Informal Employment and Poor Mental Health in a Sample of 180,260 Workers from 13 Iberoamerican Countries

The aim of this study is to estimate the association between employment conditions and mental health status in the working population of Iberoamerica. In this cross-sectional study, we pooled individual-level data from nationally representative surveys across 13 countries. A sample of 180,260 workers was analyzed. Informality was assessed by social security, health affiliation, or contract holding. Mental health was assessed using several instruments. We used Poisson regression models to estimate the contribution of informality to poor mental health by sex and country, adjusted by sociodemographic and work-related characteristics. Then, we performed a meta-analysis pooling of aggregate data using a random-effects inverse-variance model. Workers in informal employments showed a higher adjusted prevalence ratio (aPR) of poor mental health than those in formal employment in Peru (aPR men 1.5 [95% confidence intervals 1.16; 1.93]), Spain (aPR men 2.2 [1.01; 4.78]) and Mexico (aPR men 1.24 [1.04; 1.47]; women 1.39 [1.18; 1.64]). Overall estimates showed that workers in informal employment have a higher prevalence of poor mental health than formal workers, with it being 1.19 times higher (aPR 1.19 [1.02; 1.39]) among men, and 1.11 times higher prevalence among women (aPR 1.11 [1.00; 1.23]). Addressing informal employment could contribute to improving workers’ mental health.


Introduction
Paid work is a social determinant of health that can act to either promote or hinder wellness. It provides financial security, personal identity, and an opportunity to make a meaningful contribution to community life [1]. However, there are stressful characteristics of work related to demands, control, and support that can act to the detriment of workers' health [2].
Informal employment is defined as a non-regulated placement in the labor market that involves an undocumented arrangement between employee and employer in which there is no labor regulation, income taxation, social protection or entitlement to certain benefits, such as advance notice of dismissal, or unemployment and sick leave benefits [3]. Workers in informal employment may be exposed, to a greater extent, to the stressful characteristics of work and this may be reflected in their mental health status.
However, evidence of the impact of informality on mental health is limited [18] and most previous studies were conducted in upper income countries [16,[19][20][21].
The Iberoamerican community is composed of 22 Spanish-and Portuguese-speaking countries (19 Latin American countries, plus Spain, Portugal and Andorra). The relations between Iberoamerican countries have been intensifying, especially multilateral agreement on social security, which has been led by the Ibero-American Social Security Organisation (OISS by its acronym in Spanish) [22]. The current and foreseen trade agreements between the European Union and Latin America require the guarantee of conditions of equality and basic elements for sustainability, including policies that ensure decent and productive work [23]. Evidence from Iberoamerica is scarce and controversial. Previous studies showed that Brazilian workers in informal employment (categorized as self-employed and underemployed) have a higher prevalence of common mental disorders than those in formal employment, and that this association is stronger among women [16,24]. In Central America, workers lacking social security coverage showed a higher prevalence of poor mental health [25]. On the contrary, results from Colombia show that most informal workers perceive their health as normal, good, or very good and are satisfied with their quality of life. However, these results have not been compared to those of workers in formal employment [26]. Results from Chile show that the prevalence of poor mental health is higher among informal workers when compared to formal workers only in the case of male dependent workers, and that there are no differences among women [18].
The aim of this study is to assess the relationship between poor mental health and informal employment among a sample of workers from 13 Iberoamerican countries.

Study Design
Cross-sectional.

Data Source and Study Participants
Nationally representative surveys of household, health, or working conditions were identified for 13 Iberoamerican countries (Table 1). These 13 countries were selected for the study because data could be found that met the following criteria. Eligible surveys were performed after 2012 and included individual-level data on employment conditions and mental health assessed by a valid instrument. When more than one survey for a country complied with inclusion criteria, the most recent was chosen. According to methodological reports from these surveys, all interviews were performed face-to-face at the interviewees' houses. Our sample was restricted to those who had worked during the week preceding the interview, or were absent from work whether due to leave, illness, strike or attendance at a work-related course. This resulted in a sample of 180,260 workers aged 15 and over.

