Abstract
Despite the vital role migrant live-in care workers play in elder care across Europe, they often confront precarious employment conditions with limited legal protections and low visibility. Quantitative evidence on their work and health has been scarce. This study (i) describes live-in care workers’ sociodemographic characteristics and job conditions in German homecare; (ii) compares their job demands, resources and motivational, and health-related outcomes with those of geriatric nurses and the German general working population; and (iii) analyses the extent to which their demands and resources are linked to motivational and health outcomes. The Copenhagen Psychosocial Questionnaire was used to assess the psychosocial working conditions, health-related and motivational outcomes of the participating live-in care workers. The sample comprised 429 migrant live-in care workers, of whom 93% were women and 77% were older than 49 years. Most live-ins originated from Poland (50%), followed by Bulgaria (30%), Croatia (13%), and Romania (4%). Nearly half lack formal contracts. Compared to geriatric nurses and all occupations, they experience higher work–privacy conflicts (d = 0.66; d = 0.93) and emotional demands but lower quantitative and physical demands. They report lower job satisfaction, higher burnout, greater presenteeism, and stronger turnover intentions (d = 0.71; d = 0.81). Job demands and resources are significant predictors of health and motivational outcomes. Our study highlights the importance of work–privacy conflict as a central occupational stressor for live-ins. Policy actions and interventions must improve their working conditions, strengthen job resources and protect their health and well-being.
Introduction
A hallmark of sustainable development and public health is decent, safe, and equitable working conditions for all. The United Nations’ Leave No One Behind principle and Sustainable Development Target 8.8 highlight the urgency of protecting labour rights and promoting secure work settings, particularly for migrant workers, women migrants, individuals in precarious employment, and other vulnerable groups [1]. In the case of migrant live-in care workers, three aspects are considered: (i) ‘live-in’ is an umbrella term for adults mainly from Eastern and South-Eastern Europe who support and care for individuals in private households abroad and receive wages in return; (ii) the vast majority of live-ins are women [2]; and (iii) live-ins usually work under precarious circumstances such as extremely low wages, extensive working hours and insecure employment conditions [3–6].
Live-in care workers are a marginalized yet vital group in the delivery of support and care for the elderly in ageing Western societies, with nearly half a million in Germany alone [7].
Previous studies on migrant domestic workers have focused on legal and systemic analyses [8, 9], and some qualitative research explored live-ins’ working and living conditions [10–12]. In an interview study, Kriegsmann-Rabe et al. [12] identified constant availability, minimal free time, isolation from family, and difficult interactions with care recipients or their relatives as key stressors and supportive relationships, personal autonomy, job experience, and genuine commitment to caregiving as protective factors. Examining a group of Romanian live-ins, Hopfgartner et al. [4] reported that although most of the interviewees reported being satisfied with their work, they also described being treated unfairly through low pay, long hours, insecure jobs, unpaid work, and tasks beyond caregiving, among others.
Overall, quantitative data on the work characteristics and health status of migrant live-ins in Europe as well as representative comparisons with other occupational groups have been scant. To highlight this gap, Kuhlmann et al. [13] examined the situation of migrant carers in long-term care across several European Union member states, discussing their structural invisibility in public health and workforce research. The researchers called for their systematic inclusion in health workforce studies as precarious and informal labour arrangements pose health risks to both carers and the wider population. Such data are urgently needed to inform strategic decisions that address health system needs, labour market dynamics, ultimately allowing for the development of effective measures that protect labour rights and ensure safe working conditions for this marginalized group.
An established theoretical framework for a joint analysis of work characteristics and health is the job demands–resources (JD–R) model [14], which differentiates between work-related demands (e.g. emotional demands) and resources (e.g. autonomy), with the former being linked to strain or impaired health and the latter to motivation or health-enhancing effects (see Bakker et al. [15] for meta-analytical support). Applying the JD–R model to the live-in group offers a structure for investigating how job characteristics affect their health and motivation and to target occupational safety and health improvements.
In this context, the main objective of the study was to investigate the work and health situation of migrant live-in care workers in Germany, thereby generating evidence that may support public health and occupational safety policymaking. We adopted a stakeholder-driven approach through a survey of live-ins from four Eastern European countries working in Germany. The research questions are:
R1: What are the sociodemographic characteristics and general job conditions of migrant live-ins in German homecare?
R2: How do job demands, job resources, and motivational and health-related outcomes of live-ins differ from those of geriatric nursing staff and the general working population in Germany?
R3: How are job demands and resources linked to migrant live-ins’ motivational and health-related outcomes?
Our study provides three important contributions to the literature. First, it offers theory-driven, quantitative evidence on a marginalized and understudied workforce, supplementing studies on live-ins, which have been predominantly qualitative. Second, by applying the JD–R model, this study systematically investigates the associations between specific job demands and resources on health and motivation in this unique occupational group. Third, it highlights distinct risk and protective factors for live-ins through comparative insights involving geriatric nurses and the general workforce.
Methods
The present data derive from the study ‘Determining the Work and Health Situation of Live-ins in Germany’ conducted by the Federal Institute for Occupational Safety and Health (BAuA), Germany, in cooperation with the Berlin-based research organization Minor–Project Office for Education and Research.
Ethical considerations
The study’s design and methods were reviewed and approved by the independent Ethics Committee of the BAuA (approval no. 065_2023). Participants were assured of their anonymity and confidentiality.
Study design and participant recruitment
This study adopted a cross-sectional approach via a structured online questionnaire. Participants were recruited through digital street work, a flexible, low-threshold, and dialogical method that extends outreach activities into digital spaces frequented by the target population. Recruitment began by contacting potential participants via closed, project-based Facebook groups established by Minor for migrant live-in care workers (Polish, Romanian, Bulgarian, and Croatian). These groups comprised nearly 10 000 members in Germany and functioned as protected spaces restricted to individuals currently working as live-in. Within these groups, the survey was announced through multiple posts during the recruitment period. For example, in the Polish group, four posts were published, including follow-up posts linked to the initial announcement. Participants actively engaged with these posts by asking questions, which were addressed by project staff, thereby maintaining high visibility of the survey within the groups. To further diversify the sample, additional groups and stakeholder organizations (e.g. advisory services, union initiatives) were later approached to share the survey links. Interested live-ins were provided a link to the survey, which included an information sheet with details on the study’s purpose, procedures, and ethical considerations. Those who consented to participate received an expense allowance of about 15 euro. Data were collected from late June to late July 2023.
