Article Text
PDF
Abstract
Objectives A historical cohort study of cancer mortality is being conducted among workers in a chrysotile mine and its enrichment factories in the town of Asbest, Russian Federation. Because individual-level information on tobacco use is not available for Asbest Chrysotile Cohort members, a cross-sectional survey of smoking behaviours was conducted among active and retired workers.
Methods Self-administered questionnaires were completed by active workers during meetings organised by occupational safety personnel. Retired workers completed questionnaires during meetings of the Veterans Council or were interviewed via telephone or in person. Of the respondents, 46% could be linked to the Asbest Chrysotile Cohort. Among those, logistic regression models were used to assess associations between smoking and cumulative dust exposure.
Results Among men, smoking prevalence was high and relatively consistent across birth decades (average, 66%), and was similar in workers across all levels of cumulative dust exposure (p trend, 0.44). Among women, the prevalence increased from <10% in those born before 1960 to 30% in those born after 1980, and smoking was associated with exposure to dust versus not exposed to dust (p value, 0.006), but did not vary appreciably across workers in different cumulative dust exposure categories (p trend, 0.29).
Conclusions Our study suggests that cross-sectional surveys may be a useful tool for understanding the potential health impact from smoking in occupational cohorts, including possible confounding by smoking. This survey showed that adjustment at the age group level among women is needed to reduce residual confounding and account for smoking patterns, which have changed substantially over time.
- asbestos
- cross sectional studies
- smoking
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Key messages
What is already known about this subject?
Occupational cohort studies often lack data on study participants’ individual smoking habits; if smoking is related to the exposure and the outcome of interest it may confound the results.
What are the new findings?
This survey among retired and current workers at a chrysotile asbestos mine and enrichment factories in the Russian Federation shows that the potential confounding of results in men is unlikely, while it remains a concern in women as their smoking habits have changed substantially over time and differ between those working in non-production parts and production parts of the company.
How might this impact on policy or clinical practice in the foreseeable future?
This survey shows that it is useful to conduct surveys in specialised workforces to capture smoking patterns in men and women across birth cohorts.
Introduction
A historical cohort study was launched in 2013 of cancer mortality among current and former workers of the Joint Stock Company (JSC) Uralasbest, which runs the world’s largest operating chrysotile asbestos mine and its enrichment factories in Asbest, Sverdlovsk Region, Russian Federation.1 Cohort members comprise all workers in the mine, factories, autotransport, external rail, central laboratory and explosives unit at JSC Uralasbest that were employed for at least 1 year between 1975 and 2010. Detailed work histories were derived from company records, and vital status at end of follow-up was ascertained from official sources such as the Pension Fund, the Federal Migration Service and Civil Act Registration Office providing the date and cause of death of those deceased in Sverdlovsk Region. The overall aim of the cohort study is to characterise associations between occupational exposure to chrysotile asbestos and site-specific cancers.1 Most of the cancers in the study (lung, larynx, pharynx, stomach, colon and rectum, and ovary) are associated with tobacco smoking, suggesting that adjustment for tobacco smoking is needed in analyses of association between chrysotile exposure and cancer mortality.2 Smoking patterns are influenced by various factors including occupational status and tobacco control policies such as full or partial workplace bans; yet individuals may compensate for workplace smoking restrictions by smoking more before or after work.3 4 Therefore, relying on national smoking surveys to estimate the potential impacts of smoking in industrial cohorts may not be reflective of workers’ individual behaviour. As cohort members included those who were already employed in 1975 and workers who were hired between 1975 and 2010, we conducted an independent survey of smoking behaviours in JSC Uralasbest among active and retired workers which would largely represent cohort members. Hence, in this analysis, we report the results from a cross-sectional survey on smoking behaviour to assess (1) the overall smoking prevalence in this special workforce and (2) whether smoking patterns vary by (2a) age and sex among active and retired workers and (2b) by categories of cumulative occupational dust exposure among those that could be linked to the cohort via names and date of birth.
Methods
Survey administration
The smoking survey was conducted between January and June 2017. Paper questionnaires were distributed among active and retired workers of JSC Uralasbest. Before distribution, the people responsible for administering the questionnaires were trained by an occupational safety engineer to ensure consistency in the delivery and completion. Details of the survey’s distribution were recorded, including meeting dates, workplace, number of attendees, questionnaires returned, details of administering personnel and date of questionnaire transfer to the local data entry team.
