Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

According to the World Health Organization (2007), both globalization and recent changes in the nature of work are probably worsening work related stress issues in Latin America. When this WHO document was published in 2007, very few studies on work stress in Latin-American countries had been carried out and no mandatory rules or risk standards had been implemented to promote good practices at the workplace against psychosocial risk exposure. However, a recent study showed that occupational health and safety priorities in the region have changed during the last decade pointing to the need to monitor psychosocial hazards and to address work-related stress, violence, harassment, unhealthy behaviors, and other workplace hazards (Kortume and Leka 2014).

The Pan American Health Organization (2015) asserts that the labor force in the Americas is made up of 484 million people, almost half of the total population. The organization also affirms that Latin America and the Caribbean contribute 62.3 % of the labor force in the region, equivalent to more than 300 million people many of them “exposed to dangerous work conditions ranging from exposure to chemical agents, to physical and biological dangers, ergonomic and psychological stressors, and unsafe conditions” (pp. 1).

In the context of economic globalization, Effort-Reward Imbalance (ERI) at work has been used as an indicator of psychosocial work-related stress in many countries around the world. Research using the ERI model and questionnaire in Latin America could help us understand how economic globalization and the changing working conditions are affecting workers’ health in this part of the world. It could also help clarify whether economic, social and cultural characteristics of the region modify the impact of ERI on occupational health.

To contextualize the description and analysis of studies undertaken in Latin America that look into some health conditions and their relationship to ERI, this chapter first introduces a brief description of the work conditions to which workers in this subcontinent have been subject to over the last 8–10 years. We have considered it pertinent to highlight data that may serve as indirect indicators of effort and rewards. In addition, we consider it is interesting to offer a brief overview of health conditions in the sub region in order to allow readers to observe how such conditions are beginning to acquire a similar pattern to the one observed in developed countries, whereby chronic and stress-related diseases have been documented and how they are becoming increasingly associated with morbidity and mortality in the working-age population.

According to the above, the goals of this chapter are to:

  1. (a)

    Provide a short general characterization of the labor market and working conditions in Latin America (in particular those related to the concepts of effort and reward), of existing legislation and regulations associated with psychosocial risk and stress, supplemented by some core data on health and wellbeing of people of working age.

  2. (b)

    Undertake a systematic review and characterization of the studies published with samples from Latin America, which have assessed work stress using the ERI model. In addition to describing the prevalence of ERI as reported in the papers, and its association with the health of Latin American workers, in this chapter we undertake a critical analysis of published studies in order to identify their strengths and weaknesses, and offer suggestions for future research on the topic, in particular as far as based on the Effort-Reward Imbalance model.

2 Working Conditions and Employment in Latin America

Even though the information on the particular working conditions in this region is scarce, its link to the economic situation offers an approach to understanding such conditions. In this sense, the history of the changes of economic indicators in Latin America (rises and falls) is becoming increasingly similar to the patterns seen in the global economy. This is due to the fusion of the global financial markets and growing globalization in general, even though it is clear that indicators of economic growth have been lower than those of industrialized economies, especially since the 1980s. Weller (2011) mentions that during the first decade of 2000 and especially between 2003 and 2004 strong economic growth was reported that seemed to benefit the region, opening a path to a more favorable context for job creation and better job quality, at the same time as it tried to re-regulate the labor market. However, the financial crisis of 2008–2009 interrupted this process of improvement, causing a decline in the indicators, whereby some of them, such as the unemployment rate, went back to levels similar to those present in the 1990s.

Weller (2011) describes four types of employment exclusion in the region: during the 1980s, exclusion from the labor market (e.g., lack of female incorporation) and from productive employment (lack of jobs in highly productive sectors) prevails, thus increasing urban informal employment; beginning in the 1990s and up until now, exclusion from employment (e.g., typical unemployment) and from good quality employment (e.g., precarious work) has expanded.

In agreement with these ideas, a report from the International Labor Organization (ILO) in 2012 for the Latin American region (International Labor Organization 2012) concluded that despite the apparent economic growth, there are still important shortcomings in available decent work as a result of precarious employment, including income insecurity, a decline in social protection, and high labor turnover. More recently, the report on work in Latin America and the Caribbean (International Labor Organization 2014) describes that the situation is marked by a deceleration of economic growth, where uncertainty and concern prevail. More vulnerable groups have worse working conditions. Indeed, the unemployment rate is 30 % higher for women, and youths aged between 15 and 24 face unemployment rates that are two to four times greater than those of adults. The urban unemployment rate reaches an annual average of 6.1 % for Latin America and the Caribbean, representing 15 million people unemployed, only counting urban areas. It is important to note that Guatemala, Panama and Brazil have the lowest unemployment rates, whereas the Bahamas, Jamaica and Belize have the highest, with no great changes over the past few years (Fig. 11.1).