Measures
Live-ins’ psychosocial working conditions were assessed using the Copenhagen Psychosocial Questionnaire (COPSOQ) [16, 17], a multidimensional questionnaire that comprehensively evaluates psychosocial aspects of work that can be integrated into the JD–R model [18]. The version that this study used contains 68 items across multiple domains, including job demands (quantitative demands, emotional demands, demands for hiding emotions, work–privacy conflicts, role conflicts, working environment/physical demands), job resources (influence at work, influence at breaks, possibilities for development, meaning of work, role clarity, appreciation), and health-related (general health, personal burnout, presenteeism, inability detaching from work) and motivational outcomes (job satisfaction, work engagement, intention to leave). This study also used the sexual harassment [19], aggressive customer contact, and customer-related burnout [20] scales from the COPSOQ add-on modules. Respondents rated the items using Likert-type scales. Eight supplementary items were designed to capture additional aspects relevant to the research questions (e.g. weekly working days, type of employment). Their development was informed by previous research on live-ins and refined in consultation with experts in the field of occupational health and migrant care.
To ensure linguistic accessibility, the questionnaire was available in Polish, Romanian, Croatian, and Bulgarian, using official COPSOQ III translations provided by the Freiburg Research Centre for Occupational Sciences. The COPSOQ items were adopted from these official translations without modification, except for minor contextual wording adaptations where necessary. Native-speaking experts reviewed these versions, and minor wording adaptations were made to improve clarity and fit the live-in context (e.g. replacing references to ‘colleagues’ or ‘supervisors’, as live-ins usually work alone). No separate pilot test was conducted; items and instructions were kept simple to minimize language barriers.
Reliability was analysed using Cronbach’s alpha to evaluate the scales’ internal consistency (Supplementary 1), and the resulting values indicated acceptable reliability, ranging from .59 for emotional demands to .93 for work–privacy conflict [21]. The mean reliability was .78, with only a few scales falling below the acceptable threshold of .70 (emotional demands, possibilities for development, role clarity, work engagement, and customer-related burnout).
Statistical analyses
In line with the general COPSOQ procedure, the scales and single items were linearly transformed into a standardized metric from 0 to 100, with higher scores representing a greater expression of the construct [16], to retain the full variability of responses and avoid information loss resulting from categorization.
R1 was answered by calculating descriptive statistics (frequencies) for sociodemographic characteristics and general job conditions and R2 by comparing data from the live-in sample with COPSOQ reference values for geriatric nurses and the general working population (all occupations). For group comparisons, effect sizes (Cohen’s d) were derived from mean values, standard deviations, and sample sizes using the online tool Psychometrica [22] and subsequently interpreted according to Cohen’s guidelines [23], where values of >.2 indicate a small effect, >.5 a medium effect, and >.8 a large effect]. To examine R3, multiple linear regression models for each outcome were applied with all job demand and job resource indicators as simultaneous predictors. All statistical analyses, including descriptive statistics and regression models, were conducted using IBM SPSS Statistics version 29.
Given the exploratory nature of the study and the challenges associated with accessing this hard-to-reach population of migrant live-in care workers, no formal a priori power analysis was conducted. Nevertheless, the achieved sample size can be considered adequate for the planned regression analyses. With 14 predictors included in the models, the sample exceeds commonly recommended rules of thumb of approximately 10–15 observations per predictor variable.
We also conducted a posterior sample size check with the related power analysis tool in SPSS. Based on prior meta-analytic evidence, indicating a minimum explained variance of R2 = 0.13 (job engagement) for comparable outcomes [24], this corresponds to an effect size of f2 = .15 (medium-sized effect). Assuming an alpha level of .05 and statistical power between .80 and .95, the required sample size for a multiple regression model with 14 predictors ranges from approximately N = 135 to N = 194. As the present study includes a sample of N = 429, it substantially exceeds these requirements. Taken together, these considerations suggest that the study is sufficiently powered to detect effects of the magnitude reported in prior research.
Results
Sociodemographic characteristics and general job conditions (R1)
The final sample comprised 429 live-in care workers in Germany who participated in the online survey distributed via multilingual Facebook groups and related community networks. Table 1 presents their sociodemographic and general working characteristics. More than 76% of them were over 49 years old (M = 54.7 years), 93.2% reported being female. Almost half were Polish, followed by Bulgarians and Croatians. Their work experience as live-in ranged from 0 to 417 months, with most reporting between 5 and 10 years. A large proportion usually worked 7 days per week with an average of 71.5 weekly working hours. Almost half of the respondents were on call, and nearly two-thirds reported getting up at night to work. Employment and contract types differed within the sample: 38.9% were formally employed, 42.6% worked as contractors and 11.2% were self-employed, with around half holding formal employment contracts.
Table 1.