The Asbest Chrysotile Cohort is a historical cohort study where the enrolment was based on company records and the follow-up was based on record linkages from official sources. Therefore, no individual contact with or consent from study participants was required. Participation in this survey was voluntary and could be anonymous, so no consent forms were obtained.
Study population
Active workers were invited to voluntarily participate in the survey during the mandatory occupational safety meetings at JSC Uralasbest, and questionnaires were administered to retired workers during events arranged by the Veterans Council for its ~6000 members. Pensioners who did not attend any organised events were contacted by a Veterans Council assistant at home. The assistant telephoned the pensioner to arrange a time to visit them, or went directly to their home and knocked on the door. When it was not possible to make a home visit, the survey was completed over the phone (figure 1). Of the total of 8090 completed questionnaires, we excluded 665 due to incomplete information (eg, incoherent information, missing sex or age) and duplicate questionnaires (if worker participated in two meetings or events), resulting in 7425 respondents (response rate, 74%) in the overall analyses, of which 40% were active and 60% were retired workers. For the analysis of smoking in relation to cumulative dust exposure, 4583 respondents (46% of survey participants) could be linked to the individual exposure information in the Asbest Chrysotile Cohort because survey participants identified themselves by providing their full names on their returned survey questionnaires (figure 1).
Flow chart describing the data collection events in the JSC Uralasbest tobacco survey. JSC, joint stock company.
(A) Smoking prevalence among respondents (n=7425) in the JSC Uralasbest Tobacco Survey by birth decade and sex, and (B) percentage of ever-smokers (n=2183) who reported smoking an average of >10 cigarettes per day, by birth decade and sex. JSC, joint stock company.
Survey questionnaire
The aim was to determine smoking behaviour, and this was ascertained using three questions: ‘Have you smoked at least 100 cigarettes in your entire life?’, ‘Have you ever been a daily cigarette smoker, for a period of at least half a year?’ and ‘Do you currently smoke cigarettes on a daily basis?’ If a respondent answered ‘yes’ to any of the smoking questions, they were categorised as an ever-smoker. If they were an ever-smoker, starting age, stopping age and number of cigarettes smoked in a typical day were in addition ascertained. Respondents were also asked for their date of birth, sex, work history (eg, age at start, years worked in total, and main profession at JSC Uralasbest and at other enterprises) and to optionally include their full names (first, patronymic and last names).
All answers from paper questionnaires were entered into a computer database by the data entry team. To assess the data entry quality, 57 problematic (eg, containing contradictory information) and 40 randomly selected questionnaires were scanned and sent to the International Agency for Research on Cancer (IARC) for double entry and comparison with the original data entry.
Occupational dust exposure
The exposure assessment and assignment for the workers in the Asbest Chrysotile Cohort is derived from a large database of dust measurements.5 Measured occupational dust concentrations were systematically collected by the company’s central laboratory across the factories (from 1951) and the mine (from 1964). An industry-specific job exposure matrix was constructed based on the measurements. When measurements were not systematically collected or available for a given period, exposure was modelled and extrapolated to cover all relevant years for all jobs. Estimated annual dust exposures by occupation were linked to each cohort member by job history within the company.
Statistical analyses
Participants were categorised as never-smokers or ever-smokers (smoking status) based on self-reported smoking behaviour. Trend tests in ever-smoking prevalence across birth decade (≤1940, 1941–1950, 1951–1960, 1961–1970, 1971–1980, ≥1981) were assessed. Logistic regression models were then fitted to estimate odds ratios (ORs) and 95% CIs for the relationship between smoking status and cumulative dust exposure. Continuous cumulative dust exposure was categorised into four categories: non-exposed (no occupational dust exposure) and terciles (<49.8, 49.8–97.3, >97.3) of cumulative dust exposure (mg/m3-years). The terciles were assessed with cumulative dust exposure considering all professionally exposed workers. Models were adjusted for birth year (<1950, 1950–1970, >1970) and were performed separately by sex and by work area (mine/autotransport/external rail/explosion unit, enrichment factory/central laboratory, or both). Tests for linear trend in ORs were conducted treating the categorised cumulative dust exposure variable as an equally spaced ordinal variable in the logistic regression models. Associations between smoking and exposure to dust were also investigated using the exposure as a binary variable (exposed vs non-exposed). Homogeneity of ORs in those who were exposed was examined using a Wald-type test statistic comparing the model with a linearly restricted one, testing the null hypothesis of similar ORs for all exposed.
