Staffing, nursing home, and market characteristics associated with caregiver turnover in Japan: a national cross-sectional study

Abstract

Background

Caregiver turnover in nursing homes remains a critical issue in long-term care. However, few nationwide studies have distinguished between full-time and part-time caregivers, despite their differing work expectations. This study examined national turnover rates in Japan and identified staffing, facility, and market factors associated with turnover by employment type.

Methods

We conducted a cross-sectional analysis of 4,042 nursing homes operating in 2022, using data from the National Long-Term Care Service Information Disclosure System. Annual caregiver turnover was defined as the number of caregivers who left during the year divided by the average number of caregivers. Geographic variations were assessed by mapping turnover rates across prefectures and municipalities. Logistic regression models were applied to identify staffing, facility, and market factors associated with high turnover rate, with stratification by full-time and part-time employment status.

Results

The mean annual turnover rate was 12.6% overall, 11.5% for full-time, and 25.0% for part-time caregivers. Considerable geographic variation was observed (4.9%–15.0%). Lower full-time turnover was associated with higher caregiver-to-bed ratios, greater proportions of certified caregivers, and staff with ≥ 3 years’ experience. Increased night-shift staffing was linked to higher turnover for both groups. Larger facilities, shorter years in operation, and provision of sputum suction or end-of-life care were related to lower turnover. At the market level, higher unemployment and higher income were associated with higher turnover.

Conclusions

Nationwide evidence demonstrates that caregiver turnover is shaped by staffing, facility, and market factors, with distinct patterns between full-time and part-time staff. Recognizing these differences may help nursing homes and policymakers consider more tailored approaches to workforce retention and create supportive environments that reduce turnover.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14074-4.

Background

A persistent challenge in long-term care faces high caregiver turnover rates, which can have detrimental effects on both the quality of care and nursing home residents’ health status [1, 2]. The issue of caregiver turnover is particularly dire in societies characterized by a rapidly aging population and significant shortage of caregivers, such as Japan. In Japan, the turnover rate of full-time caregivers in 2017 surpassed 19.0%, significantly exceeding the average turnover rate across all industries (12.1%) [3]. Furthermore, nursing home caregivers experience a higher turnover rate than home caregivers [3].

Caregiver turnover, or caregiver turnover rate, measures the number of caregivers who leave employment during a specified period, typically one year [1, 4]. High caregiver turnover is reported to negatively impact not only nursing homes, but also caregivers’ performance and residents’ outcomes. For providers, frequent turnover necessitates continuous recruitment and training of new caregivers, which can be time-consuming and expensive [5]. When turnover occurs, the remaining caregivers may be required to assume additional responsibilities or work longer hours, leading to burnout and exhaustion [6]. Residents may not have consistent caregivers familiar with their specific needs, preferences, and routines, which can negatively impact their well-being [2, 7].

Caregiver turnover has interested researchers for decades and has led to a plethora of studies. However, many studies have only collected data from small numbers of facilities [6, 7], such as those only in a single state; moreover, few have been conducted at the national level, despite potential regional differences in caregiver turnover [2]. In addition, studies have failed to examine caregiver turnover separately by work type, such as full-time or part-time [4, 7–9]. According to the Japan Care Work Foundation, motivation and expectations are reported to differ between full-time and part-time caregivers; [10] therefore, there is a need to analyze the results separately by work type. Most studies have reported staffing and organizational factors as predictors of caregiver turnover; however, these findings are based on data from the United States and United Kingdom and have not been validated in other countries implementing universal coverage of long-term care [1, 2, 6–9].

In Japan, caregiver turnover is strongly shaped by staffing, organizational, and market environments due to the structural characteristics of the LTC labor market. Staffing shortages and an insufficient mix of certified and experienced caregivers can intensify workloads, emotional labor, and role stress, which are established predictors of turnover in care work [11]. Moreover, facilities with demanding working arrangements—such as a high resident case mix or extensive night-shift duties—may create work environments that increase burnout and job dissatisfaction. From a labor market perspective, Japan’s chronic nationwide shortage of caregivers enables high job mobility among care workers even during periods of economic stagnation, making workplace conditions a critical determinant of turnover. These structural features suggest that staffing, facility characteristics, and market conditions may jointly influence turnover in Japan’s LTC sector.