Sociodemographic and work characteristics of the live-in sample
| Frequencies |
||
|---|---|---|
| n | % | |
| Gender | ||
| Female | 400 | 93.2 |
| Male | 27 | 6.3 |
| Other | 2 | 0.5 |
| Age | ||
| <30 years | 10 | 2.3 |
| 30–39 years | 21 | 4.9 |
| 40–49 years | 68 | 15.9 |
| 49< years | 330 | 76.9 |
| Nationality | ||
| Bulgarian | 127 | 29.6 |
| Croatian | 54 | 12.6 |
| Polish | 213 | 49.7 |
| Romanian | 17 | 3.9 |
| Other | 18 | 4.2 |
| Work experience as a live-in in years | ||
| <1 year | 24 | 5.6 |
| 1–5 years | 136 | 31.7 |
| 5< years ≤10 | 168 | 39.2 |
| 10< years ≤20 | 91 | 21.2 |
| >20 years | 8 | 1.9 |
| Not reported | 2 | 0.5 |
| Work days per week | ||
| 5 | 16 | 3.7 |
| 6 | 83 | 19.3 |
| 7 | 328 | 76.5 |
| Not reported | 2 | 0.5 |
| Working hours per week | ||
| <20 hours | 0 | 0.0 |
| 21–40 hours | 41 | 9.6 |
| 41–60 hours | 138 | 32.2 |
| >60 hours | 193 | 45.0 |
| Not reported | 57 | 13.3 |
| On-call duties | ||
| Yes | 202 | 47.1 |
| No | 227 | 52.9 |
| Having to get up at night | ||
| Yes | 271 | 63.2 |
| No | 158 | 36.8 |
| Type of employment | ||
| Employee | 167 | 38.9 |
| Self-employed | 48 | 11.2 |
| Contractor | 183 | 42.6 |
| Other | 11 | 2.6 |
| No answer | 20 | 4.7 |
| Employment contract | ||
| Yes | 217 | 50.6 |
| No | 212 | 49.4 |
N = 429.
Job demands, job resources, and health and motivation (R2)
Table 2 addresses R2 by comparing job demands, job resources, and health-related and motivational outcomes between live-ins, geriatric nurses, and all occupations. Notably, we excluded influence at breaks, sexual harassment, aggressive customer contact, and customer-related burnout from these analyses as comparison data were missing.
Table 2.
Comparison of job demands, job resources and health-related and motivational outcomes of live-ins, geriatric nurses and all occupations
| Variables | Live-ins |
Geriatric nurses |
All occupations |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| vs. Live-ins |
vs. Live-ins |
|||||||||||
| M | SD | M | SD | Cohen’s d | 95% CI |
M | SD | Cohen’s d | 95% CI |
|||
| LL | UL | LL | UL | |||||||||
| Job demands | ||||||||||||
| Quantitative demands | 39.87 | 19.79 | 55.43 | 19.54 | −0.79* | −0.91 | −0.67 | 54.84 | 18.33 | −0.82** | −0.91 | −0.72 |
| Emotional demands | 55.10 | 25.01 | 74.69 | 21.57 | −0.86** | −0.74 | −0.98 | 47.11 | 24.64 | 0.32 | 0.42 | 0.23 |
| Demands for hiding emotions | 59.94 | 26.90 | 50.56 | 23.43 | 0.38 | 0.27 | 0.5 | 44.42 | 25.31 | 0.61* | 0.52 | 0.71 |
| Work-privacy conflicts | 58.38 | 29.85 | 40.10 | 26.65 | 0.66* | 0.54 | 0.78 | 35.10 | 25.08 | 0.93** | 0.83 | 1.02 |
| Role conflicts | 46.48 | 27.14 | 43.11 | 23.92 | 0.14 | 0.02 | 0.25 | 44.62 | 22.06 | 0.08 | −0.01 | 0.18 |
| Working environment/phys. Demands | 31.81 | 20.33 | 44.33 | 20.84 | −0.61* | −0.72 | −0.49 | 33.54 | 19.77 | −0.09 | −0.25 | 0.08 |
| Sexual harassmenta | 9.91 | 18.29 | – | – | – | – | – | – | – | – | – | – |
| Aggressive customer contacta | 29.48 | 25.30 | – | – | – | – | – | – | – | – | – | – |
| Job resources (influence and development) | ||||||||||||
| Influence at work | 49.61 | 24.45 | 47.32 | 21.78 | 0.10 | −0.01 | 0.22 | 44.94 | 21.8 | 0.21 | 0.12 | 0.31 |
| Influence at breaksb | 44.70 | 32.69 | – | – | – | – | – | – | – | – | – | – |
| Possibilities for development | 44.99 | 21.23 | 65.23 | 20.25 | −0.98** | −1.11 | −0.86 | 64.14 | 20.45 | −0.94** | −1.03 | −0.84 |
| Measure | Live-ins |
Geriatric nurses |
All occupations |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| vs. Live-ins |
vs. Live-ins |
|||||||||||
| M | SD | M | SD | Cohen’s d | 95% CI |
M | SD | Cohen’s d | 95% CI |
|||
| LL | UL | LL | UL | |||||||||
| Meaning of work | 76.31 | 21.18 | 85.87 | 18.78 | −0.49* | −0.61 | −0.37 | 74.82 | 20.84 | 0.07 | −0.09 | −0.23 |
| Role clarity | 74.49 | 17.96 | 76.38 | 18.07 | 0.10 | −0.22 | 0.01 | 70.87 | 18.37 | 0.20 | 0.04 | 0.36 |
| Appreciation | 57.28 | 30.05 | 52.56 | 28.85 | 0.16 | 0.05 | 0.28 | 48.64 | 28.01 | 0.31 | 0.21 | 0.40 |
| Health-related and motivational outcomes | ||||||||||||
| Job satisfaction | 52.92 | 18.29 | 63.85 | 16.30 | −0.64* | −0.76 | −0.53 | 63.85 | 16.65 | −0.66* | −0.75 | −0.56 |
| Work engagement | 64.88 | 19.47 | 69.12 | 19.23 | −0.22 | −0.34 | −0.10 | 63.5 | 19.47 | 0.07 | −0.09 | 0.23 |
| General health | 69.18 | 20.69 | 66.93 | 19.73 | 0.11 | 0.003 | 0.23 | 70.27 | 19.19 | −0.06 | −0.15 | 0.04 |
| Intention to leave | 37.70 | 30.44 | 19.64 | 22.81 | 0.71* | 0.59 | 0.83 | 19.43 | 22.43 | 0.81** | 0.72 | 0.91 |
| Personal burnout | 56.02 | 18.50 | 54.04 | 20.26 | 0.10 | −0.02 | 0.22 | 49.26 | 20.88 | 0.32 | 0.23 | 0.42 |
| Customer-related burnouta | 48.41 | 20.61 | – | – | – | – | – | – | – | – | – | – |
| Presenteeism | 57.81 | 31.00 | 48.61 | 27.55 | 0.32 | 0.2 | 0.44 | 41.48 | 26.39 | 0.62* | 0.52 | 0.71 |
| Inability detaching from work | 46.21 | 29.80 | 46.28 | 28.16 | −0.12 | −0.12 | 0.12 | 46.05 | 28.13 | −0.01 | −0.13 | 0.10 |
N live-ins = 429. Ngeriatrc nurses = 859–891. Nall occupations = 220.382–242.313. All scales ranged from 0 (never/to a very low degree) to 100 (always/to a very high degree). M = mean; SD = standard deviation; CI = confidence interval; LL = lower limit; UL = upper limit.
