Introduction

Tobacco use is the leading cause of preventable, premature death and disease worldwide1, killing over eight million people a year, including deaths directly caused by tobacco use as well as deaths attributed to exposure to environmental tobacco smoke (or ETS, also called secondhand smoke or passive smoke)2. Reports dating back to the 1950 s linked tobacco use to multiple deleterious health consequences3,4. The Global Burden of Disease study calculated that 200 million disability-adjusted life years were due to smoking tobacco use5. In the Region of the Americas, tobacco use is related to nearly one million deaths yearly6. After a slowdown in tobacco use reduction was noted in 2017, the Pan American Health Organization (PAHO) unanimously approved an action plan specific to the region6.

As noted, the burden of tobacco use is not limited to the user. The topic of nonsmokers being exposed to ETS was introduced in the early 1970s, and in 1986, the US Surgeon General7, released a report demonstrating the causal link between ETS and lung cancer, with continued supportive evidence reported decades later using meta-analysis8. Researchers estimate that more than 880,000 individuals worldwide die each year from ETS9The negative health effects of ETS exposure are also notable in children, associated with health outcomes such as asthma10,11, respiratory infections12, and an increased risk for sudden infant death syndrome13, are observed. Workplaces are also affected by ETS. For instance, the British Occupational Cancer Burden Study Group14 reported that ETS exposure is a top occupational hazard linked to cancer among nonsmokers.

As part of broader tobacco control efforts to reduce tobacco-related morbidity and mortality, an effective measure is promoting and implementing smoke-free policies to create smoke-free environments. Smoke-free environments have indeed been shown to reduce the risks associated with ETS. Multiple studies summarized in the 2006 US Surgeon General’s report15 concluded that with indoor smoking bans, there was an observed reduction in heart disease mortality16 and admissions for heart attacks17,18. The effectiveness of indoor bans and smoke-free environments also extend to workplaces. Using 2015 NHIS data, Su et al.19 reported that across all industries, respondents who work in states with smoke-free laws reported less exposure to ETS at work (8.6%) compared to those in states with no smoke-free laws (11.0%).

Despite the evidence that indoor smoking bans are effective in protecting others, including workers, regulations, and enforcement have been slower across Latin America. Since the 1990s, at least three large tobacco companies (i.e., Philip Morris International, RJ Reynolds, and British American Tobacco) have promoted the “Courtesy of Choice” program, intended to “accommodate” smokers in Latin America. To allow for continued smoking, the program endorses ventilation as a solution to ETS. In a 2006 US Surgeon General’s Report15, evidence was presented that ventilating buildings, along with cleaning the air and separating smokers from nonsmokers, does not eliminate ETS; eliminating indoor smoking is the only full protection.

People living in Latin America continue to be exposed to ETS in public spaces, including workplaces. In 2001, PAHO launched the “Smoke-Free Americas” initiative to attempt to ameliorate this ongoing issue. However, researchers found that nicotine (measured as vapor concentration in the air) was detected in 94% of locations surveyed, including hospitals, schools, government buildings, bars, and restaurants20. Policies alone are insufficient, as they require effective enforcement. For example, after the PAHO initiative, universities implemented smoke-free campus policies to align with it. Nonetheless, air-borne nicotine was detected in more than half of the samples21.

Moreover, enforcing labor regulations in Latin America, including Central America, is challenging due to the prevalence of informal employment in the Latin America and the Caribbean region, where more than half of the working population is engaged in informal employment22. This percentage has been reported as high as 81% in Central America23.

Monitoring policy implementation and enforcement is imperative to assess ETS exposure in the workplace. Further, identifying exposure by industry could be important to revise strategies where necessary or develop new approaches. However, the scarcity of reliable and widely available occupational health data in Central America has historically posed a challenge to such studies. Nevertheless, this study addresses this challenge by taking advantage of the 2018 II Central American Working Conditions and Health Survey (ECCTS, by its Spanish acronym)24, to assess ETS exposure among workers across industries in Central America. Therefore, in this study, we assessed exposure to ETS by industry in Central American workers.