Fig. 11.1
figure 1

Percentage of urban unemployment in Latin American countries between January 2013 and September 2014 (Source: ILO 2014)

Full size image

The report of the ILO (2014) recognizes that global crises affect this region in a particular manner, and that there are problems beyond that of unemployment. These problems include that of employment quality, when we consider that there are 130 million people working in informal employment, which generally implies precarious working conditions with no kind of social protection. As is well known, the average rate of informal workers in the whole region is of 47.7 %, although there seem to be substantial differences between countries. For example, countries such as Honduras, Bolivia or Peru approach a 70 % rate, whereas Chile, Venezuela and Argentina have between 40 % and 50 %: all highly alarming rates (International Labor Organization 2012).

In the report about Decent Work in the Americas (ILO 2006), the ILO points out that the labor market structure in Latin America is typically fragmented. For example, almost a third of the total labor market is found in the rural areas, whereas 50 % of the total is taken up by independent and domestic workers, unpaid family workers, or salaried workers in micro companies (with up to five staff) who tend to represent the greatest concentration of poverty, informality and the lack of decent work.

On the other hand, although informal work seems to be a particular factor of precariousness that characterizes the labor force in Latin America, formal work also has a number of very particular characteristics if we analyze the productive models and work organization systems typical to the region. For example, whereas developed countries have gradually abandoned the inherent forms of deeply entrenched hierarchical and rigid division of labor, known as Taylorism, and mass production has been transformed by means of automation and technological and communications advances into hugely flexible models of labor known as Toyotism, Total Quality Management, etc., in Latin America work organization systems continue to follow the old style Taylorist-Fordist tradition. They do, however, coexist with practices pertaining to labor flexibility and Toyotism models of the developed world which cannot fully materialize due to the technological lag in Latin America. This fosters types of particular mixed productive experiences that determine employment quality in the region (Rodríguez and Mendoza 2007), a combination that also establishes a mixture of stressful demands within each system, leading to psychosocial exposures that may be unique to Latin America and require further study.

Evidently, the employment quality indicators link up directly to the concepts pertaining to the effort/reward model, and are reflected in issues such as work overtime, long work hours, salaries, and job insecurity, among others. There are some data available that provide a complementary picture of such working conditions in Latin American countries. For example, according to data from the Organization for Economic Cooperation and Development (2015), the three Latin American countries that are members of this organization have the longest working days of all. Mexico, in first place with 2228 h; Costa Rica in second place, with 2216 h; and Chile in fourth place, with 1990 h, whereas respective numbers in Western modern economies are substantially lower (e.g. USA 1789, Japan 1729, Sweden 1609, or Germany 1371 h). Similarly, a global survey on the percentage of workers that work overtime shows Peru as top of the list, with more than 50 % of its labor force working 48 h of overtime a week, whereas Argentina and Mexico present around 30 % (International Labor Organization 2007).

The above demonstrates not only that it is no longer the Asians that work the longest hours (please note that not all Latin American countries are included in the OECD’s comparative studies), but also that globalization has encouraged developed countries to export work processes with particular demands to Latin America. For example, since 2010, Mexico occupies first place in terms of average hours worked per year, at the same time as the Mexican government highlights a significant economic growth in the industry (Expansión 2013).

With respect to wages, the trend is exactly the opposite of that for hours worked. The comparative study of salaries in countries that have been members of the OECD since 2013 shows that Mexico and Chile pay the lowest hourly rate (OECD 2015). As for 2013 hourly minimal wages in US Dollars are 0.62 and 2.32 respectively, compared e.g. to 3.02 in Turkey, 4.93 in Spain or up to 15.61 in Australia. This trend has come about in recent years and is consistent with other international surveys in which more Latin American countries take part. For example, a global salaries survey (adjusted by parity and purchasing power) applied by the International Labor Organization in 72 countries, shows that a number of Latin American countries, in particular Argentina, Chile, Brazil, Colombia and Mexico, pay wages that are located far below the median of the distribution of average global wage across the countries which is 1480 US Dollars per month (BBC 2012).

