Dealing with the exposure variable in occupational and environmental epidemiology

Assessing dose-response relationships by cumulative exposures in epidemiological studies

BACKGROUND

If the occurrence of disease monotonically increases with the degree of exposure in an epidemiologic study, a dose-response (or exposure-response) relationship is indicated and facilitates the interpretation that the exposure has a causal role. It is not uncommon, however, that there is some effect in terms of an overall increased relative risk but no clear dose-response relationship.

METHODS

Models presented here show that cumulative exposure, as involving the duration of exposure, is not an adequate parameter when more recent exposure or the intensity of the exposure plays the greater role for the disease outcome.

CONCLUSIONS

In lack of a dose-response pattern by cumulative exposure, the interpretation of an overall increased risk might well be that there is no definite effect. The proper consideration should be, however, that the measure of exposure could be inadequate, suggesting a need for further analyses and evaluations of the material studied.

Methodological issues regarding confounding and exposure misclassification in epidemiological studies of occupational exposures

BACKGROUND

Confounding and exposure misclassification are issues that concern epidemiologists because of their potential to bias results of studies and complicate interpretations. In occupational epidemiology both are routinely raised to argue that an observed result is either a false positive or a false negative finding. Although it is important to consider the potential for limitations of epidemiologic investigations, judgment regarding their importance should be based on their actual likelihood of occurrence.

METHODS

This paper is based on our experience in epidemiologic analyses and a brief review of the literature regarding confounding and exposure misclassification.

RESULTS

Examples of substantial confounding are rare in occupational epidemiology. In fact, even for studies of occupational exposures and lung cancer, tobacco-adjusted relative risks rarely differ appreciably from the unadjusted estimates. This is surprising because it seems the perfect situation for confounding to occur. Yet, despite the lack of evidence that confounding is a common problem, nearly every epidemiologic paper includes a lengthy discussion on uncontrolled or residual confounding. On the other hand, exposure misclassification probably occurs in all studies. The only question is, how much? The direction and magnitude of nondifferential exposure misclassification (the type most likely to occur in cohort studies) on estimates of relative risks can be largely predicted given knowledge on the degree of misclassification, that is, relatively small amounts of misclassification can bias relative risks substantially towards the null. The literature, however, is full of discussions implying that misclassification of exposure is an explanation for a positive finding.

CONCLUSIONS

These comments are not to suggest that all potential limitations for epidemiologic studies should not be considered and evaluated. We do believe, however, that the likelihood of occurrence and the direction and magnitude of the effect should be more carefully and realistically considered when making judgments about study design or data interpretation.

Risk factors for carpal tunnel syndrome in a general population

OBJECTIVE

To determine the individual, physical, and psychosocial risk factors for carpal tunnel syndrome in a general population.

METHODS

Population based case-control study in Marshfield epidemiological study area in Wisconsin, USA. Cases were men and women aged 18-69 with newly diagnosed carpal tunnel syndrome (n = 206 (83.1%) of 248 eligible). Controls were a random sample of residents of the study area who had no history of diagnosed carpal tunnel syndrome (n = 211 (81.5%) of 259 eligible). Cases and controls were matched by age. Telephone interviews and reviews of medical records obtained height and weight, medical history, average daily hours of exposure to selected physical and organisational work factors, and self ratings on psychosocial work scales.

RESULTS

In the final logistic regression model, five work and three non-work variables were associated with risk of carpal tunnel syndrome, after adjusting for age. For each one unit of increase in body mass index (kg/m2), risk increased 8% (odds ratio (OR) 1.08; 95% confidence interval (95% CI) 1.03 to 1.14). Having a previous musculoskeletal condition was positively associated with carpal tunnel syndrome (OR 2.54; 95% CI 1.03 to 6.23). People reporting the least influence at work had 2.86 times the risk (95% CI, 1.10 to 7.14) than those with the most influence at work.

CONCLUSIONS

Carpal tunnel syndrome is a work related disease, although some important measures of occupational exposure, including keyboard use, were not risk factors in this general population study. The mechanism whereby a weight gain of about six pounds increases the risk of disease 8% requires explanation.

Do quantitative exposure assessments improve risk estimates in occupational studies of cancer?

Quantitative assessment of exposure intensity is a difficult process, particularly for jobs held long ago. Despite difficulties, the use of this approach is growing in occupational epidemiology because it is hoped that the estimates will more closely approximate delivered dose than more traditional measures such as duration of exposure. If this assumption is correct, development and use of quantitative exposure estimates should reduce nondifferential exposure misclassification, sharpen exposure-response gradients, and enhance interpretation of study results. In this report, we used two methods to assess the value of quantitative exposure assessments in cancer epidemiology. In one, we surveyed the literature for investigations on occupational cancer that included assessments of both duration and intensity of exposure. The results of this survey indicated that exposure measures based on some measure of intensity of exposure yielded monotonically increasing exposure-response gradients and larger relative risks more often than those based on duration of exposure. Duration of exposure, however, occasionally provided the larger relative risks. In another approach, we found that different measures of exposure to formaldehyde classified subjects quite differently. For example, duration of exposure was unrelated to average exposure and was only weakly associated with exposure intensity or peak exposure. Because different measures of exposure may classify subjects quite differently and because quantitative estimates usually, but not always, yield larger relative risks and sharper exposure-response gradients than other measures of exposure, we believe that the prudent approach in epidemiologic investigations would be to develop quantitative estimates of exposure and to conduct analyses using several different measures of exposure, or combinations such as duration by intensity. Multiple comparisons would, however, increase chance findings. The value of such an approach is twofold. When a true association exists, use of several different measures decreases the chances of an unfortunate selection of an exposure measure that is poorly related to delivered dose, which would tend to produce negative results, and increases the chances of uncovering sharper exposure-response gradients. Use of several exposure measures in investigations that fail to exhibit an association between exposure and disease would be of value because such an approach would provide greater confidence that negative findings were not simply due to exposure misclassification.