Methods

Study sample

The II ECCTS was conducted between February 2018 and June 2018 by trained field workers who administered the survey through face-to-face interviews at the workers’ homes. The II ECCTS is a nationally representative sample of the working population of the six Spanish-speaking countries (Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and Panama) of Central America. The II ECCTS targeted workers 18 years and older who worked at least one hour for pay the week before the survey. The questionnaire collected information on employment and working conditions, business characteristics, resources and preventive activities, health and well-being, and socio-demographic and family characteristics. The II ECCTS used a probabilistic multistage sampling design with the country-level census sampling frame to select a nationally representative sample of 1,500 workers per country. The response rate was 27% (9,032 completed interviews out of 33,776 attempts), and the cooperation rate was 68%23. To represent the working population in each country, post-sampling weights were developed and used based on each country’s demographic population information by age, sex, and industry.

Measures

The study’s main variables were exposure to ETS in the workplace and workers’ industry. The outcome variable, ETS exposure, was assessed with a single item: “Regarding the environment of your workplace and taking as a reference a usual day of work, how often are you exposed to tobacco smoke?” A 5-point Likert scale was used, and responses were categorized into unexposed (never) and exposed (seldom, sometimes, often, and always). The information on industry was collected via an open-ended question (i.e., “What is the main economic activity of the company, business, institution, or organization where the person works or of the job the person does).” Responses were then coded into standardized classifications by trained coders using the International Standard Industrial Classification of All Economic Activities25, and then grouped into the following categories: agriculture, manufacturing, services, construction, sales, transportation, education and healthcare, and public administration and defense.

Other work-related covariates included the ___location of the main job, grouped into “house” (e.g., a worker’s home, such as a teleworker or home-based sewer; or someone else’s home, such as a domestic worker), “building” (e.g., commercial or office space), “country” (e.g., rural areas), “street” (e.g., street vendors), and “other” (e.g., a driver, whose ___location would be the means of transportation they are working in, such as a bus, taxi, etc.), and physical and psychosocial working conditions. The physical items included exposure to extreme (hot or cold) temperatures, noise exposure (i.e., “Levels of noise that are so high that have to raise voice to talk to people”), and working with a heavy object (i.e., “Throws, pulls, lifts, moves, pulls, or pushes loads, people, animals, or other heavy objects”) or performing strong or strong physical effort. All these aspects were assessed on a 5-point Likert scale and then dichotomized into always/often vs. never/seldom/sometimes.

Five psychosocial work dimensions were explored: psychological demands (7 items about quantitative psychological demands such as the volume of work in relation to the available time and the type of tasks, including emotional and psychological demands), active work and possibilities of development (9 items about job control and opportunities for employees to apply and acquire knowledge, skills, and experience), insecurity at work (4 items regarding job stability, employability prospects, and unwanted changes in working conditions), social support and quality of leadership (10 items about opportunities for interpersonal relationships, whether instrumental or emotional, and the characteristics of team management by supervisors); and esteem (4 items about the appreciation, respect, and fair treatment by the management at work). All of the psychosocial work items were drawn from two standard questionnaires: the first version of the ISTAS-21 short form26, complemented with items from the CTESLAC27. The ISTAS questionnaire is the Spanish version of the Copenhagen Psychosocial Questionnaire, which was developed to cover a broad spectrum of psychosocial work theories28. Three additional items were drawn from the CTESLAC (Spanish acronym for Basic Questionnaire on Working Conditions, Employment and Health in Latin America and the Caribbean), a series of items agreed by consensus by an international group of experts based on items available at the national surveys of working conditions and health conducted in Spanish-speaking countries. All the psychosocial work items were measured with a 5-point Likert scale (0 never to 4 always) and scored in the same direction, with higher scores indicating a less favorable situation. Each participant was assigned the average of the items’ scores on the corresponding scale if at least 80% of the factor’s items were non-missing and missing otherwise. Following a common procedure, exposure to each factor was dichotomized into the less favorable category (i.e., high for demands and low for control and support) if the average was equal or higher than the median or into the more favorable category otherwise.

The II ECCTS questionnaire also collected sociodemographic information such as gender (female/male), age in categories (18–24, 25–45, 46–65, or 66 + years) to account for non-linear effects, align with common age groupings in the literature, and improve interpretability, Central American country (Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, and Panama) where the respondent was interviewed, the highest level of education completed (no school/primary, secondary, university), and the average monthly income range over the last three months, which was collected in country-specific currency categories then grouped into ≤$200, $201-$500, or >$501 for comparison between countries. Smoking status was measured with two questions, “Do you smoke now?” (yes/no) and “Have you smoked at least 100 cigarettes during your life?” (yes/no): Current smokers were workers answering ‘Yes’ to the first question; former smokers were workers reporting not currently smoking but having smoked at least 100 cigarettes during their lives; and never smokers were workers reporting not currently smoking and not having smoked at least 100 cigarettes during their lives.