As for fundamental labor rights in Latin America, there are huge uncovered gaps in terms of freedom of union association and collective negotiation, forced labor, child labor and discrimination. For example, the region is host, according to the ILO, to 5.7 % of the global cases of complaint over freedom of union association, 29 % of the cases regarding layoffs against freedom of union association, 10 % of forced labor, 5.1 % of global child labor, and higher rates of gender or ethnic discrimination (ILO 2006).

Finally, although the global numbers of work accidents have diminished in general, there are only two exceptions: China and Latin America, where the number of fatal work accidents increased between 1998 and 2001. Workers with less protection tend to work in micro-enterprises, the informal economy and sectors such as agriculture, fishing, mining and construction (ILO 2006). In 2006 7.6 million work accidents were reported in the region, representing an area of opportunity for improvement with regards to working conditions in Latin America and the Caribbean (see Fig. 11.2).

Fig. 11.2
figure 2

Evolution of fatal work accidents in the world in the 1998–2001 period (percentage variation) (Source: ILO 2006)

Full size image

In accordance with the above, the conclusions of a report recently published by the Organizacion Panamericana de la Salud-OPS (Pan American Health Organization-PAHO) point out that the conditions of employment (salary, unemployment, social protection, among others) and work (dangers and risks in the workplace)—both considered social determinants of health—have been transformed over the past 15 years due to the changes that have taken place in the world of work, increasing the multiple inequities in terms of workers’ health in Latin America (Organización Panamericana de la Salud [OPS], 2013).

3 Progress in Legislation for Psychosocial Risk and Stress in Latin America

Despite the aforementioned areas of opportunity, all the mandatory labor laws or regulations in Latin America indubitably include the right to dignified work (which is free of health risks). Moreover, over the past 10 years, much progress has been observed in the regulations and legislation in the specific topic of the psychosocial risk of stress and the organization of work. The growing body of research on the topic of psychosocial risk in the international arena, together with the research undertaken in the Latin America region, as well as the changes in the policies and recommendations made by worldwide organizations such as the International Labor Organization (ILO) and the World Health Organization (WHO), may have had important influences on this progress.

Basically, the regulations have evolved in two ways. On the one hand, there are those regarding health and safety hazards, which as well as physical and chemical risks, and which finally also include psychosocial risks in some cases. On the other hand, independent regulations now refer to the risks associated with psychological harassment at work (bullying). In particular, the latter has enjoyed the greatest legislative progress, not just because there seem to be more laws on the topic having been implemented in more countries, but also because the criteria for determining such harassment and its prevention seem increasingly more widely disseminated. For example, there are countries which have no legislation concerning psychosocial risk in general, but where they do have laws on harassment at work (Table 11.1).

Table 11.1 Laws, articles and decrees related to psychosocial factors in Latin America
Full size table

Insofar as the regulations identified, these have a marked focus towards prevention and vigilance against psychological risk factors at work. Also the most common risks that appear among these regulations are the different types of job demands, control, participation or autonomy at work, interpersonal relationships among workmates and supervisors, leadership, organizational justice, rewards or recognition, work-family balance, and job security, among others. The theoretical base to which some of these regulations allude (for example, in Colombia and Mexico), point directly to the demand/control model or the effort/reward imbalance model as its conceptual base, although dimensions from other theoretical models and some that are not part of any of the models are also incorporated.

4 Health and Well-being in the Latin-American Region

Given that the relationship between income inequality and psychiatric disorders has been documented in both developed and developing countries (Wilkinson and Pickett 2006; Patel and Kleinman 2003), to describe the general health conditions in the Latin American region, we consider it pertinent to offer data on psychiatric/mental disorders. It is also important to describe not just some data on these disorders among the Latin American population, but also some of the other health indicators that could be related to the current labor market and its new stressful demands such as: cardiovascular diseases, other non-communicable diseases (chronic degenerative conditions), as well as mortality in general (Kivimäki et al. 2012; Landsbergis et al. 2013; Leka and Jain 2010; WHO 2003).