COPSOQ additional modules: No comparative data are available.
Only one item: No comparative data available. Effect sizes (Cohen’s d) are highlighted by magnitude (bold: small > 0.2; bold with *: medium > 0.5; bold with **: large > 0.8). Alt text: Table comparing mean scores and effect sizes for job demands, job resources, and health-related and motivational outcomes between 429 live-ins, geriatric nurses, and the general working population, including standard deviations, Cohen’s d, and confidence intervals.
Job demands
Live-ins experienced higher work–privacy conflicts and demand for hiding emotions than geriatric nurses and the general working population, with small to moderate effect sizes (e.g. for work–privacy conflicts: M = 58.38 vs. 35.10, d = 0.93 compared to all occupations). Meanwhile, live-ins reported substantially lower quantitative and emotional demands compared to geriatric nurses (e.g. quantitative demands: M = 39.87 vs. 55.43, d = −0.79) and also perceived demands from the working environment/physical demands as less strenuous. While reports of sexual harassment were rare among live-in care workers, they frequently experienced aggressive customer contact.
Job resources
Live-ins reported considerably fewer development opportunities than the comparison groups, with large effect sizes indicating substantial differences (M = 44.99 vs. 65.23 among geriatric nurses, d = −0.98). Meaning of work was slightly lower for live-ins than for geriatric nurses, while differences with the general workforce were negligible.
Motivational and health-related outcomes
With regard to both reference groups, live-ins reported lower job satisfaction and a higher intention to leave, with moderate to large effect sizes (e.g. intention to leave: M = 37.70 vs. 19.43, d = 0.81 compared to all occupations), as well as more frequent presenteeism (M = 57.81 vs. 48.61, d = 0.32 compared to geriatric nurses). Personal burnout levels among live-in care workers were comparable to those of geriatric nurses but slightly higher than those of the general working population.
Associations between work characteristics and health-related and motivational outcomes (R3)
Multiple linear regression results (Table 3) showed that job demands and resources explained between 6% (inability to detach from work) and 52% (job satisfaction) of the variance in health-related and motivational outcomes among live-in care workers. Work–privacy conflicts emerged as a particularly strong predictor of lower job satisfaction (b = −0.20) and work engagement (b = −0.14) as well as higher intention to leave (b = 0.19), burnout (b = 0.25), customer-related burnout, and presenteeism. Meaning of work and development opportunities were consistently associated with favourable outcomes, including higher work engagement (b = 0.30 for meaning of work) and job satisfaction and lower burnout and presenteeism.
Table 3.
Regression analysis results with job demands and resources as predictors of health-related and motivational outcomes
| Job satisfaction |
Work engagement |
General health |
Intention to leave |
Personal burnout |
Customer-related burnout |
Presenteeism |
Inability detaching from work |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | b | b | b | b | b | b | b | |||||||||
| Job Demands | ||||||||||||||||
| Quantitative demands | −.02 | .13 | * | −.09 | −.05 | .04 | <.01 | .07 | <.01 | |||||||
| Emotional demands | −.11 | * | −.06 | <.01 | .10 | .04 | .02 | −.05 | .08 | |||||||
| Demands for hiding emotions | .03 | −.02 | .01 | .03 | .05 | .12 | * | .04 | .07 | |||||||
| Work-privacy conflicts | −.20 | *** | −.14 | * | −.10 | .19 | *** | .25 | *** | .14 | ** | .18 | ** | .11 | ||
| Role conflicts | −.06 | −.01 | .16a | * | .05 | .07 | .24 | *** | <.01 | <.01 | ||||||
| Working environment/phys. demands | −.04 | −.03 | −.16 | ** | .09 | .15 | ** | .08 | .22 | *** | <.01 | |||||
| Sexual harassment | −.04 | −.05 | −.05 | <.01 | .02 | −.05 | .08 | .07 | ||||||||
| Aggressive customer contact | −.12 | ** | .01 | −.03 | .10 | .05 | .11 | * | −.08 | .03 | ||||||
| Job resources | ||||||||||||||||
| Influence at work | .10 | * | −.03 | .05 | <.01 | −.01 | −.01 | −.10 | * | −.03 | ||||||
| Influence at breaks | .06 | −.06 | −.01 | −.07 | .02 | .12 | ** | .07 | .10 | |||||||
| Possibilities for development | .03 | .16 | *** | .12 | * | −.08 | −.13 | ** | −.10 | * | −.17 | *** | −.03 | |||
| Meaning of work | .15 | *** | .30 | *** | <.01 | −.24 | *** | −.16 | *** | −.08 | * | −.01 | −.10 | |||
| Role clarity | .11 | ** | .04 | .16 | ** | −.04 | <.01 | −.03 | .07 | −.10 | ||||||
| Appreciation | .17 | *** | .15 | ** | .11 | −.07 | −.01 | −.19 | *** | −.10 | .09 | |||||
| F(14,414) | 34.4 | *** | 11.9 | *** | 6.6 | *** | 18.7 | *** | 17.6 | *** | 28.0 | *** | 9.1 | *** | 3.1 | *** |
| adj. R2 | .52 | .26 | .15 | .37 | .35 | .47 | .21 | .06 | ||||||||
N = 429.
