research paper in employee retention

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Healthcare (Basel)
• PMC10341299

Retaining Healthcare Workers: A Systematic Review of Strategies for Sustaining Power in the Workplace

Neeltje de vries.

1 Department of Internal Medicine, Spaarne Gasthuis, P.O. Box 417, 2000 AK Haarlem and Hoofddorp, The Netherlands

2 Spaarne Gasthuis Academy, P.O. Box 417, 2000 AK Haarlem and Hoofdorp, The Netherlands

Olivia Lavreysen

3 Centre for Environment and Health, University of Leuven (KU Leuven), P.O. Box 952, 3000 Leuven, Belgium

José Bouman

Szymon szemik.

4 Department of Epidemiology, School of Medicine in Katowice, Medical University of Silesia, 40-055 Katowice, Poland

Kamil Baranski

Lode godderis.

5 IDEWE, External Service for Prevention and Protection at Work, 3000 Leuven, Belgium

Peter De Winter

6 Department of Paediatrics, Spaarne Gasthuis, P.O. Box 417, 2000 AK Haarlem and Hoofddorp, The Netherlands

7 Leuven Child and Health Institute, University of Leuven (KU Leuven), P.O. Box 3717, 3000 Leuven, Belgium

8 Department of Development and Regeneration, University of Leuven (KU Leuven), P.O. Box 611, 3000 Leuven, Belgium

Associated Data

The authors confirm that the data supporting the findings of this study are available within the article and/or its Supplementary Materials .

The shortage of healthcare workers is a growing concern. The COVID-19 pandemic and retirement wave have accelerated turnover rates. This systematic review aimed to identify and analyse the existing interventions for job retention of healthcare workers, in terms of nurses and physicians, in a hospital setting. A comprehensive search was conducted within three electronic databases, guided by the preferred reporting items for systematic review and meta-analyses (PRISMA) and synthesis without meta-analysis (SWiM) guidelines, this resulted in 55 records that met the inclusion criteria. The intervention outcomes are categorized into substantial themes: onboarding, transition program to a different unit, stress coping, social support, extra staffing, coping with the demands of patient care, work relationships, development opportunities and department resources, job environment, work organization, recruitment approach, and technological innovations. Considering the literature, onboarding programs and mentoring for nurses and physicians are recommended. Additionally, other interventions described in this review could positively affect the retention of nurses and physicians. When selecting an intervention for implementation, managers and human resources should consider the intervention that matches the determinant of intention to leave of their healthcare workers and the hospital’s mission, vision, and values. Sharing the success stories of implemented interventions may benefit healthcare organizations.

1. Introduction

Worldwide, there is a growing concern about the number of healthcare workers, which currently suffers from a shortage of 5.9 million nurses [ 1 ] and 4.3 million doctors [ 2 ]. Turnover rates were accelerated by the COVID-19 pandemic. For example, a study in the United States revealed that 18% of healthcare workers left their jobs as a result of the pandemic [ 3 ]. Furthermore, the outflow of healthcare workers leaving the hospital will also increase in the future with the retirement of healthcare workers. Globally, about 17% of all nurses are expected to retire within the next ten years. In particular, the ageing workforce in the United States and Europe means that retirement rates will remain high over the next ten years [ 1 ].

Furthermore, the healthcare system is struggling to recruit the younger generation of healthcare workers who deem the nursing profession unattractive due to salary or low job status [ 4 ] and physicians deem the medical profession due to a lack of training positions and the lack of salary comparing to their working conditions [ 5 , 6 ]. These arguments for why younger generations of nurses and physicians are less willing to start a healthcare career also explain the push and pull factors resulting in the international migration of healthcare personnel. Lots of physicians are mentioning working in high-income countries and strained healthcare systems such as Australia, New Zealand, and Central Asia, instead of low- or middle-income countries or less-strained countries (e.g., Ireland, the United Kingdom, or sub-Saharan Africa) [ 7 , 8 , 9 ]. This migration process results in an enlarging shortage of physicians in these countries with a tremendous shortage of doctors [ 7 , 8 ].

Altogether, this looming crisis demands a coordinated response with the government, health organizations, and other stakeholders working together to ensure that healthcare workers have the support they need to remain in the field.

The turnover rate of nurses and physicians poses substantial financial and non-financial burdens for healthcare organizations [ 10 ]. Multiple studies have found an association between nurse staff turnover and patient outcomes such as patient health [ 11 ], length of stay of hospitalized patients [ 12 ], and quality of care [ 13 ]. Physician turnover has also been shown to affect patient care costs by disrupting the continuity of care and causing dissatisfaction in patients who have lost their current provider or the need to establish a new relationship with another provider [ 14 ]. Moreover, high turnover rates reduce staff productivity because there is limited personnel to complete the tasks [ 15 ]. This can lower the morale of the remaining staff [ 16 , 17 ] and may lead to additional turnover among the remaining employees [ 14 ]. As a result, healthcare organizations incur enormous costs associated with recruiting, hiring, and instructing new personnel [ 18 , 19 ]. In the United States, the recruitment cost per nurse vacancy has been estimated between USD 10,000 to USD 88,000 [ 18 ], while costs for physician recruitment are even higher, ranging from USD 88,000 to USD 1,000,000 per physician [ 14 , 19 , 20 ].

Aside from the financial problems caused by turnover, frequent staff turnover can decrease the job satisfaction of healthcare workers and trigger them to leave the profession. In addition, this process results in a loss of knowledge and experience in the healthcare profession [ 14 , 21 ].

In view of the many problems associated with turnover, it is crucial to minimize the impact of the shortage of nurses and physicians by retaining them in their hospital. Furthermore, retaining nurses and physicians will improve patient health, length of stay, and quality of care. However, an overview of interventions which are effective for retaining nurses and physicians in hospitals is lacking. To address this issue, this systematic review aims to identify and analyse the current interventions that minimize nurse and physician job retention in a hospital setting.

This systematic review constitutes the starting point of an EU-funded project named METEOR (MEnTal hEalth: fOcus on Retention of healthcare workers) [ 22 ].

2.1. Design and Population

The systematic review was carried out in accordance with the Preferred Reporting Items for systematic review and meta-analysis (PRISMA) statement [ 23 ] and the synthesis without meta-analysis (SWiM) reporting guidelines [ 24 ]. PRISMA checklist and SWiM items can be found in Supplemental S1 . At the international prospective register of systematic reviews (PROSPERO), the systematic review has been recorded, CRD42022364748.

To create homogeneity in the results, the population studied in this review included healthcare professionals in terms of nurses and physicians in a hospital setting.

2.2. Data Sources and Searches

The conducted literature search string in this systematic review was identical to the earlier published systematic review of De Vries et al. [ 25 ]. De Vries et al. [ 25 ] used the outcomes including determinants impacting retention, whereas this current study included studies on how to improve retention. The design of the search string was set up using the domain, determinant and outcome framework. The domain contained the following synonyms: ‘health personnel’, ‘healthcare workers’, ‘healthcare providers’, ‘healthcare professionals’, ‘health workforce’ and ‘health workers’, ‘nurses’, ‘nurse’, ‘nursing personnel’, ‘physicians’, ‘physician’ or ‘doctor’. Synonyms for the domain were ‘determinants’, ‘factors’, ‘predictors’, and ‘interventions’. As outcomes, the following terms were used: ‘personnel turnover [Mesh]’, ‘personnel turnover’, ‘retaining personnel’, ‘job retention’, ‘retention rates’, ‘turnover intention’, ’intention to leave’, ‘intention to quit’, ‘intention to stay’. The synonyms in selecting domain, determinant, and outcome were combined with OR. The overall domain-, determinant-, and outcome sections were combined with AND [ 25 ]. The entire search string is consultable in Supplemental S2 .

The search string was developed in Cinahl, Embase, and PubMed in the week of 18 July 2022 [ 25 ].

2.3. Screening and Data Extraction

Articles were included if they were conducted between 2012 and July 2022 and if the intervention was applied to healthcare workers, namely nurses and physicians. The included manuscript must be written in English, and the research must be conducted in a hospital setting. Study designs such as systematic reviews, thesis, guidelines, and study protocols were excluded. Furthermore, the study was excluded if the full text was unavailable. There were no restrictions in sampling choice. After screening the title and abstract the full texts were studied. Three pairs of two independent reviewers (AB, KK, OL, SS, NdV, and PdW) conducted the screenings.

Furthermore, quality assessment was conducted using the Mixed Methods Appraisal Tool (MMAT) version 2018. The MMAT was selected because of the heterogeneity of study designs included in this systematic review. The same pair of reviewers conducted the quality assessment independently to decrease the change of bias. Disagreement about study eligibility was resolved through consensus discussion or by an extra author, not a duo member. To show an overview of the quality of included articles, a quality rating was calculated showing an overall score. Answering ‘Yes’ in the MMAT tool counted for one point, whereas answering ‘No’ counted for zero points in the overall score. If a quality criterion was answered with ‘Cannot tell’, more information was needed to give a legit answer in terms of ‘Yes’ or ‘No’ [ 26 ] and the criterion was not included in the overall score. The final overall score is an overview. An overall score of zero points is labelled as a bad-quality study. All other scores are labelled as non-bad quality studies. The overall score of the quality assessment does not reveal what aspect of the assessment is questionable [ 27 ]. Therefore, it is desirable to scale the overall score with the complete quality assessment screening, which will be shown in Supplemental S3 .

