The Different Facets of Work Stress

Abstract

Work-related stress has been identified as a relevant problem leading to negative effects on health and quality of life. Using data from 844 nurses, latent profile analyses (LPA) were applied to identify distinct patterns of work stress. Several sociodemographic variables, including nurses’ working and living conditions, as well as nurses’ reactions to workload, were considered to predict respondents’ profile membership. LPA revealed three distinct profiles that can be distinguished by a low, moderate, and higher stress level. Being financially secure is positively related to the low stress profile, whereas working in an urban area and having low job satisfaction increases the chance of belonging to the higher stress profile. Our results can be used as a basis to develop interventions to create a healthy nursing home environment by supporting the balance between family and work, providing access to job resources and optimizing recovery opportunities.

Keywords

Introduction

The facts concerning an ageing society are well-known. In addition to the positive aspect of a long life, the need for care of people reaching 85 years and over is rapidly increasing (Lutz, Sanderson, & Scherbov, 2008). Elderly care in Austria and Germany is most frequently provided by family members (Bundesministerium für Arbeit, Soziales und Konsumentenschutz [BMASK], 2011; Statistisches Bundesamt, 2011). Institutional care is often seen as a last resort. Keeping in mind the individual needs of the elderly residents, the best possible care would require adequate staffing as well as interdisciplinary and intradisciplinary collaboration (Kalisch, Lee, & Rochman, 2010; Kramer, Maguire, & Brewer, 2011; Kutney-Lee et al., 2009). These standards are difficult to achieve when considering that the current nursing home work environment is characterized by increasing responsibilities and inadequate staffing and is often underappreciated (Fronteira & Ferrinho, 2011; Hsu et al., 2007; Ilhan, Durukan, Taner, Maral, & Bumin, 2008; Kayser-Jones et al., 2003; Laschinger, Finegan, Shamian, & Wilk, 2003; Piko, 2006; Ulrich et al., 2007). The last three decades of work-related stress research have provided important contributions regarding the relationship between work environment, well-being, and health (e.g., Johnson & Hall, 1988; Karasek, 1979; Siegrist, 1996a; Spoor, De Jonge, & Hamers, 2010; Vieira, 2007). In the Effort-Reward Imbalance Model Siegrist (1996a) discusses the necessity of balance between effort at work and being rewarded. Reward in this model is associated with an appropriate salary, esteem, career opportunities, and job security (Bishop, Squillace, Meagher, Anderson, & Wiener, 2009). A lack of reciprocity can lead to stress and increase an individual’s risk for physical and mental disorders (Siegrist, 1996a; Van Vegchel, De Jonge, Bosma, & Schaufeli, 2005). Another assumption within the Effort-Reward Imbalance model points to over-commitment leading to sustained stress reactions (Siegrist, 1996b). Over-committed people exaggerate their efforts beyond levels usually considered appropriate, a behavior also discussed in the personality-centered approaches of burnout (e.g., Burisch, 2006).

Some authors associate work-related stress with increased sick-leave days and decreased quality of life (Baune, Adrian, & Jacobi, 2007; Head et al., 2007). The lack of a social support at work increases the prevalence of physical and mental illness (Rösler, Jacobi, & Rau, 2006; Van Vegchel et al., 2005). Alfredsson, Karasek, and Theorell (1982) identified nurses as the occupational group most affected by stress within the work environment. Studies by De Jonge, Janssen, and Van Breukelen (1996) and Lee et al. (2009) also associated nursing work with burnout, depression, and low job satisfaction. Fronteira and Ferrinho (2011) reported that nurses in elderly care have higher burnout and dissatisfaction compared to nurses working in other areas.

Research has explored the conjunction of work- and nonwork-related stress (Edwards & Rothbard, 2000; Goode, 1960; Greenhaus & Powell, 2006; Haar, 2004). Previous studies focused on the negative accumulation of stress in job and home responsibilities (Eby, Maher, & Butts, 2010; Greenhaus & Beutell, 1985). Being involved in multiple roles can deplete a person’s resources. Negative spillover occurs when involvement in one domain hinders functioning in other domains (Edwards & Rothbard, 2000).The alternative approach, positive spillover, suggests that multiple roles can have a positive effect: occupational stress such as work time demands can be compensated through nonwork resources such as social support in leisure time (Barnett & Baruch, 1985; Greenhaus & Powell, 2006; Wayne, Grzywacz, Carlson, & Kacmar, 2007). Whereas positive spillover has the potential to positively influence people in work and nonwork domains, negative spillover produces harmful stress experiences in both domains. Despite the obvious work-related stress and high rates of staff turnover in elderly care environments, there is a lack of comprehensive studies in this field in Austria. Extending previous research, we examined nurses’ work-related stress and reactions to workload in two federal states of Austria. To take both urban and rural areas into account, the study was conducted in Vienna (Capital of Austria, approx. 1.6 million citizens) and Carinthia (rural area, approx. 560,000 citizens; Statistik Austria, 2011).

The aim of the study was to group individuals into so-called profiles, each consisting of individuals who were similar to each other. In the current study, a latent profile approach was used to classify nurses who showed similar stress response patterns and thus fit a certain profile. To develop such a typology, a person-centered modeling technique called latent profile analysis (LPA) was performed. The term “person-centered” does not refer to the relationship among the variables of interest but rather to the relationship among individuals (Muthén & Muthén, 2000). Once these profiles of work-related stress were identified, we further explored the factors that explain profile membership. Several sociodemographic variables, including the nurses’ work and living conditions, as well as nurses’ reactions to workload, were considered to predict the respondents’ profile membership and to identify potential protective factors for work-related stress.

Method

We conducted a cross-sectional study of 81 nursing homes in two federal states of Austria. To obtain comprehensive data on work-related and nonwork-related stress in the care for the elderly, we sampled and attempted to survey all registered nurses who possessed at least 1 year of professional experience. The first author gave full information about the aims, methods, and expected benefits of the study to all Carinthian (n = 52) nursing homes and to one public utility provider in Vienna, which is responsible for 32 nursing homes. Nurses were informed that participation was voluntary and that the information provided would be treated confidentially. The approach followed the Ethical Principles of the APA (American Psychological Association, 2010) and conformed to our institution’s human subjects protection policy (http://www.uni-klu.ac.at/rechtabt/downloads/mbl3b1_08_09.pdf).