To begin with, it is worth pointing out that even though Latin America is not the world’s poorest region, it is one of the least equitable, most unjust, and it has the most unequal distribution of wealth, which clearly leads to a bad “distribution” of health (CEPAL 2015). Using the historic register of the “Gini index” as an indicator of global inequality, some analyses point out that with the exception of some African countries and China, in general, the American continent seems to be the most unequal, especially Latin American countries and the United States (Hillebrand 2009; Ginni Coefficient 2015; Quandl 2015). Insofar as mental health, the World Mental Health (WMH) Survey Initiative, by the World Health Organization, reports studies in which highly unequal countries as measured by the GINI index such as Colombia, Mexico, and the United States show a prevalence of mental disorders that is among the highest of the 14 countries studied in 4 continents (17.8 %, 12.2 % and 26.3 % respectively). However, it is interesting to point out that some less unequal countries like France and Holland, also present a high prevalence of such disorders (18.4 % and 14.8 %) (Demyttenaere et al. 2004; Kessler et al. 2009).

With respect to mental health, it is necessary to point out that in all Latin American countries, mental or neuropsychiatric disorders are the most prevalent. Depressive disorders have been increasing across the globe, so much so that the WHO forecasts that by 2020, they will be the second leading cause of sick leave after complications such as heart attacks, coronary insufficiency or strokes. Mental problems are also an important cause of disability that we know is stress-related (DeVries and Wilkerson 2003;WHO and ILO 2000; Stansfeld and Candy 2006). For instance, a recent study by Kohn and Rodríguez (2009) indicates that neuropsychiatric disorders make a substantial contribution to the total number of life years lost or lived with disability in all Latin American countries, with estimated percentages ranging from 15,1 % in Bolivia to 30,5 % in Chile.

There are also considerable differences in the main causes of mortality between Latin America and other regions of the world. Compared to other countries of the American continent with different development conditions, such as the United States and Canada, among working-age people, ischemic heart disease causes less deaths in Latin America than in Canada and the United States (8,96 % vs. 13.12 % respectively). However, chronic problems such as diabetes, cerebrovascular disease and hypertensive diseases cause more deaths in Latin America than in the USA and Canada (3.27 %, 3.51 %, and 2.18 % respectively) (PAHO 2012). In fact, a recent comparative study undertaken by the OECD on diabetes among adults aged 20–79, shows that the only two Latin American member countries of OECD (Mexico and Brazil) occupy first and second place with the highest prevalence (15.9 % and 10.4 % respectively). Similarly, these countries have a 10-year lower life expectancy at birth than developed countries (73 years). Moreover, Mexico, in particular, occupies second place in terms of obesity at international level, only second to the United States (OECD 2013). The same survey shows that when comparing ischemic heart disease between countries, although Mexico doesn’t present the highest death rate for this cause, it does show the greatest percentage of change in the trend from 1990 to 2011, which means that this disease seems to be increasing significantly in the region (OECD 2013).

Likewise, given that unhealthy lifestyle habits and work stress play an important role in their origin, non-communicable diseases may be a good indicator of the health of Latin American workers. As we can see in Table 11.2, the prevalence of non-communicable diseases, among them hypertension and type II diabetes in adults, varies notably between countries. Guatemala seems to be an extreme case in which all percentages are low, whereas Argentina, Chile, Mexico and Uruguay show higher statistics in general that are probably closer to those found in European and North American countries.

Table 11.2 Mortality according to age for all non-communicable diseases (NCD), percentage of deaths caused by them, and prevalence of hypertension and type II diabetes in Latin-American adults
Full size table

It is a well-known fact that not enough studies have been carried out to explain the difference in the prevalence of non-communicable diseases among Latin American countries and the rest of the world. We could speculate that the socially and culturally determined eating patterns may play a relevant role, given that Mexico, Argentina, Chile and Uruguay are among the countries that consume the most calories per-capita in the region (Bermudez and and Tucker 2003). Despite the fact that it is not possible to make clear conclusions regarding the origin of such problems, these data are presented in this context because it is worth remembering that work stress has been associated not only with obesity but with unhealthy eating habits and low levels of physical activity as well as with both hypertension and type II diabetes (Roos et al. 2007; Devine et al. 2003; Kearney et al. 2004; Lowden et al. 2010).

5 Systematic Review: Studies Using the Effort-Reward Imbalance Model in Latin America

Within the context of the aforementioned precariousness of work conditions, over the last decade, some Latin American researchers have begun to think about the impact that these conditions could have on workers’ health, especially in terms of chronic problems, mental health indicators or on their ability to work. We followed the method described below to characterize the studies that have used the ERI model and questionnaire to study work-related stress in workers in Latin America, and its impact on their health.

5.1 Methodological Strategy

The authors of this chapter followed two paths to identify the studies that could be included in this review. The first was to write to a number of Latin American colleagues that work in the field and ask them to send information on any studies they may have undertaken using the ERI model. The second, consisted in a literature search based on the following criteria:

  1. 1.

