Recovering true risks when multilevel exposure and covariables are both misclassified.

The authors extend previous results on nondifferential exposure misclassification to the situation in which multilevel exposure and covariables are both misclassified. They show that if misclassification is nondifferential and the predictive value matrices are independent of other predictor variables it is possible to recover the true relative risks as a function of the biased estimates and the misclassification matrices alone. If the covariable is a confounder, the true relative risks may be recovered from the apparent relative risks derived from misclassified data and the misclassification matrix for the exposure variable with respect to its surrogate. If the covariable is an effect modifier, the true relative risk matrix may be recovered from the apparent relative risk matrix and misclassification matrices for both the exposure variable with respect to its surrogate and the covariable with respect to its surrogate. By varying the misclassification matrices, the sensitivity of published relative risk estimates to different patterns of misclassification can be analyzed. If it is not possible to design a study protocol that is free of misclassification, choosing surrogate variables whose predictive value is constant with respect to other predictors appears to be a desirable design objective.

Use of multiple surveys to estimate mortality among never, current, and former smokers: changes over a 20-year interval.

OBJECTIVES: This study sought to demonstrate how data from publicly available large-scale cross-sectional health surveys can be combined to analyze changes in mortality risks among never, current, and former smokers. METHODS: Data from the 1966/68 and 1986 National Mortality Followback Surveys and the 1970 and 1987 National Health Interview Surveys were used to estimate the distribution of never, current, and former smokers among the US population at risk and decedents. Standardized mortality ratios and quotients of standardized mortality ratios were used to estimate mortality risks. RESULTS: Generally, during the period from 1966 through 1986, mortality rates in the United States for most causes of death declined among all smoking groups. However, mortality rates from respiratory diseases increased for current and former smokers. CONCLUSIONS: The reported changes in never and current smoker mortality risks are similar in magnitude and direction to those reported in a previous study based on longitudinal data. The use of combined data from the National Mortality Followback Survey and the National Health Interview Survey offers several advantages as an epidemiological tool.

Linear extrapolation models of lung cancer risk associated with exposure to environmental tobacco smoke.

This paper presents a model to estimate the number of lung cancer deaths due to ETS exposure among the 1992 U.S. never-smoking population, based on downward linear extrapolation from the estimated risks of active smokers. The model uses several recently available data sources including an extensive review of the published literature on indoor concentration of ETS constituents measured under real-world conditions and data from the National Mortality Followback Survey and the National Health Interview Survey which furnish nationally representative estimates of the distribution of the U.S. population and the persons who died from lung cancer by sex, age, and smoking status. The linear extrapolation model estimates that five male and six female excess lung cancer deaths due to ETS exposure would be expected in the 1992 U.S. population of over 52 million never smokers age 35 and over. Explanations for differences between the results of our downward extrapolation model and those of others are presented.

An alternative explanation for the apparent elevated relative mortality and morbidity risks associated with exposure to environmental tobacco smoke.

Insofar as industrial and other blue collar workers are more likely to bring home toxic materials on their person, and also are more likely to smoke than those in other occupations, members of a household are more likely to be subject to paraoccupational exposure and belong to lower socioeconomic strata if the household contains a smoker than if the household does not contain a smoker. Thus observed differences in risk of mortality or morbidity ascribed to ETS on the basis of a comparison of households with and without smokers may be partly or entirely due to differences in paraoccupational exposure and socioeconomic strata. Similarly, differences in mortality and morbidity ascribed to paraoccupational exposure may be partly or entirely due to differences in ETS exposure that are also related to social class and to types of occupation. Unfortunately, there are no data now in existence that could help determine separately the effects of these major confounded variables. There exists, then, a situation in which two explanations are advanced for respiratory diseases among members of a household, each based on similar study populations but focused on different major risk variables: ETS on the one hand, socioeconomic status and paraoccupational exposure on the other. Properly focused investigations need to be initiated.

A critical examination of OSHA's assessment of risk associated with workplace exposure to environmental tobacco smoke.

In response to a request for information on indoor air quality problems, the U.S. Occupational Health and Safety Administration (OSHA) has proposed a rule addressing indoor air quality in general, and especially environmental tobacco smoke (ETS), in indoor work environments. As justification for the proposed rule, OSHA relies on a quantitative risk assessment used to provide estimates of lifetime risk of lung cancer and heart disease associated with workplace exposure to ETS. However, there are a number of concerns regarding the OSHA risk assessment. (i) The form of the underlying mathematical model used in the risk assessment is inappropriate. (ii) OSHA was highly selective in choosing what data values to use in their risk assessment. (iii) Many data values required as input to the OSHA risk assessment model are simply not known at this time. When such values are required, known, but possibly inappropriate, values were substituted. The conclusions arrived at by OSHA on the basis of this risk assessment seem unwarranted.

Correcting standardized rate ratios for imprecise classification of a polychotomous exposure variable with limited data.

