Workplace stress in nursing workers from an emergency hospital : Job Stress Scale analysis 1

This study identifies workplace stress according to the Job Stress Scale and associates it with socio-demographic and occupational variables of nursing workers from an emergency hospital. This is a cross-sectional study and data were collected through a questionnaire applied to 388 nursing professionals. Descriptive statistics were applied; univariate and multivariate analyses were performed. The results indicate there is a significant association with being a nursing technician or auxiliary, working in the position for more than 15 years, and having low social support, with 3.84, 2.25 and 4.79 times more chances of being placed in the ‘high strain job’ quadrant. The study reveals that aspects related to the workplace should be monitored by competent agencies in order to improve the quality of life of nursing workers.


Introduction
Workplace stress has been extensively studied in recent years from different perspectives aiming to identify its importance in the etiology of workers' health disorders (1)(2)(3)(4)(5)(6)(7) .Among the models used to evaluate occupational stress, as well as its implications on the health of individuals, the Demand-Control Model stands out (8)(9)(10) .It assumes that occupational stress is a result of imbalance between psychological demands and how much control one has over his/her work, which can harm the health of workers.Hence, the greater the demand and the less control one has over his/her work, the greater the risk an individual will become physically or psychologically ill (9) .Control over work encompasses aspects related to the use of intellectual skills (the extent to which the job involves learning, repetitiveness, creativity, varied tasks, and the development of individual special abilities) and decision-making authority (individual ability to make decisions over one's own job, influence held over a group's work, and managerial policies).Psychological demand refers to psychological requirements faced by workers during the performance of tasks.Social Rev. Latino-Am.Enfermagem 2011 Sep.-Oct.;19(5):1122-1131.
support in turn refers to issues related to relationships established with co-workers and bosses (8)(9)(10) .
The dimensions 'psychological demand' and 'control over job' gave origin to the four quadrants of the Demand-Control Model: 'high strain jobs', 'active jobs', 'passive jobs', and 'low strain jobs'.The quadrants 'low strain jobs' and 'active jobs' are reported by the author as the ones having the lowest association with a risk of becoming sick, while the 'high strain jobs' and 'passive jobs' are those most associated with such a risk (8)(9)(10) .
Many studies have sought associations between high strain jobs according to the Demand-Control Model and health disorders.Hypertension (3,7) , minor psychological disorders (1)(2) and musculoskeletal pain (5) are among the main studied disorders; many of these studies have found positive association between such disorders and high strain.
Two studies involving public institutions, both Brazilian (3) and international (11) , which used the Demand-Control Model, report the possibility that job stability affects the perception of workers; requirements accruing from the work process may be minimized given their