Data were extracted into multiple characteristics: type of study, country, type of healthcare worker (physicians or nurses), sample size, the department where the intervention took place, description of the intervention, and results on the micro-level, meso-level and macro-level. Micro-level: refers to the individual level of analysis, such as a person’s behaviour. Meso-level: refers to the study of groups of people and their interactions, such as organizations and communities. Macro-level: refers to the study of large-scale phenomena and the broader forces that shape society, such as political, economic, and cultural systems. Furthermore, the factors that influence the effectiveness of the intervention were described, and an additional check was done on whether a price analysis was conducted.

Due to the heterogeneity between studies regarding the study designs and outcome measures, a meta-analysis was not conducted.

The literature search resulted in 5177 articles. Before screening 1126 duplicates were removed and 178 duplicates were detected automatically by an application. The detected duplicates were checked by the author and removed by hand. The remaining papers were checked by hand for any missed duplicates by the application. This resulted in 948 extra duplicate papers which were removed. Moreover, 152 records were removed due to foreign language. In total, 3899 records were screened on inclusion and exclusion criteria, and 219 documents were assessed for eligibility. After reading full texts, 162 records were excluded for not fitting the inclusion or exclusion criteria. Two records were excluded due to bad quality [ 28 , 29 ]. For full quality assessment, Supplemental S3 can be consulted. Finally, 55 records were included in this systematic review ( Figure 1 ).

PRISMA flow diagram.

3.1. Methodological Characteristics of the Studies

Of the included records, 85.5% ( n = 47) were quantitative research, 9% ( n = 5) were qualitative research, and 9% ( n = 5) were mixed-method studies. Of the included studies, 83.6% ( n = 46) focused on nurses, 7.2% ( n = 4) on physicians, 3.6% ( n = 2) on both, and 7.2% ( n = 4) on others (including nurses and physicians). Most studies were completed in the US (43.6%) or Asian (27.3%) countries. The quality of records differed, as shown in Table 1 .

Data-extraction table and quality assessment summary of included records.

a The quality assessment was conducted using the Mixed Methods Appraisal Tool (MMAT) (version 2018). ‘Yes’ counted for one point and ‘No’ for zero points. In case a quality criterion was answered with ‘cannot tell’, more information was needed to give a legit answer in terms of ‘yes’ or ‘no’ [ 26 ]. Therefore, this criterion is not included in the overall score.

3.2. Intervention Outcomes

An overview of the data extraction of the records in terms of micro-level, meso-level, and macro-level results and factors influencing the effectiveness of the intervention are shown in Table 2 . The included interventions are subdivided into twelve themes described in the following paragraphs.

Data-extraction table micro, meso, macro results and factors influencing the effectiveness of the intervention.

3.2.1. Onboarding

Multiple records have described that new nurses feel overwhelmed in the transition from student towards their new role as nurse [ 57 , 62 , 75 , 77 ], which suggests supporting those healthcare workers during this transition period can be beneficial. In addition, onboarding, the terminology used to describe new employees joining and integrating into the organization [ 85 ], is an important item.

Four of the included studies mainly focused on the onboarding program on the transition from nursing school towards the first job as a nurse and started at the last stage of nursing school [ 34 , 57 , 62 , 77 ]. First, Tseng et al. [ 77 ] studied an extensive externship program (EP), the Corporate-Academic Cooperation Program (CACP), to bridge the gap from nursing school to a clinical setting. During the CACP there was more focus on practicum arrangement, courses (e.g., career education and seminars), and establishing a collaborative partnership between the school and hospital. The control group received the standard EP. Students who participated in the CACP achieved a statistically significant improvement in retention rates relative to those who participated in the EP ( p < 0.05) [ 77 ].

Furthermore, Kaihlanen et al. [ 62 ] studied the effect of the final clinical practicum (FCP) in Finland. The FCP focuses on student preparation for the upcoming transition to working life. FCP uses elements such as gaining learning experience mirroring the real work as graduated, being a professional team member, and receiving adequate support and supervisory relationships. They found a significant association between turnover intention and FCP (β = 0.21, p < 0.001).

To continue, Hernandez et al. [ 57 ] implemented the University of New Mexico College of Nursing (UNM CON)/University of New Mexico Hospital (UNMH) internship program in Mexico for new graduates. The focus of this internship program contained six items: focusing on organizing work and setting priorities, communicating effectively, developing clinical leadership skills, developing technical skills which are needed to provide safe care, practising quality care with actually sick patients, and learning to work in an emergency or end-of-life setting. A total of 43.3% of the participants who could have been employed for five years remained employed at the hospital after the internship program. In addition, 63.6% of the participants who remained employed at the hospital for five years or more continued to work in the exact location they had at the first year of their employment. There is no statistical test adjusted to study the retention rates.

Finally, The Academic Partnership Program (APU) of Aull et al. [ 34 ] included an evidence-based clinical education program designed to train, recruit, and retain Bachelor of Science students towards Bachelor of Science in Nursing (BSN) prepared nurses without the need for an academic faculty. The APU is a practice of students in a home-based department with a nursing preceptor serving as a clinical instructor. The students work with their instructor on multiple units as long as the program continues, which helps to integrate the student into the culture of their assigned unit. As a result, the turnover rate reduced from 23.9% nationally to 7% after the APU.

Six of the included studies focused on their onboarding program for new graduates [ 37 , 40 , 42 , 63 , 67 , 73 ]. Blegen et al. [ 37 ] studied the effect of a structured transition-to-practice (TTP) program for new graduates containing multiple online modules. The preceptor of the hospital needed to complete an online model for introduction to the TTP program, described the difference in high and low preceptor support, and the effect of this support program for new graduates. They found a difference in outcomes of new graduates getting high preceptor support (HPS) versus low preceptor support (LPS). The retention rates of HSP hospitals were higher (86%) at the end of the first-year program, whereas only 80% of the hired students at LPS hospitals were retained ( p > 0.01) [ 37 ]. This shows that the intensity of preceptor support is an important factor in a mentorship program for onboarding new graduates [ 37 ].

In addition, Kang et al. [ 63 ] developed a situational initiation-training program (SITP). SITP focuses on the preceptor aiming to reduce stress levels and intention to leave of new graduates who have support from the preceptor. SITP contained four courses: “Covered preceptor roles, functions, and responsibilities; communication skills; stress management skills and relationship maintenance skills.” [ 63 ]. During the first preceptorship year, the new graduates showed low to shallow intentions to leave their current job at month three (mean = 4.18), six (mean = 3.8), nine (mean = 4.87), and twelve (mean = 2.6) [ 63 ].

Furthermore, Rudin et al. [ 73 ] studied the effect of the mentorship programme in Malaysia. The results showed a positive impact of the mentorship program on remaining in the nursing profession (r = 312, p = 0.001), though it is unclear how this mentorship program was set up in detail [ 73 ].

Koneri et al. [ 67 ] studied the one-year residency program with six touchpoints to focus on during the program. Touchpoints are defined by Koneri et al. [ 67 ] as: “distinct points in the company-customer experience.”. Whereas, employees are customer types of a company, Koneri et al. [ 67 ] designed a six-touchpoint program including: recognizing intrinsic worth (by sending personalized cards), developing loyalty (success stories featured in the newsletter), respect and dignity (monthly coffee-and-chat opportunities), valuing (organization of development day, educational events and sharing positive experiences) and trusting (inter-professionals teams focusing on communication, leadership, situation monitoring, and newly nurses’ support). The touchpoint program had a positive effect on retention rates compared to the non-intervention cohort ( p < 0.00). The program had a cost-effective impact on retention (USD 180 versus USD 47,000) [ 67 ].

Additionally, Chen et al. [ 42 ] studied a three-month adaptive education program on learning, mental health, and work intentions. The education program led to an increase in the turnover rate of 12.6%, after three months of implementation, towards an 87.9% one-year retention rate. Unfortunately, no comparison is available for these turnover rates before implementation [ 42 ].

Lastly, Çamveren et al. [ 40 ] tested an organizational socialization model-based preceptorship program for nurses focusing on new graduates in transition. The preceptor must support the new graduate. The program contained preceptor training and support meetings for newcomer nurses. Both components contained feedback moments, which were used to improve the preceptorship program. At the end of the one-year program, there was no significant difference compared to the baseline in nurses’ intention to leave the unit or profession. Moreover, after the program, the results showed a significant increase in the nurses’ intention to leave their organization (t = −4.153, p < 0.001) compared to the year before. This study showed that not all preceptorship programs positively impact retention rates [ 40 ].

