Nurse Turnover Intention: Meta Analysis Factors
Nurse Turnover Intention: Meta Analysis Factors
ABSTRACT
Turnover intention refers to an employee's intention to leave their current job and
seek new employment within a specific time frame. High turnover of nurses in hospitals can
create obstacles in providing services to patients. This study aims to analyze workload,
salary and job satisfaction with nurse turnover intention. The data source uses secondary
data obtained from previous research online. The data collection procedure uses the
Garuda Portal and Scopus which were published in the last 5 years. Data were analyzed
using Review Manager 5.4.1. Of the 375 studies obtained according to the inclusion and
exclusion criteria, 14 studies were obtained which could be continued into the analysis. The
workload variable uses 5 studies, salary 7 studies and job satisfaction 5 studies. The results
of data analysis show that there is a relationship between workload and nurse turnover
intention with a p value of 0.04, namely p<0.05 and a pooled odds ratio value of 2.36 (95%
CI 1.06-5.28). The results of data analysis show that there is a relationship between salary
and nurse turnover intention with a p value <0.00001 and a pooled odds ratio value of 0.71
(95% CI 0.64-0.79). The results of data analysis show that there is a relationship between
job satisfaction and nurse turnover intention, namely a p value <0.00001 and a pooled odds
ratio value of 0.63 (95% CI 0.50-0.78). The findings indicate that workload, salary, and job
satisfaction are significantly related to nurses’ turnover intention in hospitals, as evidenced
by this meta-analysis study.
59
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
recruit and train new nurses. If it is not addressed related to turnover intention in nurses. In general,
immediately, it will indirectly affect the quality of it shows that there is no consistency in research
service to patients. results that there is a relationship between
Turnover intention in a worker, especially workload, salary and job satisfaction on turnover
nurses, can be influenced by various things, intention among nurses. Given these
8
including according to Abelson , there are several discrepancies, this study aims to conduct a meta-
factors that influence the occurrence of turnover analysis to synthesize evidence from various
intention, including individual factors including studies on the relationship between workload,
age, education, skills, family size, workload, salary, and job satisfaction with turnover intention
length of work, personal typology, copying stress among nurses. The goal is to provide a clearer
and organizational factors include organizational understanding of these associations and to
policies, recruitment, salary, career development, inform hospital policies aimed at reducing nurse
job design, work affiliation, supervision, turnover.
leadership and environmental factors including
competitors and geography (distance or METHOD
transportation). Apart from that, according to This study is a systematic review and
9
Mobley , the factors that influence turnover meta-analysis aimed at assessing the effect size
intention consist of external factors, namely of workload, salary, and job satisfaction on
environmental aspects and individual aspects turnover intention among nurses. Meta-analysis
(young age, gender and length of service), is a high-quality method of systematic review
internal factors include organizational culture, considered one of the most reliable sources of
leadership style, compensation, job satisfaction evidence for guiding clinical practice. The
and career. This study is a development of objective of this systematic review is to provide a
Abelson's theory regarding factors that influence comprehensive summary of primary research
turnover intention, namely workload and salary addressing the research question.
and adds another factor, namely job satisfaction, The research protocol uses the concept
8,9
which comes from Mobley's theory. of Preferred Reporting Items for Systematic
21
Research related to workload conducted by Reviews and Meta-Analysis (PRISMA). The
(Siregar (2019); Xu et al (2020); Okstoria (2022)) quality assessment of the article is based on the
obtained (p < α (0.05)) stating that there is a criteria of the Joanna Briggs Institute Checklist
relationship between workload and turnover for analytical cross-sectional studies (JBI). JBI is
10-12
intention in nurses. Meanwhile, according to an instrument to assess the methodological
research conducted by (Lee (2020); Yoon quality of a study and to determine the extent to
(2020)), based on the chi-square test, the value which a study has addressed potential bias in its
obtained (p > 𝛼� (0.05)) states that there is no design, implementation and analysis. This
relationship between workload and turnover instrument includes 8 questions with the answers
13,14
intention in nurses. Yes, No, Unclear, and Not applicable. The results
The results of research (Bolado (2023); of this assessment can then be used to inform
Sokhanvar (2018)) using the chi square test the synthesis and interpretation of research
obtained a value (p < 𝛼� (0.05)) stating that there results. Literature or articles that can be said to
is a relationship between salary and turnover be good and valid to be used as research
15,16
intention in nurses. Meanwhile research by material if they get results above 50% of the
22
(Bae (2023); Boateng et al (2022)) states that the research.
results of the analysis between salary and The 8 questions include:
turnover intention in nurses based on the chi- 1. Were the inclusion criteria for the sample
square test obtained a value of (p > 𝛼� (0.05)) clearly defined?
which means there is no relationship between 2. Were the study subjects and setting
17,18
salary and turnover intention in nurses. described in detail?