    Empirical studies published up to July 2015 in any peer-reviewed journal.

  2. 2.

    Studies undertaken involving populations from any of the Central or South American countries.

  3. 3.

    Studies that used the ERI model and its instrument for data gathering and interpretation.

The article search was undertaken using the following databases or indexes: Scielo, ISI, Redalyc, Scopus, Dialnet, Lilacs, Psicodoc, Pubmed. The key words used for the search were ERI, esfuerzo-recompensa, effort-reward imbalance, estrés laboral, job stress, factores psicosociales laborales, psychosocial working conditions, psychosocial risk factors.

The first search procedure allowed us to receive information on a large number of works, some of which are as yet unpublished, and others are in the process of being prepared for publication (for example, thesis work, and studies only presented at congresses). Our final decision was to limit ourselves, for this chapter, to articles that had been published in peer-reviewed journals or peer-reviewed chapters of books, in order to guarantee a minimum level of quality insofar as the studies we used. This decision also allowed us to be fair to all our colleagues, given that all those we did not contact directly or those that were not in the same networks as the colleagues we knew, would have otherwise had less chance of their work being identified.

Following the second search procedure, the authors of this article first carried out an independent search and then, together, confirmed the coincidences in their results. This procedure brought to light a number of studies and it allowed us to affirm that almost all the articles published and sent by colleagues that responded to our invitation were included in our literature search. Two studies, despite having been published, were not identified in the search as they had been included in a peer-reviewed book but not included in the consulted databases. We decided to include these articles in our review.

One of the studies identified had used the ERI as a model to undertake and interpret a semi-structured interview, but it did not use the ERI questionnaire (Tejada and Gómez 2009). This work was excluded, as it was not comparable to any other study.

Once the articles to be used as the object of analysis had been identified and located, they were described using the following criteria:

  1. (a)

    Studies’ country of origin.

  2. (b)

    Authors and publication information.

  3. (c)

    Sociodemographic information of the sample, region of origin within the country, participation rate.

  4. (d)

    Version of the questionnaire used.

  5. (e)

    Psychometric information of the instrument.

  6. (f)

    Other models or questionnaires on work stress used simultaneously.

  7. (g)

    Study goal.

  8. (h)

    Study design.

  9. (i)

    Main results.

  10. (j)

    Health outcome related to ERI model

  11. (k)

    Discussion.

  12. (l)

    Contextualization offered for having used the ERI model.

5.2 Results

The search identified 46 articles published between 2008 and 2015; however, we also found that some of the publications (of Brasil and Chile) refer to the same sample, but report different health outcomes in different papers. Origins of the studies were Brasil (17 studies/25 publications), Chile (6 studies/7 publications), México (6), Colombia (3), Venezuela (2), Cuba (1), Perú (1), multiple countries (1).

The studies, in general, had three goals: (a) to validate the questionnaire (19.57 %), (b) to describe the prevalence of effort-reward imbalance in a particular group of workers (6.52 %) or (c) to relate ERI with a health outcome (73.91 %).

The most relevant information from the nine studies whose central purpose was to assess the psychometric properties of the ERI questionnaire is presented in Table 11.3. It describes the origin and size of the samples and the average values found for the factors that make up the instrument. Despite the variability in the samples size, ranging from 100 to 3010, and despite the different occupations included, the internal consistency of the ERI scales reported in these studies are, on the whole, satisfactory and comparable with those reported by studies of ERI in other countries. The different validity indicators (construct, content and convergent) showed adequate adjustment indices in all cases.

Table 11.3 Description of validation studies
Full size table

As a convergent validity criterion we considered it interesting to describe the coincidence in the ERI scores with those of the instruments that measure closely related concepts. A good number of studies (n = 16) assessed in their samples simultaneously the ERI and the Job Content (JCQ) questionnaires. However, only the studies carried out in Colombia reported correlations between the Job Strain indicator and ERI, reporting values of between 0.2 and 0.6, significant in all cases, except for one. The remaining studies pointed out that the prediction capacity of the considered health variables increased when taking into account simultaneously effort-reward imbalance and Job Strain.

The over-commitment scale was not assessed in any of the studies as a moderating variable of the impact of effort-reward imbalance on health, as suggested in the original model. The studies that measured this factor described its prevalence in each sample and related it directly to the health outcome studied. However, this lack of testing the moderating effects of over-commitment is not a specific fact of Latin-American studies since it is commonly observed so far in international research on the model.