Worker’s general conditions of employment were measured and examined with the outcome variable, including formal or informal workers, occupation categories based on Standard Occupational Classification Hierarchy29, employment type (employer/entrepreneur/owner, self-employed/free-lancer, dependent/salaried worker, contributing family worker, or members of producers’ cooperatives), job contract/agreement types (fixed/permanent, temporary, or internship/scholarship/practice), average hours worked per week using an open-ended question, type of work schedule (fix/rigid or flexible), frequency of overtime compensation (i.e., “In the last twelve months, how often did you receive compensation that was owed for overtime work, either as pay or with additional time off?: always, sometimes, never, or I did not work overtime”), means of commuting transportation (on foot, bicycle, motorcycle, public transportation, taxi, private car, or company transportation), and average commute time (hours and minutes). In addition, workplace safety conditions, such as frequency of exposure to slippery floors, use of instruments, tools, or machines, which can cause damage, and cleanness of the work area, were measured with a 5-point Likert scale (0 never to 4 always).

Finally, health-related questions included self-reported general health status (very good, good, fair, poor, and very poor), average hours slept per day in the last week using an open-ended question, frequency of alcohol drinking in the past 30 days (i.e., “Considering all of your drinking times in the past 30 days, about how often did you have any beer, wine or liquor?”) with responses collected in an 11-points frequency scale including “three or more times a day, twice a day, once a day, 5 or 6 days a week, once a week, three days a month, one or two days a month”, and three options qualifying the answer ‘not at all in the last 30 days’, such as “But I do drink sometimes,” “Because I don’t drink anymore,” and “I have never drunk.” Pain was measured by asking, “In the last month, have you felt pain in the following body locations: neck, head, shoulders, upper back, mid back, elbows, wrists/hands, lower back, hips/thighs, knees, ankles/feet?” Comorbidities were collected by asking, “In the last month, have you felt any of the following problems or disorders: respiratory, dermatological, coronary, diabetes, vision, auditory, hypertension, varicose veins, or chronic kidney disease?” Finally, the 12-item General Health Questionnaire30 was used to assess the likelihood of common mental disorders and, so, as a general measure of mental well-being.

Statistical analysis

Unweighted frequencies and weighted prevalence of ETS exposure, with corresponding 95% confidence intervals (CIs), were calculated for all variables. Given sampling weights, a survey-weighted logistic regression model was used to estimate the associations between ETS exposure and all other variables. To accurately estimate variances, we used a ___domain option in the survey logistic procedure in SAS (version 9.4, SAS Institute, Cary, NC) to compute the estimates and their variances accurately, creating an indicator variable for missing values. Variables with a p-value ≤ 0.20 in the bivariate analyses were introduced into three preliminary multivariable models, adjusting separately for (i) sociodemographics, (ii) physical working conditions, and (c) psychosocial working conditions models. A priori key variables (age, ethnicity, and smoking status) and covariates with a p-value ≤ 0.05 from the three preliminary models were retained for the final model. A sensitivity analysis for missing values was conducted to ensure the model’s robustness and identify potential confounding issues. All analyses were performed using SAS (version 9.4, SAS Institute, Cary, NC).

Both The University of Texas Health Science Center at Houston Committee for Protection of Human Subjects (# HSC-SPH-16–0803) and the Ethics Committee of the Universidad Nacional in Costa Rica (# UNA-CECUNA-ACUE-09–2017) gave their approval for the study protocol. Written informed consent was obtained for all recruited participants. No questions of a sensitive nature were asked, and participants were free to refuse to answer any question and/or terminate their involvement at any time. All methods were carried out in accordance with relevant guidelines and regulations in the ethics approvals and informed consent. All data were deidentified before analysis.

Results

The sample (Table 1) was predominantly male (61.3%), with the majority aged 25–45 (43.8%), and 32.4% employed in agriculture. Over half of workers had only primary education or none (51.6%), and nearly 90% earned less than $500 monthly. Regarding smoking status, 13.2% were current smokers, 8.4% were former smokers, and 78.3% had never smoked. For job ___location, almost half (47.9%) worked in homes, 27.9% in rural areas, 15.5% in buildings, 5.6% on the street, and 3.2% elsewhere. Over half (54.3%) frequently faced extreme temperatures, and 19.8% often encountered high noise levels requiring them to raise their voices in their workplace. Heavy lifting or strenuous physical effort was common for 44.5% of workers. Additionally, 47.2% reported high psychological demands, 39.7% felt insecure at work, 54.2% experienced low social support or poor leadership, and 58.7% had low work esteem.

