Background

The coronavirus disease 2019 (COVID-19) pandemic began in late 2019, with an estimated 400 000 000 infections and 6 200 000 deaths caused by the disease as of April 2022 [1]. Several coronavirus variants surfaced throughout the pandemic, resulting in repeated waves of widespread infections in countries worldwide. The severity of COVID-19 infections, coupled with the large number of cases, placed immense pressure on healthcare systems as high volumes of patients in need of acute treatment required hospitalisation; some regions suffered from high COVID-19 mortality rates due to shortages of medical workers and equipment [2].

A major contributing factor to the shortages seen in the health workforce is employee turnover [3]. In the context of the healthcare sector, turnover intention refers to the willingness of healthcare workers (HCWs) to leave their positions of employment for other positions in either the same or different professions [4]. The COVID-19 pandemic has caused an unprecedented wave of resignations. In the United States, the healthcare sector suffered a net loss of 460 000 workers between February 2020 and November 2021 [5]. A survey of 1000 American HCWs revealed that 18% of them left their jobs over the course of the pandemic, citing it as one of the driving factors behind their resignation [6]. Resignation rates of HCWs in Singapore spiked in 2021, and this was driven by both foreign workers looking to migrate, as well as local workers experiencing severe levels of burnout [7].

High turnover rates cause difficulty in staffing healthcare facilities adequately, which has several implications on the quality of care delivered to patients. Low nurse staffing is associated with increased patient mortality rates, as low nurse-to-patient ratios result in fewer nursing care hours available for each patient[8]. Other outcomes such as patient safety and quality of care are also adversely affected by healthcare understaffing, as higher quantities of care are left undone at the end of shifts[9]. Healthcare staff may suffer from stress and burnout when being overworked to compensate for low manpower which compromises their ability to deliver care, resulting in a higher risk of medical errors [10].

While some factors are known, turnover is caused by multiple factors, and a fuller understanding of these factors must be pursued if employers seek to reduce turnover. A search for systematic reviews published in the last 10 years examining turnover intention among HCWs on PubMed, PROSPERO and Google Scholar produced nine relevant systematic reviews [11,12,13,14,15,16,17,18,19]; none of which included studies that took place during the COVID-19 pandemic. One integrative review examined the COVID-19 pandemic’s impact on predictors of nurses’ turnover intention, but it included mainly pre-COVID-19 studies, which diminished the focus of the results regarding the current situation [20]. As evidenced by the increased turnover among HCWs, the COVID-19 pandemic has likely exacerbated many previously existing factors that affected turnover intention. The challenges facing human resource management in healthcare before and during the pandemic might differ. Considering the pressing healthcare turnover issue and the lack of reviews addressing turnover intention among HCWs during the COVID-19 pandemic, this mixed studies review aimed to examine factors affecting turnover intention in the context of the highly turbulent pandemic-focused healthcare environment.

Methods

This review was conducted in accordance with the Joanna Briggs Institute (JBI) methodology for mixed methods systematic reviews using a convergent integrated approach [21].

Search strategy

An initial search of PubMed was performed with free-text words addressing the review aims to identify relevant articles. Titles, abstracts, and keywords of these articles were analysed, which informed the development of an extensive search strategy, details of which are provided in Additional file 1. Key search terms included ‘healthcare worker’, ‘turnover’ and ‘COVID-19’. Reference lists of all studies selected for critical appraisal were also screened for additional studies. The PubMed, Embase, Scopus, CINAHL, Web of Science and PsycINFO databases were searched for studies published from January 2020 up to March 2022.

Study selection

This review included studies that contained HCWs, following the definition by the World Health Organization as an occupation group consisting of doctors, nurses, and other professionals or supporting personnel, such as pharmacists, physiotherapists, and occupational therapists, that provide health services [22]. Studies that contained non-healthcare workers were also included if HCWs made up the majority of study participants. The review also included studies examining factors that affect turnover intention, as defined in the background [4], as an outcome. Quantitative, qualitative, and mixed methods studies of any design in the English language were included. Only peer-reviewed articles were considered to ensure high quality of included studies.

Following the search, all identified citations were collated and uploaded into EndNote 20 (Clavariate Analytics). Two reviewers (RP & PL) first independently screened the titles and abstracts for assessment against the inclusion criteria, followed by full-text articles. Full-text studies that did not meet the inclusion criteria were excluded. Any disagreements that arose between the two reviewers were resolved through discussion, with the assistance of a third reviewer (SL) where necessary.