Outcome
Mental health status was assessed by five different instruments. For the World Health Organization-Five Well-Being Index (WHO-5), poor mental health was defined as a score under 13 [27]; for the 9-item Patient Health Questionnaire (PHQ-9), as a score of 10 or more [28]; for the psychological domain of the Abbreviated World Health Organization Quality of Life questionnaire (WHOQOL-Bref), as a score under 60 [29]; for the 12-item General Health Questionnaire (GHQ-12) following the GHQ usual scoring method (0-0-1-1), as a score of 3 or more [30]; and for the affect domain of the Washington Group Extended Set on Functioning (WG-ES), as a score of 4 in either depression or anxiety scales [31].

Work-Related Variables
Regarding employment conditions, in order to operationalize the variable of informal employment, according to the ILO approach [32], social security or health affiliation, or having a contract, were the criteria used based on the information available. Workers were classified according to their social security affiliation as either formal (affiliated) or informal (not affiliated). This was the case for Argentina, countries in Central America, and Peru. When information on social security was not available, classification was based on affiliation with a health system: workers who were entitled to health services as holders were considered formal and those not entitled or entitled as beneficiaries for other family members were considered informal. This was the case for Brazil. In the case of Chile, Mexico, Portugal, and Spain, due to data availability, employment status was based on contract holding. Workers holding a contract were considered formal and those working without a contract were considered informal (Supplementary Table S1).
Regarding labor relationship, workers were classified as an employee, self-employed or an employer. Occupational category was first transformed from local classification into the International Standard Classification of Occupation (ISCO-08) [33] and then grouped into four categories: non-manual skilled (managers, professionals, technicians and associate professionals), non-manual non-skilled (clerical support and services and sales workers), manual skilled (skilled agricultural, forestry and fishery workers, craft and related trades workers, and plant and machine operators and assemblers), and manual non-skilled (elementary occupations). Further aggrupation into manual and non-manual was performed. Workers in the armed forces were excluded from the analysis.

Socio-Demographic Variables
Age comprised four categories (younger than 24; 25-44; 45-64; and over 65 years old). Highest educational level achieved was categorized into less than primary school; primary school; middle school; and higher education. Marital status comprised two categories: married or cohabiting; and single, separated, or widowed.

Analysis
At the country level, we describe the population distribution by frequencies and weighted proportions (N%). We estimated the prevalence of poor mental health with a 95% confidence interval (95% CI) by work-related and socio-demographic characteristics, and estimated the association between employment condition and poor mental health by crude prevalence ratios (cPR) using Poisson regression with robust variance. Models were adjusted by age, marital status, education, occupational category, and labor relationship in the case of Argentina, Chile, Peru, Brazil and countries in Central America. Models for Spain and Portugal were adjusted by age, marital status, education, and occupational category, and models for Mexico were adjusted by age, marital status, and education due to data availability. Country-specific survey weights were applied. All analyses were stratified by sex.
Second, a meta-analysis [34] pooling of aggregate data from each country using a random-effects inverse-variance model was performed to assess the overall effect of employment condition on mental health status in the region. The analysis was stratified by a mental health measurement tool.

Population Characteristics
Workers in informal employment represent more than 40% (N = 83,781) of the working population in Iberoamerica according to the surveys included in this study. . Men showed higher rates of working in manual occupations than women across all countries. Approximately 17% of workers (N = 31,135) did not finish primary school, this figure being below 3% in Spain and Portugal and above 25% for Brazil. Women represented a lower proportion of unschooled workers than men, except for in the cases of Peru and Guatemala (Supplementary Table S2).

Mental Health Distribution
Among the countries using GHQ-12, the lowest prevalence of poor mental health was found in Guatemalan men ( The least educated workers showed a higher prevalence of poor mental health than the most educated across regions, except for Panamanian and Brazilian male workers. Differences in the prevalence of poor mental health between the lowest and highest educational categories were greater for women than for men, except in the cases of El Salvador (men 4.8 pp, women 2.3 pp) and Spain (men 10.2 pp, women 7.2 pp) (Tables 2 and S3).

