Geographic variation in cancer incidence in the USSR: estimating the proportion of avoidable cancer.

BACKGROUND. The Soviet Union offers a unique frame in which to study geographic variation in cancer incidence because of its uniform registration system of all newly diagnosed cancer cases throughout its 15 republics and 162 oblasts (administrative units). Variation in cancer rates is stronger when examined by oblasts than it is when examined by republics. In 1986, the age-standardized all-site cancer incidence rate for both sexes, in the Soviet republic having the highest overall rate (Russian Soviet Federative Socialist Republic for males, Lithuania for females) was about twice that of the Soviet republic with the lowest all-site cancer rate (Georgia). Within Soviet oblasts (data are from 1979), the ratio of the highest to the lowest rates was about 4 for all sites, 144 for male oesophageal cancer, and 303 for female oesophageal cancer. The excess incidence in the entire USSR compared with the republic having the lowest all-site incidence rate was 59% for males and 49% for females. CONCLUSION. Although the observed differences between republics or between oblasts are partially due to uneven quality of data recording or to statistical variability related to population sizes, there is a strong geographic variation of cancer incidence rates in the USSR that suggests a potentially important role of environmental factors in cancer etiology.     

A flexible modeling approach to estimating the component effects of smoking behavior on lung cancer

OBJECTIVE: Despite the established causal association between cigarette smoking and lung cancer, the relative contributions of age started, duration, years since quitting, and daily amount smoked have not been well characterized. We estimated the contribution of each of these aspects of smoking behavior. STUDY DESIGN AND SETTING: A case-control study was conducted in Montreal on the etiology of lung cancer. There were 640 cases and 938 control subjects for whom lifetime smoking histories were collected. We used generalized additive models, incorporating cubic smoothing splines to model nonlinear effects of various smoking variables. We adopted a multistep approach to deal with the multicollinearity among time-related variables. RESULTS: The main findings are that (1) risk increases independently by daily amount and by duration; (2) among current smokers, lung cancer risk doubles for every 10 cigarettes per day up to 30 to 40 cigarettes per day and tails off thereafter; (3) among ex-smokers, the odds ratio decreases with increasing time since quitting, the rate of decrease being sharper among heavy smokers than among light smokers; and (4) absolute risks demonstrate the dramatic public health benefits of long-term smoking cessation. CONCLUSION: Our results reinforce some previous findings on this issue.     

A note on methods for estimating the number of patients in the nationwide epidemiological survey on intractable diseases

To find the most adequate method for estimating the number of patients in nationwide epidemiological surveys of intractable diseases, we examined six existing methods using reported data on 13 selected diseases. Estimated numbers of patients by the methods, except for methods with theoretical inadequacy or other problems, were almost equal to each other. The simple method, by which the number of patients is estimated as the reported one divided by the response rate, is recommended.     

A bootstrap approach to estimating power for linkage heterogeneity

We examined the power of detecting linkage heterogeneity when the null hypothesis is that all families are linked to one locus (A) and the two alternative hypotheses are either (1) a proportion of the families are linked to locus A and the remaining families are linked to a second locus B or (2) a proportion of the families are linked to locus A or B and a third proportion of the families are unlinked to either locus. The power of detecting linkage heterogeneity is estimated for various proportions of families linked to loci A, B or unlinked to either locus (sampling under the alternative hypothesis). To estimate the significance level, the data set is sampled under the null hypothesis. For sampling under both hypotheses, a bootstrap approach is employed, sampling the simulated pedigrees with replacement. The power to detect linkage heterogeneity is strongest when the recombination fraction is 0 and equal proportions of the families are linked to loci A and B. The power decreases as the recombination fraction increases, the propor-tion of unlinked families increases and the disparity between the proportion of the families linked to either locus A or B increases. In the data set of 32 Duke Familial Alzheimer Disease families, when equal proportions of families are linked to loci A and B, the power to detect linkage heterogeneity is 0.94 using a likelihood ratio criterion of 10:1. The p value that corresponds to the likelihood ratio of 1O:l is estimated as 0.013 with a 95% confidence interval for p ranging from 0.012 to 0.014. © 1993 Wiley-Liss, Inc.     