This study uses the 2022 National Long-Term Care Service Information Disclosure System to produce national estimates of caregiver turnover in Japan. This study aims to (i) determine the regional distribution of caregiver turnover in nursing homes and (ii) explore the factors associated with turnover separately for full-time and part-time caregivers.

Methods

Data source and sample

The data were obtained from the National Long-Term Care Service Information Disclosure System. The system, launched in April 2006, empowers users to freely search for and collect comprehensive information on long-term care providers, enabling them to compare and make informed choices regarding long-term care providers. Almost all nursing homes update their data once a year, including basic information (i.e., location, facility size, and ownership), staffing levels, and the range of services provided.

A total of 4,042 nursing homes, operating in 2022, were included, representing 94.5% of all the nursing homes in Japan. There are three types of nursing homes in Japan, and our study included only one type—“Geriatric Health Services Facility,” which is an intermediary facility between hospitals, homes, and living long-term care facilities [12].

Measures

Caregiver turnover rate

The annual turnover rate is defined as the sum of caregivers turned over in a given year divided by the average number of caregivers. The average number of caregivers was calculated by adding the numbers at the beginning and end of a given year and dividing this sum by two [5]. This method provides a more accurate estimate of the population at risk of leaving than using only the number of caregivers at a single time point, especially when staffing levels fluctuate. We did not observe whether caregivers leave nursing homes voluntarily or involuntarily. In previous studies, skewed turnover distributions have been grouped into categories [4, 5, 7], dividing turnover rates at the median into low and high groups. In our dataset, turnover also demonstrated a highly right-skewed distribution with extreme values, particularly among part-time caregivers. Using turnover as a continuous variable would therefore lead to model instability and disproportionately reflect the influence of outliers. Categorizing turnover into high and low groups helps reduce the impact of skewness, improves model stability, and facilitates interpretation from a managerial and policy perspective.

Regional distribution in the caregiver turnover rate

The regional distribution of caregiver turnover was calculated at the prefecture and municipality levels. Regional disparities are expressed in terms of the coefficient of variation and max/min ratio.

Exposures

The selection of exposures was guided by previous studies [1, 5, 6, 13] that have commonly grouped variables into three levels: staffing, nursing home, and market characteristics. In addition to ensuring comparability with prior work, these categories were chosen based on their theoretical relevance to caregiver turnover, as staffing mix and experience can influence caregiver workload and support, nursing home characteristics can shape working conditions and care burden, and market conditions can affect job mobility. Therefore, the selected covariates reflect key factors that may influence caregiver turnover from organizational and labor market perspectives.

Staffing characteristics

In previous studies, staffing level has been reported as a predictor of turnover. We included the number of nurses and caregivers per 100 beds, respectively; percentage of caregivers with at least three years of work experience; and percentage of certified caregivers. As employment status is related to turnover, we added the ratio of full-time to part-time caregivers. We added one variable to determine whether the number of night-shift staff members was higher than the standard.

Nursing home characteristics

Nursing home demographic features, such as years in business, number of beds, location, and types of care (conventional vs. unit), were included. Conventional care is mainly provided in nursing homes in a shared-room setting. By contrast, unit care refers to person-centered care for a small number of residents (< 10) living in a single unit, where care is mainly provided in private rooms. The resident case mix (i.e., percentage of moderate and severe care levels among all residents), which served as a proxy for the clinical demand of nursing homes, was included.

Our study added two business management policies [14] that were assumed to be related to caregivers’ working environments: providing end-of-life care, and caregivers providing sputum suctioning services.

Market characteristics

We included three variables representing the regional economic status at the municipal level, which influences staff turnover: the unemployment rate, per capita taxable income, and competition. We applied the Herfindahl-Hirschman index to measure market concentration, which was calculated as the sum of the squared market shares of nursing homes in the municipality [7].