The survey took place over a 20-month period in 49 Carinthian and 32 Viennese nursing homes. All nurses spoke German fluently enough to answer the self-administered questionnaire. In the province of Carinthia, data were collected from 603 registered nurses, representing a response rate of 78%. In the province of Vienna, 241 questionnaires were returned by registered nurses (60% response rate). Evidence suggests that a defining characteristic of the Viennese sample is the multiculturalism of the team. Whereas Austrians are almost exclusively employed in Carinthia, the proportion of Austrians is only 39% in Vienna.

Measures

Nurses’ work stress was measured by a standardized self-administered questionnaire (Zimber & Weyerer, 1999) that was validated in a previous study (Jenull-Schiefer, 2005). This questionnaire includes 26 items that refer to demanding aspects of the job (e.g., stress due to residents, family members, working conditions, and negative emotions). A sample item is “You are called by the residents even for minor things.” Items are scored on a 5-point Likert-type scale ranging from (1) not at all to (5) yes, exactly. Scales of nurses’ work stress were computed as the mean score of the scale-specific items (Zimber & Weyerer, 1999). The overall Cronbach’s α coefficient (26 items) was .88.

Nurses’ reactions to workload were assessed within a self-administered questionnaire (7-point scale ranging from (1) applicable to (7) not applicable) containing 38 items that measured emotional exhaustion, intrinsic motivation, dissatisfaction at work, aversion to residents, and reactive screening (Hacker & Reinhold, 1999). An example item is “I sometimes feel very exhausted.” The internal consistency of the scales was satisfactory, with Cronbach’s α varying from .76 to .83. Following Hacker and Reinhold (1999), scales were computed as the sum score of the corresponding items.

Additional information regarding sociodemographic and occupational variables (e.g., age, gender, nationality, family status, professional group affiliation, and diseases) was collected. Furthermore, nurses’ perceived burden of household chores, perceived financial situation, and perceived job relevance were assessed using three single items with a Likert-type scale ranging from (1) low to (5) high, (1) secure to (5) precarious, and (1) high to (4) low, respectively.

Statistical Analysis

To develop a typology of nurses’ work-related stress, a mixture modeling approach was performed. The application of a polytomous latent class analysis was not feasible due to the large number of response patterns resulting from 26 items that were answerable on a 1 to 5 scale. Thus, latent profile analysis (LPA; e.g., Vermunt & Magidson, 2002) was applied to identify clusters of respondents having similar values on the four subscales “stress due to residents,” “stress due to family members,” “stress due to working conditions,” and “stress due to negative emotions.” The total score of each subscale was divided by the number of scale-specific items to rescale each indicator on the interval from 1 (low stress) to 5 (high stress). LPA assumes that the underlying correlational structure of these continuous observed indicators can be explained solely by an underlying categorical latent variable (known as the local independence assumption). Each latent profile (LP) is represented by cluster-conditional mean and variance estimates for the four subscales. Vermunt and Magidson (2002) note that in most applications, standard LPA is sufficient to identify groups that differ with respect to their means, assuming equal variances of indictors across clusters. However, this assumption can be relaxed, thereby allowing heterogeneity of responses across clusters. To determine the optimal LP-model to describe nurses’ work-related stress, all models were estimated with (a) equal variances across clusters and (b) different variance estimates across clusters. To select the appropriate number of latent profiles, the Akaike information criterion (AIC; Akaike, 1987), the Bayes Information Criterion (BIC; Schwarz, 1978) and the sample size adjusted BIC (Sclove, 1987) were applied; lower values indicate a better fit of the model. Nylund, Asparouhov, and Muthén (2007) showed that the AIC tends to overestimate the correct number of latent profiles. Thus, we focused on the BIC- and adjusted BIC-values during model selection. In addition, the unadjusted and adjusted Lo-Mendell-Rubin tests (Lo, Mendell, & Rubin, 2001) were applied to assess if the k cluster solution is superior to the k—1 cluster solution. A nonsignificant test result suggests no superiority of the k cluster solution over the k—1 cluster model. The entropy and the average profile probabilities were inspected to assess the separability of each LPA solution. As suggested by Marsh, Lüdtke, Trautwein, and Morin (2009), the interpretability of the model coefficients was inspected for each solution. In addition to the sole identification of different patterns of work stress, covariates were used to predict LP membership. Previous studies showed that a two-stage approach, in which a second step LP membership is used as a manifest dependent variable in a multinomial logistic regression, might overestimate the influence of the predictor variables (Clarke & Muthén, in submission; Roeder, Lynch, & Nagin, 1999). Thus, after selecting the optimal number of latent profiles, a one-step approach was conducted in which the LP identification and profile membership prediction were performed simultaneously. Figure 1 shows the final LPA using standard path diagram conventions. Rectangles represent the observed variables (i.e., the four subscales measuring nurses’ work-related stress), and the oval represents the categorical latent variable (i.e., the latent profiles of work-related stress). The four continuous subscales are used to separate the latent patterns of work-related stress, whereas the profile membership of respondents is simultaneously explained by the following covariates, as shown in Table 1: age, gender (0 = male, 1 = female), years of employment, nationality (0 = Austria, 1 = others), province (0 = Carinthia, 1 = Vienna), financial situation (response categories (a) “secure” and (b) “rather secure” were coded as one; response categories (c) “tight (d) “rather precarious,” and (e) “precarious” were coded as zero), burden of household chores (response categories (a) “low” and (b) “rather low” were coded as one; response categories (c) “moderate,” (d) “rather high,” and (e) “high” were coded as zero), and perceived relevance of work (response categories (a) “high” and (b) “rather high” were coded as one; response categories (c) “rather low” and (d) “low” were coded as zero). In addition, the five subscales “emotional exhaustion,” “intrinsic motivation,” “dissatisfaction at work,” “aversion to residents,” and “reactive screening,” which measured nurses’ work stress reactions, were used to predict profile membership. Due to rather high numbers of nonresponses (ranging from 5 % to 22 %), the five subscales were dichotomized using the subscale-specific median value as the cut off (0 = less than or equal to median value, 1 = greater than median value; the corresponding median values are given in Table 2). Nonresponses were coded for each subscale using an additional indicator. To avoid local maxima of the log-likelihood 1,000, random starts were chosen for each model. All calculations were performed in MPlus (Muthén & Muthén, 1998-2007).

Figure 1.

