Work ability of school servants and the mediating effects between personal characteristics,work conditions,and sickness absence 1

Abstract

BACKGROUND:

School servants are professionals involved with cleaning and feeding activities in schools. Exposed to conditions of physical and psychosocial overload at work, they may face situations of morbidity and functional restrictions.

OBJECTIVE:

To evaluate personal and work characteristics and work ability index (WAI) of school servants in Brazil and test the mediating effect of WAI in the duration of sickness absence.

METHODS:

163 Brazilian school servants participated in this cross-sectional study, that including the self-completion of a sociodemographic questionnaire, information about the Work Environment, the Protocol of Psychosocial Risks at Work’s Evaluation, and the Work Ability Index. Also was collected, in a secondary health database, information about sickness absence. A linear regression model was constructed to identify the predictors of the duration of absence, followed by the analysis of mediation with the Path Analysis method.

RESULTS:

The study found that the WAI partially mediates the relation of the personal characteristics and the work conditions with the duration of absence of the school servants. The predictors included the organizational aspects of work, mental exhaustion, physical damage, health self-evaluation, availability of personal protective equipment, and the need to assume inadequate postures at work. These predictors explained 33% of the duration of absence and 53% of the WAI of these workers.

CONCLUSION:

This study indicated that inadequate working conditions and low WAI impact in the sickness absence, increasing by 33% the length of time off work of school servants.

Keywords

Occupational health sick leave work capacity evaluation mediation analysis working conditions

1 Introduction

The organization and the dynamics of work have been through transformations in recent years, with changes in the ways people work, in the profile of diseases of the worker, and on the prevalence of indicators such as indices of people on leave, with disabilities, and of returning to work [1, 2]. In Brazil, most work leaves are due to lesions, followed by musculoskeletal and mental diseases [3,4, 3,4]. Specific groups of workers with high physical and psychosocial demands at work are exposed to greater risk and vulnerability, with high rates of morbimortality, absenteeism from work activities, and disability retirement [5, 6].

This is the reality of the work of school servants, professionals responsible for cleaning, preparing meals for all students, disciplinary assistance in recreational moments and use of restrooms. In Brazil, school servants do not provide pedagogical and educational services. The Brazilian school servants typically perform work in manual labor (blue collar), and are perceived to be less educated, low-skilled, and of a lower social class. In their daily work, school servants are subjected to occupational segregation, social invisibility, gender inequality, precarious work conditions, and disease, especially due to musculoskeletal disorders [7,8, 7,8]. Investigations in workers who carry out activities that are similar to those of the school servants, describe inadequate work place, posture, overload, physical environment, equipment, and processes [9 –11]. Brazilian and international publications related with school environment focus on the general staff of schools, especially teachers [12, 13]. Although it is necessary to intervene in the work conditions of the school servants, studies about these professionals are scarce, especially with regard to the work ability.

Our study is based on the multidimensional Work Ability House Model who been proposed to explain the work ability process. This model considers that the WA depends on the balance between individual resources, work-related factors, and the macrosocial environment [14,15, 14,15]. The model is expressed by a four-floor House and a roof inserted in an encircling environment [15]. The first floor represents the base that supports the building and concerns health status and functional capacity, including physical, mental, and social aspects. The second floor relates to professional competence (knowledge and skills, training and learning at work) and its continuous development used to meet the demands of working life. The third floor represents the internal aspects of the individual, manifested as values, attitudes, and motivation. The last floor represents factors related to work such as work conditions, demands and content, organization and communitarian environment, management and supervision: this is the “heaviest” floor in the building and can affect the other dimensions. The roof of this building is the WA, resulting from the interaction and balance between the previous dimensions [15].

Many studies have investigated factors related to the work ability in different professional categories [16 –18], highlighting the sickness absence as associated to a low work ability [19,20, 19,20]. As an indicator of health at work, monitoring the duration of absence of workers generates indicators of the exposure to risks in the varied sectors of work, making it possible to monitor the prevalences and most common diseases, while predicting states of disability and reflecting investments on the health of the worker [21]. Sick leaves also represent a financial expenditure, associated to the loss of productivity, to the hiring and training of new professionals, expenses with benefits or welfare, and extra costs with health services [5].

There is evidence of associations between personal and work characteristics with work ability and with the duration of absence of the workers, but the processes involved in this interaction are not clear [20, 22]. Recently, the type of occupation was mentioned as an important moderator of work ability and its associated factors [23]. Furthermore, a longitudinal study that followed municipal workers for 15 years indicated specific determinants associated to different trajectories of absence [24]. There are few investigations about the effect of work ability as a mediator of the relation between personal characteristics of the worker, working conditions, and the duration of absence [14 , 26]. The study by Kartika and Widhiandono [25] investigated the mediation effect of work ability on the relationship between job training and job performance; von Bonsdorff et al. [26] investigated the personal strategies and high-involvement work practices and the mediation effect of company work ability and company performance. A recent literature review investigated work ability and its affects in work attitudes, performance, and retirement [14]. As can be seen, these studies focus their analysis on workers characteristics or work characteristics, without interrelating them in the mediation between work ability and sick leave, especially in school servants. Identifying these characteristics can be the key for targeted interventions to reduce the expenses with diseases in this group of workers, in addition to promoting them with safer, healthier, and more functional work conditions.

Considering the above mentioned, the aims of this study were to analyze personal characteristics and work conditions of Brazilian school servants, and to verify whether the work ability of these professionals has a mediating effect between their personal characteristics such as work conditions, and their duration of absence.

2 Methods

2.1 Participants

The cross-sectional study carried out with public servants from the Education Secretariat of the city of Nova Lima, Minas Gerais (MG), Brazil.

The study included workers who were active school servants, above 18 years old, who had been working for at least three months in the public schools of the city. Active workers were those who were exercising their functions, who were not away from work and who were not retired. The criterion for inclusion of “minimum of three months” was chosen because with this time the participant would already have contact and experience with work activities in order to respond more reliably to the work evaluation instruments. All servants who, during data collection, were distance workers, carried out administrative functions, or were young apprentice (teenagers aged between 14 and 18) were excluded.

To calculate the sample, a significance level α=0.01; testing power (1-β)=0.95; large conventional effect (f² = 0.35), and a maximum number of predictors = 15. Therefore, the sample size, for a regression model and for the analysis of the mediating variable, was n = 118, as calculated by the software G*Power 3.1. The population of active school servants in 2019 was 324. Due to protective measures adopted in schools against the coronavirus disease 2019 (COVID-19) pandemic, 193 servants continued working in-person in 2020. All the servants were invited to participate in the study and, from those, 163 school servants agreed to participate. This sample by convenience (163) totals 84.5% of active workers at the time of data collection.