Research related to job satisfaction by (Hua 3. Were the exposures measured in a valid
(2023); Liu (2023)) states that (p < 𝛼� (0.05)) and reliable manner?
there is a relationship between job satisfaction 4. What were the standard criteria used for
and turnover intention in nurses. Meanwhile, measuring the objective conditions?
according to research conducted by (Muharni 5. Were confounding factors identified?
(2020)), it is stated that (p > 𝛼� (0.05)) there is no 6. Were strategies for dealing with confounding
relationship between job satisfaction and turnover factors stated?
19,20
intention in nurses. 7. Were the outcomes measured in an
Based on the description above, it can be appropriate and reliable manner?
seen that there are differences in results between 8. Was the statistical analysis used
one study and another, where many studies have appropriate?
been carried out to determine the risk factors
60
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
61
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
The study quality of each article provisions of the systematic review and were
designated as a systematic review reference was given a quality score of 5-6 sufficient, 7-8 good,
determined based on the quality analysis of the in studies with articles using the JBI critical
JBI critical appraisal tools, so that 14 articles appraisal tools with a cross-sectional study
were obtained that were in accordance with the design.
PRISMA describes the steps taken by Screening based on duplication, journals that
researchers in conducting meta-analysis starting meet the above criteria are obtained, then a
from research identification to research included check is carried out to see if there are the
in the meta-analysis. The search or collection of same/duplicate articles. If the same article is
data in this study was carried out by conducting a found, it is discarded so that there is only one
literature search through electronic media with article with the same title and content, the
reference to database sources, including the number of articles after screening is based on
Garuda Portal and Scopus with Indonesian duplication. Screening based on abstracts,
keywords and English keywords with the help of abstracts that are not relevant to the screening
the Publish or Perish application. Furthermore, criteria will be excluded. All data (research
data screening is the filtering or selection of data articles) obtained are in the form of quantitative
(research articles) which aims to select research research articles that have met all the
problems that are in accordance with the topic or requirements and criteria that have been
title, abstract and keywords being studied. determined so that the analysis stage can be
Screening is carried out based on the carried out. Then, determine the final number of
identification of appropriate titles and a time span journals that match the inclusion criteria and
of 5 years (2018-2023), the types of journals exclusion criteria.
used are research articles and original articles.
62
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
A total of 14 research journals were workload and turnover intention in nurses, all of
reviewed and were eligible for data input using which are cross-sectional studies with a total (n)
the RevMan 5.4.1 program. There are 5 research sample of 1.049.
journals that examine the relationship between
A total of 14 research journals were pay and turnover intention in nurses, all of which
reviewed and were eligible for data input using are cross-sectional studies with a total (n) sample
the RevMan 5.4.1 program. there are 5 research of 22.533.
journals that examine the relationship between
A total of 14 research journals were job satisfaction and turnover intention in nurses,
reviewed and were eligible for data input using all of which are cross-sectional studies with a
the RevMan 5.4.1 program. There are 5 research total (n) sample of 21.457.
journals that examine the relationship between
63
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
The image above shows the results of a showing that many nurses with a low workload
meta-analysis of data from 5 research articles increase the risk of turnover intention compared
regarding the relationship between workload and to nurses with a high workload.
turnover intention among nurses in hospitals. The The results of the data analysis displayed
variance in effect size includes the variance in in the forest plot show that there is a relationship
2
the actual effect size of the research study. The I between workload and turnover intention among
value obtained is 95% and the p-value of the nurses in hospitals with a value of p=0.04,
heterogeneity test is p<0.00001, meaning that namely p<0.05 and a pooled odds ratio value of
heterogeneity between studies is high because 2.36 with a CI range of 95% (1.06-5.28). This
the p-value is <0.05 and heterogeneity is >50%, value shows that nurses with a low workload
so To assess the magnitude of the effect of increase the risk of turnover intention 2.36 times
workload on turnover intention in nurses using a greater than nurses with a high workload.
random effect model. The diamond image is Workload tends to have a large effect on turnover
more to the right of the vertical line (line no effect) intention with an ES value (2.09).