Method
This cross-sectional study was conducted at a hospital emergency service in the southern region of Brazil.The studied population was composed of nursing workers (nurses, nursing technicians and auxiliaries) from the nursing service of this facility, totaling 606 workers.The target population was invited to voluntarily participate through posters hung in care areas and in meetings specific to each field.
A total of 185 workers were excluded from the study.These comprised all temporary workers, those working at the facility for less than one year, those on leave (sick leave, maternity leave, personal problems, etc.), or formally occupying a managerial position and also those who worked in fields providing indirect care to patients.Of the 421 workers who met the inclusion criteria, 33 (7.8%) did not agree to participate in the study.Therefore, the sample was composed of 388 (92.2%) nursing workers, in a non-probabilistic convenience sample.
Data were collected through interviews and questionnaires from February to June, 2010.To evaluate occupational stress, the short version (9) of the Demand-Control Model instrument was used, the Job Stress Scale, which was translated and validated in Brazil (10) .This version has 17 questions, five to evaluate psychological demand at work (scores ranging from 5 to 20), six questions to evaluate control over job (scores from 6 to 24), and another six questions to evaluate social support (scores ranging from 6 to 24) (8)(9)(10) .
The demand-control quadrants were defined based on analyses of the dimensions 'psychological demand' and 'control over job'.Each dimension generated a score that resulted from the sum of scores obtained by each of the questions and was then classified as "low" or "high" after normal distribution was confirmed through averaging.Low demand (score ≤15) or high demand (score ≥16) was defined with an average score of 14.64 (±2.4).For dimension control, low (score ≤17) or high control (score ≥18) was defined with an average score of 16.88 (±2.3).
Based on the demand and control dimensions (low and high), the Demand-Control quadrants were categorized into 'high strain jobs' (high psychological demand and low control); 'active jobs' (high psychological demand and high control); 'passive jobs' (low psychological demand and low control) and 'low strain jobs' (low psychological demand and high control) (8)(9)(10) .
Reliability was verified by estimating internal consistency in the global scale and its respective dimensions using Cronbach's alpha coefficient: the psychological demand reached 0.649 and the control dimension reached 0.523.In the evaluation of sub-items in the control dimension, Cronbach's alpha presented greater consistency for "decision-making authority" (0.627).
Other variables were considered according to two groups: reported socio-demographic variables (ethnicity, age, education, income, marital status, and number of children) and occupational variables (position, weekly workload, time in the current position, sector, social support).The scores of social support were defined by average and standard deviation (18.16±3.0)and classified either as low social support (score ≤18) or high social support (score ≥19).
The results were descriptively analyzed through central tendency measures (average or median) and variability (standard deviation or interquartile amplitude) as well as absolute and relative distribution (n-%).

Symmetry of continuous distributions was evaluated by
the Kolmogorov-Smirnov test.The Chi-square test was used to compare proportions between categories of the same variable (univariate analysis) taking into account the theoretical distribution of homogeneity between the compared categories.In the contingency tables in which at least 25% of the cell values presented expected frequency below 5, Fisher's exact test was used, while in situations in which at least one variable had polytomous characteristics, the Monte Carlo simulation was used.
To identify related factors relevant to the demandcontrol quadrants, the Multinomial Logistic Regression model was used; Odds Ratio was used to evaluate and directly estimate the probability of an event occurring (demand-control quadrants) (12) .All the variables with a minimum level of significance below 0.250 were considered in the non-adjusted logistic regression.
In the selection of variables to be excluded from the initial model, the likelihood ratio test (-2 log likelihood), obtained through the Chi-square test, for comparisons of the explicative variables and the dependent variable (demand-control quadrants) was performed.
The low demand quadrant was considered to be a reference group (lower risk of becoming ill) for the logistic regression according to the model proposed by Karasek and Theörell (9) .Data were statistically analyzed using the Statistical Package for Social Sciences ® (SPSS) version 17.0, Chicago, IL, USA.For decision criteria, level of significance (α) was fixed at 5%.When probability was situated between 0.05 and 0.10, the level of significance was considered the limit.