This study was approved the Research Ethics Committee from the Federal University of Triângulo Mineiro on October 27, 2020 (CAAE: 36566420.0.0000.5154, opinion 4.364.500). The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for reporting observational data was followed in this study [27]. To reduce potential information bias, the evaluator was previously trained in the application of the instruments and the participants were blinded to the objectives related to the analysis of the study data.

2.2 Instruments

The first stage took place from October to December 2020. All instruments were self-administered and participants answered the research instruments at their workplace. The data collection protocol was formed by a set of instruments: a sociodemographic, professional, life habit, and health information questionnaire specifically designed for this study; a set of questions about Work Environment; the Protocol of Psychosocial Risks at Work’s Evaluation (PPRWE); and the Work Ability Index (WAI).

The sociodemographic, professional, life habit, and health information questionnaire also included self-rated health. The question “How do you rate your health in the last 30 days?” was used and the response included options ranging from very bad to very good. Subsequently, the responses were dichotomized into positive perception and negative perception [28].

The block of Questions about the Work Environment was adapted from Barbosa et al. [29], including 15 items about three factors (factorial structure with KMO = 0.727; sphericity test Bartlett’s with p = 0.000; accumulated variance = 62.21): environmental conditions (6 items: α=0,81), physical demands (6 items; α=0,73), and personal protective equipment (PPE) (3 items; α=0,83). The items about PPE had two possible responses (yes/no); the score of the others was calculated using four-point Likert-type scales. For the environmental conditions, these varied from inadequate to adequate, while for the frequency of the physical demand, it varied from never to always.

The Protocol of Psychosocial Risks at Work’s Evaluation (PPRWE) is an instrument elaborated by Brazilian researches to map the psychosocial risks at work [30]. The PPRWE has been widely used in Brazil and in studies with workers from other countries with similar socio-economic conditions [30 –32]. It includes 92 self-applicable items, distributed in 4 scales: Prescript Organization of Work Scale (POWS); Styles of Management Scale (SMS); Pathogenic Suffering at Work Scale (PSWS); and Psychosocial Damages at Work Scale (PDWS). Each PPRWE scale contains a block of questions about common issues. Thus, POWS has items on organizational aspects of the work context, including 10 items of the division of tasks (α=0,85) and 12 items of the social division of labor (α=0,90) [30]. The SMS points to the characteristic of management used by the organization, with 10 items for the individualist (α=0,85) and 11 items for the collectivist style (α=0,93) [30]. The PSWS includes questions about lack of recognition (11 items; α=0,92), meaninglessness at work (9 items; α=0,90), and mental burnout (8 items; α=0,91) [30]. Finally, the PDWS gathers questions about physical (9 items; α=0,93), psychological (7 items; α=0,94) and social (7 items; α=0,91) problems that workers assess as essentially caused by work [30]. Each scale was subjected to factorial analysis, using the PAF method with oblimim rotation. The literature indicated a satisfactory factorial structure with eigenvalues above 1.00 for all factors; items with a load greater than 0.30, KMO above 0.94, accumulated variance around 51.79 and correlations with values above 0.22 [30]. The score of the items in the PPRWE follows a Likert-type scale, varying from 1 (never) to 5 (always). The score of each scale is calculated with a simple arithmetic mean. Finally, the value of each score is classified as low, medium, or high psychosocial risk, with the exception of the SMS scale, which is classified as a little, moderate or predominant pattern in management styles [30].

The Work Ability Index (WAI) was used to measure the work ability. This is an instrument validated in Brazil with good psychometric properties, as Cronbach’s alpha α=0.72, construct validity p < 0.001 and criterion validity p < 0.001 [33]. The WAI evaluates the perception of the worker about their ability to perform their jobs in the present and in the future, considering the demands, their state of health and physical and mental abilities. The WAI has 7 items, who evaluate current work ability compared to lifetime best; work ability in relation to the mental and physical demands of the job; current number of illnesses diagnosed by the physician; estimated loss to work due to illness; absences from work due to illness in the last 12 months; and enjoyment of daily activities, feeling active, alert, and hopeful about the future [15]. The final score varies from 7 to 49 points and higher values indicate better work ability [15]. The index also can be divided in four classes: poor (7–27 points), moderate (28–36 points), good (37–43 points), and excellent (44–49 points) [15].

In the second stage, electronic data on sick leaves were collected from the Integrated Service for the Health Care of Servants (SIASS). This stage was the extraction of electronic data on sickness absence in the database of clinical and epidemiological surveillance at the SIASS. The objective of SIASS is to coordinate and integrate public servants’ health care actions and programs, official expertise, and the promotion and prevention of occupational health [34]. The data used in the study are already part of the routine of health services for public servants in Nova Lima. These data had the causes and the total duration of absence, per disease of each participant, from January 2017 to December 2020. The collection of absenteeism data was carried out up to the date of application of the questionnaires on participants. These data included the sum of all days of sick leaves certified by medical or odontological documents, with a maximum period of 15 days for each certificate. According to Brazilian labor laws [35], workers who are on leave since 15 days are sent for external expert evaluations. The link between the information in the questionnaires and the leave data was established by the full name of the participants. After this connection, the names of the participants were encoded in numbers to guarantee secrecy.

2.3 Data analysis

Data was entered in Microsoft Office Excel 2007 and analyzed using SPSS version 26.0 and R version 4.0.3. The duration of absence was considered to be the depending variable and the WAI was evaluated as a mediating variable of the relation between personal characteristics, work conditions, and duration of absence. The categoric variables were submitted to a descriptive analysis, with frequency and proportion measures. For the numerical variables, the median and quartiles were calculated, except for the PPRWE, which were analyzed using mean and standard deviation, according to application criteria of the instrument itself [30].

The means of WAI and duration of absence did not have a normal distribution, and, for the bivariate inferential analysis, Spearman’s correlation was used for numerical values; Mann-Whitney U test for categorical values with up to two independent groups; and Kruskal-Wallis, followed by Bonferroni, for the categorical co-variables with three or more independent groups. For the bivariate analysis, a significance level of α=0.20 was considered.

The multivariate analysis carried out using a “best subsets regression” (ols_step_best_subset), attempted to identify a model that worked better to explain the duration of absence, using the adjusted highest R² value as a criteria for choice. As candidates for predictors of this model, we considered the co-variables that were simultaneously correlated with WAI and with the duration of absence in the bivariate analysis. The normality for the residues and the homoscedasticity of the model selected were estimated (ols_test_normality e ols_test_f), with a result of p > 0.05.