Latency analysis in occupational epidemiology

Allowance for prolonged disease induction and latency times is an important consideration in occupational epidemiology studies of cancer and other delayed effects of exposure. Two useful approaches for assessing prolonged induction and latency periods are (1) exposure lagging and (2) considering exposures only within moving time windows. The exposure weighting scheme proposed by Jahr2 to assess exposure burdens is another method that accounts for induction and latency, although not explicitly. These three approaches, which are shown to be special cases of exposure weighting, are illustrated with an analysis of lung cancer mortality among a cohort of workers from an asbestos textile plant.

Design and conduct of occupational epidemiology studies: I. Design aspects of cohort studies

Cohort and case-control studies are two standard approaches for investigating the etiology of occupational diseases. This paper, which is the first of a four-part series, contains a review of the design features of occupational cohort studies. Topics discussed include the basic features of prospective and historical cohort studies, options for defining the cohort, disease incidence ascertainment, and considerations involved in planning an occupational cohort study. Subsequent papers in this series will focus on data analysis of occupational cohort studies and the design and analysis of occupational case-control studies.

Problems with cross-sectional data in research on working environment and health

Occupational turnover due to health related selection might introduce a bias in cross-sectional studies that tends to mask real occupational health effects. People could have changed occupation so that they, when disease occurs and/or the data collection is accomplished, are working in an environment that is irrelevant for the disease in question. The aim of this study was to determine whether there is any difference in morbidity between 'stable' workers and 'changers'. Occurrence of long-term illness was studied on four exposure levels, defined according to the physical demands at work. The study populations comprised 10,487 men and 10,058 women between 25 to 74 years of age, who were interviewed within the scope of the Statistics Sweden Survey of Living Conditions in the years 1977 and 1979-81. Considerable differences in health outcomes were found between stable workers and changers, especially when considering the degree of physical strain at work. Thus, the findings indicate the necessity of detailed recording of occupational histories within the framework of cross-sectional studies, especially when the aim of the investigation is to study and compare health outcomes for workers in occupations with different turnover rates.

Occupation and lung cancer in Shanghai: a case-control study

Occupation was evaluated as a potential risk factor for lung cancer as part of a large population based case-control study conducted in the ten urban districts of Shanghai. A total of 733 newly diagnosed cases of male lung cancer and 760 controls selected from the general population was interviewed to obtain lifetime occupational histories and information on smoking and other factors. Of the approximately 25 major industrial titles examined, significantly raised risks, adjusted for smoking, were found for employment in agricultural production (odds ratio (OR) = 1.6, 95% confidence interval (CI) = 1.0-2.6). A concomitant increase was detected for farmers (OR = 1.6, 95% CI = 1.0-2.5) when 35 major occupational titles were examined. There was a 70% excess among workers in the chemical industry (OR = 1.7, 95% CI = 0.9-3.1) and a significant decrease among textile industry workers (OR = 0.7, 95% CI = 0.5-1.0). Raised risks of 30% to 80% were associated with reported job exposures to wood and coal dusts, smoke from burning fuels, and chemical fumes. Employment categories were also examined for 672 cases and 735 controls among women, but small numbers in many of the industrial and occupational categories precluded detailed analyses. The largest excess risk among women (OR = 5.1, 95% CI 1.3-23.5) was among glass products workers. Although cigarette smoking was the dominant cause of lung cancer among men and a significant risk factor among women in Shanghai, these findings suggest the importance of certain workplace exposures and offer leads to occupational carcinogens.

Indirect methods of assessing the effects of tobacco use in occupational studies

For various reasons, data on smoking are frequently missing, or only partially available, in retrospective epidemiologic studies of occupational risk factors. In such situations, indirect methods may be used to evaluate the magnitude and direction of the potentially confounding effects of smoking. Such an evaluation can be made quantitatively or qualitatively. Here we describe both approaches. A specific problem relates to case-referent studies, where sampling variation in referent selection may limit the possibility of controlling for confounding by smoking, even when smoking data are available. We present data showing that estimates of risk from occupational exposures which are not controlled for smoking may be as accurate as estimates derived after controlling for smoking, when the number of referents is relatively small. The problem of interaction is also discussed. In the absence of smoking data, the investigator has no indication of how smoking and occupation jointly affect disease risk (eg, additively or multiplicatively). The multiplicative model is usually assumed. However, if exposure and smoking act independently (additively), rate ratios are diminished. In such situations, in the presence of negative confounding by smoking, rate ratios may actually even be less than one--also when exposure and disease are strongly related.