Results
This study's sample (n=388) was initially evaluated in relation to socio-demographic and occupational aspects.The participants presented an average age of 48.3 years old (±7.2), with a minimum age of 29 years old and a maximum of 68 years; ages between 49 to 59 years old (41.8%) were in the majority; 81.2% were women; 58% had completed secondary school; and most were married or lived in stable unions (56.4%).Monthly income was not reported by 29 (7.4%) professionals; of the 359 professionals who answered this variable, 50.6% (n=180) reported monthly per capita income up to R$ 1,200.00* ; 50% of the professionals had from two to four family members.
In relation to occupational characteristics, the sectors that had the largest number of workers were Intensive Care Units (ICUs) with 39.4% (n=153), followed by Surgical Centers with 14.4% (n=56), hospitalization units with 23.5% (n=91), and emergency units with 22.7% (n=88).More than half of the professionals (51.5%; n=200) reported they worked on the night shift and most worked 12 hours per shift (75.3%; n=292).
In relation to their position, 14.9% (n=58) were nurses and 85.1% (n=330) were nursing technicians and auxiliaries.These professionals worked in these positions between one and 37 years, with an average of 15.9 years (±6.7);54.6% of the participants worked for 16 years or more in their position; 82.9% (n=322) did not have an extra job.In relation to social support, 53.9% (n=209) of the professionals were classified as having low social support.
In relation to the psychological demand, answers presented an average score of 14.6 (±2.4), which ranged from 7 to 19; the number of professionals classified as having low psychological demand was significantly (p<0.001)higher.
An average score of 16.9 (±2.4) was obtained in the dimension 'control over work', with minimum and maximum scores of 8 and 22, respectively; the number of professionals with low control was significantly high (p=0.006).The application of adjusted and non-adjusted logistic regression is presented in Tables 4 and 5. Table 5 presents the variables that most reliably represent the association with demand-control quadrants.
Holding the low demand quadrant as a reference,   (8)(9) .Combining with these workers those classified as belonging to the high strain quadrant (21.4%), we verify that a large number of workers (57%) belong to quadrants that represent a risk to health.
Four studies conducted in Brazil used the Demand-Control Model in investigations addressing nursing workers in hospital facilities.The percentages found in the study conducted in a hospital in Bahia (1) , Brazil were lower for passive job and higher for high strain (28.4% and 27.8% respectively), but similar if compared to the sum of the two quadrants (56.2%).Another study addressing nursing workers from critical units of a hospital in Rio Grande do Sul (13) , Brazil also found a lower percentage for passive job and higher for high strain (19.7% and 44.1% respectively); thus, 63.8% of the workers was classified in these quadrants.
A study conducted with nursing workers from a public university hospital in the Midwest region of Rio Grande do Sul (6) found a lower result for passive job and similar to this study for high strain (29.9% and 21.2% respectively), that is, 51.1% of the workers belonged to quadrants that presented risk to health.The other study addressing nursing workers from a city hospital in Rio de Janeiro (2) , RJ, Brazil reports data opposite to these: Urbanetto JS, Silva PC, Hoffmeister E, Negri BS, Pinheiro da Costa BE, Poli de Figueiredo CE.
lower frequency of workers in the passive job and high strain (16.6% and 17.9% respectively), or only 34.5% of the workers belonged to quadrants that represent risk to health.
Two international studies addressing nursing workers also report important results.One of them, conducted in a Korean hospital, reported a lower percentage for passive job and higher percentage for high strain (26.1% and 29.2% respectively), totaling 55.3% of workers in these quadrants (14) .The other study reports a greater concentration of nurses in the passive job and high strain quadrants (32.1% and 25.8% respectively) (15) .
Another study reports that the low psychological demand identified among nursing workers was similar to that found among nursing technicians and auxiliaries (6) .
Such results are surprising, since the general context of the nursing job imposes an intense job pace (16) , often generated by the excessive demand of hospitalized patients (17) .The low psychological demand of the studied workers is linked to the fact that there are psychological conditions such as the time to perform tasks, fewer urgent situations, or little need to reorganize their jobs.Such a fact may be related to aspects such as continuity of job (nurses are present for 24 hours) and teamwork (46.1% of the professionals reported high social support).
There was a concentration of workers in the low control quadrant, probably because only 14.9% of the workers were nurses.A study evaluating the dimension 'control according to position', revealed that 75.3% of nurses had high control over job and 62.3% of the nursing technicians and auxiliaries had low control (6) .Low control over job is of concern because it characterizes a job process highly repetitive with limited autonomy and few opportunities to learn new things.This aspect may demotivate workers and generate low self-esteem, which according to the authors of the Demand-Control Model, are developments that are very harmful to workers' health.
Socio-demographic characteristics were not statistically associated with the demand-control quadrants, with the exception of the variable 'education'.
While professionals who were college graduates significantly associated with the active job and low strain quadrants, those who did not have a bachelor's degree were significantly associated with passive jobs and high strain.Being a nurse was also associated with the low risk quadrants, which may be related to the educational level previously mentioned, since this position requires higher education.
Other studies found a positive association between educational level and the quadrants, while nursing professionals who fit into the passive job and high strain quadrants were mainly individuals with no bachelor's degrees (1,18) .with occupational stress (19)(20) .Another study reported representativeness of data in ICU in the high strain (53.0%) and passive job (51.8%) quadrants (1) .