P ≤ .05.
P ≤ .01.
P ≤.001.
Crude model relation in opposite direction (b = −.18, P <. 01), which might indicate suppression effects. In an additional multiple regression model using a stepwise approach, this variable was no longer a significant general health predictor (significant effects of the other variables remained consistent).
Discussion
Our findings show that live-in care workers work under highly demanding conditions, working exceptionally long hours throughout the week (77% more than 40 h), frequently interrupting their sleep for nighttime care (63%), and often being continuously on call (47%). Their employment arrangements considerably vary, with half of the participants (49%) lacking formal employment contracts, reflecting well-known patterns of labour market segmentation and informal employment arrangements in migrant care work [25, 26].
Regarding job demands and resources, live-ins manage a distinct job profile than other occupational groups. They report lower quantitative and emotional demands but higher work–privacy conflicts and emotion suppression. Development opportunities are fewer, and job satisfaction is lower, accompanied by higher turnover intention, more presenteeism, and difficulty detaching from work.
Work–privacy conflicts were the most consistent predictor of poorer outcomes among live-ins, being associated with lower job satisfaction and work engagement and higher intention to leave, burnout, and presenteeism. Emotional demands were also linked to lower job satisfaction and higher intention to leave. Meaning of work proved to be a strong protective factor, promoting satisfaction and engagement, while buffering against burnout and intention to leave. Development opportunities facilitated higher work engagement and better general health among live-ins.
While most associations between job characteristics and outcomes followed the JD–R model’s expected directions, a few exceptions warrant further discussion. Quantitative demands showed a weak but significant positive association with work engagement, suggesting that live-ins, under certain conditions, may regard challenging workloads as activating or meaningful rather than draining which may be attributed to intrinsic motivation [12] or task significance. Furthermore, the positive association between role conflicts and general health may reflect a suppression effect or the presence of meaningful role diversity in some cases. Unexpectedly, influence at breaks was positively linked to customer-related burnout, suggesting a measurement artefact or an unmeasured third variable (e.g. greater autonomy in difficult care situations). Future studies should investigate these patterns further.
Interestingly, despite reporting higher demands in terms of workload, emotional labour, and physical stressors, geriatric nurses reported higher job resources such as development opportunities and meaning of work, which may help buffer the negative effects of job demands as Falco et al. [27] observed in their sample of 119 employees in an Italian public healthcare organization. Meanwhile, live-ins experience a paradox: some demands appear lower on the surface, while structural characteristics of the job such as chronic intrusion of work into their personal life (e.g. through blurred work–life boundaries, 24/7 availability and isolation from family) amplifies strain and impedes recovery. This suggests a potential decoupling between perceived workload and the extensiveness of working time and availability. This is compounded by contextual factors not fully captured in the present study, including language barriers, limited access to support systems, and living at the workplace. Unlike findings for other occupational groups such as physicians [28], where job and personal resources may help balance work–life demands, live-in care workers seem unable to offset this imbalance through available resources likely because they lack key protective factors, such as social support from supervisors or colleagues, and have limited access to development opportunities. The absence of such structural and interpersonal resources, alongside the aforementioned critical structural conditions, renders live-ins particularly vulnerable to the strain due to blurred boundaries between work and personal life. Beyond these contextual explanations, the paradoxical finding may be partly attributable to measurement-related limitations, as available standardized instruments such as COPSOQ may not fully capture the specific characteristics of live-in care arrangements. In particular, continuous availability and blurred temporal boundaries may not be perceived as ‘quantitative demands’ in the conventional sense, but rather as an inherent condition of the job. In addition, adaptation processes or shifting reference frames may lead live-ins—unlike the comparison groups—to be more likely to appraise demanding working conditions as less burdensome or as part of the normal job context, potentially resulting in comparatively lower self-reported demand levels. Finally, social desirability and dependency structures inherent in live-in arrangements may contribute to more conservative reporting of adverse working conditions.
Study limitations
Several limitations should be considered when interpreting our findings. First, the cross-sectional design precludes causal inferences. Although observed associations between job demands/resources and health or motivational outcomes align with the JD–R framework, directionality remains unclear. Recent theoretical developments of the JD–R theory [15] suggest reciprocal effects, such as between burnout and engagement, which may also influence perceived demands and resources over time. This dynamic perspective is supported by Guthier et al. [29], whose meta-analytic evidence demonstrated the reciprocal effects between job stressors and burnout. Thus, the associations in the current data may reflect not only the impact of working on health but also reverse or bidirectional influences. Longitudinal studies with multiple time points and continuous-time modelling are needed to disentangle these dynamics. Nonetheless, Spector [30] emphasizes that well-designed cross-sectional studies still offer valuable insights, especially when used alongside longitudinal data to build a more pragmatic and resource-efficient evidence base.
Second, reliance on self-report measures introduces potential for common method bias or social desirability effects, possibly inflating associations [31]. The absence of objective health or employment data also limits triangulation. However, evidence from instrumental-variable analyses suggests that associations based on self-reported job demands and job strain can reflect causal relations with employee well-being even when considering potential reporting biases [32].
Third, selection bias may have resulted from the use of digital outreach channels to recruit the sample [33], which may reflect live-ins with access to digital platforms and some autonomy possibly underrepresenting more vulnerable or isolated individuals. Moreover, because recruitment occurred through existing social media networks, the total number of individuals who viewed or received the survey link remains unknown, preventing the calculation of a precise response rate and limiting generalizability to the population of approximately half a million live-in care workers in Germany. In addition, this population estimate itself is subject to considerable uncertainty due to the often informal and poorly documented nature of live-in care arrangements, which further complicates any assessment of sample representativeness.
Finally, while the COPSOQ has been widely used and validated, it was developed for standard employment settings; hence, its applicability to the highly specific and hybrid work–living arrangements of live-ins may be limited in some domains. For instance, live-ins may have a substantially different idea of ‘workplace influence’ compared to traditional employees.