Statistical analysis

The regional distributions of caregiver turnover were calculated for each prefecture and municipality. Regional disparities in caregiver turnover were expressed in terms of the coefficient of variation and max/min ratio. A descriptive analysis was conducted to review the distribution of outcomes and exposures. Multicollinearity and collinearity levels among the independent variables were examined before conducting a multivariate analysis. We performed a logistic regression to identify the staffing, nursing home, and market characteristics associated with high caregiver turnover rates (defined as turnover rates above the median), separately for all caregivers, full-time caregivers, and part-time caregivers. Data management and analysis were performed using STATA version 16.0, and P-values < 0.05 were regarded as statistically significant.

Results

Descriptive statistics

Table 1 presents the descriptive statistics and definitions of the variables used in this study. The average turnover rates were 12.6%, 11.5%, and 25.0% for all caregivers, full-time caregivers, and part-time caregivers, respectively. On average, the nursing homes consisted of 87.8 beds and had been in operation for 20.8 years. Regarding staffing, there was an average of 31.9 caregivers and 11.8 nurses per 100 residents. The distributions of turnover rates for all caregivers, full-time caregivers, and part-time caregivers are shown in Supplementary Figure S1. In addition, the background characteristics of facilities between the high and low turnover groups are summarized in Supplementary Table S1.

Table 1.

Distribution and measurement of dependent and independent variables used in the analysis

	Definition	Mean (SD) or percent
Dependent variables
Turnover among all caregivers (%)	Number of caregivers who left in 2022/ [(number of caregivers at the beginning + number of caregivers at the end)/2]	12.6% (13.3)
Turnover among full-time caregivers (%)	Number of full-time caregivers who left in 2022/ [(number of full-time caregivers at the beginning + number of part-time caregivers at the end)/2]	11.5% (13.5)
Turnover among part-time caregivers (%)	Number of part-time caregivers who left in 2022/ [(number of part-time caregivers at the beginning + number of part-time caregivers at the end)/2]	25.0% (65.7)
Independent variables		N = 4,042
Staffing characteristics
Nurse staffing	FTE nurses per 100 beds	11.8 (4.1)
Caregiver staffing	FTE caregivers per 100 beds	31.9 (8.7)
% of certified caregivers	Percentage of certified caregivers among total caregivers	68.3% (21.0)
% of caregivers with at least three years of experience	Percentage of caregivers with at least three years of experience among total caregivers	75.7% (16.2)
% of full-time caregivers	Percentage of full-time caregivers among total caregivers	82.3% (14.3)
Increased night-shift staffing	Percentage of facilities that number of night-shift staff members is higher than the standard	12.0%
Nursing home characteristics
Facility size	Number of beds	87.8 (31.7)
Years in business	Number of years in business	20.8 (8.2)
Unit care facilities	Percentage of unit care facilities	12.1%
Resident case mix	Percentage of moderate and severe care levels among all residents	68.2% (11.8)
Providing sputum suctioning services	Caregivers providing sputum suctioning services	75.8%
Providing end-of-life care	Caregivers providing care to residents who are near the end of life	34.6%
Market characteristics
Unemployment rate	The number of unemployed people as a percentage of the labor force	3.9 (0.8)
Competition Herfindahl-Hirschman Index	The sum of each nursing home’s squared percentage share of beds in the municipality for all nursing homes in the municipality	35889.0 (2833.5)
Per capita taxable income	The total municipality taxable income divided by the number of people in the municipality	323.6 (57.7)

Note: n = 4,042 for all caregivers; n = 4,039 for full-time caregivers; n = 3,572 for part-time caregivers; SD: standard deviation; FTE: full-time equivalent

Regional distribution

The annual caregiver turnover rates varied largely across prefectures, with a mean of 9.34%, ranging from 4.87% to 14.96% (max/min ratio 3.07), and prefectures that included the three largest cities in Japan (Tokyo, Oska, and Nagoya) had high turnover rates (Fig. 1). Regional differences in caregiver turnover by municipality ranged from 0% to 142.9% (Supplementary Figure S2).

Fig. 1.