Path diagram of the final latent profile analysis.

Table 1.

Overall Sample Characteristics and Covariates.

		Province
Variable		Carinthia(n = 603)		Vienna(n = 241)		Total(n = 844)^a		Significance^b
Age	M (SD)	37.02	(9.3)	40.02	(8.6)	37.84	(9.18)	t(428.5) = −4.34, p < .001
Gender
Female	n (%)	560	(93.5)	213	(89.1)	773	(92.2)	χ²(1) = 4.56, p = .033
Male	n (%)	39	(6.5)	26	(10.9)	65	(7.8)
Years of employment	M (SD)	11.5	(6.3)	14.5	(7.2)	12.31	(6.72)	t(376.0) = −5.64, p < .001
Nationality
Austria	n (%)	564	(95.3)	91	(39.1)	655	(79.4)	χ²(1) = 322.95, p < .001
Others	n (%)	28	(4.7)	142	(60.9)	170	(20.6)
Financial situation
Precarious	n (%)	346	(58.4)	140	(60.3)	486	(59.0)	χ²(1) = 0.25, p = .618
Secure	n (%)	246	(41.6)	92	(39.7)	338	(41.0)
Burden of housework
High	n (%)	202	(35.1)	65	(30.0)	267	(33.7)	χ²(1) = 1.89, p = .169
Low	n (%)	373	(64.9)	152	(70.0)	525	(66.3)
Perceived relevance of work
Low	n (%)	103	(17.2)	81	(34.6)	184	(22.1)	χ²(1) = 29.81, p < .001
High	n (%)	497	(82.8)	153	(65.4)	650	(77.9)
Nurse’s work stress due to…
…Residents	M (SD)	2.80	(0.65)	3.06	(0.69)	2.88	(0.67)	t(388.5) = −4.83, p < .001
…Working condition	M (SD)	3.64	(0.50)	3.75	(0.54)	3.67	(0.52)	t(402.6) = −2.67, p = .008
…Family members	M (SD)	2.97	(0.56)	3.20	(0.59)	3.04	(0.58)	t(406.9) = −5.12, p < .001
…Negative emotions	M (SD)	2.83	(0.55)	2.96	(0.55)	2.87	(0.55)	t(417.0) = −2.99, p = .003
Nurse’s reactions
Emotional exhaustion	M (SD)	35.5	(12.58)	37.1	(13.07)	35.9	(12.74)	t(400.9) = −1.65, p = .100
Intrinsic motivation	M (SD)	23.6	(8.65)	26.6	(7.79)	24.5	(8.51)	t(365.2) = −4.08, p < .001
Work dissatisfaction	M (SD)	18.4	(8.71)	21.7	(9.20)	19.3	(8.97)	t(375.9) = −4.61, p < .001
Aversion to residents	M (SD)	11.8	(5.54)	12.8	(6.16)	12.1	(5.67)	t(283.2) = −1.92, p = .055
Reactive screening	M (SD)	25.1	(4.82)	24.9	(4.80)	25.1	(4.81)	t(404.0) = −0.60, p = .552

Note: M = mean, SD = standard deviation, n = frequency, ^aThe total number of respondents does not equal 844 due to missing values. ^bWelch t-test for continuous measures and χ²-test for categorical covariates.

Table 2.

Parameter Estimates for the Three-Cluster Solutions Without Simultaneously Incorporating Covariates.

	Model I: Equal variances across profiles
	LP1 (n = 174)		LP2 (n = 562)		LP3 (n = 101)
Stress due to . . .	Mean	Variance	Mean	Variance	Mean	Variance
. . . Residents	2.11	0.17	2.95	0.17	3.88	0.17
. . . Working conditions	3.37	0.23	3.73	0.23	3.94	0.23
. . . Family members	2.49	0.20	3.09	0.20	3.72	0.20
. . . Negative emotions	2.49	0.23	2.89	0.23	3.39	0.23
	Model II: Free variances across profiles
	LP1 (n = 170)		LP2 (n = 564)		LP3 (n = 103)
. . . Residents	2.10	0.18	2.94	0.17	3.88	0.17
. . . Working conditions	3.36	0.25	3.72	0.22	3.94	0.25
. . . Family members	2.47	0.14	3.09	0.21	3.72	0.19
. . . Negative emotions	2.48	0.19	2.89	0.23	3.38	0.30

Note: LP = latent profile

Results

Table 1 provides an overview of the characteristics of the present sample. Nurses from the Federal State of Vienna were significantly older, had a significantly higher number of years of employment, and were less likely to be female. Furthermore, nurses employed in Vienna were significantly less likely to be born in Austria and tended to experience lower job relevance. In addition, there were significant differences across the provinces with respect to the nurses’ stress measures and the nurses’ stress reactions: nurses employed in Vienna experienced significantly higher stress due to residents, family members, working conditions, and negative emotions, experienced higher intrinsic motives, and were less satisfied with their jobs. Overall, these results underline the necessity to account for provincial differences in the prediction of LP membership.

Table 3 gives the correlations among nurses’ stress measures and the stress reactions. Due to the large number of observations, we focused on the magnitude of the Pearson correlation coefficients instead of p-values. In general, the highest correlations were observed among nurses’ stress reactions and experienced stress due to residents (correlations ranged between .309 and .504).

Table 3.

Correlations Among Measures of Nurses’ Stress and Stress Reactions.

	Stress due to . . .
Stress reactions	Residents	Working conditions	Family members	Negative emotions
Emotional exhaustion	.504***	.108**	.364***	.262***
Intrinsic motivation	.309***	−.122**	.200***	.093*
Work dissatisfaction	.448***	−.083*	.346***	.190***
Aversion to residents	.425***	.034	.306***	.253***

Note: *p < .05. **p < .01. ***p < .001

Typology of Nurses’ Work-Related Stress

Table 4 gives the model fit indices for the 2-, 3-, 4-, and 5-cluster solutions for the models in which the variance estimates are equal across profiles (in Table 4 referred to as Model I) and the models where the parameter estimates of variances were free to vary across profiles (see Model II in Table 4). Within the models with equal variances across profiles, the BIC and both Lo-Mendell-Rubin tests identified the three-profile solution as the best model. The adjusted BIC favored the four-profile solution. However, differences between the adjusted BIC-scores for the three- and four-cluster solution were minimal. According to Raftery (1995), this suggests “weak evidence” favoring the three-cluster solution. In addition, the three-cluster model showed the highest entropy, which also gave empirical evidence that supports a model consisting of three latent profiles. Similar results were obtained for the models allowing free variance across clusters. The BIC, adjusted BIC, and both Lo-Mendell-Rubin tests favored the three-cluster solution (see Table 4, Model II). To select the final LP-model for further analyses, the potential heterogeneity of indicators across clusters was examined by inspecting the variance estimates of Models I and II (see Table 2). The results suggest that none of the indicators showed a higher dispersion within a certain cluster in Model II. In addition, all of the information criteria identified Model I as being more parsimonious. Thus, the three-cluster solution of Model I (equal variance across clusters) was selected as the final model.