Out of the publications mentioned, 26 studies analyze the relationship between ERI and health or describe the prevalence of ERI. These investigations demonstrate a great deal of variability in the size of the samples studied, as can be seen in Table 11.4. One study assessed only 15 workers that were compared to a control group of similar demographic characteristics; 5 studies included samples of less than 100 workers (57–99 range); 15 studies involved between 100 and 1000 participants, and only 5 studies had samples between 1000 and 3010 participants.

Table 11.4 Description of studies with interest on prevalence or relationships ERI-health indicators
Full size table

Several samples had similar occupations, mostly nurses and health workers, (which was the case in 13 of the 26 studies), military personnel, teachers, and professional drivers. Other samples included people with different occupations, usually from the same company (bank, mining company, electrical company) but also from multiple companies. Only one Chilean study reported having used a national sample of salaried workers (3010 participants).

Another important characteristic of the samples is that they include a majority of women. Eighteen of 26 studies included only women, or women represented at least 75 % of the samples. Three studies included only men and in one study men were the majority (95 %). Five studies included more or less equal numbers of both sexes, and in one there is no clear information of the distribution.

Importantly, all the studies involved a cross-sectional design with the exception of two that were carried out in Brazil. The first of the two exceptions was a case-control study that included 385 participants, 160 cases (that required sick leave benefits for over 15 days after having been diagnosed with a “mental and behavioral disorder”) and 225 controls (who took sick leave for other diseases). Women constituted 43.3 % of this group (Martinez et al. 2015). The second was a cohort study which related work stress with work ability carried out 3 years later that included 1022 participants in the first wave and in which 41.4 % of these participants responded to the second wave of questionnaires. The participants were all workers in a private hospital, of which 72.1 % were female (Silva-Junior and Fischer 2014).

The samples of the studies whose purpose was to determine the prevalence of ERI and/or it relationship to health outcomes (analytic studies) were selected fundamentally from Rio de Janeiro, Sao Paulo and Minas Gerais (Brazil); Santiago and a national sample (Chile), Cuernavaca, Xochimilco, León, Torreon and Mérida (Mexico), Bogota (Colombia), Caracas (Venezuela) and Habana (Cuba). Brazil has carried out the most studies (and has the most publications) using ERI. Twenty-nine of the publications (63 %) have been carried out in English, 14 (30.5 %) in Spanish and 3 (6.5 %) in Portuguese.

Most of the study designs are cross-sectional (44 of 46), and although this does not characterize only Latin American studies, the fact of this design does limit their contribution to clarifying the causal relationships between work conditions and health.

With only a few exceptions, the studies used the 23-item version of the ERI questionnaire and the two-step answer format. The exceptions are two Brazilian studies that reported having used the 46-item version (Fogaça et al. 2009, 2010), and the studies led by Ansoleaga (Ansoleaga and Toro 2010; Ansoleaga and Castillo-Carniglia 2011; Ansoleaga et al. 2013a, b; 2014, Ansoleaga and Castillo-Carniglia 2011) in Chile, using the short 10-item version.

The averages of some or all of the ERI scales are reported in only 11 of the 37 studies (see Table 11.5). We consider that not reporting the means of the scales is an important deficiency, given that it is the means that can most easily be compared between samples from different countries and groups (assuming that there is no differential item functioning-DIF.). To make possible any kind of comparison between studies, we summarize the ERI-data reported by them in Table 11.5.

Table 11.5 Analytic studies compared by the type of data reported (means, ERI prevalence or extreme tertiles)
Full size table

As can be appreciated, the averages of the effort scale, that can range between 6 and 24, were between 8.07 (physicians and nurses in Brazil) and 17.6 (school teachers in Colombia). In general, the values for effort and reward are within the range of data reported by Siegrist et al. (2004) in various European countries, and the same holds true for over-commitment.

Eleven of the 37 studies reported prevalence rates indicating the percentage of people in the sample whose value of the effort and reward ratio is greater than 1. The values reported range from a low of 0.8 % to a high of 75 %. Prevalences of less than 20 % were mentioned in three studies (0.8; 7.8; 12.6); values between 25 % and 50 % were reported for three studies (22.3 24.6 and 33); and values between 50 % and 75 % were reported in five studies (50; 65.8; 66.9; 67; 75). It is worth pointing out that the lowest value reported corresponds to the study that used the 46-item version, and the highest corresponds to one of the studies that used the short version.