The
Characteristics related to work that place hospitalization units in the same quadrant as ICUs need to be investigated in this specific hospital.An element to take into account is that the number of patients cared for by a health provider in a hospitalization unit is usually higher than is the case in ICUs.The latter has its own regulations concerning the maximum number of patients to be cared for by nurses and nursing technicians and auxiliaries.
The findings related to position corroborate the organization of nursing work, in which actions are distributed according to a hierarchy and complexity of care.Nurses are responsible for planning, controlling and evaluating the work process and manual activities of greater complexity while nursing technicians or auxiliaries are responsible for manual activities of lower complexity and those prescribed by nurses (21) .This fact may explain the concentration of nurses in quadrants that do not present harm to health (low strain and active job) and the concentration of nursing auxiliaries and technicians in quadrants that represent risk to health (high strain and passive job).These data were also found in other studies (1,6) .
Working in the same position more than 15 years was associated with the quadrant high strain, with a 2.25 times greater chance when compared to the low strain quadrant.This fact is extremely relevant because there are a high number of professionals working in public facilities who remain in their positions until they retire.Even though this study does not provide evidence for an association between age and demandcontrol quadrants, other studies associate old age with illness.Another study found association between perception.Even though studies addressing health disorders in nursing workers have considerably developed in recent years, the relationship of such disorders with stressing factors in the workplace still requires greater clarification.In this context, identifying the profile of nursing professionals working in emergency hospital services in relation to their perceptions concerning psychological demands and control over their jobs can support the planning and implementation of actions toward a better quality of life and quality of life at work, both on the part of nursing workers and managers of health facilities.In this context and following the line of investigation of other studies, this study identifies occupational stress through the Job Stress Scale (JSS) and verifies its association with socio-demographic and occupational aspects of emergency service nursing workers.
Committee at the Pontifical Catholic University of Rio Grande do Sul (PUCRS) (Protocol No. OF.CEP-997/09) and the City Health Department (CEP 399Protocol Nº 001.044987.09.9) and each participant signed free and informed consent forms.
being a nursing auxiliary or technician, working in the position for more than 15 years, and having low social support were positively associated and indicated increased chances of an individual to be classified in the high strain quadrant.A nursing technician or auxiliary had a greater chance of belonging to the passive job quadrant while workers with more than 15 years in the position had a greater chance of belonging to the active job quadrant; workers from hospitalization units or ICUs had a lower risk of belonging to the active job quadrant.
occupational characteristics statistically associated with the demand-control quadrants were sector (not confirmed in the logistic regression), position, time in position, and social support.In relation to sector, the high strain quadrant most frequently concentrated professionals working in the hospitalization and emergency sectors.The hospitalization unit being classified as high strain was a surprise, because ICUs and the emergency sectors are usually the ones associated
presents the dimensions 'psychological demand' and 'control' (low and high) and the JSS demand-control quadrants.* Exchange rate at the time was US$ 1.00 = R$ 1.80

Table 1 -
Frequency of psychological demand, control over job and the JSS demand-control quadrants.Porto Alegre, RS, Brazil, 2010The socio-demographic profile of the nursing workers classified in the demand-control quadrants is presented in Table2.The variables gender, age, marital status and per capita income were not statistically associated with the demand-control quadrants.

Table 2 -
Distribution of the socio-demographic variables according to the JSS demand-control quadrants.Porto Alegre, RS, Brazil, 2010 * Pearson's Chi-square.

Table 4 -
Association of demand-control quadrants and co-variables generated in raw regression taking low strain as reference.Porto Alegre, RS, Brazil,

Table 5 -
Association of the demand-control quadrants and co-variables generated in the adjusted regression, taking low strain as the reference.Porto Alegre, RS, Brazil, 2010