Table 4.

Model Fit Indices for the Latent Profile Models Without Simultaneously Incorporating Covariates (Indices suggesting best model fit are marked bold).

Model I: Equal variances across profiles
No. of clusters	AIC	BIC	adjBIC	Entropy	LMR	adjLMR
2	5254.8	5316.3	5275.0	0.569	<.0001	<.0001
3	5109.4	5194.6	5137.4	0.728	<.0001	<.0001
4	5100.6	5209.4	5136.4	0.626	0.178	0.1874
5	5095.0	5227.5	5138.5	0.688	<.001	<.001
Model II: Free variances across profiles
2	5256.0	5336.4	5282.4	0.591	<.001	<.001
3	5113.2	5236.2	5153.6	0.739	<.0001	<.0001
4	5100.9	5266.4	5155.3	0.785	0.078	0.082
5	5091.1	5299.2	5159.5	0.815	0.103	0.107

Note: AIC = Akaike Information Criterion. BIC = Bayes Information Criterion. adjBIC = adjusted BIC. LMR = p-values of the Lo-Mendell-Rubin test. adjLMR = p-value of the adjusted LMR

Predicting LP Membership

In the next step, covariates were entered into the selected model to predict cluster membership. Nonsignificant predictors were deleted to obtain the final LP membership predictions. The total number of respondents was reduced to 820 due to missing values. However, the inclusion of covariates did not substantially change the latent profiles. Figure 2 displays the profiles of nurses’ work-related stress of the final Model I incorporating significant covariates.

Figure 2.

Work stress profiles based on the final LP-model including covariates.

Overall, work-related stress due to residents and family members showed the highest separability between the extracted latent profiles. LP 1 accounted for 26 % of the sample (n = 217) and was distinguished from the other profiles by showing a low stress level on all four subscales. All profile-conditional means were lower than the overall sample means for every work stress indicator given in Table 1. LP 2 (n = 509, 62 % of the total sample) included nurses who showed moderate stress on all subscales. For each work stress indicator, the profile-conditional mean closely matched the overall sample mean. In addition, the final LPA identified a smaller subgroup of nurses (n = 94, 12 %) suffering from higher stress due to residents, family members, working conditions, and negative emotions. Each profile-conditional mean of this group exceeded the corresponding overall sample mean.

As shown in Figure 1, the final LP-model included a logistic regression model where the latent profiles were regressed on the above mentioned covariates. The following section reports the results of the final logistic regression model for predicting the LP membership after deleting nonsignificant covariates. To identify potential protective factors of work-related stress, the “higher stress” profile was used as the reference group. Table 5 gives the sample characteristics for each latent profile. The odds ratios of profile membership, as well as the 95 % confidence intervals of the regression model, are given in Table 6. A secure financial situation was positively related to profile memberships when comparing the “higher stress” and the “low stress” profiles (OR = 2.96, 95% CI = 1.36—6.44). Working in Vienna was inversely related to profile membership when comparing the higher and low stress profiles (OR = 0.17, 95% CI = 0.08—0.35) and when comparing the higher and the moderate stress profiles (OR = 0.30, 95% CI = 0.16—0.53). Higher scores for emotional exhaustion and not responding to at least one of the items measuring emotional exhaustion were inversely related to profile membership when comparing the higher and low stress groups (OR = 0.08, 95% CI = 0.03—0.21 and OR = 0.17, 95% CI = 0.04—0.71, respectively). Comparing the higher and the moderate stress profiles showed similar results for higher emotional exhaustion scores (OR = 0.32, 95% CI = 0.14—0.72), whereas nonresponse was not significantly related to membership. Higher work dissatisfaction was inversely related to higher versus low stress profile membership (OR = 0.22, 95% CI = 0.09—0.50); the relationship was not significant for the moderate stress profile. Higher aversion to residents showed a significant inverse relation when comparing the low and higher stress memberships (OR = 0.15, 95% CI = 0.07—0.32) and when comparing the moderate and higher stress profiles (OR = 0.45, 95% CI = 0.24—0.85).

Table 5.

Sample Characteristics of the Three Latent Profiles.