Table 1 Distribution of sample characteristics and prevalence of exposure to environmental tobacco smoke (ETS) in the II ECCTS* (N = 9,032).
Full size table

ETS exposure was highest among workers in transportation (40.1%) and construction (30.9%), and lowest in the education/healthcare (11.5%) industry. El Salvador had the highest ETS exposure prevalence (24.0%), while Guatemala had the lowest (15.1%). Males had nearly double the ETS exposure (22.7%) compared to females (12.0%). The youngest workers (18–24) had the lowest prevalence (14.6%). Current smokers had the highest ETS exposure (45.5%), double that of former smokers (22.9%), and three times higher than never-smokers (14.2%). ETS exposure varied by job ___location, with the lowest prevalence among those working in homes (15.0%) and the highest among street workers (35.5%). Workers frequently exposed to extreme temperatures had a higher ETS exposure (21.7%) than those less exposed (14.9%). Similarly, workers always/often exposed to loud noise had a higher prevalence of ETS exposure (29.9%) than those less exposed (16.0%). Workers frequently handling heavy objects or performing strenuous tasks also had higher ETS exposure (25.5%) than those less involved in such activities (13.0%). ETS exposure was higher among workers with high psychological work demands (25.6%) than those with low demands (12.3%). Similarly, 23.5% of workers with high job insecurity were exposed to ETS, versus 15.3% with low insecurity. Conversely, workers with high social support/quality of leadership had lower ETS exposure (15.0%) compared to those with low support/leadership (21.2%). Additionally, workers with high work esteem had lower ETS exposure (14.9%) than those with low esteem (21.5%).

Regarding smoking status by industry (Fig. 1), the transportation industry had the highest percentage of current smokers (25.1%), followed by construction (21.1%) and public administration/defense (20.8%). Conversely, the education/healthcare industry had the highest percentage of never-smokers (87.4%), followed by sales (87.2%) and manufacturing (79.4%).

Fig. 1
figure 1

Prevalence (%) of smoking status and 95% confidence intervals by industry in the II ECCTS* (N = 9,032).

Full size image

Results from the multivariable models (Table 2) showed that males had higher ETS exposure odds than females (OR = 1.33; 95% CI: 1.01,1.76). Among occupation-related factors, workers in transportation (OR = 2.33; 95% CI: 1.24,4.39), public administration/defense (OR = 2.03; 95% CI: 1.14,3.63), construction (OR = 1.88; 95% CI: 1.08,3.29), services (OR = 1.65; 95% CI: 1.06,2.58), and sales (OR = 1.64; 95% CI: 1.05,2.58) had higher odds compared to education/healthcare workers. Main job ___location also impacted ETS exposure, with street workers having higher odds (OR = 1.92; 95% CI: 1.28,2.88) compared to workers who worked in homes. For physical conditions, exposure to noise (OR = 1.35; 95% CI: 1.04,1.74) and heavy physical effort (OR = 1.55; 95% CI: 1.20,2.02) had higher odds compared to those who reported less of these conditions. Regarding psychosocial working conditions, those with high psychological work demands (OR = 1.92; 95% CI: 1.49,2.47), low social support/leadership (OR = 1.57; 95% CI: 1.2,2.02), and low work esteem (OR = 1.35; 95% CI: 1.06,1.72) had higher odds of exposure. ETS exposure also varied by worker’s smoking status, with former smokers having 1.70 (95% CI: 1.18,2.45) times and current smokers 4.46 (95% CI: 3.27,6.08) times the odds of ETS exposure than never-smokers.

Table 2 Odds ratios* for exposure to tobacco smoke in II ECCTS‡ by selected sample characteristics (N = 9,032).
Full size table

Discussion

In this study, there were prominent differences between industries concerning ETS exposure. Compared to education and healthcare settings, workers in several industries reported ETS exposure, including services, construction, sales, transportation, and public administration and defense. Notably, exposure to ETS does not appear to be fully explained by the prevalence of smoking, whether current or former, for some of these industries either. While education and healthcare had the highest prevalence of never-smokers (87.4%), industries including sales (87.2%) and services (78.7%) had comparable rates of never-smoking although higher odds of ETS exposure. Whereas several measures of physical and psychological conditions were also highly predictive of ETS exposure, accounting for these differences and demographic variables in the final model still resulted in important differences by industry.