Figure 3. The Funnel Plot of the Analysis of the Relationship between Workload with turnover
Intention in Nurses
In the funnel plot above, it can be seen that continued using meta-analysis because
the studies are spread asymmetrically which according to Sastroasmoro (2011) in his book,
indicates that there is publication bias for studies meta-analysis is a combination of two or more
that examine the relationship between workload studies. So it can be concluded that with a
22
with turnover intention in nurses. minimum of two studies, quantitative analysis can
The results of journal searches from the be carried out with meta-analysis. 5 studies with
Garuda Portal and Scopus ultimately produced 5 workload variables were analyzed involving 1,049
articles that met the inclusion criteria that could research samples from the total number of
be analyzed into a meta-analysis with the samples in each research study. Analysis results
RevMan 5.4.1 application software. This can be for 3 research studies which stated there was a
64
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
relationship between workload and nurse high standards and achievements so that the
turnover intention (Okstoria (2022); Siregar workload is felt to be high. Apart from that,
(2019); Xu (2020)) and 2 studies which stated inappropriate division of tasks causes high
there was no relationship between workload and workloads and physical and emotional fatigue in
nurse turnover intention (Lee (2020); Yoon nurses which then gives rise to turnover intention
10-14
(2020). Results of the heterogeneity test in nurses. The results of this study are supported
2
analysis of 5 study articles. The I value obtained by Xu et al's research, there is a significant
was 95% and the heterogeneity test p-value was relationship between workload and the desire for
p<0.00001, meaning that heterogeneity between turnover intention in nurses (p-value=0.001). The
studies was high. The results of the data analysis results of the analysis show an OR value=5.33.
displayed in the forest plot show that there is The This determines that nurses with a low workload
relationship between workload and turnover have 5.33 times more likely to want to make
intention among nurses in hospitals with a value turnover compared to nurses with a high
of p=0.04, namely p<0.05 and a pooled odds workload.
ratio value of 2.36 with a CI range of 95% (1.06- Based on findings from research articles,
5.28). shows that nurses with a heavy workload workload is a factor that can influence turnover
increase the risk of turnover intention 2.36 times intention. Research shows that excessive
greater than nurses with a light workload. workload can lead to the need to work an
Workload tends to have a large effect on turnover excessive number of hours, which is an
intention with an ES value of (2.09). additional source of stress. This can cause
The funnel plot appears asymmetrical turnover intention. Fatigue and exhaustion can
indicating that there is publication bias. The occur when nurses work more than 80% of their
results of this meta-analysis research are in line work time. This means that the nurse's
with research by Okstoria which states that there productive time is less than or only reaches 80%,
is a relationship between workload and turnover if it is more than 80% then the nurse's workload
12
intention in nurses (p-value= 0.002). Excessive is said to be high or inappropriate so that
workload will affect the productivity of health excessive working hours cause stress for nurses
workers and of course the productivity of the and greatly affect the level of turnover intention.
hospital itself. High workload is obtained due to
Figure 4. The Results of Forest Plot the Analysis of the Relationship between Pay with Turnover
Intention in Nurses
The image above shows the results of a that many nurses with low salaries or who are
meta-analysis of data from 7 research articles dissatisfied can increase the risk of turnover
regarding the relationship between salary and intention compared to nurses with high salaries
turnover intention among nurses in hospitals. The or who are satisfied.
variance in effect size includes the variance in The results of the data analysis displayed
2
the actual effect size of the research study. The I in the forest plot show that there is a relationship
value obtained is 47% and the p-value of the between salary and turnover intention among
heterogeneity test is p=0.08, meaning nurses in hospitals with a value of p<0.00001,
heterogeneity between studies is low because namely p<0.05 and a pooled odds ratio value of
the p value is>0.05 and heterogeneity is <50%, 0.71 with a CI range of 95% (0.64-0.79). This
so To assess the magnitude of the effect of salary value shows that nurses with high salaries or
on turnover intention for nurses using a fixed satisfaction can prevent the risk of turnover
effect model. The diamond image is more to the intention. Salary tends to have a large effect on
right of the vertical line (line no effect) showing turnover intention with an ES value of (6.57).
