies.

The results of the subgroup analysis by region indicated that the prevalence of turnover intention among emergency nurses in Asia was somewhat higher than that in other regions. This phenomenon may be intricately linked to regional economic advancement and the development and establishment of healthcare infrastructure. In numerous developing and resource-poor countries, skilled and adequately trained nurses tend to migrate to developed nations in search of improved salaries, working conditions, and professional growth opportunities, thereby influencing the mobility of nurses within a particular geographical region [48]. The prevalence of turnover intentions among emergency nurses in Asian countries was significant. Economic underdevelopment and the development of healthcare systems in Asia, compounded by its large population, pose substantial challenges for providing adequate medical and nursing services [49, 50]. Moreover, Asia has been the epicenter for most emerging and reemerging infectious diseases, including rotavirus, SARS, and the novel coronavirus (COVID-19) [51, 52]. These health threats have further taxed an already strained healthcare system, exacerbating the need for robust and resilient medical and nursing services [53, 54]. Additionally, compared with the Americas, Europe, Oceania, and Africa, there is a relative abundance of research on the turnover intentions of emergency nurses in Asia, which may introduce a certain degree of bias. Therefore, future studies should aim to carry out more multicenter cross-sectional studies, thereby increasing the number of original studies and exploring the variations among different geographical regions or countries.

Despite the use of a range of tools with established reliability and validity to measure turnover intentions, there is currently no consensus on the assessment tools for turnover intentions. Among them, the Turnover Intention Questionnaire (TIQ) developed by Michaels and Spector in 1982 and the Leiden Quality of Work Life Questionnaire for Nurses (LQWLQ-N) developed by Maes and others in 1999 are the most widely used. For subgroup analysis, this study categorizes turnover intentions into two distinct groups: those assessed with published scales that have been validated and are widely used and those evaluated through self-devised questionnaires. The results showed that, compared with studies using published scales (41%), studies using self-developed questionnaires had a slightly greater prevalence of turnover intention (52%). The possible reason for this result may be that the homemade questionnaire may not have undergone rigorous testing for reliability and validity, thus leading to a biased calculation of the prevalence of turnover intention among this population. Owing to the uncertainty in the definition of intention to turnover and the differences in evaluation tools, more generalizable and reliable tools may be developed in the future [55]. This scale accurately assesses the turnover intention of the emergency nurse population.

The subgroup analysis results by age indicated that the turnover intention prevalence among emergency nurses in studies older than 35.6 years was 30%. In contrast, in studies younger than 35.6 years, the prevalence was 61%, with a significant difference observed (P < 0.05). These findings are similar to those of a previous study on the prevalence of turnover intention among nurses in the intensive care unit [56]. A survey indicated that young, newly graduated nurses consistently exhibit a high intention to turnover, with potential reasons including heavy workload; poor communication with patients, family members, or team members; and a lack of skills and knowledge [57]. For young nurses, balancing work and personal life is equally important. Suppose that work interferes with their family life, social activities, or other personal interests. This can lead to decreased job satisfaction and increased turnover intention [58]. Young nurses often harbor high expectations for professional development and advancement opportunities early in their careers. If they perceive a lack of developmental opportunities or unclear avenues for promotion in their current work environment, they may contemplate leaving to pursue more favourable platforms for career growth [59]. This penchant for exploring career advancements may contribute to the higher prevalence of turnover intentions observed among this demographic. If many young nurses leave the nursing profession early in their careers, the hospital will suffer significantly because of unfavourable staff turnover and the high costs it entails [60]. Therefore, nursing leaders should pay greater attention to this phenomenon and implement affirmative measures.