2.3.1 Analysis of the mediating effect

The mediation is observed when, in a relation between predictors (X) and response (Y), a third variable (Med) can intervene in the strength of this relation, entirely or partially reducing it [36]. The path analysis is the statistical method traditionally used to test the mediating effect [36, 37]. According to Fig. 1, for this effect to be proved in the regression equation, the relation must be significant in each stage of the path analysis: when estimating the relation between predictive variables and the outcome (step 1, path c), of the predictive variables with the mediator (step 2, path a), and of the mediator with the outcome (step 3, path b). Furthermore, when the mediator variable is inserted in the relation between predictive variables and outcome (step 4, path c’), it should be able to reduce or nullify the strength of this relation. If the coefficient estimated for Med is significantly different from zero, it shows an effect of total mediation, in which the entire magnitude of the relation of X with Y is nullified by the insertion of Med. However, in both cases of mediation, there is a reduction in the determination coefficient (R²) from the path c to the path c’. This study adopted a significance level of α=0.05. For each step of the path analysis, post hoc tests were performed in the G*power software, to check the power of each regression model considering the missing data, and maintaining the analysis only for a power greater than 80%.

Fig. 1

Analysis diagram of the path analysis to test the mediating effect.

3 Results

A total of 163 school servants participated in the study. Table 1 shows information about the personal characteristics of participants, including sociodemographic, occupational, life habits and health data. The characterization of physical and psychosocial work conditions is described in Table 2.

Table 1
Sociodemographic, occupational, life habit, and health characteristics (N = 163)

Categorical variables¹ n (%) Median DA (Q1–Q3) p value DA Median WAI (Q1–Q3) p value WAI

Sex Female 162 (99.4%) 26 (12–57) – 34 (30–38) –

Male² 1 (0.6%) – –

Marital status Lives with a partner 93 (57.8%) 38 (15–66) 0.054^* 33 (27–38) 0.251

Does not live with a partner 68 (42.2%) 19.5 (8.75–34.25) 36 (30.75–38.5)

Educational level Up to 8 years of formal education 34 (21.3%) 23.5 (6.5–47.75) 0.845 34.5 (27.5–38.75) 0.893

More than 8 years of formal education 126 (78.8%) 26 (12.5–62) 34 (30.5–38)

Other paid jobs No 153 (95.0%) 26 (12–59) 0.599 34 (29.75–38) 0.617

Yes 8 (5.0%) 19 (13–35) 33 (32–41)

Workplace Nursery 41 (25.2%) 33 (18–66) 0.261 34 (29–37) 0.577

Elementary education I 66 (40.5%) 20.5 (8–40.75) 36 (30.25–39.75)

Nursery &Elemental Education I 51 (31.3%) 25 (13.5–56.5) 34 (28.5–28.5)

Specialized Teaching School 5 (3.1%) 13 (9–57) 38 (25–41)

Weekly workload 35 hours³ 92 (56.8%) 33 (13–66) 0.132^* 34 (29.25–37.75) 0.281

40 hours³ 64 (39.5%) 19.5 (8.25–47.25) 36 (30–40)

Other⁴ 6 (3.7%) 27 (16.75–41) 32 (28.5–37)

Practices physical activities No 100 (61.7%) 24 (12–44.25) 0.396 34 (30–37.75) 0.865

Yes 62 (38.3%) 32.5 (12.5–65.25) 34.5 (29.25–38.25)

Frequency of physical activities 1 to 2 a week 30 (49.2%) 21.5 (13–58 + 5) 0.691 33 (27.75–37.75) 0.325

3 or more a week 31 (50.8%) 33 (8–69) 36 (29–38.5)

Regular alcohol consumption No 83 (53.2%) 31 (12.5–61) 0.200 35 (30–38) 0.994

Yes 73 (46.8%) 24 (12–51.25) 34 (29.75–38.25)

Regular alcohol consumption in standard drinks Up to 3 standard drinks 73 (83.9%) 25 (12–58) 0.225 34 (29.5–38) 0.336

More than 3 standard drinks 14 (16.1%) 16 (8–32) 37 (31–39.5)

Smoking No 140 (88.1%) 26 (13–59) 0.675 34.5 (30–38) 0.956

Yes 19 (11.9%) 19 (11–41) 33 (29–42)

Health self-evaluation Positive perception 100 (62.5%) 22 (9–41) 0.070^* 37 (32–41) 0,000^*

Negative perception 60 (37.5%) 35 (16–69.5) 31 (26.5–35)

Numerical variables¹ median (Q1–Q3) p value DA p value WAI

Age (years) 47.5 (40–53) 0.271 0.141^*

Children (number) 2 (1–2) 0.003^* 0.343

Family income (dollars –US$) 463.69

(370.95–581.47) 0.027^* 0.042^*

Time working as a school servant (years) 9 (8.0–16.0) 0.260 0.702

Time working in the school (years) 9 (6.0–13.8) 0.041^* 0.465

Body Mass Index (kg/m²) 28.9 (24.8–32.2) 0.031^* 0.016^*

Categorical variables¹	n (%)	Median DA (Q1–Q3)	p value DA	Median WAI (Q1–Q3)	p value WAI
Sex	Female	162 (99.4%)	26 (12–57)	–	34 (30–38)	–
	Male²	1 (0.6%)	–		–
Marital status	Lives with a partner	93 (57.8%)	38 (15–66)	0.054^*	33 (27–38)	0.251
	Does not live with a partner	68 (42.2%)	19.5 (8.75–34.25)		36 (30.75–38.5)
Educational level	Up to 8 years of formal education	34 (21.3%)	23.5 (6.5–47.75)	0.845	34.5 (27.5–38.75)	0.893
	More than 8 years of formal education	126 (78.8%)	26 (12.5–62)		34 (30.5–38)
Other paid jobs	No	153 (95.0%)	26 (12–59)	0.599	34 (29.75–38)	0.617
	Yes	8 (5.0%)	19 (13–35)		33 (32–41)
Workplace	Nursery	41 (25.2%)	33 (18–66)	0.261	34 (29–37)	0.577
	Elementary education I	66 (40.5%)	20.5 (8–40.75)		36 (30.25–39.75)
	Nursery &Elemental Education I	51 (31.3%)	25 (13.5–56.5)		34 (28.5–28.5)
	Specialized Teaching School	5 (3.1%)	13 (9–57)		38 (25–41)
Weekly workload	35 hours³	92 (56.8%)	33 (13–66)	0.132^*	34 (29.25–37.75)	0.281
	40 hours³	64 (39.5%)	19.5 (8.25–47.25)		36 (30–40)
	Other⁴	6 (3.7%)	27 (16.75–41)		32 (28.5–37)
Practices physical activities	No	100 (61.7%)	24 (12–44.25)	0.396	34 (30–37.75)	0.865
	Yes	62 (38.3%)	32.5 (12.5–65.25)		34.5 (29.25–38.25)
Frequency of physical activities	1 to 2 a week	30 (49.2%)	21.5 (13–58 + 5)	0.691	33 (27.75–37.75)	0.325
	3 or more a week	31 (50.8%)	33 (8–69)		36 (29–38.5)
Regular alcohol consumption	No	83 (53.2%)	31 (12.5–61)	0.200	35 (30–38)	0.994
	Yes	73 (46.8%)	24 (12–51.25)		34 (29.75–38.25)
Regular alcohol consumption in standard drinks	Up to 3 standard drinks	73 (83.9%)	25 (12–58)	0.225	34 (29.5–38)	0.336
	More than 3 standard drinks	14 (16.1%)	16 (8–32)		37 (31–39.5)
Smoking	No	140 (88.1%)	26 (13–59)	0.675	34.5 (30–38)	0.956
	Yes	19 (11.9%)	19 (11–41)		33 (29–42)
Health self-evaluation	Positive perception	100 (62.5%)	22 (9–41)	0.070^*	37 (32–41)	0,000^*
	Negative perception	60 (37.5%)	35 (16–69.5)		31 (26.5–35)
Numerical variables¹	median (Q1–Q3)		p value DA		p value WAI
Age (years)		47.5 (40–53)		0.271		0.141^*
Children (number)		2 (1–2)		0.003^*		0.343
Family income (dollars –US$)		463.69
		(370.95–581.47)		0.027^*		0.042^*
Time working as a school servant (years)		9 (8.0–16.0)		0.260		0.702
Time working in the school (years)		9 (6.0–13.8)		0.041^*		0.465
Body Mass Index (kg/m²)		28.9 (24.8–32.2)		0.031^*		0.016^*