Assessment of methodological quality

The JBI checklist for analytical cross-sectional studies and the JBI checklist for qualitative research were used to appraise the included quantitative and qualitative studies, respectively [23]. The Mixed Methods Appraisal Tool (MMAT) version 2018 [24] was used to appraise mixed methods studies. Critical appraisal was performed by two independent reviewers, and disagreements were resolved through discussion. For this review, a low methodological quality refers to a score assigned to a study of less than 50%, a medium quality refers to one between 50 and 75%, and a high quality refers to one greater than 75%.

Data extraction and synthesis

Full-text articles of eligible studies were retrieved and reviewed. To obtain relevant information that assisted in answering the review question, a customised data extraction template that included the origin and year of publication, study methodology and objective, occupations of participants and primary findings was used. Data was extracted independently by two reviewers (RP & PL), and any discrepancies observed were resolved through discussion with the assistance of a third reviewer (SL). Adhering to the JBI approach to mixed methods systematic reviews, a convergent integrated approach was adopted, where both quantitative and qualitative data were combined and synthesised simultaneously [21]. Quantitative data was first coded and presented in a textual descriptive form to allow for integration with qualitative data. A three-step thematic synthesis was then conducted [25]. Initial inductive codes were generated using line-by-line coding. These codes were organised into categories, forming descriptive themes. The reviewers then compared these descriptive themes with textual data from the studies, allowing analytical themes to emerge which were finalised through discussion among the two reviewers (RP & PL). All synthesised findings were presented in a narrative summary and categorised thematically.

Results

Search outcomes

A total of 1,082 articles were retrieved. After removing 631 duplicates, 451 records were screened, based on titles and abstracts. Irrelevant records were removed, and 71 full-text articles were screened based on eligibility. A total of 43 articles [26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68] met the inclusion criteria and were included for the synthesis. The flow of the selection process is illustrated in Fig. 1, the modified Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) format.

Fig. 1
figure 1

Modified PRISMA flow diagram

Full size image

Study characteristics

Among the 43 included studies, there were 39 quantitative studies, two qualitative studies and two mixed methods studies. Eighteen studies were conducted in the Middle East, ten in the Americas, nine in the Asia–Pacific region and six in Europe. Nurses (n = 35) were included in a vast majority of the studies, while physicians (n = 13), allied health workers (n = 11) and healthcare administrative or management staff (n = 7) were included in a smaller proportion. Quantitative studies measured turnover intention with the Turnover Intention Scale (TIS-6) [69] or modified versions, Likert scales, or yes–no questions. Qualitative studies conducted semi-structured interviews with individual participants. Mixed methods studies used online questionnaires with both closed questions—Likert scales and yes–no questions—and open-ended questions. The sample sizes for quantitative and mixed methods studies ranged from 72 to 5,088 participants, while the sample sizes for qualitative studies were comparatively smaller and ranged from 10 to 19 participants. Refer to Table 1 for the study characteristics and summaries of their findings. The overall critical appraisal quality ratings of all included studies ranged from 71 to 100%, indicating medium to high methodological quality. Most studies that were unable to achieve high methodological quality did not use valid and reliable tools to measure turnover intention. Refer to Additional file 2 for the results of the quality appraisal.

Table 1 Study characteristics and findings from included studies
Full size table

Data was categorised into five themes that emerged during the extraction process: (1) fear of COVID-19 exposure, (2) psychological responses to stress, (3) socio-demographic characteristics, (4) adverse working conditions, and (5) organisational support.

Data synthesis

Fear of COVID-19 exposure

The most prevalent theme was fear of COVID-19 exposure, which had the greatest number of studies reporting about it. A total of 12 studies revealed a positive correlation or association between fear of COVID-19 and turnover intention. Six quantitative studies used the Fear of COVID-19 Scale (FCV-19S) as an instrument to measure the psychological impact of exposure to COVID-19 on HCWs [26, 36, 37, 41, 42, 49], while one study used the Coronavirus Anxiety Scale (CAS) [39], with all studies finding a positive correlation between fear of COVID-19 exposure and turnover intention. Two studies from Iran and the United Kingdom also found fear of COVID-19 exposure to be associated with turnover intention through qualitative results [64, 66]. Nurses were at a greater risk of experiencing turnover intention during the COVID-19 pandemic compared to pre-pandemic times [53]. Nurses who were infected or had team members who were infected with COVID-19 [45] or received COVID-19 positive patients [35] were more likely to experience turnover intention. Within nursing home staff, facing increased COVID-19-related challenges, such as a lack of personal protective equipment (PPE) and increased risk of COVID-19 transmission, were indirectly and positively associated with turnover [33].