Nevertheless, this study offers important insights into live-in care workers’ employment conditions and health. It recommends that future studies (a) employ mixed-method or longitudinal designs to better capture causal mechanisms and complex job–health dynamics, (b) include vulnerable subgroups through diverse recruitment strategies, and (c) adapt current measurement tools to live-in settings or develop context-specific instruments.
Implications
Our findings emphasize work–privacy conflict as a relevant job demand among live-in care workers with its strong association with adverse health and motivational outcomes. Future research should explore buffers and intervention strategies that target this stressor. A recent meta-analysis and theoretical integration by von Allmen et al. [34] identified the components of effective work–nonwork interventions, highlighting the potential of nonwork-based strategies particularly those enhancing personal and social resources to improve employee well-being and offering promising opportunities for application in live-in care work. A particularly relevant approach involves strengthening social embeddedness and providing structured opportunities for recovery. For example, examining the influence of peer network access, leisure-time activities, and regulated rest periods on perceptions of role clarity, autonomy and work–life boundaries may help identify factors that strengthen psychological well-being. From a public health perspective, Ho et al. [35] further highlighted the importance of social connectedness and identified social isolation as a structural health risk for live-in migrant domestic workers. Their review emphasized the value of stronger peer networks and informal social support systems as crucial community-level resources that mitigate mental distress and promote resilience in precarious care arrangements. These resources may not only reduce conflict between work and private life but also enhance positive work experiences and improve overall health.
In an increasingly globalized and mobile labour market, migrant live-in care workers’ health is frequently compromised by risky working conditions, insufficient regulation, and limited access to healthcare or social protection, reflecting Benach et al.’s [36] concept of precarious employment—an emerging social determinant that worsens health inequities through job insecurity, lack of rights, and psychosocial stressors. As a largely invisible occupational group, live-ins often fall outside the scope of standard labour protections, creating legal and ethical blind spots [37]. These systemic disadvantages intersect with social determinants of health, such as legal status, income insecurity and housing conditions, placing live-in care workers at heightened risk for both physical and mental health problems. This pattern reflects and reinforces broader migrant health inequalities in precarious employment contexts [38]. These disparities must be addressed to promote health equity and fulfil the objectives of the 2030 Agenda for Sustainable Development especially SDG 8.8, which aims to protect labour rights and ensure safe and secure working environments for all workers.
Policymakers must establish clear, enforceable regulations defining the responsibilities of families, agencies, and care workers, particularly regarding working hours, rest periods, and contracts. In line with the Health in All Policies approach, public health actors and live-in care workers should be actively involved in shaping labour and migration policies. In Germany, recent debates on ‘legal certainty’ in live-in care have often centred on the interests of agencies and families, neglecting the fundamental incompatibility of 24-hour care with existing labour protections [39]. Efforts that fail to address the rights and realities of migrant care workers risk perpetuating precarious conditions and deepening health inequities. Ensuring equitable access to social security and health services must be a fundamental component of health-promoting employment.
From a practice perspective, work–privacy conflict has emerged as a target for workplace interventions in live-in care as it was not only significantly higher among live-ins but was also consistently associated with burnout, presenteeism, and intention to leave. To reduce this conflict, practical measures for live-ins may include clear scheduling policies, designated rest times, private spaces, and support mechanisms to navigate their dual role as workers and quasi-household members. A critical foundation for these interventions constitutes fair, clearly defined, and legally viable employment contracts [3, 40]. Interventions that specifically reduce role ambiguity and strengthen work autonomy could further mitigate the psychological strain of blurred work–life boundaries.
Conclusions
This survey study offers novel insights into the working conditions, indicators of work-related health and job attitudes of migrant live-in care workers in Germany. Using the JD–R model, this study identified work–privacy conflict as a primary job demand and meaning of work as a key protective resource. Despite reporting lower quantitative and emotional demands compared to other care professions, live-ins confront significant health and motivational risks because of unclear work–life boundaries and limited job resources. The findings highlight the urgency of developing tailored interventions and policy measures that address the unique challenges of this marginalized workforce. Ensuring decent working conditions for live-ins promotes both sustainable care provision and workers health—and thus a key public health concern in ageing populations.
Supplementary Material
Contributor Information
Judith E Beike, Faculty of Psychology, TU Dresden, Dresden, Germany.
Ulrike Rösler, Unit 3.3 Designing Service Work, Federal Institute for Occupational Safety and Health, Dresden, Germany.
Agnieszka Skwarek, Minor—Projektkontor für Bildung und Forschung GmbH, Berlin, Germany.
Johannes Wendsche, Unit 3.3 Designing Service Work, Federal Institute for Occupational Safety and Health, Dresden, Germany.
Christian Pfeffer-Hoffmann, Minor—Projektkontor für Bildung und Forschung GmbH, Berlin, Germany.
Acknowledgements
The survey was approved by the German Federal Institute for Occupational Safety and Health. We thank Doan Nguyen for her assistance with data collection.
Supplementary data
Supplementary data are available at EURPUB online.
Conflict of interest: None declared.
Funding
This study was financed by our own research institute (Federal Institute for Occupational Safety and Health, Germany) and did not receive any external funding.
Data availability
The data underlying this article will be shared on reasonable request to the corresponding author.
Key points.
Most migrant live-in care workers in Germany are women over 49 years old, commonly from Poland, Bulgaria, Croatia, or Romania working extremely long hours (71.5 hours/week on average) with diverse employment contracts and often providing on-call or nighttime care.
Compared to geriatric nurses and general workers, live-in care personnel experience higher work–privacy conflicts and demands to hide their emotions but lower quantitative, emotional and physical demands.
Live-in care workers report significantly fewer development opportunities, lower job satisfaction, higher intention to leave and more presenteeism as well as burnout levels similar to those of geriatric nurses but higher than those of the general workforce.
Work–privacy conflict is a strong predictor of negative outcomes such as burnout, intention to leave and presenteeism, while job resources such as meaning of work and development opportunities promote better job satisfaction and engagement.
Public health policies must urgently address live-in care workers’ precarious work conditions and strengthen support systems to protect their health and well-being.
