3628 An investigation of mortality and cancer incidence in United Kingdom oil refinery workers, 1951-2016

Overview:

This technical publication was commissioned by the Energy Institute’s (EI’s) Health Technical Committee. The aim of this technical publication is to investigate the mortality and cancer morbidity experienced by a cohort of 28 548 oil refinery workers. The cohort included male employees first employed at one of eight United Kingdom (UK) oil refineries in the period 1946–1974; all subjects had a minimum of 12 months employment with some employment after 1 January 1951.

Observed numbers of cause-specific deaths were compared with that expected to estimate standardised mortality ratios extended from 1951 to 2016; this is an additional five years of follow-up compared to the previous report (EI 3300). Standardised incidence ratios were calculated for specific types of cancer using an additional four years of follow-up for the cohort (to 2015).

What the report tells us

• The report presents the results of an analysis of deaths and cancer diagnoses amongst a defined cohort of male oil refinery workers during the period 1951-2016.
• The findings are expressed as standardised mortality ratios (SMRs) and standardised incidence ratios (SIRs), which compare the experience of the cohort against that of the general population. The findings are also compared with that of the previous research.
• The use of multivariable Poisson regression models allows for a more detailed examination of the variation in risk by factors such as year of death/cancer registration, period from commencing employment, decade of hire, job type, and level of exposure to company operations. Compared to the previous report, this approach provides a greater level of detail and may help better understand where within the cohort any excess risk may be present.

What the report does not tell us

• The findings relate specifically to the cohort of workers studied over the period covered by the study. They reflect the conditions, exposures and working practices that were prevalent during that period and should be interpreted with care when considering other work environments, other industries, countries or workers employed under more recent conditions.
• Exposure information is limited to job title and a broad classification of exposure to company operations. Individual-level quantitative exposure data including lifestyle and other personal risk factors are excluded from the analysis. Each worker in this study has been assigned a single job type, which is a simplification of what may have been a varied working history. This limits the precision with which job-specific risks can be attributed to particular exposures.
• Where only a small number of cases of a particular cause of death or cancer type are observed within the cohort, the resulting SMR, SIR or relative risk estimate should be treated with caution, because in such cases, the associated confidence intervals will typically be wide, which leads to statistical uncertainty.  Elevated SMRs and SIRs for specific job types/exposure groups may help identify where within the workforce any risk may be located but should not be generalised to the workforce within or outside the cohort.
• Comparisons between the workers in the cohort and the general population are subject to the ‘healthy worker effect.’ Because the general population includes both those who are and those who are not healthy enough to work, whereas workers in the cohort will have been assessed as fit for work, observed mortality and cancer incidence rates in occupational cohorts are typically lower than expected. This means that where results appear close to, or no different from, the general population, this should not automatically be taken to indicate the absence of any occupational risk.

Interpreting changes from the previous report

• This report covers a longer observation period than the previous report (EI 3300), which covered the period 1951-2011. This report extends follow-up by a further five years to the end of 2016. Changes in SMRs or SIRs between reports do not necessarily represent a change in risk.
• As the period of follow-up increases, new findings may emerge that were not apparent in earlier analyses (especially with the introduction of exposed vs minimally exposed groups in this edition) findings. Similarly, findings that appeared significant in a previous report may become less pronounced as the cohort ages and its composition changes relative to the general population.
• Differences in SMRs and SIRs between this and the previous report should not be interpreted as evidence that the situation has necessarily got better or worse; instead, the findings of successive reports should be read together as a progressively more complete picture of the cohort's health experience over time.

What the findings mean for the industry

• The findings should be considered in the context of historical working conditions that were prevalent during the period covered in the study. They should not be interpreted as a true reflection of the risks faced by oil refinery workers employed under current conditions. Changes in regulatory frameworks, occupational hygiene standards, health surveillance and engineering controls should be considered.
• SMRs and SIRs that do not reach statistical significance should not be dismissed as a lack of statistical significance does not necessarily mean the absence of risk, particularly in job roles where the number of workers is relatively small.
• Where there is an indication of elevated (i.e. statistically significant) risk for specific job roles or exposure groups, these should be considered carefully, taking into account the specific operational context and the limitations of the study outlined in this note.
• Employers should consult occupational health professionals to determine whether the exposures commonly associated with potentially higher-risk roles have been adequately assessed and controlled.