Journal of epidemiology and biostatistics最新文献

Background: To increase the precision and power of case-control studies, controls are often matched on age and sex, but rarely on other known risk factors. Expensive tests of genetic susceptibility in a case-control study of laryngeal cancer made us examine the effect of matching for smoking and alcohol consumption on the power and potential size reduction of the required control sample.

Methods: According to published smoking and alcohol consumption distributions in laryngeal cancer cases and population controls, we simulated 10000 frequency-matched and unmatched studies. The exposure of interest was distributed according to different scenarios concerning its relation with the disease and with smoking and alcohol consumption. Studies were analysed with multivariable logistic regression.

Results: Matching increased the precision and power in all scenarios. The gain was most pronounced in scenarios assuming moderate confounding by smoking and alcohol consumption. In such scenarios, equivalent precision or power was only obtained with three times as many unmatched as matched controls.

Conclusions: Matching on strong risk factors may increase the precision and power of case-control studies considerably. In studies employing expensive biologic testing, matching on known strong risk factors may be cost-effective more often than previously thought.

Background: To use IgG antibody measurements to detect infection with Helicobacter pylori (H. pylori). one typically defines a cut-off value based on samples of persons presumed to be infected or uninfected. When there are no good 'gold standard' tests to determine infection status, or when laboratory conditions vary, it is useful to have a method based on the IgG measurements themselves to determine infection status.

Methods: We present a two component mixture model to analyse serologic data on H. pylori infection. The mixing proportions correspond to the probability that a latent variable, the true, unknown infection status I of a person, is either 0 (uninfected) or 1 (infected). By using a logistic model for these probabilities, we are able to incorporate covariate information.

Results: The model is applied to IgG data from Shandong, China. The distribution of the true infection status given the IgG value and a set of covariates is calculated using the IgG distribution function. An optimal cut-off point is found by minimising the probability of misclassification for the Shandong data. The optimal cut-off point is slightly lower than the pre-defined one.

Conclusions: We contrast results from the mixture model with results from tabulations and from standard logistic regression that are based on fixed cut-points. The mixture model yields information on the probability that a person is truly infected as a function of IgG levels and covariates. In our data, the mixture model indicates that a slightly lower cut-off value than the pre-defined cut-point 1.0 can reduce misclassification rates.

Background: Policies and measures for the prevention of human immunodeficiency virus Type 1 (HIV-1) transmission require adequate information about the risk profile of AIDS which is time-, place- and population-dependent. We have studied the risk factors for AIDS among men in Greece, a country with relatively low incidence of AIDS.

Methods: A case-control study of all male patients with incident disease, who have been diagnosed in the major university-affiliated, AIDS Unit from February 1995 through August 1997 was conducted in Athens, Greece, a country with relatively low incidence of AIDS. Eighty-three AIDS patients were enrolled and an equal number of orthopaedic patients as controls. All interviews were conducted by the same physician and took place in the hospital.

Results: There were no differences among heterosexual men with AIDS, homo- or bi-sexual men with AIDS, and controls with respect to any socio-economic variable. The odds ratio for AIDS among homo- or bi-sexual men, in comparison with heterosexual men, was 51.5 (95% confidence intervals 21.6-122.7). Blood transfusion, intravenous drug abuse and haemophilia were less important risk factors for AIDS in this study. Condom use was generally very low and there was a tendency for lesser use among men at highest risk for HIV transmission, that is, those with a preference for receptive anal intercourse.

Conclusions: AIDS among men in Greece is mainly driven by homosexual behaviour, but the relatively high proportion of bisexual men and the relatively low frequency of condom use are warning signs for the potential of the epidemic to expand in the future. The relatively low incidence of AIDS in Greece, in comparison with other European populations, may be due to a phase difference in the epidemic, but it could also be due to the traditional role separation of homosexuals in this geographical area, and the easy accessibility of disposable syringes and needles in Greece.

Background: Central and Eastern European countries offer opportunities for studying the health effects of historical and present exposures, as well as the transition to a market economy. A prerequisite for research is the availability of good-quality information. This study was undertaken to describe sources of individual data that are available for epidemiological research in Estonia. Particular attention was paid to the methods of operation of health registries.