Psychological responses to stress

A variety of psychological responses to stress displayed by HCWs experiencing turnover intention was observed in several studies Two studies from Saudi Arabia and the United States discovered that high levels of psychological stress and anxiety were linked to HCWs exhibiting higher degrees of turnover intention [31, 34], while a study from Germany linked higher levels of depression to greater turnover intention [60]. Burnout was another key element that was associated with increased turnover intention [54, 58, 61, 63]. Furthermore, two Chinese studies found that HCWs were at a greater chance of experiencing turnover intention if they were suffering from poor mental health or a combination of psychosocial issues [40, 68]. A study that surveyed nurses from the Philippines found that turnover intention was positively correlated with compassion fatigue [43]. In addition, COVID-19-associated discrimination was positively correlated with turnover intention, as it contributed to aggravating factors, such as poor mental health and burnout [44]. On the other hand, resilience within HCWs was found to be negatively correlated with turnover intention [27,28,29, 52], with two studies concluding that it acted as a protective mechanism against factors that contributed to turnover intention [43, 44]. Professional commitment and job satisfaction were other psychological factors that were associated with reduced turnover intention within nurses and physicians, respectively [26, 64].

Socio-demographic characteristics

Three studies noted that certain socio-demographic characteristics significantly influenced the likelihood of HCWs experiencing turnover intention [34, 51, 67]. Two studies reported that married nurses were more likely to experience turnover intention [34, 51]. Mirzaei et al. [51] also noted that male nurses experienced turnover intention at a higher rate. While two American studies found that seniority in nurses was associated with greater turnover intention [34, 50], this was in contrast to findings in a study from Peru [67], which reported that younger HCWs were more prone to experiencing turnover intention. Social affiliation was also identified by some studies to impact turnover intention, as nurses with strong ties to friends and family perceived greater social support. Three studies determined that social support was able to act as a mediating factor in reducing turnover intention among nurses [31, 39, 68]. A direct correlation between social support and reduced turnover intention was also established by an Iranian study [51]. Conversely, an Australian study found that nurses who did not feel connected to their colleagues or team were at greater risk of experiencing turnover intention [35].

Adverse working conditions

The COVID-19 pandemic placed heavy pressure on healthcare systems, resulting in excessive job demands and tumultuous work environments. Increases in workload were widely found to be positively correlated with turnover intention among HCWs [38, 45, 55, 56]. In addition, HCWs who worked long hours, or faced an increase in working hours due to the pandemic, were more prone to experiencing turnover intention [40, 50, 55, 59]. Mirzaei et al. [51] found that job stressors were positively correlated with turnover intention. Two studies from Turkey and China found that being subjected to workplace violence was associated with increased turnover intention among HCWs, as traumatic experiences accelerated burnout and mental health deterioration [55, 68]. HCWs were also more likely to experience turnover intention if they faced changes at work [63] or deployment to other departments [55, 60]. A study from Iran found that the perceived safety climate was negatively correlated with turnover intention [46]. In addition, nurses who perceived greater issues with patient safety and quality, as well as work environment, were at a greater risk of experiencing turnover intention, as they were more substantially affected by moral distress [62]. Two Italian studies found that HCWs who put in higher levels of effort in their work also had a higher chance of experiencing turnover intention [47, 48]. Other factors that were associated with turnover intention among HCWs include poor job resources [45], low staff morale [45] and perceived high pandemic impact on practice [57]. Conversely, a study from Saudi Arabia found that a positive nursing practice environment was negatively correlated with turnover intention [30].

Organisational support

Several domains of organisational support were examined by studies and were found to impact turnover intention among HCWs in various ways. Organisational trust and perceived organisational support were core factors that protected nurses from increased turnover intention [63]. Other factors, such as quality employer communication and job preparedness, were also associated with decreased turnover intention among nursing home staff [33]. Moreover, leadership support was linked to decreased turnover intention among nurses [41, 65]. A qualitative study from Iran discovered that both a positive organisational atmosphere and organisational motivation were associated with reduced turnover intention among nurses [64]. In contrast, Australian nurses who did not feel connected to their organisation were more likely to experience turnover intention [35]. HCWs who perceived low rewards from work were also more prone to experiencing turnover intention [47, 48, 60]. A study from the United Kingdom did not find any significant association between turnover intention and access to well-being centres [32].