¹ Spearman’s correlation for numerical variables; Mann-Whitney U test for categorical values with up to two independent groups; Kruskal-Wallis, followed by Bonferroni, for the categorical variables with three or more independent groups; *p<0.20; ²DA and WAI were constant in the male sex; ³There was a difference only in the groups with 35 and 40 hours; 4 Other includes employees entitled to reduced working hours due to legal issues in the municipality, such as, for example, having children with disabilities; DA= duration of absence; WAI= work ability index; Q1= 1st quartile; Q3= 3rd quartile.

Table 2

Physical and psychosocial conditions (N = 163)

Categorical variables¹	n (%)	Median DA (Q1–Q3)	p value DA	Median WAI (Q1–Q3)	p value WAI
Ventilation	Inadequate	119 (73.0%)	26.5 (13–58.25)	0.606	33 (27–37)	0,001^*
	Adequate	44 (27.0%)	19 (7.5–41)		37 (33–41)
Temperature	Inadequate	126 (78.3%)	27.5 (13.25–58.75)	0.364	33.5 (27–38)	0,010^*
	Adequate	35 (21.7%)	18 (7.5–41)		37 (32.5–41)
Lighting	Inadequate	67 (41.1%)	27 (13–55)	0.747	33 (27–37)	0,155^*
	Adequate	96 (58.9%)	24 (12–59)		35.5 (31–38.75)
Chairs and benches	Inadequate	104 (65.0%)	27 (13–58)	0.355	33 (27–37)	0,001^*
	Adequate	56 (35.0%)	19 (8–50)		37 (33–40)
Technical resources and equipment	Inadequate	108 (67.9%)	26 (13–58.75)	0.346	33 (27–38)	0,176^*
	Adequate	51 (32.1%)	23 (13–58.5)		36 (32–38)
Availability of PPE³	No	13 (8.3%)	53.5 (12–113.25)	0.151^*	30.5 (25–37.25)	0,150^*
	Yes	144 (91.7%)	50 (25–80.4)		35 (30–38)
Use of PPE³	No	16 (10.4%)	34 (14–61.5)	0.109^*	32 (28–39)	0,906
	Yes	138 (89.6%)	25 (12–55.5)		34.5 (30–38)
Relation between the demands from tasks and the resources available	Inadequate	111 (71.2%)	25.5 (12–56.5)	0.636	33 (27–37)	0,014^*
	Adequate	45 (28.8%)	25.5 (13–60.75)		36.5 (32–40.25)
Adoption of attitudes that can generate pain or discomfort	Never/rarely	26 (16.7%)	23 (10–50)	0.059^*	37 (33.5–38.25)	0,082^*
	Sometimes/always	130 (83.3%)	39.5 (22.25–67.75)		34 (29–38)
Standing	Never/rarely²	1 (0.6%)	26 (12.5–57.5)	–	34.5 (30–38)	–
	Sometimes/always	159 (99.4%)	–		–
Sitting	Never/rarely	143 (92.3%)	26 (13–59)	0.773	34 (29–38)	0,245
	Sometimes/always	12 (7.7%)	23 (11–45)		39 (31–41)
Walking	Never/rarely	9 (5.6%)	23 (10.75–37.75)	0.686	38.5 (32.75–44)	0,054^*
	Sometimes/always	152 (94.4%)	26 (13.5–36.5)		34 (29–38)
Lifting, raising, or pushing excessive weight	Never/rarely	21 (13.0%)	23 (9–38)	0.576	37 (30–43)	0,108^*
	Sometimes/always	141 (87.0%)	26 (13–59)		34 (30–38)
Not having breaks	Never/rarely	73 (46.2%)	25 (12–41)	0.306	36.5 (32–41)	0,004^*
	Sometimes/always	85 (53.8%)	26.5 (12.75–63.5)		33 (27–37)
Noise	Never/rarely	92 (57.9%)	26 (13–57.25)	0.841	34 (30–38.25)	0,916
	Sometimes/always	67 (42.1%)	24 (13–57)		34 (28.5–38)
Numerical variables¹			mean (SD)		p value DA	p value WAI
Prescript Organization of Work Scale			3.34 (0.64)		0,010^*	0.003^*
Styles of Management Scale			2.92 (0.38)		0.880	0.863
Pathogenic Suffering at Work Scale			2.13 (0.67)		0.124^*	0.000^*
Psychosocial Damages at Work Scale			2.39 (0.68)		0.017^*	0.000^*

The PRWE scales presented the following values: 3.34 (SD = 0.64) for POWS; 2.92 (SD = 0.38) for SMS; 2.13 (SD = 0.67) for the PSWS; and 2.39 (SD = 0.68) for the PDWS. These values indicate low risk for experiences of suffering related to work; moderate pattern in management styles, individualistic and collectivist; medium risk for harm at work, especially physical harm; and medium risk for work organization, with a weak point for division of tasks, number of workers, sufficient resources and flexibility in deadlines.