Three studies focused on a mentorship program regarding the onboarding [ 51 , 80 , 83 ]. Fleig et al. [ 51 ] described the effect of interpersonal mentoring as support for healthcare workers. Healthcare workers who received more interpersonal mentoring were more affectively committed to the organization (r 2 = 0.35, F (3, 144) = 25.83, p < 0.01). This affective commitment moderated the effect of knowledge transfer and turnover intentions. Respondents who reported higher levels of knowledge transfer considered leaving the organization when their affective commitment was low. Though, knowledge transfer showed no significant direct relation with turnover intention (r 2 = 0.09, F (3, 141) = 4.57, p < 0.01). The direct impact of the mentoring program on turnover intention was lacking [ 51 ].

Secondly, Williams et al. [ 80 ] focused on one-to-one mentorship, which was defined as “where a single mentor is assigned to a mentee”. The participants who received one-to-one mentoring rated the experience in helping transition to practice, professional development, and stress management higher than their colleagues. There was no significant relationship found between turnover intention and the two types of mentoring [ 80 ].

Last, Zhang et al. [ 83 ] investigated the one-to-one mentorship program for one year, where the mentee and mentor mainly focused on individual career development and the relationship, social support, and role modelling between both. They compared this mentorship program with a basic preceptorship program. For the one-to-one mentorship program, the mentor received an orientation program of four hours that focused on developing mentoring skills. The one-to-one mentorship program resulted in a significantly lower turnover rate (3.77%) in the first year than the control group (14.07%). The rates in the second and third years were not different [ 83 ].

Most of the above articles have affirmed the positive impact of onboarding programs on retention rates [ 37 , 42 , 51 , 57 , 62 , 67 , 73 , 77 , 80 , 83 ].

3.2.2. Transition Program to a Different Unit

Three studies focused on the transition to a different unit [ 54 , 71 , 75 ]. Morphet et al. [ 71 ] studied the Transition to Specialty Practice Program (TSPP) for novice nurses entering a nursing specialty. TSPP offers a formal education and clinical support program combining “extended orientation, theoretical preparation, supernumerary time, preceptorship, and clinical support” [ 71 ]. Qualitative interviews indicated that the TSSP positively affected nursing recruitment in a studied emergency department. The organization and emergency ward became more attractive for the new nurses by focusing on education and support [ 71 ].

In addition, The Bridging Program of Gilroy et al. [ 54 ] focused on experienced paediatric nurses who wanted to develop and specialize in paediatric critical care. Gilroy et al. [ 54 ] did not execute the statistical analysis. However, they noticed that the external turnover rate of the participants of the program was 9%, which was lower than the overall unit turnover at that moment (12%) [ 54 ]. This outcome supports the positive outcomes of the other transition programs.

Finally, Schroyer et al. [ 75 ] focused on their Academy of Medical-Surgical Nurses (AMSN) Mentoring Program for nurses newly entering a specialty unit within critical care service, another transition during the career and stage of onboarding at another department and team. During the AMSN Mentoring Program, every newly entering nurse is paired with a mentor (experienced nurse) who provides guidance and nurturing. In the not-mentored group, 66% of nurses were retained, whereas 91% of the mentored nurses were retained ( p = 0.001, chi 2 = 6.873, 95% CI). Apart from that, nurses and trainees explained it was sometimes difficult to catch up with their mentors due to different shifts [ 75 ].

3.2.3. Stress Coping

Healthcare workers are dealing with high-stress levels. Seven of the included studies revealed interventions focusing on coping skills to reduce the intention to leave [ 43 , 50 , 58 , 60 , 72 , 74 , 78 ]. Im et al. [ 60 ] set up a Huddling Programme in Korea for new nurses. The Huddling Programme contains four sessions within nine weeks of peer group activities focusing on empowerment. The programme focused on the mechanism of the group dynamic of nurses, which could help them cope with job stress and related problems [ 60 ]. Analyses revealed that turnover rates during the study period were lower for the intervention group (4.2%) than the control group (20.0%); however, they were not statistically tested [ 60 ].

El Khamali et al. [ 50 ] designed a five-day course for nurses. This course is intended to reduce job strain by improving the ability of ICU nurses to cope with stress by complementing medical knowledge and facilitating role-plays. The course led to significantly better numbers of retention than the control group ( p = 0.04). The intervention costs the employer approximately EUR 2000 per nurse [ 50 ].

In the US, Hines et al. [ 58 ] implemented the American Nurses Association’s (ANAs) self-care guidelines in a small sample at the women’s and new-born service department. The ANAs guideline services tools to assist the nurses by selecting the appropriate self-care activities based on the particular stress in their workplace. The guideline resulted in a non-significant stress reduction (z = 0.58, p = 0.564) and a non-significant reduction in intent to leave (z = 1.13, p = 0.257) [ 58 ].

Additionally, Vardaman et al. [ 78 ] supported nurses with two computerized training sessions for Change-related Self-Efficacy (CSE). Self-efficacy is one’s belief in their ability to perform capably during any change [ 78 ]. Results showed that for every unit CSE increases, the turnover intention decrease by 0.46 ( p < 0.01) [ 78 ].

Rushton et al. [ 74 ] set up the Mindful Ethical Practice and Resilience Academy (MEPRA), which enhanced a culture of mindfulness, ethical competence, and resilience. This was cultivated by six experimental workshops of four hours with six different elements, daily technology-enabled mindfulness and reflective practice, and reflective questions. MEPRA resulted in decreasing turnover intention (F = (3, 83), p = 0.05) [ 74 ].

Furthermore, Chu et al. [ 43 ] studied the use of nurses’ strength, which is defined as: “the characteristics of a person that allow them to perform well or at their personal best” [ 86 ]. Using nurses’ strengths was fulfilled by the included nurses using positive psychology and positive organizational behaviour [ 43 ]. They found out that nurses’ strengths use had a significant positive relationship with job crafting. Job crafting is defined as: “Nurses spontaneously changing the boundaries of cognition, tasks, and relationships of their work resulting in improving the job fit.” [ 43 ]. It seemed that strength use was significant positive related to job crafting (β = 0.68, p < 0.001). Job constructing was negatively correlated with turnover intention (β = −0.27, p < 0.01). This suggests that nurses’ strength use would decrease turnover intention, though this relationship was not significant (β = −0.01) [ 43 ].

In terms of self-care, a study conducted in the US studied the effect of four creative arts therapy programs [ 72 ]. The study aimed to allow healthcare workers to gain control over their psychological stress using visual art, musical practice, creative writing, or physical dance or movement. The intervention program showed improvements in turnover intentions ( p = 0.001) [ 72 ].

3.2.4. Social Support

Four of the included studies have examined social support, which may help with managing stress and possibly impact job retention [ 44 , 53 , 55 ]. Forde et al. [ 53 ] gave the nurses a moment to speak up by implementing and testing a ‘listening to staff’ event (L2S). After implementation, the turnover numbers decreased from 18.9% in October 2017 to 10.2% in October 2020.

Concilio et al. [ 44 ] exchanged a six-week digital intervention using text messaging. The text messages of the control group only contained medical facts in the experimental group. The text messages in the interventional group contained emotional, esteem, and networking support. The digital intervention derived an increasing sense of social support in the control group. Though, the intention to leave (BF = 2.459) did not change in the control or the experimental group [ 44 ].

Additionally, Guo et al. [ 55 ] had a valuable result with the WeChat Three Good Things That Happened (3GT). During this six-month intervention, the nurses were asked to record three good things that happened. Afterwards, they had to discuss why these good things happened and their role in making them happen. Using the WeChat 3GT intervention resulted in a significantly decreased turnover intention (F = 11.0323, p = 0.001) [ 55 ]. It should be noted that the WeChat 3GT intervention was tested on burnout nurses exclusively.

Lastly, Jensen et al. [ 61 ] set up a study to research the effect of facility dogs on healthcare workers. For this study, the participants had to work for at least six months with the facility dogs. The presence of facility dog had a significant association with turnover intention Healthcare workers who work with a facility dog reported reduced intentions to quit their jobs than the control group (β = −0.27, p = 0.002, d = −0.50) [ 61 ].

3.2.5. Extra Staffing

With the shortage of healthcare workers, two studies revealed interventions contracting other personnel with a healthcare background (e.g., unlicensed assistive personnel) to support the nurse staffing and prevent them from leaving [ 49 , 81 ]. Winslow et al. [ 81 ] constructed a partnership Care Delivery Model (CDM) in a Magnet hospital in the US. The project was designed using a dyad or triad comprised of two nurses, one nurse and one unlicensed assistive personnel, two nurses, and one unlicensed assistive personnel taking care of a team of patients. The partnership CDM did not result in significant differences in the nurse turnover [ 81 ].

Furthermore, at an Acute Care department in Australia, Duffield et al. [ 49 ] added unregulated nurses support workers (unlicensed) to existing nurse staffing. Wards where nurse support was added, had non-significant higher numbers of nurse intent to leave [ 49 ].

3.2.6. Coping with the Demands of Patient Care

The primary responsibility of healthcare workers is patient care. Two included studies revealed tools which may support novice nurses in coping with the demands of patient care [ 76 , 84 ].

Zhong et al. [ 84 ] tested the humanistic care teaching model. Paediatric nurses practised this patient-care model by writing various clinical cases and practising with organized role-plays. It showed that the turnover intention in the intervention group was significantly lower than in the control group ( p < 0.001) [ 84 ].