The prevalences obtained using tertiles were reported in only six studies. A couple of studies report data for high effort, low rewards or high over-commitment but they do not indicate the criteria used to decide what constitutes high and low levels, so their data are therefore not included here. The authors of these studies used the calculation of the tertiles to explore a dose-response relationships between the model components, the ratio between effort and reward and health indicators. The huge variability in the reported data leads us to think that not all the studies follow the same procedure to calculate their cut-off points, making the values difficult to compare.

The data above makes it appropriate to point out the need to encourage researchers in Latin America to always report the averages of each scale, prevalences and as much information as possible about all the procedures followed as well as about the reasons for not following the suggestions made by the original authors of the ERI scale, for example to use the ER-ratio either as continuous variable or as categorical variable based on the quartiles of the distribution. This would make it more feasible to carry out more informative comparative analyses.

ERI-related health outcomes in the analytical studies are described in Fig. 11.3.

Fig. 11.3
figure 3

Percentage of type of studies and health outcomes associated to ERI indicators in analytical studies

Full size image

As we can see, mental health was the most evaluated outcome, followed by perceived health. In general, there is a lot of variability among the health indicators selected, and in the use of different strategies of coefficients for statistical analysis (Odds Ratios, Betas, Pearson coefficients, etc.). This makes it complicated to compare the relationships between ERI factors and each health outcome. All the outcomes showed a positive and significant relationship with the global indicator (ERI) and with each of its components, with the exception of blood pressure. Greater values on the ERI scales and a greater E/R imbalance are related to greater proportions of mental health problems, more depression, less work ability, lower quality of life, worse self-rated health, more mental disorders, alcohol and drug use, sick leave, menopause symptoms and fatigue, among others. It seems that the perception and reporting of these problems increases as levels of effort, over-commitment and E/R imbalance increase, and as perceived rewards decrease. Blood pressure could only be predicted when both ERI and Job Strain were used in conjunction. Considering the data on morbidity and mortality in Latin American countries, it is surprising that there aren’t more studies that use cardiovascular problems, hypertension, diabetes, obesity and burnout (given their association to depression) as health indicators. The data described supports the idea that the ERI scale allows us to identify health risks clearly, but it could also indicate a possible publication bias.

6 Conclusions and Recommendations

The first important conclusion of this chapter is that the results of the research on work stress and its impact on health in Latin America reflect an early stage of scientific inquiry with clear methodological limitations. The subcontinent is lagging behind in terms of studying this topic, and this is reflected in a significant scarcity of research and in the fact that the existing publications have multiple limitations. The above is particularly true for studies that use ERI, but is not exclusive to research done with this model.

We believe that the main reasons for this scarcity of studies may include, on the one hand, the difficulty of obtaining funding and of having access to samples (especially large or national ones), which are closely related issues. On the other hand, we also consider the hindrances surrounding publishing in specialized high-impact journals.

Countries’ economic problems together with professionals’ and stakeholders’ insufficient knowledge in terms of the models and the relevance of such studies are at the root of this scarcity. However, it is also fair to point out that many of the individuals that are interested in this problem do not enjoy the necessary conditions to be able to access much of the research published in international journals, many of them written in English. Researchers in different countries should consider to augment publishing and reading in different languages in order to increase the exchange and enrichment of everyone’s knowledge and practice.

The difficulties highlighted may explain why most of the studies are concentrated in certain countries, cities, sectors or occupations. To overcome some of the problems, we would have to think about options such as establishing different types of alliances, perhaps with private advisors or state agencies, with researchers from different countries, or additional strategies can be imagined. It is important to increase the number and level of studies carried out in our subcontinent and to publish their results.

A number of other issues add to making publication difficult. One of them is that the editors of a number of recognized journals in this area are not aware of the gaps in regional knowledge, thus disregarding or rejecting respective reports submitted for publication, often labeling them as: “not contributing to existing knowledge”. Furthermore, given the difficulties surrounding publication, there may be a substantial publication bias as researchers are likely to submit studies with positive findings.

Overcoming the limitations to the current studies will require research to improve in a number of ways, some of which we suggest here:

  1. 1.

    Most of the health indicators used in these studies were of a psychological nature and self-reported. It is necessary to increase the use of health and performance objective indicators (e.g. blood pressure) and to increase the size and occupational variability of the samples. These considerations refer to aspects that would increase our chances of more appropriately assessing the impact of work stress on health in the Latin American subcontinent. However, it also seems necessary to use more complex designs including conducting longitudinal studies as well as utilizing advanced strategies for statistical analyses.