		LP1 (n = 217, 26%)		LP2 (n = 509, 62%)		LP3 (n = 94, 12%)
Variable		“Low stress”		“Moderate stress”		“Higher stress”
Latent profile probability	M (SD)	0.876	(0.148)	0.884	(0.134)	0.892	(0.136)
Age	M (SD)	38.5	(9.4)	37.5	(9.1)	38.0	(9.4)
Gender: Female	n (%)	197	(90.8)	466	(92.5)	91	(96.8)
Years of employment	M (SD)	12.1	(7.0)	12.2	(6.4)	13.2	(7.0)
Foreign nationality	n (%)	27	(12.7)	91	(18.2)	44	(49.4)
Province: Vienna	n (%)	37	(17.1)	136	(26.7)	57	(60.6)
Financial situation: Secure	n (%)	131	(60.4)	182	(35.8)	24	(25.5)
Burden of housework: Low	n (%)	149	(73.0)	302	(62.4)	64	(71.9)
Relevance of work: High	n (%)	193	(89.8)	389	(77.0)	55	(58.5)
Stress due to . . .
. . . Residents	M (SD)^a	2.15	(0.40)	2.99	(0.40)	3.88	(0.40)
. . . Working conditions	M (SD)^a	3.41	(0.49)	3.73	(0.49)	3.92	(0.49)
. . . Family members	M (SD)^a	2.56	(0.44)	3.11	(0.44)	3.73	(0.44)
. . . Negative emotions	M (SD)^a	2.53	(0.48)	2.91	(0.48)	3.38	(0.48)
Reactions . . .
Emotional exhaustion	M (SD)	25.9	(8.2)	37.8	(11.0)	48.7	(14.2)
> MD = 35	n (%)	24	(11.1)	261	(51.3)	74	(78.7)
≤ MD = 35	n (%)	178	(82.0)	206	(40.5)	13	(13.8)
Nonresponse	n (%)	15	(6.9)	42	(8.3)	7	(7.5)
Intrinsic motivation	M (SD)	20.8	(8.8)	25.0	(7.7)	29.2	(8.7)
> MD = 24	n (%)	46	(21.2)	202	(39.7)	54	(57.4)
≤ MD = 24	n (%)	112	(51.6)	177	(34.8)	22	(23.4)
Nonresponse	n (%)	59	(27.2)	130	(25.5)	18	(19.1)
Work dissatisfaction	M (SD)	13.4	(5.6)	20.2	(8.2)	27.7	(10.7)
> MD = 18	n (%)	33	(15.2)	253	(49.7)	73	(77.7)
≤ MD = 18	n (%)	162	(74.7)	222	(43.6)	16	(17.0)
Nonresponse	n (%)	22	(10.1)	34	(6.7)	5	(5.3)
Aversion to residents	M (SD)	8.7	(3.6)	12.3	(5.1)	17.3	(7.2)
> MD = 11	n (%)	25	(11.5)	184	(36.1)	59	(62.8)
≤ MD = 11	n (%)	126	(58.1)	177	(34.8)	13	(13.9)
Nonresponse	n (%)	66	(30.4)	148	(29.1)	22	(23.4)
Reactive screening	M (SD)	25.3	(4.7)	24.9	(4.8)	25.8	(4.8)
> MD = 25	n (%)	95	(43.8)	214	(42.0)	46	(48.9)
≤ MD = 25	n (%)	112	(51.6)	262	(51.5)	42	(44.7)
Nonresponse	n (%)	10	(4.6)	33	(6.5)	6	(6.4)

Note: M = mean, SD = standard deviation, n = frequency

estimated model parameters

Table 6.

Final Multinomial Logistic Regression to Predict Latent Profile Membership.

	Latent profile: “Low stress”				Latent profile: “Moderate stress”
	95 % CI				95 % CI
Variables	OR	Lower	Upper	p-value	OR	Lower	Upper	p-value
Secure financial situation	2.96	1.36	6.44	0.006	1.48	0.71	3.07	0.293
Province: Vienna	0.17	0.08	0.35	<.0001	0.30	0.16	0.53	<.0001
High emotional exhaustion	0.08	0.03	0.21	<.0001	0.32	0.14	0.72	0.006
NA in emotional exhaustion	0.17	0.04	0.71	0.015	0.28	0.07	1.10	0.069
High work dissatisfaction	0.22	0.09	0.50	<.0001	0.50	0.24	1.04	0.066
High aversion to residents	0.15	0.07	0.32	<.0001	0.45	0.24	0.85	0.014

Note: Reference = Latent Profile 2 (“higher stress”)

OR = Odds Ratio, CI = Confidence Interval, NA = Missing Value

Discussion

Taking the established working stress models into consideration, the results suggest that employees in inpatient care for the elderly are a high-risk group for burnout. The experience of an imbalance between high work demands, marked by a large number of distressing experiences, on one side versus an inappropriate valuation of these efforts on the other side, as described in the Effort-Reward Imbalance model by Siegrist (1996a, 1996b), is relevant for the nurses analyzed in this study.

The LPA divided the data of 844 nurses into three profiles. The three-cluster model proved to be the best solution based on the model accuracy indices and with regard to reflections on the content. The three profiles can be separated from each other by a low, moderate, or higher stress of workload. Nurses with low stress factors represented a quarter of the total sample. Persons in this group are financially secure national residents living in rural areas. The majority of these nurses describe their work as satisfying, varied, and interesting. Nurses with this characteristic represent a very stable group of employees, and they are the most likely to remain in the job and even act as a stabilizing factor for other staff members. They should be supported and encouraged by care managers to promote a positive work environment. The group situated between the low and highly stressed nurses represents the largest group (509 persons). They are mainly Austrians and reside mostly in rural housing areas. The largest group could benefit highly from health promoting interventions and general improvements of the work environment. The highly stressed nurses predominately work in the Federal State of Vienna, and half of them are not of Austrian origin. The combination of these two factors is characteristic of the demands nurses in inpatient care for the elderly experience in major cities. Globalization and the general labor market situation have led to an increase of different nationalities within nursing teams working in Vienna. Due to language barriers, misunderstandings arise in everyday work (Iecovich, 2011). Discrepancies in the understanding of nursing practice often lead to long discussions about responsibilities and authorizations to perform certain tasks. These causes for conflict often depend on the different educational levels or training of the nurses (Pedersen, 2006; Piko, 2006). A high level of education can be counted as a substantial resource (Laubach & Brosig, 1998; Schwirian, 1984; Yung, 1996), although these competencies in particular, especially within the field of patient-centered care, may be felt as a contradiction in daily nursing services (Stankova, 1994). The specific challenge regarding holistic and needful inpatient care for the elderly can be seen in the necessary combination of highly professional nursing activities with every day care such as foot, nail, or hair care, simple assisting duties normally performed by family members. In addition to problems arising from the diversity of vocational training, the impact of living conditions on foreign nurses must be considered in the sense of a negative spillover with regard to perceived working stress. Many foreign nurses, mostly from Eastern Europe, participating in this study and working in Austria spend most of their leisure time in their home countries. Necessary adjustment for them, in addition to general difficulties perceived by migrants, such as deficiencies in language acquisition and insecure or lacking social contacts (Pedersen, 2006), also consist of changes or modifications of values, norms, and sometimes even very basic local routines (Papadopoulos, 2003). Knowing the elder persons in the nursing homes personally and having private contacts with other staff members is another possible factor of influence in the sense of a positive spillover in rural areas. This community connectedness, together with a lack of job alternatives, and a higher quality of life, may be positively associated with job satisfaction (Philips & McLeroy, 2004). One important finding of our study was that nurses working in a rural area with a secure financial situation are threefold less likely to experience higher stress. Staff, including persons from different cultures, might need appropriate culturally sensitive techniques (Berry, 1997).