A comparison of the generalized estimating equation approach with the maximum likelihood approach for repeated measurements

Liang and Zeger proposed an extension of generalized linear models to the analysis of longitudinal data. Their approach is closely related to quasi-likelihood methods and can handle both normal and non-normal outcome variables such as Poisson or binary outcomes. Their approach, however, has been applied mainly to non-normal outcome variables. This is probably due to the fact that there is a large class of multivariate linear models available for normal outcomes such as growth models and random-effects models. Further-more, there are many iterative algorithms that yield maximum likelihood estimators )MLEs( of the model parameters. The multivariate linear model approach, based on maximum likelihood )ML( estimation, specifies the joint multivariate normal distribution of outcome variables, while the approach of Liang and Zeger, based on the quasi-likelihood, specifies only the marginal distributions. In this paper, I compare the approach of Liang and Zeger and the ML approach for the multivariate normal outcomes. I show that the generalized estimating equation )GEE( reduces to the score equation only when the data do not have missing observations and the correlation is unstructured. In more general cases, however, the GEE estimation yields consistent estimators that may differ from the MLEs. That is, the GEE does not always reduce to the score equation even when the outcome variables are multivariate normal. I compare the small sample properties of the GEE estimators and the MLEs by means of a Monte Carlo simulation study.     

A flexible approach for estimating the effects of covariates on health expenditures

Our estimation strategy uses sequences of conditional probability functions, similar to those used in discrete time hazard rate analyses, to construct a discrete approximation to the density function of an outcome of interest conditional on exogenous explanatory variables. Once the conditional density function has been constructed, we can examine expectations of arbitrary functions of the outcome of interest and evaluate how these expectations vary with observed exogenous covariates. We demonstrate the features and precision of the conditional density estimation method (and compare it to other commonly used methods) through Monte Carlo experiments and an application to health expenditures using the RAND Health Insurance Experiment data. Overall, we find that the approximate conditional density estimator provides accurate and precise estimates of derivatives of expected outcomes for a wide range of types of explanatory variables.     

A hierarchical Bayesian approach to age-specific back-calculation of cancer incidence rates.

We propose a Bayesian hierarchical model to estimate age-specific cancer incidence per year from age-specific cancer mortality. The model is based upon the empirical Bayesian approach of Liao and Brookmeyer (1995) and extends that model by consideration of the dependence on age. The incident cases per year are considered as observations from a discrete-time stochastic process following an autoregressive structure within a Poisson regression model. The model assumes that the survival probability among those with cancer is known. We have investigated the sensitivity of the model to the choice of this distribution and have found that this is the most sensitive part of the model. By comparison the predictions of the model are relatively robust to changes in other key areas, such as the number of years an incident case contributes before death, assumptions about parameter equality for identification and the initial prior distributions. The proposed methodology has been investigated using lung cancer mortality data from Scotland. Parameter estimates were obtained through Markov chain Monte Carlo methods, implemented using BUGS.     

EURRECA's approach for estimating micronutrient requirements

In Europe, micronutrient dietary reference values have been established by (inter)national committees of experts and are used by public health policy decision-makers to monitor and assess the adequacy of diets within population groups. The approaches used to derive dietary reference values (including average requirements) vary considerably across countries, and so far no evidence-based reason has been identified for this variation. Nutrient requirements are traditionally based on the minimum amount of a nutrient needed by an individual to avoid deficiency, and is defined by the body's physiological needs. Alternatively the requirement can be defined as the intake at which health is optimal, including the prevention of chronic diet-related diseases. Both approaches are confronted with many challenges (e. g., bioavailability, inter and intra-individual variability). EURRECA has derived a transparent approach for the quantitative integration of evidence on Intake-Status-Health associations and/or Factorial approach (including bioavailability) estimates. To facilitate the derivation of dietary reference values, EURopean micronutrient RECommendations Aligned (EURRECA) is developing a process flow chart to guide nutrient requirement-setting bodies through the process of setting dietary reference values, which aims to facilitate the scientific alignment of deriving these values.     

A simple unified approach for estimating natural direct and indirect effects

An important problem within both epidemiology and many social sciences is to break down the effect of a given treatment into different causal pathways and to quantify the importance of each pathway. Formal mediation analysis based on counterfactuals is a key tool when addressing this problem. During the last decade, the theoretical framework for mediation analysis has been greatly extended to enable the use of arbitrary statistical models for outcome and mediator. However, the researcher attempting to use these techniques in practice will often find implementation a daunting task, as it tends to require special statistical programming. In this paper, the authors introduce a simple procedure based on marginal structural models that directly parameterize the natural direct and indirect effects of interest. It tends to produce more parsimonious results than current techniques, greatly simplifies testing for the presence of a direct or an indirect effect, and has the advantage that it can be conducted in standard software. However, its simplicity comes at the price of relying on correct specification of models for the distribution of mediator (and exposure) and accepting some loss of precision compared with more complex methods. Web Appendixes 1 and 2, which are posted on the Journal's Web site (http://aje.oupjournals.org/), contain implementation examples in SAS software (SAS Institute, Inc., Cary, North Carolina) and R language (R Foundation for Statistical Computing, Vienna, Austria).     