Tang et al. [ 76 ] suggested that novice nurses have a hard time prioritizing and managing the health problems of their patients. The Humanoid Diagram Teaching Strategy (HDTS) was implemented to help novice nurses reintegrate their knowledge and skills to make decisions. The training started after the first month of pre-service training and was conducted three times a week for four weeks consecutively. During this training, the patient’s appearance was drawn in three parts: the head and neck, trunk, and limbs. The clinical preceptor encourages the novice nurse to employ association thinking and use guidance and discussion. The goal of the HDTS was to identify the primary patient problems and solutions, which resulted in learning how to manage specific cases. This HDTS resulted in a significant difference in the retention rate of the intervention group (β = −0.33, p < 0.005) [ 76 ].

3.2.7. Work Relationships

Work relationships have an impact on the retention rates of healthcare workers, which was shown in five of the included studies [ 31 , 35 , 56 , 64 , 65 ]. In 2017, Kang et al. [ 65 ] published the effect of the cognitive rehearsal program for nurses on interpersonal relationships with ten topics: nonviolent communication, withholding information, backbiting, sabotage, disgracing, undermining activities, failure to respect privacy, physical aggression, verbal affront, and self-empathy. After the intervention, the intervention and comparison groups showed significant differences in intention (F = 5.55, p = 0.024), which continued up to four weeks after the intervention program [ 65 ].

Secondly, Harris et al. [ 56 ] studied the effect of communication strategies. Specifically, the communication strategy contained clinician training using situation, background, assessment, and recommendation (SBAR) twice daily shift huddles with the BAR method and a monthly clinician meeting over three months. The communication strategies led to a decrement in unit turnover from 7.84% to 2.33% at the end of the three-month project. The increased cost for this project occurred with staff meetings being held once a month. Half of these individuals were paid an extra hour for attending this meeting, and the other half were already present for their shifts [ 56 ].

Then, Baik et al. [ 35 ] set up five four-hour interprofessional team training: a team intervention including team strategies and tools to enhance performance and patient safety (TeamSTEPPS) and communication training. Furthermore, the team followed quarterly leadership workshops. Lastly, a structured bedside rounding was implemented. The six months turnover rate before the interventions was 5.74%. Post-intervention, this turnover rate decreased to 5.3%. The retention rates were not statistically tested [ 35 ].

Kang et al. [ 64 ] developed a smartphone application to cognitively train nurses to cope with bullying situations in the workplace. The application consists of an introduction to nonviolent communication as the standard, six digital comic drawings of workplace bullying situations and nonviolent communication strategies, and a board for questions and answers. The intervention effectively decreased nurses’ person-related bullying and the experiences of work-related bullying. Pre-measurement the mean (SD) was 3.26 (0.81). The smartphone application decreased retention rates at four-week implementation by 3.13 eight-week measurement 3.36 (0.77). Whereas, the mean (SD) score of the control group was 3.59 (0.84), 3.66 (0.84), and 3.67 (0.71), respectively [ 64 ].

Lastly, Al Sabei et al. [ 31 ]) researched the impact of interprofessional teamwork. This practice is characterized by shared team identity, clarity, shared responsibility, integration, and independence, on the intention to leave of nurses. Interprofessional teamwork was directly associated with nurses’ intention to leave and indirectly mediated by job satisfaction and burnout [ 31 ]. The study did not reveal how the interprofessional teamwork intervention was precisely examined in practice [ 31 ].

3.2.8. Development Opportunities and Department Resources

Opportunities in the development of the workforce and resources may help through the retention of nurses and physicians. Three included studies discussed certain interventions [ 46 , 47 , 69 ]. In Australia, Dawson et al. [ 47 ] studied supportive strategies. The strategies contained providing employment options, rewarding performance, enhancing professional development and training, and improving management practice [ 47 ]. However, Dawson et al. [ 47 ] did not describe concrete results of these strategies.

Furthermore, Dawood et al. [ 46 ] set up a qualitative study to discuss the effect of dual roles: working as a nurse and an emergency nurse practitioner (ENP), as an intervention to improve retention. If the dual role was not available, most part-time ENPs did not consider leaving nursing altogether. However, full-time participants without dual roles considered leaving nursing, confirming that dual roles could force retention [ 46 ].

Moreover, Melnyk et al. [ 69 ] focused on the idea that implementing evidence-based practice (EBP) will result in renewing the nurses’ professional spirit and giving them a voice [ 87 , 88 ] which may have a positive impact on job satisfaction [ 87 ]. Melnyk et al. [ 69 ] used the Advancing Research and Clinical Practice through Close Collaboration (ARCC) model to implement EBP. EBP culture and EBP mentorship resulted in being key variables that significantly positively impact the intention to stay among nurses ( p = 0.02) [ 69 ].

Brabson et al. [ 38 ] focused on the three EBP training models for physicians: “Cascading model, learning collaborative, and distance education.” [ 38 ]. Results showed no differences in turnover rates at the 12-month measurement point (χ 2 (2, n = 96) = 2.10, p = 0.35, Cramer’s V = 0.15) or at the end of the study (χ 2 (2, n = 95) = 0.51, p = 0.77, Cramer’s V = 0.07) [ 38 ].

3.2.9. Job Environment

Six studies demonstrated interventions by influencing the job-/work environment to impact the retention rates [ 30 , 32 , 39 , 48 , 70 , 79 ]. Brewer et al. [ 39 ] studied the effect of a Magnet hospital. The Magnet Recognition Program ® acknowledges healthcare institutions that offer exceptional nursing care and working environments through an inventive program. It seemed that working in a Magnet hospital did not significantly impact turnover intentions (coefficient (CI) = 0.039 (−0.150 to 0.227), p = 0.687) [ 39 ].

Mohamadzadeh et al. [ 70 ] compared the outcomes of excellence-awarded hospitals to the outcomes of hospitals that do not have an excellence plan. The European Foundation for Quality Management (EFQM) is an excellent plan which has three levels. At the first level, eight criteria have been considered to evaluate the performing hospitals. These eight criteria were: leadership, policy and strategy, employees (human resources), participations and resources, customers’ results, employees’ results, and societies and performance key results. At the second level, the criteria were described in detail using subsets. At the third level, a list of specific guidelines regarding more explanation of each subset was available. The score means of intention to leave the organization in performing and non-performing organizations of the excellence plan showed a significant difference ( p = 0.004). Performing the organizational excellence plan reduced the intention to leave [ 70 ].

In terms of environment, Alvaro et al. [ 32 ] tested the impact of the architectural design of the hospital on patient and staff outcomes using a pretest-posttest quasi-experimental study. The new design mainly focused on creating an architecture of wellness containing communal dining spaces on each floor, public spaces, multiple outdoor terraces, and a rooftop terrace with views of the skyline, lake, and green environment. Workplace satisfaction of healthcare workers did improve ( p = 0.000). There was no significant difference in intention to quit staff. Though, staff with favourable impressions of the new architectural design and a greater sense of belonging to the neighbourhood showed a decreased intention to quit ( p < 0.01) [ 32 ].

Furthermore, Walker et al. [ 79 ] studied the effect of the integration of an electronic health record on retention rates. Quality of care did improve significantly in terms of infections, pressure ulcers, and falls ( p < 0.01). Though, the analysis of data revealed no significant model effect (F (2, 42) = 2.09, p > 0.05, r 2 = 0.07), nor did the model explain the variance in the nurse turnover [ 79 ].

Adams et al. [ 30 ] explored the impact of the Cultural Change Toolkit on the nursing work environment. The toolkit provides information and tools that encourage positive practice changes. It mainly focuses on meaningful recognition, shared decision-making, and increasing leadership support and involvement. The implementation of the toolkit led to a reduction in the anticipated turnover scale (mean rate pre-implementation = 3.133, post-implementation 2.989), though this reduction was not significant [ 30 ].

Finally, Deng et al. [ 48 ] studied the comprehensive reform in a hospital in China. The government implemented new policies on personnel, compensation, management, and diagnosis and treatment. Details can be found on the website of the Beijing Municipal Health Commission Information Centre [ 89 ]. Four years after implementation, the average annual growth rate was 9.1% for nurses and physicians in Beijing public hospitals. The turnover intention thought of 61.4% had decreased [ 48 ].

3.2.10. Work Organization

Four included records studied the impact of the work organization on retention rates [ 36 , 45 , 68 ].

Daniels et al. [ 45 ] studied the effect of the ‘70% Full-Time Commitment’. A provincial government in Ontario, Canada, developed this strategy where at least 70% of the nurses work full-time, and the other 30% work part-time or casually. It aimed to stimulate working full-time. Results showed that the ‘70% Full-Time Commitment’ seemed to be no effective intervention in retaining part-time and casual nurses [ 45 ].

In a qualitative study in the UK by Baillie et al. [ 36 ] nurses changed from twelve-hour to eight-hour day shifts. It appeared that the eight-hour day shifts negatively affected recruitment and retention, mainly because an increased amount of staff members were needed to cover the eight-hour day shift pattern [ 36 ].