References
- 1. United Nations. Transforming Our World: The 2030 Agenda for Sustainable Development (A/RES/70/1). New York: United Nations, 2015. [Google Scholar]
- 2. Habel S, Tschenker T. Reduktion der Arbeitszeit in der Live-In-Pflege - Eine interdisziplinäre Untersuchung von Maßnahmen der Vermittlungsagenturen (Reduction of working hours in live-in care - An interdisciplinary study of measures taken by placement agencies). Report No.: 471. Düsseldorf: Hans-Böckler-Stiftung, 2022.
- 3. Emunds B. Menschenunwürdige pflegearbeit in deutschen privathaushalten. Sozialethische bemerkungen zu den arbeitsverhältnissen mittel-und osteuropäischer live-ins (Inhumane care work in German private households. Social-ethical remarks on the working conditions of Central and Eastern European live-in workers). Jahrbuch Für Christliche Sozialwissenschaften 2016;57:199–224. [Google Scholar]
- 4. Hopfgartner L, Seubert C, Sprenger F. et al. Experiences of precariousness and exploitation of Romanian transnational live-in care workers in Austria. J Ind Relat 2022;64:298–320. 10.1177/00221856211063923 [DOI] [Google Scholar]
- 5. Schilliger S. Transnationale Care-Arbeit: Osteuropäische Pendelmigrantinnen in Privathaushalten von Pflegebedürftigen (Transnational care work: Eastern European commuter migrants in private households of people in need of care). In: Schweizerisches Rotes Kreuz (ed.), Who Cares? Pflege und Solidarität in der alternden Gesellschaft. Zürich: Schweizerisches Rotes Kreuz, 2013, 142–61.
- 6. van Bochove M, zur Kleinsmiede D.. Broadening the scope of live-in migrant care research: how care networks shape the experience of precarious work. Health Soc Care Community 2020;28:51–9. 10.1111/hsc.12837 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Aulenbacher B, Lutz H, Schwiter K.. Gute Sorge ohne gute Arbeit? Live-in-Care in Deutschland, Österreich und der Schweiz (Good Care Without Good Work? Live-in Care in Germany, Austria, and Switzerland). Weinheim: Beltz Juventa, 2021.
- 8. Green O, Ayalon L.. Violations of workers’ rights and exposure to work-related abuse of live-in migrant and live-out local home care workers - a preliminary study: implications for health policy and practice. Isr J Health Policy Res 2018;7:32. 10.1186/s13584-018-0224-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Mullally S, Murphy C.. Migrant domestic workers in the UK: enacting exclusions, exemptions, and rights. HRQ 2014;36:397–427. 10.1353/hrq.2014.0021 [DOI] [Google Scholar]
- 10. Bruquetas‐Callejo M. Long‐term care crisis in The Netherlands and migration of live‐in care workers: transnational trajectories, coping strategies and motivation mixes. Int Migr 2020;58:105–18. 10.1111/imig.12628 [DOI] [Google Scholar]
- 11. Hoens S, Keer R-L, Safuta A. et al. Live-in care workers for older persons: navigating the challenges of transnational migration. Int J Care Caring 2025;1–20. 10.1332/23978821Y2025D000000119 [DOI] [Google Scholar]
- 12. Kriegsmann-Rabe M, Maus K, Hiebel N. et al. Live-in migrant home care workers in Germany: stressors and resilience factors. PLoS One 2023;18:e0282744. 10.1371/journal.pone.0282744 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kuhlmann E, Falkenbach M, Klasa K. et al. Migrant carers in Europe in times of COVID-19: a call to action for European health workforce governance and a public health approach. Eur J Public Health 2020;30:iv22–7. 10.1093/eurpub/ckaa126 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bakker AB, Demerouti E.. The job demands‐resources model: state of the art. J Manag Psychol 2007;22:309–28. 10.1108/02683940710733115 [DOI] [Google Scholar]
- 15. Bakker AB, Demerouti E, Sanz-Vergel A.. Job demands–resources theory: ten years later. Annu Rev Organ Psychol Organ Behav 2023;10:25–53. 10.1146/annurev-orgpsych-120920-053933 [DOI] [Google Scholar]
- 16. Burr H, Berthelsen H, Moncada S. et al. ; International COPSOQ Network. The third version of the Copenhagen Psychosocial Questionnaire. Saf Health Work 2019;10:482–503. 10.1016/j.shaw.2019.10.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Nübling M, Vomstein M, Haug A. et al. COPSOQ 3: internationale weiterentwicklung und deutsche standardversion (COPSOQ 3: international development and German standard version). PPmP - Psychotherapie · Psychosomatik · Medizinische Psychologie 2018;68:e41. 10.1055/s-0038-1667980 [DOI] [Google Scholar]
- 18. Berthelsen H, Hakanen JJ, Westerlund H.. Copenhagen Psychosocial Questionnaire - a validation study using the job demand-resources model. Plos One 2018;13:e0196450. 10.1371/journal.pone.0196450 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lincke H-J, Häberle N, Lindner A. et al. Analyse sexueller belästigung am arbeitsplatz in der gefährdungsbeurteilung psychischer belastungen (Analysis of sexual harassment in the workplace in the risk assessment of psychological stress). Zentralblatt für Arbeitsmedizin, Arbeitsschutz Und Ergonomie 2021;71:167–77. [Google Scholar]
- 20. Lincke H-J, Theiler A, Nübling M. Psychische Belastungen im Kontakt mit Kunden: Die Bestimmung arbeitsplatzspezifischer Konfliktpotentiale mit dem Zusatzmodul des COPSOQ (Psychological stress in contact with customers: determining workplace-specific conflict potential with the COPSOQ add-on module). In: Deutsche Gesellschaft für Arbeitsmedizin und Umweltmedizin (ed.), 54 Wissenschaftliche Jahrestagung der DGAUM: Dokumentation der 54 Jahrestagung der DGAUM 2014. Stuttgart: Gentner Verlag, 2014, 292–5.