A quantitative approach for estimating exposure to pesticides in the Agricultural Health Study

We developed a quantitative method to estimate long-term chemical-specific pesticide exposures in a large prospective cohort study of more than 58000 pesticide applicators in North Carolina and Iowa. An enrollment questionnaire was administered to applicators to collect basic time- and intensity-related information on pesticide exposure such as mixing condition, duration and frequency of application, application methods and personal protective equipment used. In addition, a detailed take-home questionnaire was administered to collect further intensity-related exposure information such as maintenance or repair of mixing and application equipment, work practices and personal hygiene. More than 40% of the enrolled applicators responded to this detailed take-home questionnaire. Two algorithms were developed to identify applicators' exposure scenarios using information from the enrollment and take-home questionnaires separately in the calculation of subject-specific intensity of exposure score to individual pesticides. The 'general algorithm' used four basic variables (i.e. mixing status, application method, equipment repair status and personal protective equipment use) from the enrollment questionnaire and measurement data from the published pesticide exposure literature to calculate estimated intensity of exposure to individual pesticides for each applicator. The 'detailed' algorithm was based on variables in the general algorithm plus additional exposure information from the take-home questionnaire, including types of mixing system used (i.e. enclosed or open), having a tractor with enclosed cab and/or charcoal filter, frequency of washing equipment after application, frequency of replacing old gloves, personal hygiene and changing clothes after a spill. Weighting factors applied in both algorithms were estimated using measurement data from the published pesticide exposure literature and professional judgment. For each study subject, chemical-specific lifetime cumulative pesticide exposure levels were derived by combining intensity of pesticide exposure as calculated by the two algorithms independently and duration/frequency of pesticide use from the questionnaire. Distributions of duration, intensity and cumulative exposure levels of 2,4-D and chlorpyrifos are presented by state, gender, age group and applicator type (i.e. farmer or commercial applicator) for the entire enrollment cohort and for the sub-cohort of applicators who responded to the take-home questionnaire. The distribution patterns of all basic exposure indices (i.e. intensity, duration and cumulative exposure to 2,4-D and chlorpyrifos) by state, gender, age and applicator type were almost identical in two study populations, indicating that the take-home questionnaire sub-cohort of applicators is representative of the entire cohort in terms of exposure.     

Problems connected with estimating the incidence of tuberculosis infection

Many problems have to be faced in the estimation of an apparently simple but valuable index—namely, the incidence of tuberculosis infection. Very little attention seems to have been paid to these problems so far.     

Observations on the application of EPI cluster survey methods for estimating disease incidence

The present study attempted to assess the incidence of target diseases of the Expanded Programme on Immunization (poliomyelitis, tetanus, measles, pertussis, neonatal tetanus, diphtheria), using cluster samples and a household interview form. The results suggest that this method can indeed serve to estimate the incidence of these diseases with reasonable precision and may also be used to demonstrate reduction in incidence for the more common diseases. Analysis of 37 surveys for poliomyelitis and neonatal tetanus in India revealed a relative uniformity in the design effect (i.e., the ratio of the variance for the cluster estimate to the variance for the binomial estimate) for diseases with low incidence and prevalence. Diseases with higher prevalence tend to have a larger design effect, which may be indicative of the epidemic and “clustered” nature of the disease. A large design effect, therefore, does not necessarily indicate a need for a larger sample size, particularly if precision is acceptable. There is no one single design that is ideal for all surveys of disease incidence and decisions must be made in the light of local conditions and available resources.     

Survey approach to estimating demand for physician assistants

A national probability sample of physicians was surveyed by mail to provide an estimate of the number of job openings for physician assistants (PAs). The responses were used to undertake an empirical analysis of the factors which influence a physician's decision to hire a PA. Estimates were adjusted to account for survey non-response. This approach may be useful for estimating the employment demand for other types of health workers employed by physicians.It was found that in 1976 there was an unconditional demand for 20.338 PAs and a further conditional demand for 3417 PAs. This contrasts sharply with the fact that as of March 1977 only 4845 PAs had graduated from training programs. Approximately 53% of this employment demand was from physicians in the primary care specialties (general/family practice, obstetrics/gynecology, internal medicine, and pediatrics). Of note was the finding that physicians expected to pay PAs a lower salary than was being received in the market place. Further, 27% of physicians not willing to hire PAs indicated lack of knowledge about them as the main reason for not hiring.     