The analysis of exposure misclassification has received considerable attention in the epidemiologic literature, with the result that methods for correcting many summary risk estimates for such misclassification are well known. However, the application of such methods typically requires more data than are usually published (for example, the complete set of exposure- and age-specific mortality rates). The authors show, under the assumption that exposure misclassification occurs independently of disease status and confounder level, that it is possible to obtain estimates of standardized rate ratios corrected for a given pattern of misclassification from only the published standardized risk ratios and the misclassification matrix. This technique allows readers of scientific literature to perform post hoc sensitivity analysis of published risk estimates.

Mortality from respiratory cancers (including lung cancer) among workers employed in formaldehyde industries.

A joint study on effects of formaldehyde exposure in industrial populations by the National Cancer Institute and the Formaldehyde Institute, Inc. (Blair et al. [1986]: J Natl Cancer Inst 76: 1071-1084; Blair and Stewart [1989]: J Occup Med 31: 881, Blair et al. [1990]: Am J Ind Med 17:683-700) reported no significant elevation in risk ratios standardized to the general population. Using the same data as Blair et al., we compared more exposed to less exposed workers to compute relative risk for respiratory and lung cancers using a multivariate, log-linear model incorporating factors for job type (hourly vs. salaried), cumulative exposure (0.1-0.5, 0.5-2, 2+ vs. < 0.1 ppm/years), length of exposure (1-5, 5-10, 10+ vs. < 1 years), and age. Models were fit for all workers, all males, all workers less than 65 years of age, and for all males less than 65 years of age. Hourly workers have a significantly elevated relative risk when compared to salaried workers. While only high levels of cumulative exposure show a significant elevation in relative lung cancer risk, trend analyses of the coefficients of a log-linear model show a significant trend of increasing risk with increasing formaldehyde exposure. The significantly elevated respiratory and lung cancer risk for workers younger than 65 may indicate a shift of respiratory cancer mortality towards younger ages among those occupationally exposed to formaldehyde.(ABSTRACT TRUNCATED AT 250 WORDS)

Comments on the Health Effects Institute-Asbestos Research (HEI-AR) report: "Asbestos in public and commercial buildings," with emphasis on risk assessment methods used.

A Health Effects Institute--Asbestos Research Report calculates the risk of exposure to environmental asbestos fibers (EAF) by downward extrapolation from the mortality of workers exposed for 20 years. This extrapolation is improper because 1) relative risks of asbestos exposure very likely are not linearly progressive; 2) the composition of EAF may not be equivalent to that in mining or fabricating; 3) the same environmental asbestos concentration probably represents different exposure doses for different populations; and 4) health effects of asbestos exposure on children, seniors, patients, the institutionalized, the handicapped, and the chronically ill may not be the same as those of healthy workers. Evidence of asbestos-related disease among family members of exposed workers demonstrates that the risk observed for EAF is substantially larger than that estimated from downward extrapolation and suggests a basis for an alternative approach to estimating asbestos-related health risks. Such epidemiologic procedures are well established and ought to form the basis for detecting the health effects of EAF. It is also unclear which industry supports HEI-AR.

Risk attribution and tobacco-related deaths.

The number of deaths that would not have occurred had an exposure or trait been absent is generally estimated by observing mortality rates in sample populations of exposed and nonexposed persons and applying them to the population of interest. Three methods used to estimate deaths due to tobacco use are evaluated. Each method requires estimates of certain absolute and relative risks, and the published estimates based on them assume that the absolute and relative risks observed in the two large American Cancer Society prospective studies can be applied to the US population or to populations in developed countries. Computations using large representative samples of US decedents and of the entire US population for these methods result in estimated numbers of deaths for the US population that are substantially lower than those based on Cancer Prevention Survey-I or Cancer Prevention Survey-II. Computations also showed that controlling for confounding from two smoking-related variables results in still lower estimates of the number of excess deaths. Consequently, published results that ignore confounding and are based on nonrepresentative data overstate the contribution of smoking. It is imperative that estimates of excess deaths be based on representative data and control for relevant confounders.

Computation of relative risk based on simultaneous surveys: an alternative to cohort and case-control studies.

If the same information on the distribution of risk factors is available for both the general population and a subset distinguished by some disease outcome, it becomes possible to derive relative risk estimates applicable to the entire population with the assurance that the data upon which the estimates are based is representative of that population. To illustrate this approach, data from the 1986 National Mortality Follow-back Survey and the 1987 National Health Interview Survey were used to compute rate ratios for several causes of death for work in dirtyier as compared with cleaner occupations by three methods commonly employed in cohort and case-control studies: the usual standardized rate ratio, the Mantel-Haenszel estimate of the rate ratio, and a multiplicative model fit to an appropriate cross-classification. Properly placed questions in appropriate surveys might very well serve as a substitute for cohort studies and could be performed at less cost and with less overall effort, and completed in a shorter time. Moreover, this approach is less subject to problems of representativeness than cohort and case-control studies.