In accord with other studies (Jenull & Brunner, 2008; Kanai-Pak, Aiken, Sloane, & Poghosyan, 2008; Mackintosh, 2006; McHugh, Kutney-Lee, Cimiotti, Sloane, & Aiken, 2011), our results show that job dissatisfaction and burnout are highly associated with deficiencies in the work environment. Understaffed, emotionally exhausted, and dissatisfied nurses are not able to provide patient-centered care. We found the highest correlations between nurses’ stress reactions and experienced stress due to residents` behavior. This is not surprising, because nursing home populations with complex, multimorbid complaints have a high prevalence of psychiatric disorders, for example, major depression or anxiety disorders (Luppa et al., 2010; Seitz, Purandare, & Conn, 2010).

The work and conflicts with the residents and their relatives represent other important key aspects of work-related stress and are well suited to separate the three profiles from each other. Contact with family members represents an area of duty with high ambivalence. Relatives are in a position to facilitate everyday nursing as important reference persons of the institutionalized people but can also add to stress for nurses if helping them is not in agreement with the nursing procedures. Nurses need consolation, support, or help themselves to deal with such ageing relatives. These comprehensive concerns cannot be covered by the nursing staff alone, because these concerns do not require nursing qualifications but psychological education and knowledge and the ability to master critical situations. Schooling in techniques for crisis intervention might be helpful.

In conclusion, this article points to the importance of taking work- and nonwork-related stress into account and its practical implications. The identified patterns of work stress can be a starting point for generating ideas for workplace improvements. It should become a major goal to invest in care quality and adequate nurse staffing (Aiken, Clarke, Sloane, Sochalski, & Silber, 2002; Aiken et al., 2012). Nurses are key figures for taking care of seriously ill people and their relatives. They need supportive supervision, emotional support in difficult situations, and coordination with other health care professionals (Spoor et al., 2010). Managers can enhance the quality of work life and the motivation to stay by supporting the balance between family and work and by providing the opportunity for part-time work or job sharing (Laschinger et al., 2003). Interventions for multicultural teams focus primarily on language acquisition (e.g., http://www.meet-the-need-project.eu/englisch/ [25.3.2013]). We think that there is a further need to reduce xenophobia and promote acculturation. A tailored intervention program should also focus on recovery (e.g., individual stress prevention) during both work and leisure time. Being exhausted and dissatisfied is a serious problem, and it affects everyone including the organization, nurses, residents, and their relatives.

Limitations and Suggestions for Future Research

The current study demonstrated that nursing homes are very interested in supporting our research, 81 of the 83 nursing homes took part. In the province of Carinthia, we reached a response rate of 78%, and in the province of Vienna, a response rate of 60% was achieved. We cannot give any information about the reasons for nonparticipation. Furthermore, nurses’ perceived financial situation, perceived burden of household chores, and perceived job relevance were assessed using three single items. Thus, no detailed psychometric information (e.g., Cronbach’s α) is currently available. Future studies might use validated instruments, such as the Organizational Commitment Questionnaire (OCQ; see Meyer, Allen & Smith, 1993), to further explore the associations between work-related stress and various components of organizational commitment and the incharge financial distress/financial well-being scale (IFDFW; Prawitz et al., 2006) to explore the impact of financial security on work-related stress. Our findings provide clear evidence of the spillover of work- and nonwork-related stress. Further research is needed to consolidate the identified patterns of stress and to replicate the study in other areas. Moreover, resilient nurses, who are mainly working in rural areas, could be studied to learn about their living and working conditions. Another avenue for further research concerns multicultural teams. We noted that East European nurses move frequently between their place of work and their home countries. Residential mobility and its association with poor health are only documented for younger people (Busacker & Kasehagen, 2012; Jelleyman & Spencer, 2008). Future research may explore the problems and benefits of multicultural teams.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Brigitte B. Jenull, PhD, is an associate professor of psychology at the Alpen-Adria-Universität of Klagenfurt, in Austria. Her research interests include caregiving, health, clinical psychology, and psychotherapy of people in crisis situations.

Wolfgang Wiedermann, PhD, is an assistant professor of psychology at the University of Vienna, in Austria. He received a doctorate in psychology from the Alpen-Adria-Universität Klagenfurt. His research focuses on applied and computational statistics, latent variable modeling, and methods for causality research.

References

Aiken

L. H.

Clarke

S. P.

Sloane

D. M.

Sochalski

Silber

J. H.

(2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. Journal of the American Medical Association, 288, 1987-1993.

Aiken

L. H.

Sermeus

Van den Heede

Sloane

D. M.

Busse

McKee

Kutney-Lee

(2012). Patient safety, satisfaction, and quality of hospital care: Cross sectional surveys of nurses and patients in 12 countries in Europe and the United States. British Medical Journal, 344. e1717. doi:10.1136/bmj.e1717

Akaike

(1987). Factor analysis and AIC. Psychometrika, 52, 317-332.

Alfredsson

Karasek

Theorell

(1982). Myocardial infarction risk and psychosocial work environment: An analysis of the male Swedish working force. Social Science & Medicine, 16, 463-467.

American Psychological Association (2010). Ethical principles of psychologists and code of conduct. Including 2010 amendments. Retrieved from http://www.apa.org/ethics/code/index.aspx

Barnett

R. C.

Baruch

G. K.

(1985). Women’s involvement in multiple roles and psychological stress. Journal of Personality and Social Psychology, 49(1), 135-145.

Baune

B. T.

Adrian

Jacobi

(2007). Medical disorders affect health outcome and general functioning depending on comorbid major depression in the general population. Journal of Psychosomatic Research, 62, 109-118.

Berry

J. W.

(1997). Immigration, acculturation, and adaptation. Applied Psychology: An International Review, 46, 5-34.

Bishop

C. E.

Squillace

M. R.

Meagher

Anderson

W. L.

Wiener

J. M.

(2009). Nursing home work practices and nursing assistants’ job satisfaction. The Gerontologist, 49, 611-622.

10.

Bundesministerium für Arbeit, Soziales und Konsumentenschutz (BMASK). (2011). Österreichischer Pflegevorsorgebericht [Austrian care report]. Retrieved from http://www.bmask.gv.at/cms/site/attachments/5/0/4/CH2094/CMS1313493260454/pflegevorsorgebericht_2011%5B1%5D.pdf

11.