Two selected studies checked the effect of the self-scheduling [ 68 , 82 ]. It was suggested that self-scheduling created a better work-life balance [ 68 ] and ensured more flexibility [ 82 ], possibly resulting in decreased turnover rates. Kullberg et al. [ 68 ] compared fixed scheduling with self-scheduling. Self-scheduling was significantly associated with more requests from management for short notice shift changes, whereas fixed scheduling was associated with less overtime. Self-scheduling showed overall relatively low to moderate levels of staff turnover compared to the fixed scheduling [ 68 ]. No significant calculations were executed.

In the US, a study with a larger group of nurses ( n = 1497) in four hospitals was conducted by Wright et al. [ 82 ] to study the effect of self-scheduling. Two hospitals showed an absolute increase in turnover rates (1.5% and 1.4%), and two other hospitals reported an absolute decrease in turnover rates (−5.3% and −5.4%) [ 82 ]. The isolated effect of self-scheduling on retention rates was not described by Wright et al. [ 82 ] because no other variables were not studied.

3.2.11. Recruitment Approach

Three studies concentrated on the recruitment approach as an intervention to retain nurses and physicians [ 33 , 52 , 66 ]. Two included studies focused specifically on the recruitment of physicians [ 33 , 52 ]. Firstly, Fleming et al. [ 52 ] studied the effect of provisional licensing to attract international medical graduated physicians who are without the licensing unable to work in Canada. The study showed international medical graduates started practice as a result of the provisional licensing but did not result in long-term retention [ 52 ].

Secondly, Arora et al. [ 33 ] set up a special rural recruitment track for physicians in the rural area of Thailand. In Thailand, the Collaborative Project to Increase Production of Rural Doctors (CPIRD) and the One District One Doctor (ODOD) project, were set up to increase the number of doctors in rural areas. Arora et al. [ 33 ] studied the long-term effect of these two recruitment projects. It seemed that doctor retention was higher in areas where the initiatives were implemented than in the regular tracks ( p < 0.05) and medical were 2.4-fold more likely to remain working for the area for a minimum period of three years (OR (95% CI) = 2.44 (2.19–2.72)) [ 33 ].

Finally, the Prospective Staffing Model researched by Kester et al. [ 66 ] focused on the implementation of a model to predict preventable and potential turnover at a thoracic surgery department. Restructure of the recruitment strategy was included in the implementation of the prediction model. It involved engaging current workers in the interview process and prioritizing the candidates regarding desirable characteristics. Furthermore, an internal nurse recruiter organized interviews and had weekly meetings with the nurse manager to improve the partnership. The hospital empowered local academic partners such as colleges and universities to improve the knowledge about the new graduates. Additionally, the length of the orientation of the newly hired nurses was enlarged towards eight weeks for experienced nurses and towards 12 weeks for new graduates. The implementation of the prediction model led to a 17.6% decrease in turnover in a four-year period. The cost of the 12-week orientation was $11,066.40 in 2018, which is still less than the average cost for the replacement of a new employer (about $52,100) [ 66 ].

3.2.12. Technological Innovations

There are technological healthcare interventions innovated to increase retention, two of the included records focused on the use of robots in healthcare [ 41 , 59 ]. Chang et al. [ 41 ] set up robots to help nurses focus on professional task engagement. They found that robot-enabled focus (“nurses’ perception that robots enable nurses to concentrate on conducting major nursing jobs” [ 41 ]) on professional task engagement positively impacted overall job satisfaction (r = 0.31, p < 0.05) and perceived health improvement (r = 0.34, p < 0.05). Robot-reduced nonprofessional task engagement (“nurses’ perception that robots help share the workload of auxiliary jobs” [ 41 ]) was positively related to only perceived health improvement (r = 0.26, p < 0.05). Furthermore, Chang et al. [ 41 ] noticed that job satisfaction and perceived health improvement were negatively related to turnover intention (r = −0.41, p < 0.05 and r = −0.18, p < 0.05) [ 41 ]. These findings suggest that, by using robots, the increased focus of nurses on professional task engagement and reduced focus on nonprofessional task engagement could help to improve job satisfaction and job retention of nurses [ 41 ].

Huang et al. [ 59 ] tested the effect of effort ensuring smooth operation (EERSO), “the time and energy needed to keep robots operating as designed”. EERSO was positively associated with time pressure (β = 0.16, p = 0.007) and missed care (β = 0.13, p = 0.003). Using robots may help reduce nurses’ workload by focusing on nurses’ saved time and, therefore, turnover intention. However, it also requires nurses’ efforts to maintain EERSO, which may adversely impact nursing professional workplaces [ 59 ].

4. Discussion

This systematic review resulted in an overview of the existing interventions for job retention of nurses and physicians in a hospital setting. The included records resulted in twelve themes on which management could focus on in terms of job retention: onboarding, transition program to a different unit, stress coping, social support, extra staffing, coping with the demands of patient care, work relationships, development opportunities and department resources, job environment, work organization, recruitment approach, and technological innovations.

The positive impact of the onboarding program [ 34 , 37 , 42 , 57 , 62 , 63 , 67 , 77 ] and mentorship [ 51 , 73 , 80 , 90 ] is in line with earlier published systematic reviews [ 91 , 92 ]. Kakyo et al. [ 93 ] explored the benefits of the informal mentoring program for nurses and confirmed that built on the reciprocal relationship between mentee and mentor; there is a substantial benefit of the mentoring program.

Furthermore, the onboarding program shows the importance of supporting the new graduates within the first two years of their working life [ 37 , 42 , 63 , 67 ]. More than 50% of newly graduated nurses leave their job within the first year due to culture shock [ 94 ]. To prevent them from leaving early in their working life and negatively impacting the staff long-term, it seems important to focus on and maintain this specific group. Stevanin et al. [ 95 ] described the difference in stress reporting between generations (e.g., baby boomers, generation x, and generation y). It showed that generation y reports more psychological stress than previous generations and requires support in their workplace [ 95 ]. It is suggestible that new generations (generation y and subsequent) have substantial needs to support them in the overwhelming transition toward their new role [ 57 , 62 , 75 , 77 ], than their previous colleagues. It makes the importance of onboarding programs, focusing on new graduates and new generations starting their careers and dealing with stress due to the transformation from students towards registered nurses, even more clear.

In this systematic review, none of the included records studied onboarding programs for physicians. It illustrates this content is missing in research and makes it questionable if physicians could profit from an onboarding program. A systematic review published in 2021 affirmed the relevance of early clinical contact during medical school and the early postgraduate period for the retention of physicians in a rural setting [ 96 ]. Additionally, Kumar et al. [ 96 ] also underscored the impact of professional and personal support on the retention rates for this group. Hence, onboarding programs, that focus on early clinical contact and support, could be beneficial for physicians, same as for nurses.

In addition, this systematic review highlighted the importance of tools for stress coping [ 43 , 50 , 72 , 74 ], though all of them focused specifically on nurses. It seems reasonable that physicians are dealing with stressful situations, likely as nurses. Unfortunately, interventions focusing on physicians coping with these stressful situations are lacking in this review. A review by Darbyshire et al. [ 97 ] confirmed that physicians in an acute care setting have a need for stress management techniques. These techniques could positively impact retention rates [ 97 ]. These findings make it highly likely that, for example, copings tools for stress management or mentorship programs could also be effective for physicians.

Interestingly, the interventions included in this systematic review do not mention salary as a solution for upgrading the retention rates for nurses and physicians. Earlier research showed that the migration of healthcare workers is, among other things, caused by the lower salaries in low- or middle-income countries [ 5 , 6 ]. A literature review by Okafor et al. [ 98 ] explained that the migration of nurses in Nigeria is affected by the worse payment and pushed nurses towards countries with better working conditions and better pay. Due to the withdrawal effects of healthcare workers’ migration to low- or middle-income countries, higher salaries may help reduce the intention to leave and migration [ 5 , 6 , 98 ], though it may not be the most cost-effective intervention [ 99 ]. An earlier systematic review revealed that salary is not the most common reason nurses and physicians leave their jobs in high-income countries; job satisfaction, work-life balance and social support are frequently named determinants that impact the intention to leave [ 25 ]. This suggests that salary impact may vary per low-, middle-, or high-income country. It is suggestible that salary is not a primary reason for leaving healthcare in high-income countries. Nevertheless, it is an important basis from wherefore leaving and thus a vital basis managers can build to rice retention. Managers must implement specific retention interventions that match the determinants that apply to the concerning culture or country.

However, implementing cost-effective retention interventions must likely overcome some barriers before success. For example, structural barriers such as staff workload and lack of time are commonly described as barriers to the implementation of hospital-based interventions [ 100 ] To overcome these barriers, it seems essential to enhance commitment and motivation of the staff by convincing them of the advantages for the staff themselves and sharing success stories [ 100 ].

Although a great effort was made to create a funded systematic review, there were some limitations. Firstly, a meta-analysis is not conducted due to the heterogeneity of the included records. Secondly, the authors may have missed some studies as a result of the exclusion of grey literature. The grey literature was excluded because the extensive search led to a large number of results and a comprehensive results paragraph. Lastly, it is feasible that the chosen themes of interventions overlap, which can create bias. This overlap demonstrates that the interventions affect multiple determinants that could positively impact retention rates. To maintain the retention intervention impacts all possible determinants, it is crucial to implement it on various organizational levels.