All analyses were performed using R statistical software V.3.4.3 (R Foundation for Statistical Computing, Vienna, Austria). P values are two sided, and the significance level was set to 0.05.
Results
Smoking prevalence by birth decade and sex
Table 1 describes the characteristics of survey participants (3692 men and 3733 women), among whom 66% of men and 9% of women were ever-smokers. Among men, smoking prevalence was highest (>70%) in those born in 1961–1980, and slightly lower in earlier and later birth decades (p trend, 0.00015). Among women, the prevalence of ever-smokers increased steadily by birth decade, from <10% in women born before 1960 to 30% in those born after 1980 (p trend <0.0001) (Figure 2A, online supplementary table 4).
Supplemental material
Selected characteristics of respondents included in the analysis of the JSC Uralasbest tobacco survey
Among ever-smokers reporting average number of daily cigarettes (1989 men and 194 women), the majority of men (>70% in men born after 1950) reported smoking >10 cigarettes per day, whereas women generally reported smoking ≤10 cigarettes per day (Figure 2B, online supplementary table 4). The proportion of women smoking >10 cigarettes per day was higher among younger women than among older women.
Quality of data entry
Assessment of the quality of the data entry, that is the double data entry with comparison, demonstrated 100% consistency for questions related to smoking behaviour. Although work history was also correctly entered, this information was not sufficiently detailed to allow linkage to the cohort. Consequently, we restricted the analyses on smoking in relation to occupational dust exposure to respondents who could be linked to the Asbest Chrysotile Cohort via their full name and date of birth.
Smoking prevalence in relation to dust levels
Table 2 reports associations between levels of cumulative dust exposure and the risk of being an ever-smoker, and includes only respondents who revealed their identity on the questionnaire and thereby could be linked to the Asbest Chrysotile Cohort (n=4583 workers), for whom we estimated individual cumulative occupational exposure to dust while working at the company.
OR and 95% CIs of smoking by cumulative dust exposure, using tercile cut-off points, in the JSC Uralasbest tobacco survey
There was a significant positive linear trend between cumulative dust exposure and smoking (p trend, 0.002) for women, but not for men (p trend, 0.44) (table 2). The observed difference in the proportion of smokers in women was between those that were not professionally exposed (3%) versus those professionally exposed to dust (from 7.1% to 7.5%) (p value, 0.006). Among the professionally exposed women, no (positive) trend was seen (p trend, 0.29).
The analyses on smoking in relation to cumulative dust exposure were repeated by work area, that is, for workers in the mine only, enrichment factories only, and both. The results were similar to the overall results but less precise; see online supplementary file, tables 1–3.
When restricting the model to heavy smokers (>10 cigarettes/day) in men (data not shown); (1) no association between dust exposure and smoking was found (p value, 0.23); and (2) no support for heterogeneity of ORs across dust categories in those who were professionally exposed was found (p value, 0.51). Women who reported smoking >10 cigarettes per day were too few to perform this analysis (n=17 ever-smokers).
Discussion
Principal findings
This cross-sectional survey on tobacco smoking among 2993 current and 4432 retired workers of JSC Uralasbest showed consistently high smoking prevalence in the male workforce (ranging from 56% to 72% by birth decade), particularly among those born in 1961–1980. Among women, there was lower but increasing smoking prevalence in recent birth decades (ranging from 4% in early to 30% in late birth decades). Smoking prevalence in women differed significantly between those professionally exposed and those not professionally exposed, and this needs to be accounted for in the forthcoming risk analysis in the Asbest Chrysotile Cohort. Nevertheless, smoking was not associated with levels of cumulative dust exposure in either women or men.
Comparison with other studies
Our results are consistent with three national general population surveys in the Russian Federation conducted in 1996 and 2004 for the New Russia Barometer programme, and in 2008–2010 for the Global Adult Tobacco Survey.6–8 Similar to our findings, smoking prevalence was high and relatively stable among men, with a peak in smoking prevalence (74%) in men born in the 1960s. Among women, increasing smoking prevalence was observed among younger generations. The Russia Longitudinal Monitoring Survey (1992–2003) confirmed that smoking among women has increased, especially in the least educated and in women living in rural areas.9 Although not representative of the Russian population, our results show similar patterns in this specialised workforce, namely employees of a chrysotile-producing industry, which is as such an interesting finding. Both our results and the national surveys, however, should be interpreted with caution for the elderly, because they are more likely to have died if they were heavy smokers compared with never-smokers at the same age, and therefore, smokers may be under-represented in any survey at old age.