The British Occupational Cancer Burden Study Group14 reported that ETS exposure is a top occupational hazard linked to cancer among nonsmokers, and multiple studies16,17,18 concluded the reduction of disease and mortality when indoor smoking bans are implemented, these findings reveal a gap for workers in the Central Americas region. In the United States, Su et al.19 reported that across all industries, respondents who work in states with smoke-free laws reported less exposure to ETS at work (8.6%) compared to those in states with no smoke-free laws (11.0%). Notably, the US Environmental Protection Agency (EPA; nd) reports that exposure, both outside and inside, may pose health risks or worsen existing health issues, thus including all occupations, regardless of place.

As previously noted, PAHO launched the “Smoke-Free Americas” initiative in 2001 to attempt to ameliorate this ongoing issue. Counter efforts are in place in the region as at least three large tobacco companies (i.e., Philip Morris International, RJ Reynolds, and British American Tobacco) have promoted the “Courtesy of Choice” program since the 1990 s, intended to “accommodate” smokers in Latin America. To allow for continued smoking, the program endorses ventilation as a solution to ETS, which has already been shown to be insufficient15.

Targeting specific industries allows for a structural level intervention that could positively impact multiple workers. Rather than identify individual employees for cessation programs or target specific worker conditions or roles, implementing and enforcing indoor smoking bans for all workplaces will yield more impactful results. Specific industries with higher exposures can also be targeted for more industry-related interventions. For example, for the construction or transportation industries, targeted ads and programs regarding the norms of smoking and how to change them could prove effective. The indoor ban policy with effective enforcement is the best-known way to reduce exposure, thus improving workers’ health. Nonetheless, despite the widespread implementation of tobacco control and smoke-free laws31, and ongoing calls for expanded clean indoor air policies32, workers’ exposure to ETS may still vary by industry, influenced by workplace culture and social norms. For example, smoking may still be permitted in certain outdoor areas (e.g., restaurant patios) or go unenforced in private homes (e.g., where a smoking homeowner employs a domestic worker). Thus, continued surveillance and targeted interventions remain crucial to addressing these disparities in exposure.

There are weaknesses to our study, but the strengths contribute to outweighing the limitations. First, due to the cross-sectional nature of our study, we cannot establish directionality and temporality between ETS exposure and the variables we examined. Future studies should consider the use of prospective longitudinal designs. Second, our study is based on self-reports, so recall bias may be present regarding the participants’ responses. Third, participation was voluntary, and the response rates differed by country; this could have produced unintended selection bias. Still, the cooperation rate was 68% (from 59% in Honduras to 75% in Guatemala), indicating that once an eligible contact was established, the interview was completed over two-thirds of the time. The sample was selected using randomization to address this bias and then sampling weights by country were used. Fourth, the information bias and potential differences by country may be present. Therefore, validated measures were used, and the comparison was made by industry, regardless of country, but adjusted for country. Fifth, differences between countries may exist regarding the interpretation of the questions. However, the II ECCTS questionnaire was tested to ensure that questions were adequately understood analogously by respondents from all the countries.

The ECCTS provides reliable and timely information from representative samples of workers in Central America. The findings can thus be generalized to the working population in this region. The data can also be used to assess the application of smoke-free policies in the workplace, identifying areas for improvement in working conditions. Continued surveillance of workers in Central America should be used to not only obtain smoking status but also broader ETS exposure, given the known negative health effects. Implementation of existing policies will reduce morbidity and mortality related to tobacco use.

Conclusion

Use of and exposure to tobacco products have had known adverse health consequences for many years. Past studies have demonstrated the positive influence of implementing an indoor smoking ban. There needs to be more uptake and enforcement of such policies in other countries, resulting in poorer health outcomes for large populations. This study demonstrated that many differences remain for ETS exposure depending on the industry, even after controlling for working conditions and sociodemographic variables. A workers’ industry should not impact or contribute further to preventable adverse health outcomes. Global attention to the negative influence of tobacco companies and resulting health risks for large populations is needed. Policies for indoor smoking bans have proven effective and are important for broad application.