WAI had a median of 34 (Q1 = 30; Q3 = 38), a moderate WAI for the group, with scores varying from 19 to 47 (mean = 34.07; SD = 6.21). The total duration of absence due to sickness was 6.170 days, varying from 2 to 222 days per participant. The median duration of absence throughout the 4 years was 26 days (Q1 = 12; Q3 = 57), with a mean of 37.85 (SD = 36.54) days, per participant, during the 4 years. The three main cause for sick leaves were Musculoskeletal System and Connective Tissue Diseases (19.5% of the total sick leaves), followed by Factors that Influence the Health State and the Contact with Health Services (9.5%) and Respiratory Tract Diseases (9.5%). The order between these three main causes of sick leave was similar between 2017 and 2019, but in 2020, during the COVID-19 pandemic, the Factors that Influence Health State and the Contact with Health Services group represented the first cause and included, mainly, leaves for isolation due to contact with patients contaminated by COVID-19. Among the total of sick leaves, 23.7% did not indicate their cause. Tables 1 2 presented the results of the inferential bivariate analysis of the personal characteristics and work conditions with the WAI and the duration of absence.

The independent variables that showed a simultaneous significant association (p < 0.20) with WAI and duration of absence were: family income, health self-evaluation, Body Mass Index (BMI), PPE availability, assuming postures that can generate pain or discomfort, and the PPRWE scales “organization of work”, “pathogenic suffering”, and “damages at work”. These variables were submitted to multiple linear regression modeling to explain the duration of absence. However, the model found was not significant (p = 0.47). Through the testing of the algorithm “best subsets regression”, the new model included, grouped, six variables predictive of the duration of absence: POWS, PSWS, PDWS, health self-evaluation, availability of PPEs, and assuming postures that can generate pain or discomfort at work.

3.1 Mediating effect of WAI

Figure 2 shows the results of the analysis of the mediating factor (path analysis). The regression model that describes the relation between the predictor variables and the duration of absence (path c, step 1), explained 33% of the duration of absence (R² =0.3331; p < 0.05). The same set of predictor variables explained 53% (R² =0.5301; p < 0.05) of the WAI (path a, step 2). Then a simple linear regression model was carried out between the mediator variable (WAI) and the outcome variable (duration of absence) (step 3, path b), indicating that the WAI explained 9% of the duration of absence (R² =0.093; p < 0.05). In the fourth step (step 4a) of the path analysis (path c’), it was possible to verify that adding WAI to the set of predictor variables weakened the explanatory power for duration of absence. Comparing the adjusted R², there was a reduction from 24% to 20%, which indicates partial mediation of WAI over duration of absence. To certify the hypothesis of partial mediation of work ability with regard to the duration of absence, the set of variables (POWS, PSWS, PDWS, self-evaluation of health, availability of PPE, assuming postures that can generate pain or discomfort, and WAI) was treated using the algorithm stepwise forward (step 4b). This analysis indicated that the estimated coefficient of WAI, together with the variables availability of PPE and assuming postures that can generate pain or discomfort at work is significantly different from zero (p < 0.01), explaining nearly 18% of the duration of absence (R² =0.177; p < 0.01).

Fig. 2

Model of the mediating effect of work ability in the relation between personal and work variables with the duration of sickness absence from work.

Table 3 shows the steps of the mediating analysis with all standardized coefficients and significance levels. It should be noted that, when WAI is input (step 4a) into the initial model, formed by the set of independent and predictor variables for the duration of absence, there is a reduction in the size of the effect (adjusted R²), provoked by the variables of the model. Consequently, there is a diminution in the significance level of each predictor variable (increased p value), proving that the presence of WAI can weaken this relation.

Table 3

Standardized coefficients and their significance level (in parenthesis) in the steps of the path analysis

	Step 1c	Step 2a	Step 3b	Step 4a c’	Step 4b c’
IV/DV	DA	WAI	DA	DA	DA
Prescript Organization of Work Scale	–0.45 (0.004)*	0.22 (0.084)	–	–0.45 (0.007)*	–
Pathogenic Suffering at Work Scale	–0.51 (0.022)*	0.09 (0.615)	–	–0.47 (0.043)*	–
Psychosocial Damages at Work Scale	0.44 (0.046)*	–0.41 (0.034)*	–	0.33 (0.167)	–
Health self-evaluation	0.10 (0.494)	–0.39 (0.003)*	–	0.10 (0.545)	–
Availability of PPE	0.18 (0.176)	–0.10 (0.374)	–	0.18 (0.195)	0.14 (0.09)*
Postures pain/discomfort	–0.22 (0.103)	–0.09 (0.407)	–	–0.17 (0.220)	–0.22 (0.01)*
WAI	–	–	–0.30 (0.000)	–0.06 (0.759)	–0.35 (0.00)*
Power	99.8%	99.9%	94.7%	99.5%	88%
R ²	0.33*	0.53*	0.09*	0.32*	0.18*
R² adjusted	0.24*			0.20 *

^*p < 0.05; IV = independent variable; DV = dependent variable; DA = duration of absence; WAI = work ability index; PPE = personal protective equipment.

4 Discussion

Occupational health indicators, such as the WAI and absenteeism due to illness, are associated with individual and contextual factors of school staff. The results of the proposed mediation model showed that personal characteristics associated with worse self-rated health and precarious working conditions can reduce work ability and predict absences from work. Specifically, precarious working conditions were the adoption of uncomfortable postures, unavailability of safety equipment, work organization flaws, mental exhaustion and the occurrence of physical harm. The mediation of the WAI suggests that the proposed methodological course represents an advance in understanding the predictors of sick leave among school servants, especially when working conditions are analyzed.

The demographic and occupational characterization in the results of this study were similar to those found by workers in the fields of cooking, cleaning, and education [7 , 39]. Most workers were middle-aged women with low incomes, who had children and lived with a partner. These characteristics are historically and culturally associated with the disqualification of the work of females and to domestic overload, with the insertion of women in less valued economic activities, with lower pay or in poorer conditions [7, 40]. In the field of education, this phenomenon is clear [7 , 40]. As opposed to other researches [11, 39], the school servants evaluated had a higher educational level. This could be associated to financial incentives towards the career progression of Brazilian public workers.

Their life habits corroborate national findings from researches on the health of the Brazilian population, with a high percentage of sedentary lifestyles, overweight, and obesity [41]. The BMI values were associated with the WAI and the absenteeism. These results are in accordance with findings about the relation between lifestyle and absenteeism at work, especially related to musculoskeletal causes in women [42]. Interventions aimed at promoting healthy lifestyle habits in school staff are described in the literature [12], mainly strategies focused on improving nutritional knowledge and reducing sedentary lifestyle [43]. However, the current literature focuses primarily on assessment of teachers and have included a small number of schools and staff [12]. In this sense, the promotion of healthy habits becomes necessary and should be structured as a strategy in occupational health programs aimed at all education workers, including school servants.