Analysis of the relationship between smokeless tobacco and cancer based on data from the National Mortality Followback Survey.

This study investigates the potential link between the use of smokeless tobacco and oral cancer and cancer of digestive organs. The combined data of the National Mortality Followback Survey (NMFS), a probability sample of the U.S. deaths, and the coincident National Health Interview Survey (NHIS), a probability sample of the living, non-institutionalized U.S. population, are used to compute risk estimates for cancer, oral cancer, and cancer of the digestive organs associated with use of smokeless tobacco based on a cross sectional study design, simultaneously controlled for potential confounding from active smoking, alcohol consumption, and occupational exposure. Use of smokeless tobacco (either as snuff or chewing tobacco) does not increase the risk of oral cancer or cancer of the digestive organs. Alcohol emerges as a major risk factor for oral cancer with a strong dose-response relationship between the amount of drinking and risk. The same is true to a lesser extent for cancer of the digestive organs. Smoking is associated with increased risk of oral cancer but not of cancer of the digestive organs. Blue collar, technical, and service workers have significantly increased risk of cancer of the digestive organs relative to professional, managerial, and clerical workers, but not of oral cancer. Differences between findings based on the NMFS/NHIS and those obtained from other data very likely are due to inadequate control for confounding. Other reasons for differences between the NMFS/NHIS data and other studies are discussed.

Bias in the attribution of lung cancer as cause of death and its possible consequences for calculating smoking-related risks.

Most published calculations of mortality risk, especially those for lung cancer associated with smoking, are based almost exclusively on the underlying cause as recorded on death certificates. Such risk calculations implicitly assume that the conditional probability of recording lung cancer as the underlying cause of death, given that it really is the underlying cause, is the same for all exposure groups. If these probabilities are not equal for all exposure groups, we call the resulting bias a cause of death attribution bias. We analyzed the 1986 National Mortality Followback Survey, a sample of 18,733 U.S. death certificates, and the 1954-1962 Dorn study, a follow-up study of approximately 250,000 holders of U.S. Veterans Life Insurance. Both data sets include information on the smoking habits of decedents and on the underlying and contributing causes of their deaths. We found that lung cancer as an underlying cause is recorded with a much smaller relative frequency if the decedent is known to be a never-smoker and with a much larger relative frequency when the decedent is known to be a smoker. On the other hand, lung cancer as a contributing cause is recorded with a much larger frequency if the decedent is known to be a never-smoker and with a much smaller frequency when the decedent is known to be a smoker. The reverse is true for cancers other than of the lung. There is no similar pattern related to smoking for other causes of death (specifically for myocardial infarction, other chronic ischemic heart disease, diabetes, or cerebrovascular disease).(ABSTRACT TRUNCATED AT 250 WORDS)

A graphical approach to the interpretation of age-period-cohort data.

Two major obstacles to the routine application of age-period-cohort models are (1) the identification problem, and (2) the fact that separate interpretation of the coefficients of the model is seldom possible. We offer a practical solution to these obstacles that involves plotting the relation between the variable of interest and the age, period, and cohort variables in such a manner that nontrivial age, period, or cohort effects are readily recognized as particular types of features in the graph. These features remain recognizable in the presence of normal sampling variability. Examples are given for applying the technique to previously published mortality data.

A practical approach to estimating the true effect of exposure despite imprecise exposure classification.

Accurate information on actual exposure to some possibly toxic agent usually is not available in long-term occupational studies. Any strategy for assigning exposure levels or categories necessarily results in misclassification, where some individuals classified as exposed have no real exposure and some individuals classified as not exposed have some exposure to the agent. Both misclassification errors serve to reduce the estimate of risk associated with exposure. The question arises, "How much does the true risk depart from the observed estimate given an assumed level of misclassification?" This paper quantifies the effect of such misclassification on several forms of standardized risk ratios. Our results express the true risk as a function of the apparent risk based on imprecise exposure classification and parameters representing the proportion of each of the groups that are correctly classified. In any practical situation, the apparent risk can be computed based on whatever classification scheme is being used. On the other hand, the proportions of the imprecisely classified groups actually exposed cannot. However, the investigator may have information or may make assumptions for likely ranges of values for these proportions. Given the apparent risk, estimated true risks can be calculated and plotted or represented in tabular form as a function of the proportions of actual exposure. The resulting graph or table enables the investigator to read off the range of possible true risk values based on what he is prepared to believe or what other information indicates about the range of proportions of misclassified subjects. For instance, results for a typical value of apparent risk of 1.8 show that the true risk may be twice the apparent risk with only 23% misclassification in each exposure group. The value of the true risk that would be necessary to be consistent with a given apparent risk increases rapidly as the extent of misclassification increases. We also show that, if the extent of misclassification is large, the apparent relative risk is close to 1.0 regardless of the actual value of the true risk. Therefore, a small apparent risk does not necessarily indicate that there is no occupational hazard.