Burisch

(2006). Das Burnout-Syndrom. Theorie der inneren Erschöpfung (3. Aufl.) [The burnout syndrome. Theory of inner exhaustion]. Heidelberg, Berlin: Springer.

12.

Busacker

Kasehagen

(2012). Association of residential mobility with child health: An analysis of the 2007 National Survey of Children’s Health. Maternal and Child Health Journal, 16(1), 78-87.

13.

Clarke

S. L.

Muthén

(2009). Relating latent class analysis results to variables not included in the analysis. Manuscript submitted for publication.

14.

De Jonge

Janssen

van Breukelen

(1996). Testing the Demand-Control-Support Model among health-care professionals: A structural equation model. Work & Stress, 10, 209-224.

15.

Eby

L. T.

Maher

C. P.

Butts

M. M.

(2010). The intersection of work and family life: The role of affect. Annual Review of Psychology, 61, 599-622.

16.

Edwards

J. R.

Rothbard

N. P.

(2000). Mechanisms linking work and family: Clarifying the relationship between work and family constructs. Academy of Management Review, 25, 178-199.

17.

Fronteira

Ferrinho

(2011). Do nurses have a different physical health profile? A systematic review of experimental and observational studies on nurses’ physical health. Journal of Clinical Nursing, 20, 2404-2424.

18.

Goode

W. J.

(1960). A theory of role strain. American Sociological Review, 25, 483-496.

19.

Greenhaus

J. H.

Beutell

N. J.

(1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76-88.

20.

Greenhaus

J. H.

Powell

G. N.

(2006). When work and family are allies: A theory of work family enrichment. Academy of Management Review, 31, 72-92.

21.

Haar

J. M.

(2004). Work-family conflict and turnover intention: Exploring the moderation effects of perceived work-family support. New Zealand Journal of Psychology, 33(1), 35-39.

22.

Hacker

Reinhold

(1999). Beanspruchungsscreening bei Humandienstleistungen (BHD – System) [Stress screening at the Human Services]. Frankfurt, Germany: Swets Test Service.

23.

Head

Kivimäki

Siegrist

Ferrie

Vahtera

Shipley

Marmot

(2007). Effort-reward imbalance and relational injustice at work predict sickness absence: The Whitehall II study. Journal of Psychosomatic Research, 63, 433-440.

24.

Hsu

H.-C.

Kung

Y.-W.

Huang

H.-C.

P.-Y.

Lin

Y.-Y.

Chen

W.-S.

(2007). Work stress among nursing home care attendants in Taiwan: A questionnaire survey. International Journal of Nursing Studies, 44, 736-746.

25.

Iecovich

(2011). What makes migrant live-in home care workers in elder care be satisfied with their job. The Gerontologist, 51, 617-629.

26.

Ilhan

M. N.

Durukan

Taner

Maral

Bumin

M. A.

(2008). Burnout and its correlates among nursing staff: Questionnaire survey. Journal of Advanced Nursing, 61, 100-106.

27.

Jelleyman

Spencer

(2008). Residential mobility in childhood and health outcomes: A systematic review. Journal of Epidemiology and Community Health, 62, 584-592.

28.

Jenull-Schiefer

(2005). Stressoren und Ressourcen in der stationären Altenpflege [Common stressors and resources in the elderly care]. Klagenfurt, Austria: Alpen-Adria-Universität.

29.

Jenull

Brunner

(2008). Death and dying in nursing homes: A burden for the staff? Journal of Applied Gerontology, 27(2), 166-180.

30.

Johnson

Hall

(1988). Job strain, work place social support, and cardiovascular disease: A cross-sectional study of a random sample of the Swedish working population. American Journal of Public Health, 78, 1336-1342.

31.

Kalisch

B. J.

Lee

Rochman

(2010). Nursing staff teamwork and job satisfaction. Journal of Nursing Management, 18, 938-947.

32.

Kanai-Pak

Aiken

L. H.

Sloane

D. M.

Poghosyan

(2008). Poor work environments and nurse inexperience are associated with burnout, job dissatisfaction and quality deficits in Japanese hospitals. Journal of Clinical Nursing, 17, 3324-2329.

33.

Karasek

(1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285-308.

34.

Kayser-Jones

Schell

Lyons

Kris

Chan

Beard

(2003). Factors that influence end-of-life care in nursing homes: The physical environment, inadequate staffing, and lack of supervision. The Gerontologist, 43, 76-84.

35.

Kramer

Maguire

Brewer

B. B.

(2011). Clinical nurses in Magnet hospitals confirm productive, healthy unit work environments. Journal of Nursing Management, 19, 5-17.

36.

Kutney-Lee

McHugh

M. D.

Sloane

D. M.

Cimiotti

J. P.

Flynn

Neff

D. F.

Aiken

L. H.

(2009). Nursing: A key to patient satisfaction. Health Affairs, 28, 669-677.

37.

Laschinger

H. K. S.

Finegan

Shamian

Wilk

(2003). Workplace environment as a predictor of nurse burnout in restructured healthcare settings. Longwoods Review, 1(3), 2-11.

38.

Laubach

Brosig

(1998). “Die ganze Zeit arbeiteten wir Hand in Hand.” Eine vergleichende Studie über die Auswirkungen von Professionalisierung am Beispiel von deutschen und US-amerikanischen Intensivpflegekräften [“The whole time we worked hand in hand.” A comparative study of the effects of professionalization on the example of German and U.S. intensive care nurses]. Pflege, 11(1), 15-20.

39.

Lee

T.-N.

Lin

J.-D.

Yen

C.-F.

Loh

C.-H.

Hsu

S.-W.

Tang

C.-C.

Chu

(2009). Extrinsic high-effort and low-reward conditions at work among institutional staff caring for people with intellectual disabilities in Taiwan. Research in Developmental Disabilities, 30, 284-293.

40.

Mendell

N. R.

Rubin

D. B.

(2001). Testing the number of components in a normal mixture. Biometrika, 88, 767-778.

41.

Luppa

Sikorski

Luck

Ehreke

Konnopka

Wiese

Riedel-Heller

S. G.

(2010). Age- and gender-specific prevalence of depression in latest-life—Systematic review and meta-analysis. Journal of affective disorders, 136, 212-221.

42.

Lutz

Sanderson

Scherbov

(2008). The coming acceleration of global population ageing. Nature, 451, 716-719.

43.