Informality and Poor Mental Health
Informal workers showed a higher prevalence of poor mental health than their counterparts across most countries and both sexes. However, formal male workers from Panama (27.0% vs. 19.2%), Portugal (11.1% vs. 10.7%), and Brazil (4.8% vs. 4.5%) and formal female workers from El Salvador (14.5% vs. 10.5%) showed a higher prevalence of poor mental health than informal workers (Tables 2 and S3).
The prevalence ratios of poor mental health between informal and formal workers were significant in the case of male workers from Costa Rica (cPR 1 Figure 1).

Discussion
Our results show that, overall, Iberoamerican workers in informal employment have a higher prevalence of poor mental health than workers in formal employment. However, the results are heterogeneous among countries and, in most cases, we could not establish a robust association between informality and poor mental health. These results may be due to various reasons: The inaccurate measurement of mental health status; the inaccurate measurement of informality; unobserved confounding factors concealing a real association; or a non-existing association between informality and poor mental health. It is important to note that informal employment is related not only to poor working and employment conditions, but also to poor living conditions, social and economic vulnerability, and poor environmental conditions. All of these factors could affect the association. Previous studies showed that national indicators of GDP, CO 2 emissions, educational attainment, life expectancy, rates of poverty, and women's labor market participation are strongly associated with the rate of informality [35].

Mental Health Status Measurement
Regarding the prevalence of poor mental health, the information found was heterogeneous, with poorly standardized records, making it difficult to establish accurate regional comparisons. The mental health questionnaires used in each survey measure different concepts of health. Take, for example, PHQ-9 and WHO-5. The former scores each of the nine DSM-IV criteria for depression with questions such as "Have you felt depressed or hopeless?", while the latter assesses psychological well-being with statements such as "I have felt cheerful and in good spirit". However, other studies have established measurement similarities between PHQ-9 and GHQ-12 [36,37], WHO-5 and GHQ-12 [38], WHO-5 and PHQ-9 [39], and PHQ-9 and the depression domain of WG-ES [40]. Besides differences across questionnaires, differences in social, cultural and historical background across countries may influence workers' perception of their mental health and quality of life, both being closely related to one's expectations. Thus, even when using the same questionnaire, mental health evaluation requires a country-specific analysis, and questionnaire validation in each country is encouraged [41].
In addition, estimates of poor mental health may be underestimated for two reasons: first, the prevalence of somatic symptom in Latin America is high, while mood symptoms are underrepresented [42]. Hence, we would expect low rates of self-identified depression. Second, GHQ-12, although used extensively, may not be the best screening tool [43].
The prevalence of poor mental health in Mexico is remarkably low; however, the case definition refers to mental conditions causing high disability. Other studies have already reported similar rates when severity is considered [44,45].

Exposure Measurement
Informality was defined according to ILO specifications, and rates of informal employment described hereby are similar to those depicted by ILO [46], implying general measurement accuracy, except for El Salvador and Brazil where we found higher rates, and Spain and Chile where we found lower rates. Nevertheless, other authors have already argued that this definition may be insufficient to analyze the context of the Latin American working population [47]. Furthermore, employment conditions on the main job are used to depict informality, but a person can simultaneously have two or more formal and/or informal jobs. The proportion of workers holding multiple jobs varied in the sample, from 3% in Mexico to around 19% in Argentinian women and Peruvian men, and was not available for Chile. The characteristics of these other jobs could not be ascertained.
Formal workers may not be as protected as we would expect, diminishing the differences between formal and informal workers. Workers at formal jobs may also be exposed to other work-related risk factors for poor mental health, such as an increased number of working hours, unpaid overtime, poor job security, poor satisfaction with one's work culture and a feeling of a lack of support, low income, shift work, and night work [48,49].
Moreover, the provision of mental health services represents less than 1% of government health spending in low-and middle-income countries, and rates of availability and uptake for mental health services remain very low [50]. Therefore, health and social services may not meet the mental health needs of formal workers in the region. On the other hand, workers may have other support systems outside social security and state resources.