65
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
Figure 5. The Results of Funnel Plot the Analysis of the Relationship between pay with Turnover
Intention in Nurses
In the funnel plot above, it can be seen that to nurses greatly determines the nurse's desire to
26
the studies are spread asymmetrically which change jobs. Therefore, the greater the salary
indicates that there is publication bias for studies given, the higher the satisfaction with salary will
that examine the relationship between salary and be. The higher the level of salary satisfaction, the
22 27
turnover intention in nurses. lower the turnover intention rate and vice versa.
Analysis results for 4 research studies The results of this meta-analysis are in line with
which stated there was a relationship between research conducted by Tang (2023) which stated
salary and nurse turnover intention (Bolado that there was a relationship between salary and
25
(2023); Gebregziabher (2020); Sokhanvar nurse turnover intention (p-value=0.001). This
(2018)) and 3 studies which stated there was no research is in accordance with research
relationship between salary and nurse turnover according to Bae (2023) which states that there is
intention (Bae (2023); Tang (2023); Boateng a relationship between basic salary and nurse
15-18,24-25
(2022); Liu (2023))). The results of the turnover intention (p-value=0.001) with an OR=
analysis of 7 studies showed that the I2 obtained 0.69 meaning that nurses with low salaries are
was 47% and the p-value of the heterogeneity 0.69 times more likely to want to make a turnover
17
test was p=0.08, meaning that heterogeneity than nurses. with high salaries.
between studies was low, so the fixed effect According to research articles, it was found
model was used. The results of the data analysis that turnover intention was caused by salary
displayed in the forest plot show that there is a dissatisfaction or low salaries, mismatches in
relationship between salary and turnover salary and allowances received caused high
intention among nurses in hospitals with a value turnover intention of nurses in hospitals. Nurses
of p<0.00001, namely p<0.05 and a pooled odds with lower salaries show higher turnover
ratio value of 0.71 with a CI range of 95% (0 .64- intentions than employees with higher salaries.
.79). This value shows that nurses with high One of the hopes of a person working is to fulfill
salaries or satisfaction can prevent the risk of their own and their family's living needs. If a
turnover intention. Salary tends to have a large nurse feels that the salary they are given cannot
effect on turnover intention with an ES value of fulfill their living needs, they will feel dissatisfied
(6.57). The funnel plot appears asymmetrical with their salary and potentially the employee will
indicating that there is publication bias. leave the hospital and look for another hospital
According to research conducted by with an appropriate salary. with his and his
17
Suyono et al (2018) on nursing staff in hospitals, family's living needs.
it was stated that the salary given by the hospital
66
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
Figure 6. The Results of Forest Plot the Analysis of the Relationship between Job Satisfaction with
Turnover Intention in Nurses
The image above shows the results of a nurses with low job satisfaction increase the risk
meta-analysis of data from 5 research articles of turnover intention compared to nurses with
regarding the relationship between job high job satisfaction.
satisfaction and turnover intention among nurses The results of the data analysis displayed
in hospitals. The variance in effect size includes in the forest plot show that there is a relationship
the variance in the actual effect size of the between job satisfaction and turnover intention
research study. The I2 value obtained is 99% and among nurses in hospitals with a value of
the p-value of the heterogeneity test is p<0.00001, namely p<0.05 and a pooled odds
p<0.00001, meaning that heterogeneity between ratio value of 0.63 with a CI range of 95% ( 0.50-
studies is high because the p-value is <0.05 and 0.78). This value shows that nurses with high job
heterogeneity is >50%, so To assess the satisfaction can prevent the risk of turnover
magnitude of the effect of job satisfaction on intention. Job satisfaction tends to have a large
turnover intention in nurses using a random effect effect on turnover intention with an ES value of
model. The diamond image is more to the right of (4.09).
the vertical line (line no effect) showing that many
Figure 7. The Results of Funnel Plot the Analysis of the Relationship between Job Satisfaction with
Turnover Intention in Nurses
In the funnel plot above, it can be seen that Analysis results for 4 research studies
the studies are spread asymmetrically which which stated there was a relationship between
indicates that there is a publication bias for job satisfaction and nurse turnover intention (Liu
studies that examine the relationship between job (2023); Hua (2023); Bae (2023); Tang (2023))
22
satisfaction with turnover intention in nurses. and 1 study which stated there was no
67
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
68
Qadrinnisa R. et al. Meta Analysis: Analysis of… Jurnal Berkala Kesehatan
DOI: 10.20527/jbk.v11i1.19192 2025; 11(1): 59-69
69