This systematic review studied extant literature on both physicians and nurses, which constitutes a key strength of this study. Though, it stands out that limited interventions that were included in this systematic review contained interventions for physicians. Other systematic reviews of this topic focused on early careers or only experienced nurses [ 91 , 92 ]. To the best of our knowledge, there is no systematic review available exploring interventions for improving retention in a hospital setting for both nurses and physicians, which makes this systematic review unique. Nevertheless, a higher number of the included records included nurses instead of physicians. Applying the results of this study to physicians in a hospital setting can create bias due to a lack of research concerning physicians. Accordingly, the outcomes should be handled with care implemented for physicians. Nonetheless, numbers showed that the shortage of physicians leaving healthcare is just as alarming as the nursing rates [ 1 , 2 ]. The literature describes high numbers of physicians dissatisfied with their jobs and burnout symptoms related to higher turnover rates [ 7 ]. A possible explanation could be that researching the intention to leave is taboo in medical culture among physicians. This might result in a minimal to not accessible target for research in retention interventions or implementation of retention strategies. Hence, the authors of this review suggest focusing on enlarging the importance of researching physicians’ intentions to leave.

Moreover, this review has thus far focused on one-factor interventions that impact the intention to leave or stay. However, job retention showed inter-correlation with other determinants (such as job satisfaction, burnout symptoms, job demands and job resources) that could also be impacted using interventions [ 28 ]. This effect is minimally studied. Hence, research on this topic could help to adjust the impact on a broader level.

Lastly, the transition from school to work seemed a vital deal breaker for nurses [ 57 , 62 , 75 , 77 ]. This raised the question about the extent to which nursing school actually prepares students for the skills they need to start work as a nurse. More research is undoubtedly desirable to prevent new graduates from leaving their workforce.

5. Conclusions

The outflow of nurses and physicians leaving hospitals is enormous. The impact of COVID-19 increases the urgency in preventing nurses and physicians from leaving. This systematic review resulted in multiple interventions that can be used to upgrade retention rates. Additionally, the implementation of organizational change and the establishment of mentorship programs are important interventions.

When selecting an intervention for implementation, managers and human resources should focus on the characteristic intervention that matches their healthcare workers and the hospital’s mission, vision, and values statements. Sharing the success stories of implanted interventions may be advantageous for all healthcare organizations.

In summary, this review can contribute to implementing retention interventions in hospitals, which can aid in maximizing retention, especially for nurses. Furthermore, this review can contribute to planning future studies containing more physician-specific interventions.

Acknowledgments

We want to thank the partners and project staff of the METEOR project for their assistance.

Abbreviations

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/healthcare11131887/s1 , S1: PRISMA Checklist and SWiM items; S2: Literature search; S3: Quality assessment MMAT.

Funding Statement

This research was funded by European Union, Chafea—3rd Health Programme, Multi-beneficiary Project Grant (HP-PJ, HP-JA), Topic: PJ-01-2020-1, Type of action: HP-PJ, SEP-210693712: Project called METEOR (MenTal hEalth: fOcus on Retention of healthcare workers). The source of funding did not influence the design of the study, the data collection, data-analysis, the manuscript writing, and the decision to submit the manuscript for publication. The views expressed and any errors or omissions are the sole responsibility of the author.

Author Contributions

Conceptualization, N.D.V., L.G. and P.D.W.; data curation, N.D.V. and J.B.; formal analysis, N.D.V., O.L., A.B., S.S., K.B. and P.D.W.; funding acquisition, L.G. and P.D.W.; investigation, N.D.V., J.B. and P.D.W.; methodology, N.D.V. and P.D.W.; project administration, N.D.V. and P.D.W.; resources, J.B.; supervision, N.D.V. and P.D.W.; validation, N.D.V., O.L., A.B., S.S., K.B. and P.D.W.; visualization, N.D.V. and P.D.W.; writing—original draft preparation, N.D.V.; writing—review and editing, N.D.V., O.L., A.B., S.S., K.B., L.G. and P.D.W. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Data availability statement, conflicts of interest.

The authors declare no conflict of interest.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Examining employee retention and motivation: the moderating effect of employee generation

Evidence-based HRM

ISSN : 2049-3983

Article publication date: 18 April 2022

Issue publication date: 20 September 2022

This study explored moderating effects of employee generations on factors related to employee retention and motivation in the workplace.

Design/methodology/approach

The authors developed a survey instrument and collected the survey data via Amazon Mechanical Turk. After filtering out bad responses, the authors ended up with 489 sample cases for this study. The authors used structural equation modeling for data analysis.

Evidence showed that only transformational leadership was significantly related to retention of Generation X employees and only work–life balance had a significant relationship with intrinsic motivation. For Generation Y employees, transformational leadership was the only factor affecting their retention while both transformational leadership and autonomy showed significant impacts on their intrinsic motivation. Generation Z employees reported that only transformation leadership affected their retention while transformational leadership, corporate social responsibility and autonomy were significantly related to their intrinsic motivation in the workplace. All three generations showed statistical significance between intrinsic motivation and employee retention.

Practical implications

This study could help business practitioners increase employees' work motivation and retention.

Originality/value

First, our results revealed interesting similarities and differences between generations in terms of the factors that affected employees' retention and motivation. Second, this study proved that employees' generation affects the impacts of transformational leadership, CSR, autonomy, WLB and technology on their motivation and retention in the workplace. Third, the results of our study also showed that employees of different generations are intrinsically motivated by different factors, proving the importance of considering generational differences in motivation literature.

• Employee generation
• Generational differences

Lee, C.C. , Lim, H.S. , Seo, D.(J). and Kwak, D.-H.A. (2022), "Examining employee retention and motivation: the moderating effect of employee generation", Evidence-based HRM , Vol. 10 No. 4, pp. 385-402. https://doi.org/10.1108/EBHRM-05-2021-0101

Emerald Publishing Limited

Copyright © 2022, Emerald Publishing Limited

1. Introduction

For the past several decades, employee retention has been an important topic to both scholars and practitioners because employees, the most valuable assets of an organization, are the ones who add to its value, quantitatively and qualitatively ( Anitha, 2016 ). Therefore, employers have taken steps to ensure that employees stay with the organization for as long as possible ( Alferaih et al. , 2018 ). Doing so is challenging because the workforce is becoming more confident and demanding due to changes in markets and demographics ( Anitha, 2016 ). A disengaged workforce leads to higher turnover rates that increase the costs of recruiting and selecting new employees ( Malinen et al. , 2013 ).

The objective of this study is to examine the moderating effects of employee generations on factors related to employee retention and motivation in the workplace. In doing so, this study makes a significant contribution to literature in several ways. First, although there have been numerous studies on factors that affect employees' retention such as a manger's leadership style (e.g. Khan and Wajidi, 2019 ), a firm's commitment to corporate social responsibility (e.g.  Valentine and Godkin, 2017 ), autonomy (e.g. Kim and Stoner, 2008 ), work–life balance (e.g.  Koubova and Buchko, 2013 ) and technology (e.g. Haar and White, 2013 ), there are no studies that have examined the effect of these five factors on employee retention and the underlying mechanism of these relationships. Second, few studies have examined effects of these five factors on different generations of employees – Gen X, Gen Y (also known as the Millennials) and Gen Z. Studies have focused on certain generations such as Gen Y (e.g. García et al. , 2019 ) or Gen X (e.g. Westerman and Yamamura, 2007 ), but no studies have been conducted to understand the different effects of the five factors on employee retention spanning three different generations. According to the US Bureau of Labor Statistics (2021) , while 40% of the 2020 American workforce comprises Gen X and 44% of Gen Y, Gen Z represented 15% of the American workforce. This indicates that Gen Z has also become an important generation to consider when examining generational differences of employee retention. Finally, no studies have reported the effects of these variables on retaining employees from these various generations during the pandemic. Retaining employees is a challenge at the best of times, but it has become even more challenging during the pandemic. A recent survey of working age people in various industries found that about 40% of respondents expressed strong intention to quit their current job in the next three to six months ( De Smet et al. , 2021 ).

2. Literature review and hypotheses

2.1 employee generations.

Based on the generational theory originated from the work of Mannheim (1970) , generations refer to groups of individuals (i.e. cohorts) born in the same period, sharing similar historical events and social experiences. This means that a cohort of individuals who shared common historical and social experiences are more likely to share similar characteristics, attitudes and behaviors ( Strauss and Howe, 1991 ). Given that the main objective of this study is to examine generational differences in effects of leadership styles, corporate social responsibility, autonomy, work–life balance and technology on intrinsic motivation and employee retention, we will use the generational theory as our theoretical framework to develop hypotheses in the next sections.