Discussion

Findings of this study identified multiple factors influencing turnover intention among HCWs during the COVID-19 pandemic. The five emerged themes encompassed factors ranging from individual, interpersonal, job-related, and organisational determinants, and many of which were known factors prior to the pandemic [11,12,13,14,15,16,17,18,19]. However, the theme ‘fear of COVID-19 exposure’ was unique and specific to this pandemic.

HCWs’ fear of COVID-19 infection emerged as the most prevalent theme, especially during the initial stage of uncertainty and limited understanding of the virulence, transmission, and health management of COVID-19. Prior to the widespread deployment of vaccines, COVID-19 was potentially a life-threatening infection. Some HCWs were at high risk of exposure while tending infected patients in the face of shortages of fundamental resource, such as PPE [70]. The possibility of spreading COVID-19 to family members also created a concerning overlap between nurses’ professional and personal lives [71]. Several countries, such as Singapore and Hong Kong, which experienced the 2002–2004 severe acute respiratory syndrome (SARS) outbreak, reported higher pandemic preparedness and response resources, e.g., PPE and negative pressure rooms in intensive care units (ICU), and this could have better mitigated HCWs’ fears regarding COVID-19 [72]. Other strategies for addressing fear include clear, trustworthy, and timely COVID-19-related interpersonal, institutional and systemwide communication, to avoid disparities in understanding and reduce work-related stress among HCWs [73, 74].

The adverse working conditions as a result of pandemic emergency responses contribute to HCWs’ turnover intention. With the increase in patients requiring ICU care and mass testing services, HCWs faced increased workloads, higher nurse-to-patient ratios and deployment to areas requiring more staffing [75]. In particular, mass staff deployment means that members of a healthcare team are not familiar with each other, which can impact interprofessional collaboration [76]. This results in continuous tensions between healthcare professionals attempting to maintain patient safety under trying conditions [77]. In addition, the limitations posed by public health measures forced HCWs to quickly adapt new and frequently changing protocols and establish new workstreams, such as delivering care through telehealth [78]. Expectations of maintaining high standards of quality care remained, despite HCWs having to adapt to fluid and demanding working environments, facing new challenges and learning new skills [79]. Inevitably, HCWs experienced immense pressure and eventually burnout, which is a major contributor to turnover intention [55]. Burnout accompanies the progression of emotional overburdening, deteriorating mental well-being, and job dissatisfaction, which may drive HCWs to conclude that resignation is their best option [80]. In hindsight, these phenomena could be alleviated by judicious considerations of the consequences and the initiation of mitigating health policy measures.

Resilience within HCWs was observed to be a vital protective factor against turnover intention, as it enabled them to better respond to the disruptions that occurred during the pandemic. Staff who exhibit resilience are able to effectively use coping skills that reduce the psychological burden of treating COVID-19 patients [81]. However, when confronted with extended turmoil, HCWs found it increasingly difficult to remain unaffected while carrying out their everyday duties. High levels of psychological and moral distress were experienced by HCWs as they witnessed patient death and suffering on a massively increased scale during the pandemic [82]. To make matters worse, frontline HCWs around the world endured episodes of harassment and violence at the hands of members of the public who held irrational beliefs about the transmissibility of COVID-19 [83]. HCWs of Chinese and other Asian ethnicities were subjected to harsh COVID-19-related racial discrimination, adding to the difficulties experienced during the pandemic [68, 84]. Future studies could look into the impact of racism on turnover intention among HCWs. It is important to provide HCWs with opportunities to ‘let off steam’, obtain peer support, allow access to keep in contact with family and friends, and perform daily quick check-ins and check-outs to monitor their health status, support them emotionally and bolster their resilience [85]. At the organisational level, ensuring the provision of accessible and optimal professional psychosocial support, such as by having a multi-disciplinary psychosocial team, 24/7 hotline and efficient referral system could also be facilitated [85]. Measures to provide psychosocial support and mitigate secondary stressors related to the basic needs of life (e.g., childcare, grocery shopping) for HCWs in isolation or quarantine should also not be neglected [86].

Several studies in this review investigated different domains of organisational support. In essence, any form of organisational support that can be perceived by HCWs will empower them to adapt to the demands of their work and motivate them to perform their duties to the best of their ability [87]. During a health crisis leadership is challenging, but it is in times of crisis that the visibility and roles of leaders become apparent and provide opportunities for healthcare teams to grow and develop stronger relationships. Apart from communication and empowerment, Walton and colleagues [86] recommended the importance of understanding the humanity of the situation and exercising humility in demonstrating role responsibilities among healthcare leaders in supporting their teams through the pandemic. Similarly, being present was especially powerful in boosting nurses’ morale—high visibility of nurse leadership was evident during instances, where nurse leaders were physically present in COVID-19 treatment units to assist in various roles, building confidence and encouraging staff nurses to continue working [88].