An approach for estimating exposure to ambient concentrations

The degree of certainty in epidemiological studies is probably limited more by estimates of exposure than by any other component. We present a methodology for computing daily pollutant concentration fields that reduces exposure uncertainty and bias by taking account of spatial variation in air quality. This approach, using elliptical influence functions, involves the optimum blending of observations from a monitoring network with gridded pollution fields predicted by the complex air quality model TAPM. Such fields allow more information to be incorporated in the exposure fields used in epidemiological studies, rather than having to assume that ambient exposure is the same across a whole city and/or that individuals remain at the one location for the duration of a study.     

A nationwide study of asthma incidence rate and its determinants in Swedish pre-school children

While many studies on asthma prevalence have been published, the number of studies on asthma incidence in pre-school children is limited. In this project, a nationwide sample of pre-school children was followed with the aim of estimating cumulative 5-year asthma incidence and its determinants. Parents of 5,886 children 1–6?years of age, sampled from day-care centres in 62 municipalities all over Sweden, responded in 2002 to a baseline postal questionnaire with questions regarding symptoms indicating possible asthma or atopic conditions, and a number of other variables. In 2007, parents of 4,255 children responded to an almost identical follow-up questionnaire. Of these, the 3,715 children who were free from asthma at baseline constitute the study population for this report. A large number of potential baseline determinants for cumulative 5-year asthma incidence were identified. Of these, food allergy, rhinitis, incomplete asthma diagnosis criteria (wheezing last 12?months, and ever had asthma but no current symptoms), parental rhinitis, parental asthma, age, and eczema, in ranking order of importance, remained significant in multivariate logistic regression analysis. The asthma incidence rate was highly dependent on presence or absence of these variables, the average annual rate ranging from 2/1,000/year in 6-year-olds with no determinants to154/1,000/year in 1-year-olds with all determinants, corresponding to 11/1,000/year based on the whole study population.     

The validity of estimating the incidence of diabetic retinopathy from prevalence

A method for estimating incidence rates of a disease from prevalence proportions with differential mortality was validated for diabetic retinopathy in a large population of diabetic persons followed for 4 years, 1980-1982 to 1984-1986, in southern Wisconsin. Four-year age- or duration-specific incidence rates of retinopathy were calculated from prevalence and mortality in persons with and those without retinopathy. These rates were lower than the actual rates obtained at the 4-year follow-up examination by 50% or more in most age intervals in both younger and older onset persons. Estimated duration-specific rates were up to 29 and 57% lower than the actual rates in younger and older onset persons, respectively. These discrepancies could not be accounted for by violations of the assumptions underlying the method. In applying the method to diabetic retinopathy, age- and duration-specific incidence may be more appropriate. Such estimated rates agree well with actual rates except in older onset persons with diabetes of short duration.     

A nationwide cohort study on the incidence of meningioma in women using postmenopausal hormone therapy in Finland

The authors conducted a nationwide cohort study to evaluate the association between postmenopausal hormone therapy and meningioma incidence in Finland. All women who had used hormone therapy at least for 6 months at the age of 50 years or older during 1994-2009 were included. Women who had used postmenopausal hormone therapy were identified from the medical reimbursement register of the Social Insurance Institution (131,480 estradiol users and 131,248 estradiol-progestin users), and meningioma cases were identified from the Finnish Cancer Registry. During the average 9 years of follow-up, 289 estradiol users and 196 estradiol-progestin users were diagnosed with meningioma. Ever use of estradiol-only therapy was associated with an increased risk of meningioma (standardized incidence ratio = 1.29, 95% confidence interval: 1.15, 1.44). Among women who had been using estradiol-only therapy for at least 3 years, the incidence of meningioma was 1.40-fold higher (95% confidence interval: 1.18, 1.64; P < 0.001) than in the background population. In contrast, this risk was not increased in users of combination therapy (standardized incidence ratio = 0.93, 95% confidence interval: 0.80, 1.06). There was no difference in risk between continuous and sequential use of hormone therapy. Estradiol-only therapy was accompanied with a slightly increased risk of meningioma.