Unobserved Confounders
Other life circumstances may have a greater impact on mental health than employment condition and may act as unobserved confounders. Exposure to poverty, food insecurity, trauma, housing instability, and decreased social support have all been proven to affect one's physical and mental well-being [51,52]. Race and migrant status have also been proven to highly affect mental health. Information on these circumstances could not be ascertained for each country and was, therefore, not included in the analysis. Future studies should explore the effect of these living conditions on the association between informality and poor mental health.

Education
When adjusting for the highest education level achieved, the magnitude of the association between informality and poor mental health decreased, implying that part of that association was explained by education. Low education may be associated with poor mental health in two ways: early onset disorders have been shown to impact school dropout, and school dropout has been proposed as a cause of internalizing disorders [53][54][55].

Gender
The association between informality and poor mental health was higher and stronger for men than for women when adjusting for sociodemographic and other work-related variables. The unequal involvement of women in unpaid labor can wield deleterious effects on their mental health and may explain some of the differences found between the sexes [56]. Information on care load and domestic work was only available for questionnaires from Brazil and Chile. In Brazil, women were in charge of caring for dependents (children, the elderly, the sick or people with special needs) more frequently than men, and those in charge of care showed a higher prevalence of poor mental health. In Chile, the results were similar, with a larger percentage of women in charge of caring for dependents and domestic work than men. Women with double presence showed a significantly higher prevalence of poor mental health.

Strengths and Limitations
The findings should be considered in light of the limitations of this study. First, in all cases, survey participants were limited to those with permanent residence and sufficient proficiency in the official language (Spanish or Portuguese), and therefore the prevalence of poor mental health may be underestimated by the exclusion of the most vulnerable workers, such as immigrants. Second, the surveys included relied solely on retrospective self-reports; hence, the data are subject to recall bias and the willingness to disclose information truthfully. Given the generally stigmatizing nature of mental health disorders, it is likely that reporting bias leads to an underestimation of prevalence. The cross-sectional design of the study prevented us from making inferences regarding causality or directionality between socio-demographic and work-related correlates and mental health status. To operationalize informal employment, and according to the ILO approach [32], different criteria were used for each country, depending on the information available: affiliation with social security, with a health system or with having a contract. However, in many countries, workers with a contract must be affiliated with a pension or health system, and thus not being affiliated with social security or a health system is tantamount to the same concept: informality. Another important limitation is that mental health status was assessed using five different instruments. However, all instruments used were validated, widely recognized and reliable measures of mental health. In addition, we measured the prevalence ratio of poor mental health between formal and informal workers within countries using the same instrument. Finally, given that the data come from different countries and therefore from different sources, the methodologies might have differed slightly. In addition, the surveys were conducted in different years, between 2012 and 2018. However, this study used data from each country to measure the difference between the mental health of informal and formal workers within countries. All these limitations encourage caution when interpreting the results.
Despite these limitations, to our knowledge this is the first study to report on the mental health status of the working population of Iberoamerica as a whole and to highlight the deleterious effect of informality on workers' mental health, regardless of other sociodemographic and work-related characteristics. In this study, we provide evidence, based on the most recent data available, for local occupational health planning and a starting point for further surveillance in the region.

Conclusions
Addressing informal employment could contribute to improving the mental health of workers in Iberoamerica. Work is a key determinant of workers' mental health, but other determinants such social and living conditions should be included in prevention policies. Given the great heterogeneity among surveys in the region, we recommend further research to incorporate common indicators of mental health and informality in order to enable comparisons. Additionally, variables such as hours dedicated to unpaid work at home and care for dependents should be included to better analyze the interaction between work, gender, and mental health.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/ijerph19137883/s1. Table S1: Definition of employment condition by questionnaire. Table S2: Population distribution by questionnaire, country and sex. Table S3: Prevalence of poor mental health (% (95% CI)) according to work and sociodemographic variables, by questionnaire, country, and sex. Institutional Review Board Statement: Ethical review and approval were waived for this study due to its characteristics. The datasets analyzed were properly anonymized and informed consent was obtained at the time of the original data collection, in which the researchers took no part.
Informed Consent Statement: Informed consent was obtained at the time of the original data collection.