The findings of this study reveal a significant upwards trend in the prevalence of nurses’ intention to leave in recent years, increasing from 37 to 55%. This surge reflects the complex interplay between individual, professional, and systemic factors that warrant comprehensive examination. First, as professional demands for mastering new technologies increase, some nurses may feel that their intrinsic motivation and sense of professional accomplishment are being overlooked, directly affecting their job satisfaction and willingness to remain in the nursing profession [61]. Second, the COVID-19 pandemic has highlighted systemic weaknesses in healthcare and exacerbated existing stressors within the nursing industry [62]. The dramatic increase in workload, the complexity of patient conditions, heightened individual infection risks, and increased emotional stress are realities that nurses faced during the pandemic [63]. This global health crisis has subjected nurses to prolonged exposure to high-stress environments, leading to burnout, emotional exhaustion, and a greater propensity to leave their jobs [64, 65]. Moreover, chronic understaffing in many healthcare facilities forces nurses to work longer hours with heavier patient loads, lowering the quality of patient care and increasing job dissatisfaction and physical fatigue. These conditions make it increasingly difficult for many nurses to continue in such demanding roles [66, 67]. Additionally, demanding schedules, including night shifts and overtime, disrupt nurses’ personal and family lives. The inability to achieve sustainable work‒life balance may drive nurses to seek employment in less demanding, more flexible environments [65, 68]. Furthermore, despite the pivotal role that nurses play in healthcare services, many nurses feel undercompensated and undervalued in their work, coupled with limited career advancement opportunities, which in turn affects their professional commitment and intention to stay [66, 69‐71]. Finally, the generational shift in the workforce has led to different expectations and career aspirations. Compared with previous generations, the current cohort of younger nurses, including Millennials and Generation Z nurses, typically places greater emphasis on work‒life balance, job satisfaction, and career advancement opportunities [69, 72, 73]. If these needs are not met, they are more likely to seek better opportunities elsewhere by leaving their current positions [66].

Subgroup analysis by publication date revealed a significant increase in nurses’ intent to leave, with a prevalence rate of 33% in studies published before 2020, which increased to 53% after 2020. This increase may be attributed to the intensified pressures of the COVID-19 pandemic, which thrust nurses into managing surging patient loads and the emotional distress of patient families. The high-pressure context has led to a surge in professional burnout, anxiety, and depression, driving the heightened intent to turnover [74]. The early pandemic phase’s PPE shortage forced nurses to work at increased infection risk, impacting their health and mental well-being and furthering their consideration of leaving the profession [75]. Prolonged and irregular work hours, a consequence of pandemic-induced uncertainties, have disrupted nurses’ family lives, causing feelings of guilt and dissatisfaction that contribute to their intent to turnover [76]. Despite being lauded as ‘heroes’, nurses may have experienced a shortfall in tangible support and recognition, leading to isolation, reduced job satisfaction, and a stronger desire for turnover [77]. Rapid medical policy and protocol adjustments have added to nurses’ workload, increasing the risk of burnout and turnover [78]. Moreover, a job’s physical demands, including chronic fatigue and musculoskeletal disorders from sustained high-intensity work, have exacerbated the situation, prompting more nurses to contemplate turnover [79].

Addressing these issues necessitates a comprehensive approach. This approach encompasses but is not limited to improving working conditions, offering competitive compensation, distributing reasonable workloads, ensuring adequate rest periods, fostering a healthy work environment, and providing opportunities for professional growth and development [80, 81]. Furthermore, societal perceptions and attitudes toward nursing work also need to evolve to better recognize that nursing work also needs to evolve to better recognize and respect the professional value of nurses [82]. Additionally, relevant departments should formulate contingency plans for special situations such as pandemics to alleviate nurses’ workload, provide necessary personal protective equipment, and implement infection control measures. Policymakers, healthcare administrators, and society must work together to increase nurses’ job satisfaction, reduce the intention to leave, and ensure the stability of the healthcare system and the continuity of nursing services by developing and implementing effective strategies [69, 83]. Personalized nursing management programs that align with the developmental trends of young nurses should be implemented [73]. This is not only an investment in the individual career trajectories of nurses but also an investment in the sustainable development of the entire healthcare system.