Methods: Information was collected during site visits, interviews with registry personnel and from published reports. For health registries, information was specifically requested on data collection, scope of recorded data, quality control, electronic linkage capability and use of data in research.

Results: The authors describe 35 data sources containing individual information on vital status, mortality, morbidity, natality and women's health, health and health care, and occupation. The most important health registries are the cancer registry, with data from 1968, and the medical birth registry, with data from 1992. Computerised cause-of-death information is available from 1983. Electronic linkage can be done with most of the data sources, the main matching variable being the eleven-digit personal identification number. Factors potentially affecting data-quality in health registries are undefined legal basis, scarcity of funding and staff, poor acknowledgement of problems, and rare scientific use of registry

Discussion: Various data sources are available for epidemiological research in Estonia. Thus far, collected data have largely been an under-used scientific resource. In health registries, more attention should be paid to quality control and continuous involvement of researchers.

Background: Organochlorine compounds, including organochlorine pesticides, have been suggested by some, but not all, studies to be associated with female breast-cancer risk. So far, studies relating organochlorine compounds and breast-cancer risk have mainly focused on polychlorinated biphenyls (PCBs) and dichlorodiphenyltrichloroethane (DDT) as risk factors for female breast cancer. This paper examines the hypothesis that environmental exposure to trans-nonachlor (TNC) and oxychlordane (OCD), a major metabolite of the insecticide chlordane, increases the

Methods: A total of 304 histologically confirmed, incident primary breast-cancer patients and 186 histologically confirmed incident benign breast-disease controls were included in the study between 1994 and 1997. Breast adipose tissue not needed for diagnostic purposes was collected and analysed for TNC, OCD and other organochlorine compounds. A standardised, structured questionnaire was used to obtain information on major known, or suspected, risk factors for breast cancer.

Results: The age and lipid-adjusted geometric mean adipose-tissue levels of OCD were similar between the cases [36.4 p.p.b., 95% confidence interval (CI) 34.7-38.2 p.p.b.] and controls (38.0 p.p.b., 95% Cl 35.7-40.6 p.p.b.). The age and lipid-adjusted geometric mean adipose-tissue levels of TNC between the cases (55.5 p.p.b., 95% CI 52.6-58.5 p.p.b.) and controls (58.1 p.p.b., 95% CI 54.2-62.3 p.p.b.) were also similar. There was no association between breast-cancer risk and mean adipose-tissue levels of OCD and TNC. The covariate-adjusted odds ratio (OR) was 0.7 (95% CI 0.4-1.3) for OCD and 1.1 (95% CI 0.6-1.9) for TNC, when the highest quartile was compared with the lowest. The risk also did not vary based on oestrogen or progesterone receptor status or menopausal status.

Discussion: We found no significantly increased risk of breast cancer associated with breast adipose-tissue levels of OCD or TNC; this is consistent with recent epidemiological studies, indicating that environmental exposure to organochlorine compounds does not have an overall significant impact on breast-cancer risk.

Background: Alport syndrome (AS) is a severe hereditary disease usually transmitted as an X dominant trait and involving a mutation of the COL4A5 gene. It leads to end-stage renal failure (ESRF), but this progression is heterogeneous. Mutations of the COL4A5 gene have been characterised in numerous families using molecular biology. Our objective was to evaluate the interfamilial heterogeneity of the disease and to study relationships between mutation types and progression to ESRF in the European Community Alport Syndrome Concerted Action group (ECASCA) registry database.

Methods: We used the frailty model framework. Frailty models have been developed to analyse censored data with non-independent observations. Random effects are introduced in a Cox proportional regression model to take into account the intracluster correlations. In this study, ESRF is considered a censored event and the intrafamilial correlations are taken into account in the frailty models.

Results: These approaches allow us to demonstrate the existence of an interfamilial heterogeneity; the role of the mutation type explains the interfamilial variability. In particular, the results suggest that some mutation types are associated with a higher risk of ESRF for males.

Conclusions: This study shows the importance of characterising the mutation at the molecular level in genetic studies, to understand the relationship between genotype and phenotype. The frailty models constitute an attractive approach in this context, when the phenotype is characterised by a censored end-point.