Annual caregiver turnover rate in nursing homes, by prefecture, in 2022. Note. SD: standard deviation

Correlates of the caregiver turnover rate

Table 2 present the odds ratio estimates for the probability of a high turnover among all caregivers, full-time caregivers, and part-time caregivers. Regarding full-time caregivers, several staffing characteristics were significantly associated with turnover. A higher number of caregivers per 100 beds, a higher proportion of certified caregivers, and a higher proportion of caregivers with at least three years of experience were all related to lower turnover. In contrast, facilities with more night-shift staff than the national standard related to higher turnover. With respect to nursing home characteristics, a longer duration of business and the provision of end-of-life care were associated with lower turnover. In terms of market characteristics, nursing homes in areas with higher unemployment rates or higher per capita taxable income had higher turnover.

Table 2.

Staffing, nursing home, and market characteristics associated with caregiver turnover

	Turnover among all caregivers (n = 4,043)		Turnover among full-time caregivers (n = 4,039)		Turnover among part-time caregivers (n = 3,572)
	OR	95%CI	OR	95%CI	OR	95%CI
Staffing characteristics
Nurse staffing	1.001	(0.984–1.020)	1.003	(0.985–1.021)	0.980	(0.959–1.001)
Caregiver staffing	0.986	(0.977–0.995)	0.986	(0.977–0.995)	1.013	(1.003–1.023)
% of certified caregivers	0.988	(0.984–0.993)	0.990	(0.986–0.994)	0.996	(0.992–1.0003)
% of caregivers with at least three years of experience	0.978	(0.973–0.982)	0.978	(0.973–0.982)	0.991	(0.986–0.996)
% of full-time caregivers	0.999	(0.994–1.004)	1.004	(0.9995-1.010)	0.949	(0.943–0.955)
Increased night-shift staffing	1.356	(1.097–1.678)	1.290	(1.045–1.591)	1.385	(1.088–1.763)
Nursing home characteristics
Facility size	1.002	(1.00007–1.005)	1.002	(1.0001–1.005)	1.015	(1.013–1.018)
Years in business	0.989	(0.980–0.998)	0.989	(0.980–0.999)	0.995	(0.985–1.005)
Unit care facilities	1.051	(0.825–1.341)	1.079	(0.848–1.373)	1.128	(0.860–1.478)
Resident case mix	1.001	(0.996–1.007)	0.999	(0.993–1.004)	1.000	(0.994–1.007)
Providing sputum suctioning services	0.847	(0.725–0.990)	0.961	(0.823–1.121)	0.811	(0.682–0.965)
Providing end-of-life care	0.805	(0.699–0.928)	0.810	(0.704–0.932)	0.927	(0.792–1.083)
Market characteristics
Unemployment rate	1.180	(1.084–1.285	1.199	(1.102–1.305)	1.044	(0.952–1.144)
Per capita taxable income (Ten thousand yen)	1.004	(1.002–1.005)	1.004	(1.002–1.005)	1.001	(1.000-1.003)
Competition Herfindahl-Hirschman Index	0.99997	(0.99995–1.000001)	0.99998	(0.99996–1.00001)	0.99997	(0.99994–1.000001)

Note. OR: odds ratio; 95%CI: 95% confidence interval

For part-time caregivers, the pattern differed in several respects. Nursing homes with a larger number of caregivers and a lower proportion of full-time staff were associated with higher turnover, in contrast to the findings for full-time staff. As with full-time caregivers, a higher proportion of caregivers with at least three years of experience was associated with lower turnover, whereas excess night-shift staffing above national standards was linked to higher turnover. Larger facility size was also associated with increased part-time turnover. In addition, facilities providing sputum suctioning services showed lower turnover among part-time caregivers.

Discussion

This is the first study to examine caregiver turnover at the national level, differentiating between full-time and part-time positions. The annual turnover rate of caregivers in Japan is 12.6%. We observed substantial regional variations in caregiver turnover rates across prefectures. Specific staffing, nursing home, and market characteristics were differentially associated with full-time and part-time caregiver turnover.