- 21. Loewenthal KM, Lewis CA.. An Introduction to Psychological Tests and Scales, 3rd edn. London: Routledge, 2020. [Google Scholar]
- 22. Lenhard W, Lenhard A.. Berechnung von Effektstärken. https://www.psychometrica.de/effektstaerke.html; 10.13140/RG.2.2.17823.92329 (26 May 2026, date last accessed). [DOI]
- 23. Cohen J. Statistical Power Analysis for the Behavioral Sciences, 2nd edn. London: Routledge, 1988. [Google Scholar]
- 24. Crawford ER, LePine JA, Rich BL.. Linking job demands and resources to employee engagement and burnout: a theoretical extension and meta-analytic test. J Appl Psychol 2010;95:834–48. 10.1037/a0019364 [DOI] [PubMed] [Google Scholar]
- 25. Shutes I, Chiatti C.. Migrant labour and the marketisation of care for older people: the employment of migrant care workers by families and service providers. J Eur Soc Policy 2012;22:392–405. 10.1177/0958928712449773 [DOI] [Google Scholar]
- 26. Schmidt AE, Winkelmann J, Rodrigues R. et al. Lessons for regulating informal markets and implications for quality assurance–the case of migrant care workers in Austria. Ageing Soc 2016;36:741–63. 10.1017/S0144686X1500001X [DOI] [Google Scholar]
- 27. Falco A, Dal Corso L, Girardi D. et al. The moderating role of job resources in the relationship between job demands and interleukin-6 in an Italian healthcare organization. Res Nurs Health 2018;41:39–48. 10.1002/nur.21844 [DOI] [PubMed] [Google Scholar]
- 28. Mache S, Bernburg M, Vitzthum K. et al. Managing work-family conflict in the medical profession: working conditions and individual resources as related factors. BMJ Open 2015;5:e006871. 10.1136/bmjopen-2014-006871 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Guthier C, Dormann C, Voelkle MC.. Reciprocal effects between job stressors and burnout: a continuous time meta-analysis of longitudinal studies. Psychol Bull 2020;146:1146–73. 10.1037/bul0000304 [DOI] [PubMed] [Google Scholar]
- 30. Spector PE. Do not cross me: optimizing the use of cross-sectional designs. J Bus Psychol 2019;34:125–37. 10.1007/s10869-018-09613-8 [DOI] [Google Scholar]
- 31. Podsakoff PM, Podsakoff NP, Williams LJ. et al. Common method bias: it’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annu Rev Organ Psychol Organ Behav 2024;11:17–61. 10.1146/annurev-orgpsych-110721-040030 [DOI] [Google Scholar]
- 32. Elovainio M, Heponiemi T, Kuusio H. et al. Job demands and job strain as risk factors for employee wellbeing in elderly care: an instrumental-variables analysis. Eur J Public Health 2015;25:103–8. 10.1093/eurpub/cku115 [DOI] [PubMed] [Google Scholar]
- 33. Bethlehem J. Selection bias in web surveys. Int Statistical Rev 2010;78:161–88. 10.1111/j.1751-5823.2010.00112.x [DOI] [Google Scholar]
- 34. von Allmen N, Hirschi A, Burmeister A. et al. The effectiveness of work-nonwork interventions: a theoretical synthesis and meta-analysis. J Appl Psychol 2024;109:1115–31. 10.1037/apl0001105 [DOI] [PubMed] [Google Scholar]
- 35. Ho KHM, Wilson IM, Wong JY. et al. Overlooked by nurses: a scoping review on health stressors, problems and coping of migrant domestic workers. Nurs Open 2023;10:1166–79. 10.1002/nop2.1391 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Benach J, Vives A, Amable M. et al. Precarious employment: understanding an emerging social determinant of health. Annu Rev Public Health 2014;35:229–53. 10.1146/annurev-publhealth-032013-182500 [DOI] [PubMed] [Google Scholar]
- 37. Phan-Warnke L, Freitag N.. Ending Live-in Care Workers’ Labour Exploitation in the European Union. Lessons from Germany. Berlin: German Institute for Human Rights, 2021.
- 38. Koseoglu Ornek O, Waibel J, Wullinger P. et al. Precarious employment and migrant workers’ mental health: a systematic review of quantitative and qualitative studies. Scand J Work Environ Health 2022;48:327–50. 10.5271/sjweh.4019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Kocher E. “Legal certainty” for live‐in work in Germany: a strategy for formalization? Int Labour Rev 2024;163:417–33. 10.1111/ilr.12434 [DOI] [Google Scholar]
- 40. Emunds B, Habel S.. Von der Schwarzarbeit zum „grauen Markt “– und darüber hinaus? Neuere und künftig notwendige Entwicklungen der sog. 24-Stunden-Pflege (From illegal employment to the “gray market” – and beyond? Recent and future developments in 24-hour care). In: Jacobs K, Kuhlmey A, Greß S, Klauber J, Schwinger A (eds), Pflege-Report 2020: Neuausrichtung Von Versorgung Und Finanzierung. Berlin, Heidelberg: Springer Berlin Heidelberg, 2020, 111–21. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article will be shared on reasonable request to the corresponding author.
Key points.
Most migrant live-in care workers in Germany are women over 49 years old, commonly from Poland, Bulgaria, Croatia, or Romania working extremely long hours (71.5 hours/week on average) with diverse employment contracts and often providing on-call or nighttime care.
Compared to geriatric nurses and general workers, live-in care personnel experience higher work–privacy conflicts and demands to hide their emotions but lower quantitative, emotional and physical demands.
Live-in care workers report significantly fewer development opportunities, lower job satisfaction, higher intention to leave and more presenteeism as well as burnout levels similar to those of geriatric nurses but higher than those of the general workforce.
Work–privacy conflict is a strong predictor of negative outcomes such as burnout, intention to leave and presenteeism, while job resources such as meaning of work and development opportunities promote better job satisfaction and engagement.
Public health policies must urgently address live-in care workers’ precarious work conditions and strengthen support systems to protect their health and well-being.