Interestingly, this review did not identify financial renumeration as a factor contributing to turnover intention in times of the COVID-19 pandemic. Similarly, in another systematic review that identified barriers to manpower retention during health emergencies, poor leadership communication, emotional support and family worries were most commonly reported, while lacking budget in training, salaries and compensation of personnel were least reported [89]. While it is unclear what contributes to such a finding, financial renumeration is an important factor that impacts the livelihood of HCWs. Reimbursing HCWs to continue their professional education, establishing career ladders with attractive compensation progression, and maintaining salaries that are reasonably comparable with other local healthcare facilities are recommended forms of financial expenditure that can reduce turnover intention [90].

This review also identified that certain socio-demographic characteristics were associated with turnover intention. Married nurses struggled to achieve a work–life balance, especially those with children, as the pandemic caused increased childcare needs as a result of the implementation of virtual learning during lockdowns [91]. The inconsistent finding on impact of age on turnover intention was likely attributed by the different pandemic-related challenges faced by young and senior HCWs. Older HCWs were more susceptible to severe COVID-19 infection and faced age-related discrimination, while the less experienced younger HCWs had lesser personal resources and might be not as capable of protecting their well-being via self-regulation [92]. Special attention should thus be given to young/inexperienced and conversely, older HCWs. Social support protects against mental health stressors, acting similar to resilience in helping HCWs cope better during such difficult times [93]. Other vulnerable groups not identified in our review included the migrant HCWs who have been separated from their families since the start of pandemic and were unable to visit their country of origin due to travel or hospital administration restrictions [94].

Prior to the pandemic, many countries were facing healthcare workforce shortages, and the pandemic brought added challenges for healthcare stakeholders in retaining the current workforce. Nonetheless, the pandemic has also stirred sympathy and gratitude toward the plight of HCWs among citizens, providing a crucial opportunity for policymakers to justify and commit the resources required to achieve meaningful healthcare reform [95]. Ultimately, it is in the interest of public health stakeholders to capitalise on this opportunity to re-evaluate the support and compensation of HCWs, particularly in countries that face an increased demand for health services due to ageing populations.

Implications for future research

While it would be valuable to healthcare leaders and policymakers for identifying the significant contributing factors impacting HCWs’ turnover intention in times of pandemic crisis, it was not the intent of this review to identify the strength of relationships between them, or the changes in weighting of these factors. Some of these existing factors might have become more important or less important during the COVID-19 pandemic. Future quantitative works could examine this. Future research conducted in times of crises could also focus on specific factors, such as working conditions or burnout, to elucidate the main drivers that influence them and how to better support or incentivise HCWs to stay in their jobs. In addition, specific vulnerable population groups, e.g., migrant HCWs and healthcare profession minorities, could be examined as they may face different and unique challenges in their personal lives and lines of work, respectively. As border controls inevitably ease over time, international travel will likely return to pre-pandemic levels and researchers could also investigate global labour market trends, such as migration, when analysing data related to HCWs turnover. While the fear of COVID-19 is a pandemic-specific factor, the remaining factors identified in this review were already present before the pandemic and were exacerbated by the extreme conditions of the pandemic; it is unclear whether these factors will diminish as the pandemic wanes, and thus future research can also serve to investigate the persistence of these factors.

Strengths and limitations

This review captured studies conducted across a wide range of countries with different cultural and social contexts. However, it did not include grey literature and studies published in non-English languages. The methodological quality of the studies included in this review ranged from medium to high, but some studies did not account for confounding factors in their analyses, which would likely influence the reliability of their results. While the included studies focused on turnover intention instead of actual turnover, there is evidence that these are correlated [69]. Due to the lack of heterogeneity across the included studies, a meta-analysis could not be performed.

Conclusions

In this review, we have provided an extensive overview of factors contributing to turnover intention among HCWs during the COVID-19 pandemic. Although it is unclear if some of the pandemic-specific factors identified will diminish over time as the pandemic ebbs, our findings highlighted the importance of acknowledging and addressing these factors to prevent further aggravation of the turnover issue. In the wake of the overwhelming pressures experienced by the health workforce in the past 2 years, this turnover may worsen, and proactive measures should be taken to retain HCWs. Future research should be more focused on specific factors, such as working conditions or burnout, and specific vulnerable groups, including migrant HCWs and healthcare profession minorities, to aid policymakers in adopting strategies to support them and incentivise them to retain them in their healthcare jobs.