The average turnover rate in Japan (12.6%) is lower than that in other countries. However, international comparisons are difficult because caregiver definitions vary by country. Even for the same country, such as the United States, surveys have found varying average turnover rates for nursing assistants, from 18% to 81%, mainly because of the different ways of measuring turnover. In Japan, the turnover among caregivers in 2022 decreased from that in 2017; however, it was still somewhat higher than the turnover rate in all industries (11.1%) [15].

There were significant regional differences in caregiver turnover rates according to prefecture and municipality. The regional differences shown on the map (Fig. 1) are helpful for local governments to understand and evaluate the current status of their regions. Although this study did not identify the causes of regional differences, we believe that the revealed factors related to caregiver turnover at the facility level may help policymakers target turnover. Future studies are needed to analyze the factors that contribute to regional differences.

Staffing level, qualifications, and experience of caregivers were associated with a low caregiver turnover among all caregivers and full-time caregivers.

Consistent with previous studies [4, 6], nursing homes with more caregivers per bed were associated with a lower turnover. Indeed, when there are more caregivers, the workload of each caregiver can be reduced, alleviating their burden. Facilities with more qualified and skilled caregivers tended to experience lower turnover rates.

Care worker certification is a national qualification in nursing care that can be obtained by passing a national exam after 1,800 h of training [16]. A previous study has concluded that highly qualified caregivers help decrease residents’ problematic behaviors and maintain physical function [17]. Hence, qualified caregivers may be better equipped to handle challenges in their profession, thereby reducing the risk of burnout. Lower stress levels contribute to lower turnover rates. The presence of more experienced caregivers in a facility is associated with a lower turnover rate among caregivers. Experienced staff may act as preceptors for recently appointed staff, providing knowledge [18] and helping new caregivers resolve challenges smoothly, leading to increased job satisfaction and decreased stress.

However, the factors associated with turnover differed partially among part-time caregivers. Facilities with more caregivers per bed were associated with a higher turnover among part-time caregivers. One possible interpretation is that having a significant number of caregivers could reduce the number of part-time shifts available, which may decrease the income of part-time caregivers. Meanwhile, a higher percentage of full-time caregivers was associated with a lower turnover among part-time caregivers. One explanation for this may be that full-time caregivers generally have a stable and continuous work schedule that creates a supportive and reliable environment for part-time caregivers. Additionally, when an organization relies heavily on part-time staff, it can lead to irregular and fluctuating work schedules, which may affect part-time caregivers’ income and job satisfaction.

Nursing homes with more night staff than the national standard were associated with a higher probability of a high turnover. Although reports on caregivers are scarce, the adverse effects of night shifts on nurses’ health have been reported in many previous studies [19]. Night shifts reportedly lead to poor sleep quality and increased risk of depression [20]. A night shift disturbs cortisol and melatonin rhythms and can cause health problems [21, 22]. Furthermore, the higher physical demand placed on caregivers who work night shifts may make them consider leaving their jobs. Recently, the Ministry of Health, Labour, and Welfare has been actively developing care robots, and nighttime monitoring devices are becoming increasingly popular in nursing care homes [23]. According to a report by the Ministry of Health, Labour and Welfare, the introduction of nighttime monitoring devices is likely to reduce the amount of time that nighttime staff spend patrolling and directly providing care [24]. Therefore, utilizing monitoring devices that lessen the burden on night staff and appropriately decrease the number of night staff would be beneficial for reducing caregiver turnover rates in nursing homes.

Consistent with previous studies [4, 6, 13], larger nursing homes were associated with a high turnover. Nursing home size is likely to influence the available resources and management practices [13]. Larger nursing homes may face communication and coordination challenges linked to care deficiencies and high turnover rates among caregivers [25].