Regarding work conditions, a recent systematic review indicated that physical factors, such as physical environment, burden and psychosocial factors, like decision making and social support, are important mediators of socioeconomic inequalities in the health of workers [6]. In our study, the physical characteristics of the work of school servants were classified as poor, confirming the association between work conditions and risks for occupational diseases. Furthermore, the participants showed a medium psychosocial risk in dimensions of the organization at work, such as number of workers and resources and flexible deadlines. A recent systematic review and meta-analysis found that persistent precarious employment to be associated with poorer health for a number of outcomes [44]. Our results show that structural and organizational interventions at work aiming the reduction of psychosocial risks and promoting ergonomic improvements can be important actions to restore work ability and avoid absenteeism among school servants. These interventions should focus on minimize physical demands, such as inappropriate postures; adequacy in the number of workers and distribution of tasks; and greater offer of PPE’s. These proposals are expected to avoid, mainly, physical harm, mental exhaustion and sick leave among school servants.

Musculoskeletal disorders were the main cause for sick leaves in school servants. This is similar to the prevalence of diseases in workers with similar activities [5, 45]. Groups with more vulnerability at work tend to present a higher prevalence of these disorders, be it due to individual characteristics or to a greater exposure to excessive physical and psychosocial burdens at work [5 , 47]. Similarly, the evaluation of the psychosocial factors indicated the presence of physical damage and mental exhaustion in workers and the evaluation of the physical conditions at work showed that most participants (83.3%) show pain or muscle discomfort. Accordingly, an ergonomic study carried out in a unit for school nutrition indicated that the work activities demand moderate effort, with discomfort and pain in the legs [48].

The mean duration of absence of the school servants was higher than that of public municipal servers from the West of Brazil [49], while lower than the duration of absence in the South [50] and then the mean national data [47]. The WAI is a widely used measure in services of health and occupational safety, since it is related to higher risks for future prolonged, frequent, or permanent absenteeism [19, 51]. In this study, the mean WAI of the group was moderate, explaining 9% of the duration of absence of the servants.

There is extensive scientific literature about the need to preserve the WAI of individuals, aiming to guarantee a better functionality of workers, with wellbeing and productivity [23]. An important finding of this research was the fact it understands which characteristics are related with the WAI and with the duration of absences due to disease. The model constructed to predict the duration of absence of the servants included a combination of aspects of work organization, mental exhaustion, physical damage, health self-evaluation, availability of PPE, and postures that can generate pain or discomfort. This result shows that the main contribution to the absenteeism of school servants is due to their working conditions. In agreement with these findings, a study with health professionals showed a greater association of sick leaves with psychosocial factors at work than with personal aspects [22]. Another research with university workers found that the duration of absence is predicted better by a combination of work and personal factors [52].

Since studies about mediation are scarce in this field, the advantage of this research was finding that the WAI has a partial mediating effect on the relation between personal characteristics and work conditions of school servants with the measures of duration of absence due to disease. The planning of actions that seek more balance between work demands and individual factors is needed to promote WAI and potentially reduce sick leaves in school servants. Directive interventions in health and work safety, aimed at minimizing physical damage and mental exhaustion, especially related with work organization, PPE availability, and minimization of situations that provoke pain or discomfort at work, are a potential strategy.

Limitations of this study include its cross-sectional design, which limits causal inferences; the memory bias associated with the use of self-reported measures; and the effect of the healthy worker. Two reasons for this were the maximum duration of each absence period being limited to 15 days in the database on absenteeism, and the fact that the research was carried out during a pandemic, moment in which many school servants with comorbidities were not present for in-person activities. The count of days off work in 2020 also considered the days of preventive isolation related to contact with people contaminated by COVID-19, not representing an illness of the worker. Despite this limitation, this research evaluated 84.5% of the total population of school servants in the municipality, active at work. The sample size of 163 school servants can be considered small when compared to other occupational groups, however, this sample size is high when analyzing the total percentage of the studied population. This fact is even more relevant when considering the precarious working conditions, the invisibility of this worker in society and the scarcity of studies with school servants, an important group of workers who support educational services.

This study gains relevant as it analyzes the mediating effect of WAI on working conditions and the sick leaves in an occupational group that has seldom been addressed by national and international literature. The sample power of the study was above 80% and the robustness of the statistical method used can be attributed both to the use of the best subset found using the regression algorithm and its submission to the recognized path analysis method [36, 37]. This analysis made it possible, thus, to evaluate complex constructs, such as the work ability, and test their association with the duration of absence and predictor variables. Although this study considers the influence of the macrosocial context on work ability and on the duration of absence of school servants, aspects such as the family and sociocultural context, work and welfare legislation were not addressed, and should be explored further in later studies.

5 Conclusion

The results indicated that the negative health perception, combined with poor work conditions, such as organizational problems, unavailability of PPE, the need to assume postures that generate pain or discomfort, mental exhaustion, and physical damage, responded for 33% of the duration of absence of the school servants analyzed. The WAI was identified as a partial mediating variable in this relation, capable of reducing in 4% the strength of the association between personal characteristics and work conditions with the time on sick leaves.

Sick leave due to illness at work reflects the working and health conditions of the population and brings social and financial losses arising from the loss of productivity, lack of provision of public services and additional overloads on the work team. In a practical way, the proposed model indicates the need to reduce the exposure of school servants to physical and psychosocial risks at work. Thus, we suggest ergonomic adaptations to minimize physical demands, such as inappropriate postures; adequacy in the number of workers, in the distribution of tasks and greater offer of PPE’s. Furthermore, it is suggested the adoption of health promotion programs that focus on disease prevention, especially musculoskeletal disorders and mental exhaustion.

These findings highlight the importance of developing actions to promote WAI and lead to better work conditions. These actions aim to reduce absenteeism and the economic, functional, and social impact of work leaves to workers, organizations, and governments.

Ethics statement

The study was approved by the Research Ethics Committee of the Federal University of Triângulo Mineiro on 27 October 2020 (CAAE: 36566420.0.0000.5154, opinion 4.364.500). The procedures followed were in accordance with the ethical standards of institutional and national bodies and consistent with the revised Declaration of Helsinki. Participants were informed about the goals of the research, the study’s risks and benefits, that their participation was voluntary, and that their identity would not be disclosed. Those who accepted to participate signed the Free and Informed Consent Form. The authors confirm that confidentiality and data security was appropriately handled.

Conflicts of Interest

The authors declare that there are no conflicts of interest. The authors are the sole responsible parties for the content and writing of the article.

Footnotes

Acknowledgments

The authors would like to thank all school servants from Nova Lima, Brazil, who contributed to this study.

Funding

The authors report no funding.

References

Schulte

, Streit

JMK

, Fatima

, Delclos

, Felknor

, Tamers

, et al. Potential Scenarios and Hazards in the Work of the Future: A Systematic Review of the Peer-Reviewed and Grey Literatures. Ann Work Expo Health. 2020;64(8):786–816. DOI: 10.1093/annweh/wxaa051.