Strengths and limitations
Strengths of our survey include a large number of respondents, and the linkage with individual exposure estimates of cumulative dust for almost half of the respondents as they were successfully linked to the Asbest Chrysotile Cohort study. Only 10% of questionnaires from active workers were not used, due to incomplete data. Because these self-administered questionnaires were filled in during mandatory meetings at the workplace, it cannot be ruled out that some of the workers were conservative in their responses due to the perception that documenting unhealthy behaviours could lead to negative consequences from their employer. This may explain why only 17% of the invited active workers identified themselves by putting their names on the questionnaire, enabling linkage to the Asbest Chrysotile Cohort. Data collection among retired workers took place at meetings, in homes, or via telephone interviews, so the completeness and consistency were generally better and only 3% of the pensioners’ questionnaires were excluded due to poor quality.
Assessing smoking behaviour retrospectively may be subject to recall bias, and perhaps even more so among the eldest age groups who in this survey showed the lowest prevalence of smoking. However, Brigham et al have examined the reliability of retrospective tobacco use measures and concluded that ‘few differences in the reliability of recall were apparent between sexes and between age groups’.10 So we interpret underrepresentation of smokers in the oldest age group rather as a survivor effect because smokers in these age groups might have died more often prior to survey administration than the never smokers.
We analysed the association between smoking status and mode of data collection, finding that women who smoked were more often interviewed at home than those who never smoked (p value, 0.009). One potential explanation could be that smokers experience worse health and therefore do not participate in events of the Veterans Council. A large proportion (~60%) of people included in our survey did not answer the question about current smoking status, which is why we based the’“ever-smoker’ status on a group of questions (see the Methods section).
These survey findings should be used to inform tobacco prevention programmes at JSC Uralasbest, especially because the proportion of female smokers is increasing.
Conclusion
In this survey on smoking behaviours in an industrial workforce, we confirm similar patterns as in published national surveys from the Russian Federation. In addition, we observed that smoking behaviour in current and retired workers at JSC Uralasbest does not appear to be associated with levels of cumulative dust exposure. Hence, in the risk analyses of cumulative dust exposure and cancer mortality, confounding by smoking may not be a major issue even for the tobacco-related cancers. However, adjustment at the age group level is needed to reduce residual confounding and account for smoking patterns among women, which have changed substantially over time. Furthermore, identification of age, sex and calendar time-specific patterns in this workforce being similar to national trends suggests that these patterns are likely to apply to other collectives of workers in the Russian Federation. The observed high smoking prevalence spanning over decades suggests that in general Russian male workforces have high tobacco-related disease burdens.
Acknowledgments
The work by Schonfeld on this manuscript was supported by the Intramural Research Program of the US National Cancer Institute, National Institutes of Health. The authors are grateful to Karen Muller for language editing.
Footnotes
Contributors Material preparation was performed by EVK, SJS, EO, MM, EF and JS; data collection was coordinated by EVK, SVK and IVB; the dust exposure assessment was conducted by EF and HK; and data management and analysis were planned and performed by JS, MT, GB, AO, MM and LB. The first draft of the manuscript was written by AO, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding This survey is funded by the Ministry of Health of the Russian Federation in the framework of the Federal Target Program 'National System of Chemical and Biological Safety of the Russian Federation' of 2009–2014 and of 2015–2020.
Disclaimer Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article, and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization.
Competing interests EVK and SVK reported receiving, on behalf of their institutes and personally through consulting firms, payments from companies to evaluate exposure to asbestos and risk of asbestos-related disease in those workplaces. All other authors have no competing interests to declare. For full transparency, EVK reported participation as an occupational and environmental health expert as part of the delegation of the Russian Ministry of Health at multiple World Health Assembly meetings as well as at the Conference of the Parties to the Basel and Rotterdam Conventions. EVK and SVK reported attending meetings organized by the International Chrysotile Association and reported that all expenses for attendance were paid by their respective institutes.
Patient consent for publication Not required.
Ethics approval The cohort was approved by the IARC Ethics Committee (IEC) in September 2012 (No. 12–22) and the IEC monitors the progress of the cohort study on an annual basis.This survey was approved by IEC in July 2016 (No. 16–24) and by the Federal State Budget Scientific Institution Research Institute of Occupational Medicine in Moscow on 19 October 2016.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement The data are not publicly available.