Two business management policies, regarding providing end-of-life care and sputum suction services, were related to a lower caregiver turnover. In Japan, end-of-life care necessitates a strong emphasis on multidisciplinary collaboration, and caregivers closely partner with physicians and nurses to ensure comprehensive care for residents [26]. This collaborative approach fosters a supportive and cohesive environment, resulting in increased motivation to work [27]. Sputum suction has been defined as a medical procedure that can only be performed by physicians and nurses. However, in April 2012, a registration system for sputum suctioning services was established, and caregivers were able to perform sputum suctioning after training when nursing homes had registered in the system [26]. Training and skill-development programs for caregivers and staff enhance their knowledge, expertise, and confidence in providing high-quality care. This may result in better job performance, [28] which may reduce turnover intention.

High unemployment rates were associated with a high probability of turnover. Although higher unemployment typically reduces job opportunities in the general labor market, the mechanism may function differently in the long-term care sector. One possible explanation is that Japan has experienced a persistent shortage of care workers regardless of macroeconomic conditions [29], resulting in sustained demand within the sector. Under such conditions, even in municipalities with high unemployment rates, caregivers may still perceive multiple job opportunities and may be more likely to change jobs in search of better compensation, working conditions, or career prospects. In addition, because unemployment was measured at the municipal level, the observed association may partly reflect geographical mobility rather than a direct causal mechanism. For example, caregivers working in municipalities with high unemployment may move to regions with more favorable labor conditions, leading to separation from their current workplace and contributing to higher turnover. These interpretations are speculative and cannot be directly tested with the current data; therefore, they should be interpreted with caution.

In addition, because unemployment was measured at the municipal level, the observed association may partly reflect geographic mobility rather than a direct causal mechanism. For example, caregivers working in municipalities with higher unemployment rates may relocate to areas with more favorable labor market conditions, leading to separation from their current workplaces and contributing to higher observed turnover. These interpretations are speculative and cannot be directly tested with the current data; therefore, they should be interpreted with caution.

High-income areas were associated with a high caregiver turnover. Higher-income areas often have a higher cost of living, which can affect caregivers’ household finances. According to the Nippon Care Service Craft Union, in 2021, caregivers’ annual income was significantly lower than that in all industries, and more than half of caregivers were dissatisfied with their current wages [30]. When caregivers feel that they are underpaid, they may be inclined to seek employment in areas with more manageable living costs.

Limitations

Our study has several limitations. First, the cross-sectional approach to the dependent variables limited our ability to make causal inferences from our findings. Second, several important factors were not controlled because of the limited information in our data. For example, the leadership styles of the administrators and trained supervisors were not considered. Further, we did not consider the financial situation of the nursing homes. Third, cases in which a part-time caregiver leaves their current job and becomes a full-time caregiver at the same facility were not considered. Fourth, we used facility-level data; thus, caution is needed before applying our results to individuals to avoid ecological fallacies. Fifth, our study only included nursing homes that serve as intermediate facilities, and its application in other types of nursing homes, for example, as places for daily living, requires further verification. Finally, the dataset did not allow us to distinguish between permanent and temporary or dispatched caregivers. The turnover rate analyzed in this study reflects only caregivers registered as regular staff within each facility. Although dispatched workers themselves were not included in the turnover measure, their increasing use in Japan may still influence workplace conditions and job satisfaction of regular staff and could indirectly affect turnover. Future studies with more detailed workforce information are needed to examine this mechanism.

Conclusions

Our national investigation revealed that caregiver turnover varies considerably across nursing homes and regions. This study will empower nursing homes and local governments to gain a comprehensive understanding of the status and challenges related to caregiver turnover in their region. High caregiver turnover is associated with specific staffing, nursing home, and market characteristics. These findings offer practical measures to effectively reduce caregiver turnover rates in Japan.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

XJ contributed to writing the original draft, methodology, resources, funding acquisition, formal analysis, and conceptualization. AK, TN, TNa, and SO contributed to writing, review, and editing. TS contributed to writing, review, editing, and supervision. All authors read and approved the final manuscript.

Funding

This work was supported by the Institute for Health Economics and Policy (27th FY 2022 Research Grant) and by a Grant-in-Aid from the Japan Society for the Promotion of Science (JP24K23738). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study is based on institutional-level data and does not include any identifiable personal information. We submitted the study to the Ethics Committee of National Institute of Public Health, and the committee confirmed that the study is exempt from ethical review. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.