  2. 2.

    It is important to unify criteria in order to determine what should be considered a risk value in the work stress questionnaires (JCQ and ERI). For now, categorizations within each sample based on predefined scores (such as upper quartile of the score distribution) are used, but it is a well-known fact that these cut-off points are sample—related and are not defined on the basis of clinical evidence. Therefore, we don’t yet know enough about how they are related to specific risks for concrete diseases. There is also a need to further explore the best form of analyzing the exposure variable, including the possibility that there exist nonlinear relationships between exposure and various outcomes, e.g. as has been found for job strain and blood pressure (Landsbergis et al. 1994).

  3. 3.

    It is important to complement the studies with the use of qualitative strategies. The theoretical models could be used as guides to carry out interviews and explore additional aspects of work or ones that are not totally described with respect to the work stress of different cultural and occupational groups. Additionally, when studying the stressful conditions of informal workers—that constitute practically half of the working population in various Latin America countries—they probably can’t be studied using the questionnaire in its current form because questions typically refer to working conditions of formally employed workers. However, the theoretical model proposed by ERI and supplementing qualitative strategies could be used to explore psychosocial working situations not clearly described so far.

  4. 4.

    Insofar as the validity of the ERI questionnaire, the overall conclusion based on a review of all the studies is that its psychometric characteristics are satisfactory and that it is valid for use in the assessment of work stress here in Latin America as in other regions of the world. Nevertheless we highlight the need for more studies to further validate the ERI questionnaire using advanced psychometric measuring techniques such as confirmatory factor analyses (CFA), differential item functioning (DIF) or the demonstration of time-invariant stability of the ERI questionnaire, using pre- and post-test methodologies (Choi et al. 2014). A further challenge concerns the improved test of criterion validity in studies analyzing associations of ERI with health outcomes (especially those which are among the most important public health problems). In addition to controlling for socio-demographic variables further potential confounding variables need to be taken into account, specifically personality traits such as negative affectivity that affect the way people answer self-reported questionnaires.

  5. 5.

    The validity of the short ten-item ERI scale with the new one-step answer format and four response options (from totally disagree to totally agree) so far has only been analyzed in Chile. This new version has been improved psychometrically and it strengthens the response rate (Tsusumi et al. 2009; Siegrist et al. 2014). The previous two-step procedure of item response resulted in a reduced response rate in some studies, as also mentioned in some studies reviewed here.

  6. 6.

    Of the nine Latin American validity studies, three revealed an overlap of some items as evident from factor analyses. In fact, we also identified some items that presented low factorial loads and were problematic. In this context, at least four of the nine validity studies identified the same problematic item of the scale ‘over-commitment’: “I can easily relax and ‘switch off’ from work.” Future studies should clarify whether this problem reflects a methodological effect (as this is the only item that needs to be reversely coded), or whether the item fails to indicate the underlying theoretical construct. Moreover, it is worth highlighting that the only study that explored the factorial invariance of the ERI scales between six Latin American countries, found that the reward scale does not seem to be totally equivalent in all samples under study. This could imply that the cultural meaning of the different reward dimensions included in the model may not be identical in the region’s different countries, even given that people speak the same language and have some further cultural communalities. Again, future studies should address this question (See Juárez-García et al. 2015).

Although the ERI model and its components are constructs based on the principles of social equity and are therefore assumed to have rather universal validity, the role of culture in the notions of the ERI model should be studied more thoroughly in the future. In particular, the meaning of obtaining rewards may vary among cultures (Kim and et al. 1990). Other studies have looked into the moderating role of cultural features such as collectivism/individualism on the effect of the rewards (Hui et al. 1991), an effect that was also verified in a Colombian study on the relationship between job strain and health (Cendales and Gomez 2013). Such cultural features related to the difference in the concept of reward have been studied in North American and other cultures such as the Latin-American, whereby for the former, success, personal achievements and hard work are important motivators in themselves, whereas for Latinos, the social benefit granted by work and its equilibrium with the family dimension seem to be the main sources of reward in the job (Díaz and Szalay 1993).

In conclusion, in view of the cultural differences mentioned specifically in the context of this research review in Latin America, and in view of the methodological limitations mentioned, more evidence on the relationship of effort-reward imbalance at work with major public health problems will be required.