2.2 Effects of leadership across generations

Transformational leadership is defined as transforming the values and priorities of followers and motivating them to perform beyond their expectations ( Kark et al. , 2003 ). Concurrently, Wilkesmann and Schmid (2014) reported that one characteristic of strong leaders is the ability to motivate and influence people. Motivation was also found to be a complex act that had several factors involved. Employees, who were proactive both at work and in their personal lives, were positively affected by both their employer's leadership style and ability to foster a team and showed stronger motivation ( Felfe and Schyns, 2014 ; Khan and Wajidi, 2019 ). Gerhold and Whiting (2020) explored the motivations of employees over several generations, from Boomers to Gen Z, and the leadership skills that inspired them. They found no significant differences among generations. Rather the differences were driven more by an employee's stage of life and career than age. They reported that leadership fundamentals were a constant. These fundamentals, building strong teams, providing feedback and understanding employees' motivations, were multi-generationally relevant skills. In addition, Diskiene et al. (2019) found that the relationship between a leader's emotional and social intelligence and an employee's motivation to work was undeniable, although there was some variance depending on the latter's age. Interestingly, younger workers relied less on their leader's emotional stability to motivate them than older, more experienced workers.

The generation of employees will moderate the effect of leadership on their intrinsic motivation in the workplace, such that transformational leadership is more positively related to the workplace motivation of younger generation employees.

The generation of employees will moderate the effect of leadership on employees' retention in the workplace, such that transformational leadership is more positively related to the workplace retention of younger generation employees.

2.3 Effects of corporate social responsibility across generations

The generation of employees will moderate the effect of CSR on employees' intrinsic motivation in the workplace, such that greater CSR is more positively related to the workplace motivation of younger generation employees.

The generation of employees will moderate the effect of CSR on employees' retention in the workplace, such that a CSR policy is more positively related to the retention of younger generation employees.

2.4 Effects of autonomy across generations

The generation of employees will moderate the effect of autonomy on employees' motivation in the workplace, such that greater autonomy in one's job role is more positively related to the workplace motivation of younger generation employees.

The generation of employees will moderate the effect of autonomy on employees' retention in the workplace, such that greater autonomy is more positively related to the retention of younger generation employees.

2.5 Effects of work–life balance across generations

The generation of employees will moderate the effect of WLB on employees' motivation in the workplace, such that greater WLB is more positively related to the workplace motivation of a younger generation of employees.

The generation of employees will moderate the effect of WLB on employees' retention in the workplace, such that greater WLB is positively related to the retention of a younger generation of employees.

2.6 Effects of technology across generations

The generation of employees will moderate the effect of technology on employees' motivation in the workplace, such that more technology is more positively related to the workplace motivation of younger generation employees.

Companies competent with IT knowledge, objects and entrepreneurship had better chances of attracting loyal prospects and retaining their employees, especially those of Gen Z ( Haar and White, 2013 ). In addition to attracting employees, digital communication created two-way channels of dialogue and helped employees understand how their roles were helping the company. This increases possible retention rates ( Kick et al. , 2015 ).

The generation of employees will moderate the effect of technology on employees' retention in the workplace, such that more technology is more positively related to the retention of younger generation employees.

2.7 Motivation and retention

Employees' intrinsic motivation is positively related to their retention in all generations.

3.1 Sample data and questionnaire

The definition of generations in terms of the birth year varies across studies. As a compromise, we used the middle value. Thus, for the purpose of this study, three generations (Gen X, Y and Z) are defined based on the age as of August 2020. Specifically, Gen X is between 40 and 55 years old; Gen Y is between 25 and 39 years old; and Gen Z is between 18 and 24 years old. We created a survey questionnaire with the items that measured our variables and posted it on Google Forms. To take the survey, we required members of Amazon's Mechanical Turk to be employed and ages 18–55 years old. The survey was first run for a week in the third week of April 2020 and received 570 responses. We deleted 9 responses due to repeat responses and 24 due to multiple missing values, which reduced the total number of valid responses to 537. Furthermore, 48 responses were deleted due to poor response quality. Poor responses were identified using items that were reverse coded. After removing the poor responses, we were left with 489 useable and valid sample cases for this research. Regarding the sample size per each generation group, Gen Z is 120 (24%), Gen Y is 278 (56%) and Gen X is 91 (18%).

3.2 Measures

Our participants indicated their responses to all items on a 7-point Likert-type scales, ranging from 1 (strongly disagree) to 7 (strongly agree). Note that Cronbach's alpha for all variables exceeded the 0.70 cutoff value ( Greco et al. , 2018 ), indicating that all of the variables were reliable and could be used in the analysis. Examples of each item for each category are in Table 1 .

3.2.1 Retention

We used three items from Armstrong-Stassen and Schlosser (2008) to measure the employees' intention to remain with their company.

3.2.2 Transformational leadership

We used the Vera and Crossan (2004) 12-item scale to assess transformational leadership, consisting of four dimensions – charismatic leadership, inspirational motivation, intellectual stimulation and individual consideration.

3.2.3 Corporate social responsibility

We chose items from Woo (2013) to measure, which assessed five dimensions of CSR: environment, human rights and labor issues, product responsibility, society and economics. We excluded the product responsibility category because of low factor loading problems.

3.2.4 Autonomy

We used three items from Hackman and Oldham (1976) to assess autonomy.

3.2.5 Work–life balance

To measure this variable, we used five items from Brett and Stroh (2003) .

3.2.6 Technology

To measure this variable, we picked three items from Nambisan et al. (1999) .

3.2.7 Intrinsic motivation

We used items from Grant (2008) to measure intrinsic motivation.

3.3 Analytical models

In this study, we created three analytical models to test our hypotheses that examine the generational differences in the relationships between five independent variables and three dependent variables. The first model was the intrinsic motivation model in which intrinsic motivation was the dependent variable and transformational leadership, CSR, autonomy, WLB and technology were the independent variables. Y 1 = β 0 + β 1 X 1 + β 2 X 2 + β 3 X 3 + β 4 X 4 ( w h e r e   Y = I n t r i n s i c   M o t i v a t i o n ;   X 1 = T r a n s f o r m a t i o n a l   L e a d e r s h i p ;   X 2 = C o r p o r a t e   S o c i a l   R e s p o n s i b i l i t y ;   X 3 = A u t o n o m y ,   X 4 = W o r k - L i f e   B a l a n c e ;   X 5 = T e c h n o l o g y )

The second model was the retention model in which retention was the dependent variable and the five independent variables were the same as the first model. Y 2 = β 0 + β 1 X 1 + β 2 X 2 + β 3 X 3 + β 4 X 4 ( w h e r e   Y 2 = R e t e n t i o n ;   s a m e   f o r   X 1   t o   4 )

The third model examined whether intrinsic motivation affects retention. Retention was the dependent variable and intrinsic motivation was an independent variable. Y 2 = β 0 + β 1 Y 1 ( w h e r e   Y 2 = R e t e n t i o n ;   Y 1 = I n t r i n s i c   M o t i v a t i o n )

Figure 1 describes our analytical models with the results. When conducting three analytical models, we used a subsample analysis instead of a two-way interaction design to examine generational differences in the relationships as we hypothesized. This method allows us to compare the impact of each independent variable on dependent variables among different generations of employees. This approach is preferable because it reduces the possibility that noise will be introduced into the model ( Stone-Romero and Anderson, 1994 ).

4.1 Descriptive statistics and correlations

Table 2 summarizes descriptive statistics for the variables used in our study.

4.2 Measurement model

To evaluate the fit of our measurement model, we conducted a series of confirmatory factor analyses (CFA). We used several fit indices such as chi-square ( χ 2 ) values, the Comparative Fit Index (CFI), the Root Mean Square Error of Approximation (RMSEA) and the Standardized Root Mean Square Residual (SRMR). As shown in Table 3 , the results of CFA suggest an excellent fit ( χ 2  = 757.41, p  < 0.01; CFI = 0.95, RMSEA = 0.06, SRMR = 0.07) for our hypothesized seven-factor model ( Hooper et al. , 2008 ). In addition to our focal seven-factor model, we further assessed the fit of alternative models. The results proved that the hypothesized seven-factor model fits the data significantly better than the other possibilities.

Several statistical indictors were used to assess the reliability and the convergent and discriminant validity of our constructs. As shown in Table 4 , composite reliability (CR) estimated our constructs to be from 0.885 to 0.946, which were all above the threshold value of 0.7 ( Fornell and Larcker, 1981 ). Therefore, internal consistency was validated. Convergent validity of the constructs was also acceptable. All estimated factor loadings were significant at p  < 0.001, and all estimates are above 0.6 and most estimates are above 0.7. Furthermore, average variance extracted (AVE) for all constructs are above 0.5, the acceptable threshold level ( Fornell and Larcker, 1981 ). Given that the AVE for each construct was greater than the squared correlations between two constructs ( Fornell and Larcker, 1981 ), discriminant validity of the constructs was achieved. Hence, these results provided support for using the seven constructs as reliable and distinctive variables in our analysis.