Mackintosh

(2006). Protecting the self: A descriptive qualitative exploration of how registered nurses cope with working in surgical areas. International Journal of Nursing Studies, 44, 982-990.

44.

Marsh

H. W.

Lüdtke

Trautwein

Morin

A. J. S.

(2009). Classical latent profile analysis of academic self-concept dimensions: Synergy of person- and variable- centered approaches to theoretical models of self-concept. Structural Equation Modeling, 16, 191-225.

45.

McHugh

M. D.

Kutney-Lee

Cimiotti

J. P.

Sloane

D. M.

Aiken

L. H.

(2011). Nurses’ widespread job dissatisfaction, burnout, and frustration with health benefits signal problems for patient care. Health Affairs, 30, 202-210.

46.

Meyer

J. P.

Allen

N. J.

Smith

C. A.

(1993). Commitment to organizations and occupations: Extension and test of a three-component conceptualization. Journal of Applied Psychology, 78, 538-551.

47.

Muthén

B. O.

Muthén

L. K.

(2000). Integrating person-centered and variable-centered analysis: Growth mixture modeling with latent trajectory classes. Alcoholism: Clinical and Experimental Research, 24, 882-891.

48.

Muthén

L. K.

Muthén

B. O.

(1998-2007). Mplus user’s guide (5th ed.). Los Angeles, CA: Muthen & Muthen.

49.

Nylund

K. L.

Asparouhov

Muthén

B. O.

(2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14, 535-569.

50.

Papadopoulos

(2003). The Papadopoulos, Tilki and Taylor model for the development of cultural competence in nursing. Journal of Health, Social and Environmental Issues, 4, 5-7.

51.

Pedersen

(2006). Knowledge gaps about stress and coping in a multicultural context. In Wong

Wong

(Eds.), Handbook of multicultural perspectives on stress and coping (pp. 579-595). New York, NY: Springer.

52.

Philips

C. D.

McLeroy

K. R.

(2004). Health in rural America: Remembering the importance of place. American Journal of Public Health, 94, 1661-1663.

53.

Piko

(2006). Burnout, role conflict, job satisfaction and psychological health among Hungarian health care staff: A questionnaire survey. International Journal of Nursing Studies, 43, 311-318.

54.

Prawitz

A. D.

Garman

E. T.

Sorhaindo

O’Neill

Kim

Drentea

(2006). The InCharge financial distress/ financial well-being scale: Development, administration, and score interpretation. Financial Counseling and Planning, 17(1), 34-50.

55.

Raftery

A. E.

(1995). Bayesian model selection in social research (with discussion by Andrew Gelman, Donald B. Rubin and Robert M. Hauser. In Marsden

P. V.

(Ed.), Social methodology (pp. 111-196). Oxford, UK: Blackwells.

56.

Roeder

Lynch

K. G.

Nagin

D. S.

(1999). Modeling uncertainty in latent class membership: A case study in criminology. Journal of the American Statistical Association, 94, 766-776.

57.

Rösler

Jacobi

Rau

(2006). Work and mental disorders in a German national representative sample. Work and Stress, 20, 234-244.

58.

Schwarz

(1978). Estimating the dimension of a model. Annals of Statistics, 6, 461-464.

59.

Schwirian

(1984). Research on nursing students. In Werley

Fitzpatrick

(Eds.), Annual review of nursing research (Vol. 2., pp. 211-237). New York, NY: Springer.

60.

Sclove

L. S.

(1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52, 333-343.

61.

Seitz

Purandare

Conn

(2010). Prevalence of psychiatric disorders among older adults in long-term care homes: A systematic review. International Psychogeriatrics, 22, 1025-1039.

62.

Siegrist

(1996a). Adverse health effects of high effort-low reward conditions. Journal of Occupational Health Psychology, 1, 27-41.

63.

Siegrist

(1996b). Soziale Krisen und Gesundheit [Social Crises and Health]. Göttingen, Germany: Hogrefe.

64.

Spoor

De Jonge

Hamers

(2010). Design of the DIRECT-project: Interventions to increase job resources and recovery opportunities to improve job-related health, well-being, and performance outcomes in nursing homes. BMC Public Health, 10, 293-303.

65.

Stankova

(1994). Entwicklungsphasen der Tschechischen Pflege [Development phases of the Czech Care]. In Seidl

Stankova

(Eds.), Ende der Pflegekrise? Ein interkultureller Vergleich von Arbeitssituationen im Krankenhaus [End of care crisis? An intercultural comparison of work situations in hospital] (pp. 13-56). Wien, Austria: Wilhelm Maudrich.

66.

Statistik Austria (2011). Statistical yearbook of Austria. Wien, Austria: Verlag Österreich.

67.

Statistisches Bundesamt (2011). Pflegestatistik 2011 [Care statistics 2011]. Retrieved from https://www.destatis.de/DE/Publikationen/Thematisch/Gesundheit/Pflege/PflegeDeutschlandergebnisse5224001119004.pdf?__blob=publicationFile

68.

Ulrich

O’Donnell

Taylor

Farrar

Danis

Grady

(2007). Ethical climate, ethics stress, and the job satisfaction of nurses and social workers in the United States. Social Science & Medicine, 65, 1708-1719.

69.

Van Vegchel

De Jonge

Bosma

Schaufeli

(2005). Reviewing the effort-reward imbalance model: Drawing up the balance of 45 empirical studies. Social Science & Medicine, 60, 1117-1131.

70.

Vermunt

J. K.

Magidson

(2002). Latent class cluster analysis. In Hagenaars

McCutcheon

(Eds.), Applied latent class analysis (pp. 89-106). Cambridge, UK: Cambridge University Press.

71.

Vieira

E. R.

(2007). Why do nurses have a high incidence of low back disorders, and what can be done to reduce their risk? Bariatric Nursing and Surgical Patient Care, 2(2), 141-147.

72.

Wayne

J. H.

Grzywacz

J. G.

Carlson

D. S.

Kacmar

K. M.

(2007). Work-family facilitation: A theoretical explanation and model of primary antecedents and consequences. Human Resource Management Review, 17, 63-76.

73.

Yung

(1996). Role conception and role discrepancy. A comparison between hospital based and degree nursing students in Hongkong. Journal of Advanced Nursing, 23, 184-191.

74.

Zimber

Weyerer

(1999). Arbeitsbelastungen in der Altenpflege [Work stress in the elderly care]. Göttingen, Germany: Hogrefe.