Nazarov

, Manuwald

, Leonardi

, Silvaggi

, Foucaud

, Lamore

, et al. Chronic Diseases and Employment: Which Interventions Support the Maintenance of Work and Return to Work among Workers with Chronic Illnesses? A Systematic Review. Int J Environ Res Public Health. 2019;16(10):1864. DOI: 10.3390/ijerph16101864.

Acompanhamento Mensal do Benefício Auxílio-Doença Previdenciário Concedido Segundo os Códigos da CID-10 –Janeiro a Dezembro de 2019. Brasília: Ministério da Economia, 2019. http://sa,previdencia.gov.br/site//03/Acompanhamento-Mensal_Auxilio-Doenca-Previdenciario__completo_CID-10.pdf

Santa-Marinha

, Teixeira

, Maciel

EMGS

, Moreira

MFR

. Epidemiological profile of sickness absenteeism at Oswaldo Cruz Foundation from through 2016. Rev Bras Med Trab. 2020;16(4):457–65. DOI: 10.5327/Z1679443520180307.

Santi

, Barbieri

, Cheade

MDFM

. Sickness absenteeism within the Brazilian public service: integrative literature review. Rev Bras Med Trab. 2018;16(1):71–81. DOI: 10.5327/Z1679443520180084.

Dieker

ACM

, Ijzelenberg

, Proper

, Burdorf

, Ket

JCF

, van der Beek

, et al. The contribution of work and lifestyle factors to socioeconomic inequalities in self-rated health –a systematic review. Scand J Work Environ Health. 2019;45(2):114–25. DOI: 10.5271/sjweh.3772.

Nunes

The meaning of work to school lunch cooks and servents in readaptation situation in Rio de Janeiro public schools [Dissertação]. Rio de Janeiro: Escola Nacional de Saúde Pública Sergio Arouca, 2000. https://www.arca.fiocruz.br/handle/icict/4817.

Cardillo

, Gemma

SFB

, Fuentes-Rojas

. An interdisciplinary look at the work of outsourced school cooks from state schools in the city of Campinas, state of São Paulo, Brazil. Rev Bras Saúde Ocup. 2021;46:e3. 10.1590/2317-6369000004819.

, Park

, Jeong

. Risk Assessment of Educational Support Staff’s Work in Schools. Work. 2021;69(3):945–55. DOI: 10.3233/WOR-213526.

10.

Park

, Lee

. Insufficient Rest Breaks at Workplace and Musculoskeletal Disorders Among Korean Kitchen Workers. Saf Health Work. 2021;12(2):225–9. DOI: 10.1016/j.shaw.2021.01.012.

11.

Rocha

MRAR

, Marin

MJS

, Macias-Seda

. Living, working conditions and mental health: a study with Brazilian and Spanish workers who work in hospital cleaning services. Cien Saúde Colet. 2020;25(10):3821–32. DOI: 10.1590/1413-812320202510.35912018.

12.

Hill

, Alston

, Needham

, Peeters

, LaMontagne

, Nichols

. Systematic review of the effectiveness of health promotion interventions targeting obesity prevention in school-based staff. Health Promot Int. 2022;37(3):daac061. DOI: 10.1093/heapro/daac061.

13.

Assunção

, Abreu

MNS

. Pressure to work, health status, and work conditions of schoolteachers in Basic Education in Brazil. Cad Saude Publica. 2019;35(Suppl. 1):e00169517. DOI: 10.1590/0102-311X00169517.

14.

Cadiz

, Brady

, Rineer

, Truxillo

. A Review and Synthesis of the Work Ability Literature. Work, Aging and Retirement. 2019;5(1):114–38. DOI: 10.1093/workar/way010.

15.

Gould

, Ilmarinen

, Järvisalo

, Koskinen

Dimensions of work ability: Results of the Health 2000. Helsinki: Finnish Institute of Occupational Health. 2008. https://sivusto.kykyviisari.fi/wp-content/uploads//09/Gould-Dimensions-of-work-ability.pdf

16.

Converso

, Sottimano

, Guidetti

, Loera

, Cortini

, Viotti

. Aging and Work Ability: The Moderating Role of Job and Personal Resources. Front. Psychol. 2018;8:2262. DOI: 10.3389/fpsyg.2017.02262.

17.

Cotrim

, Ribeiro

, Teles

, Reis

, Guerreiro

, Janicas

, et al. Monitoring Work Ability Index During a Two-Year Period Among Portuguese Municipality Workers. Int J Environ Res Public Health. 2019;16(19):3674. DOI: 10.3390/ijerph16193674.

18.

Rothmore

, Gray

. Using the Work Ability Index to identify workplace hazards. Work. 2019;62(2):251–9. DOI: 10.3233/WOR-192860.

19.

Palmlöf

, Skillgate

, Talbäck

, Josephson

, Vingård

, Holm

. Poor work ability increases sickness absence over 10 years. Occup Med [Lond]. 2019;69(5):359–65. DOI: 10.1093/ocmed/kqz083.

20.

Hosseininejad

, Ghafaari

, Labbafinejad

, Mirzamohammadi

, Moslemi

, Mohtasham

, et al. Influence of short term and long-term sickness absence on work ability index. Iran Occupational Health. 2019;16(3):1–12.

21.

Notenbomer

, van Rhenen

, Groothoff

, Roelen

CAM

. Predicting long-term sickness absence among employees with frequent sickness absence. Int Arch Occup Environ Health. 2019;92(4):501–11. DOI: 10.1007/s00420-018-1384-6.

22.

van den Berg

, Burdorf

, Robroek

SJW

. Associations between common diseases and work ability and sick leave among health care workers. Int Arch Occup Environ Health. 2017;90(7):685–93. DOI: 10.1007/s00420-017-1231-1.

23.

Brady

, Truxillo

, Cadiz

, Rineer

, Caughlin

, Bodner

. Opening the black box: Examining the nomological network of work ability and its role in organizational research. J Appl Psychol. 2020;105(6):637–70. DOI: 10.1037/apl0000454.

24.

Leino-Arjas

, Seitsamo

, Nyård

, K

C P

, Neupane

. Process of work disability: from determinants of sickness absence trajectories to disability retirement in a long-term follow-up of municipal employees. Int J Environ Res Public Health. 2021;18(5):2614. DOI: 10.3390/ijerph18052614.

25.

Kartika

, Widhiandono

The Mediation Effect ofWork Ability on The Relationship between Job Training, Work Experience, and Job Performance. Proceedings of the 3rd ICBAE. 2022;1-7. DOI: 10.4108/eai.10-8-2022.2320796.