4.3 Test for the potential common method bias

Given the nature of our data using a single source of information, we tried to control for common method bias with both procedural and statistical remedies. In terms of procedural remedies, we ensured respondent anonymity, provided a guidance with detailed instruction, added reversed items and minimized the length of the survey following guidelines provided by Podsakoff and Organ (1986) . In terms of statistical remedies, we conducted Harman's single-factor test to examine potential common method bias ( Harman, 1967 ; Podskoff and Organ, 1986 ). Our results of the Harman's single-factor test indicated that the single factor accounted for 43.37% of the total variance, not exceeding 50% ( Podsakoff and Organ, 1986 ). Thus, common method bias does not appear to be an issue in this study.

4.4 Testing hypotheses using structural equation model

The results of the testing using structural equation modeling showed that the hypothesized model yielded an excellent fit ( χ 2  = 2,194.49). In order to further assess the validity of the hypothesized model, we tested a more parsimonious model that removed the direct paths from the independent variables to retention. This would be an alternative model. According to the principle of model parsimony, an alternative model would fit the data better if the χ 2 value of the hypothesized model did not drop significantly. If the χ 2 value of the hypothesized model dropped significantly, however, the hypothesized model would fit the data better. Although the alternative model also yielded an excellent fit ( χ 2  = 2,316.78), our hypothesized model provided a significantly better model fit compared to the alternative model (Δ χ 2  = 122.29). Table 5 presents a summary of the fit indices for the hypothesized and alternative models.

Consistent with H1a , the effects of transformational leadership on the employees' intrinsic motivation were different among the generations. They were significant for Gen Y ( β  = 0.50, p  < 0.01) and Gen Z ( β  = 0.37, p  < 0.01), but not for Gen X ( β  = 0.18, n.s. ), supporting H1a . Although transformational leadership had a significant effect on all employees' retention ( β  = 0.30, p  < 0.01 for Gen X; β  = 0.33, p  < 0.01 for Gen Y; β  = 0.29, p  < 0.01 for Gen Z), the effects were not different across generation groups. Therefore, H1b was not supported.

Consistent with H2a , CSR was positively and significantly related to employees' intrinsic motivation for Gen Z ( β  = 0.23, p  < 0.05), but not for Gen X ( β  = 0.03, n.s. ) or Gen Y ( β  = −0.02, n.s. ). However, the effects of CSR on employees' retention were not significant in any of the generation groups ( β  = 0.03, n.s. for Gen X; β  = 0.03, n.s. for Gen Y; β  = 0.02, n.s. for Gen Z), failing to support H2b .

Supporting H3a , autonomy was positively and significantly related to employees' intrinsic motivation for Gen Y ( β  = 0.28, p  < 0.01) and for Gen Z ( β  = 0.24, p  < 0.05), but not for Gen X ( β  = 0.18, n.s. ). H3b was not supported because autonomy did not have a significant impact on employees' retention in any of the generation groups ( β  = 0.01, n.s. for Gen X; β  = 0.01, n.s. for Gen Y; β  = 0.01, n.s. for Gen Z).

Hypotheses 4a proposed that the effect of WLB on employees' intrinsic motivation would be more significant to younger generations while Hypothesis 4b proposed that the effect of WLB on employees' retention would be more significant to younger generations. The results revealed that WLB was positively and significantly related to employees' intrinsic motivation for Gen X ( β  = 0.42, p  < 0.01), but not for Gen Y ( β  = 0.04, n.s. ) and Gen Z ( β  = 0.05, n.s. ). However, WLB did not have a significant effect on employees' retention in any of the generation groups ( β  = 0.01, n.s. for Gen X; β  = 0.04, n.s. for Gen Y; β  = 0.02, n.s. for Gen Z). H4a was not supported because the effect of WLB on intrinsic motivation was not significant among younger generations, Gen Y and Gen Z. In addition, H4b was not supported because no significant difference was found among the three generations.

Hypotheses 5a and 5b proposed that the effect of technology on employees' intrinsic motivation ( H5a ) and their retention ( H5b ) would differ by generation. However, technology had no significant effect on employees' intrinsic motivation in any generation groups ( β  = 0.08, n.s. for Gen X; β  = 0.07, n.s. for Gen Y; β  = 0.05, n.s. for Gen Z). Furthermore, technology had no significant effect on employees' retention in any generation groups ( β  = 0.03, n.s. for Gen X; β  = 0.05, n.s. for Gen Y; β  = 0.04, n.s. for Gen Z). Based on these findings, neither H5a nor H5b was supported.

Hypothesis 6 proposed that employees' intrinsic motivation would be positively related to their retention in all generations. Our findings supported this contention ( β  = 0.54, p  < 0.01 for Gen X; β  = 0.48, p  < 0.01 for Gen Y; β  = 0.49, p  < 0.01 for Gen Z).

5. Discussion

5.1 theoretical contributions.

The results of this study provided several theoretical contributions to management literature. First, our results revealed interesting similarities and differences between generations in terms of the factors that affected employees' retention and motivation. For Gen X employees, transformational leadership was significantly related to retention and only WLB had a significant relationship with their intrinsic motivation. For Gen Y employees, transformational leadership was also the only factor affecting their retention, while both transformational leadership and autonomy had a significant impact on their intrinsic motivation. Finally, for Gen Z employees, only transformation leadership also mattered for their retention while transformational leadership, corporate social responsibility and autonomy were significantly related to their intrinsic motivation. For all three generations, there was a statistically significant relationship between intrinsic motivation and employee retention.

Second, this study proved that employees' generation affects the impacts of transformational leadership, CSR, autonomy, WLB and technology on their motivation and retention in the workplace. As motivating and retaining employees becomes more challenging and workforces become more diverse in terms of generation, understanding generational differences in employee motivation and retention becomes a very important topic to explore. Only a few studies looked at generational differences in either employee motivation ( Andrade and Westover, 2018 ) or employee's retention ( Roman-Calderon et al. , 2019 ) and no studies have examined the different effects of transformational leadership, CSR, autonomy, WLB and technology on employee motivation and retention spanning three different generations.

Third, the results of our study also showed that employees of different generations are intrinsically motivated by different factors, proving the importance of considering generational differences in motivation literature. However, our results did not provide empirical support for generational differences in retaining employees. Interestingly, only transformational leadership significantly affected employees of all generations. This finding would emphasize the critical role of leadership in retaining employees regardless of their generation.

5.2 Practical implications

The retention of an employee, especially younger generation employees, is pivotal in ensuring that organizations will be able to maintain sustained competitive advantages during the period of the pandemic since many companies have been experiencing serious younger generation employee retention issue. For instance, major retail companies, such as Target and Walmart, have been confronted with challenging managerial decisions because of the workforce shortage and have been forced to decrease their operation hours. To resolve this challenge, many companies have tried to increase the retention rate of their employees, especially those of the younger generation, by offering competitive financial and non-financial packages such as signing bonuses, healthcare benefits and/or opportunities for a college education. Despite all these endeavors, many companies have still been experiencing serious employee retention problems, which they have never experienced before. The findings of this study could be highly useful for organizations that are experiencing serious employee retention issues, many of whom are younger generation employees who are quitting their jobs during the pandemic.

First, these findings suggest reasons why so many organizations have had a challenging time managing low employee retention rates by showing that the impact of major factors (transformational leadership, CSR, autonomy, WLB and technology) on employee retention could vary depending on an employee's generation. For instance, our study's findings show that organizations actively implementing CSR policies may positively affect the retention of younger generation employees relative to older generations by intrinsically motivating younger generation employees more. Therefore, organizations should consider generational differences in employee motivation and retention when implementing employee retention strategies since an effective strategy for one employee generation may not be effective (or even harmful) for another employee generation.

Second, these results illustrate that employee retention is not a simple function, but rather a result of interactions between employee motivation and the specific generation. For instance, for Gen X, even though job autonomy does not directly affect employee retention, job autonomy still plays a crucial role in affecting employee retention by affecting employee motivation. Therefore, organizations should take care of factors affecting employee motivation as well because employee motivation works as a significant pathway to boost employee retention.

5.3 Limitations and future research directions

While the study advances our understanding in these areas, it has several limitations that future studies could explore. First, given that our study was a cross-sectional study with all responses collected from a single period, strong causality arguments cannot be made. Considering that our study collected the sample during the period of pandemic, it would have been a more interesting study if we deployed a longitudinal design because longitudinal data would have allowed us to examine how our independent variables affected employee motivation and retention as employee worked through the pandemic. Future studies should implement longitudinal design by collecting samples at different time points to provide greater insight into the causality argument as well as into the impact of the pandemic. Second, although this study examined the impact of employee motivation on employee retention as a significant pathway, we didn’t test the mediating effect of employee motivation on employee retention. Further studies could be done to investigate the mediating effect of employee motivation on employee retention in the context of different employee generations. Furthermore, regarding employee generation as a moderator, even though we used sub-sample design to examine the moderating impact of an employee's generation on employee motivation and retention, further studies could test the effect of the interaction between our independent variables and employee generation. Third, future studies could extend our study by examining whether our findings could change depending on the industry (e.g. retail, manufacturing) as well as firm characteristics (e.g. size). For instance, stronger impact of CSR on employee motivation in Gen Z may not exist in the financial industry wherein competitive environment and culture dominates.

Research framework

Summary of measure

Means, standard deviations and correlations of the variables