26.

von Bonsdorff

, Zhou

, Wang

, Vanhala

, von Bonsdorff

, Rantanen

. Employee Age and Company Performance: An Integrated Model of Aging and Human Resource Management Practices. JOM. 2018;44(8):3124–50. DOI: 10.1177/0149206316662314.

27.

Vandenbroucke

, von

Elm E

, Altman

, Gøtzsche

, Mulrow

, Pocock

, et al. STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4(10):e297. DOI: 10.1371/journal.pmed.0040297.

28.

Liss

, Uchida

, Wilkes

, Radakrishnan

, Linder

. General Health Checks in Adult Primary Care: A Review. JAMA. 2021;325(22):2294–306. DOI: 10.1001/jama.2021.6524.

29.

Barbosa

REC

, Assunção

, Araújo

. Musculoskeletal disorders among healthcare workers in Belo Horizonte, Minas Gerais State, Brazil. Cad. Saúde Pública. 2012;28(8):1569–80. DOI: 10.1590/S0102-311X2012000800015.

30.

Facas

PROART: Riscos Psicossociais Relacionados ao Trabalho. Porto Alegre/RS: Editora 858 Fi, 2021. http://biblioteca.cofen.gov.br/wp-content/uploads/2022/02/riscos-psicossociais-relacionados-trabalho.pdf

31.

Neto

GAT

Validação exploratória portuguesa do protocolo de validação dos riscos psicossociais no trabalho em trabalhadores do setor do turismo, [Dissertação de Mestrado]. Universidade Portucalense, Portugal. 2017. https://hdl-handle-net.web.bisu.edu.cn/11328/2058

32.

Matamala

, Barrera

Desgaste y Riesgos Psicosociales en profesionales de la salud durante la pandemia en Chile, [Tesis de Doctorado]. Pontificia Universidad Católica de Valparaíso, Chile. 2020. DOI: 10.13140/RG.2.2.31942.70721.

33.

Martinez

, Latorre

MRDO

, Fischer

. Validity and reliability of the Brazilian version of the Work Ability Index questionnaire. Rev Saúde Pública. 2009;43(3):e6525–32. DOI: 10.1590/S0034-89102009005000017.

34.

Torres

, Silva da

. The Brazilian Integrated Health Care Subsystem for Civil Servants (SIASS) from the perspective of federal government employees from Higher Education Institutions. Rev Bras Saúde Ocup. 2022;47:e6. DOI: 10.1590/2317-6369/36220PT2022v47e6.

35.

Brasil. Decreto n° 3048/99, de 06 de maio de 1999, aprova o Regulamento da Previdência Social e dá outras providências. Braíilia: DOU, 1999.

36.

Lee

, Cashin

, Lamb

, Hopewell

, Vansteelandt

, VanderWeele

, AGReMA group,et al. A guideline for reporting mediation analyses of randomized trials and observational studies: The AGReMA Statement. JAMA. 2021;326(11):1045–56. DOI: 10.1001/jama.2021.14075.

37.

Baron

, Kenny

. The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–82. DOI: 10.1037//0022-3514.51.6.1173.

38.

Rodrigues

HSJ

, Izquierdo

. The value of work and the value of women’s work. Revista Ártemis. 2017;23(1):211–23. DOI: 10.22478/ufpb.1807-8214.2017v23n1.35800.

39.

Gemma

, Fuentes-Rojas

, Soares

. Outsourced cleaning workers: between resentment and the recognition. Rev Bras Saúde Ocup. 2017;42:e4. 10.1590/2317-6369000006016.

40.

Baltar

, Omizzolo

. Women’s participation in the Brazilian labour market from 2014 to 2019. TEC. 2020;23(1):1–17. DOI: 10.5007/2175-8085.2020.e71522.

41.

IBGE Pesquisa Nacional da Saúde Antropometria, Rio de Janeiro: IBGE; 2019. https://sidra.ibge.gov.br/pesquisa/pns/pns-2019

42.

Svärd

, Lahti

, Mänty

, Roos

, Rahkonen

, Lahelma

, et al. Weight change among normal weight, overweight and obese employees and subsequent diagnosis-specific sickness absence: A register-linked follow-up study. Scand J Public Health. 2020;48(2):155–63. DOI: 10.1177/1403494818802990.

43.

Kupolati

, MacIntyre

, Gericke

, Becker

. A Contextual Nutrition Education Program Improves Nutrition Knowledge and Attitudes of South African Teachers and Learners. Front Public Health.258. 2019;7–10.3389/fpubh.2019.00258.

44.

Pulford

, Thapa

, Thomson

, Guilding

, Green

, Leyland

, Popham

, Katikireddi

. Does persistent precarious employment affect health outcomes among working age adults? A systematic review and meta-analysis. J Epidemiol Community Health. 2022;76(11):909–17. DOI: 10.1136/jech-2022-219292.

45.

Valença

JBM

, Alencar

MCB

. Working aspects and musculoskeletal disorders of workers in a school’s lactary. Cad Ter Ocup UFSCar. 2017;25(1):137–46. DOI: 10.4322/0104-4931.ctoAO0787.

46.

Fantini

AJE

, Assunção

AÁ

, Machado

. Musculoskeletal pain and occupational vulnerability in municipal public sector workers in Belo Horizonte, Brazil. Ciênc Saúde Colet. 2014;19(12):4727–38. DOI: 10.1590/1413-812320141912.02872013.

47.

Haeffner

, Kalinke

, Felli

VEA

, Mantovani

, Consonni

, Sarquis

LMM

. Absenteeism due to musculoskeletal disorders in Brazilian workers: thousands days missed at work. Rev Bras Epidemiol. 2018;21:e180003. DOI: 10.1590/1980-549720180003.

48.

Sandoval

LCN

, Spinelli

MGN

. Ergonomic conditions of school food and nutrition units under the same management. Nutrivisa. 2017;3(3):155–60.

49.

Leão

ALM

, Barbosa-Branco

, Neto

, Ribeiro

CAN

, Turchi

. Sickness absence in a municipal public service of Goiânia, Brazil. Rev Bras Epidemiol. 2015;18(1):262–77. DOI: 10.1590/1980-5497201500010020.

50.

Daniel

, Koerich

CRC

, Lang

. The absenteeism profile among municipal public servants of the city of Curitiba, from 2010 to 2015. Rev Bras Med Trab. 2017;15(2):142–9. DOI: 10.5327/Z1679443520176021.

51.

Klein

, Reini

, Saarela

. Sickness absence and disability pension in the very long term: A Finnish register-based study with 20 years follow-up 2021;9(142):556648. DOI: 10.3389/fpubh.2021.556648.

52.

Schouteten

. Predicting absenteeism: screening for work ability or burnout. Occup Med [London]. 2017;67(1):52–7. DOI: 10.1093/occmed/kqw161.