Re: Ecologic studies--biases, misconceptions, and counterexamples

Changing geographical distribution of diabetes mellitus type 1 incidence in Austrian children 1989--2005

OBJECTIVES

We investigated whether the significant spatial West/East trend of incidence of Diabetes mellitus Type 1 in Austrian children changed between 1989 and 2005. Additionally we tested, whether population density and mean body mass index of newborns is associated with incidence.

STUDY DESIGN AND SETTING

All newly diagnosed cases aged 0-14 years in Austria (n = 2644), prospectively registered from 1989 to 2005 were allocated to 99 districts. Ecological regression and analysis was done by Bayesian hierarchical models including spatially correlated risks.

RESULTS

In Austria, the incidence rose from 9.1/100.000 in 1989--1994 to 14.9/100.000 in 2001--2005. The spatial trend in the period 1989--1995, showing high incidence in the Eastern parts of Austria, disappeared. In the period 2001--2005 a significant inverse negative association of population density with incidence was found. Mean body mass index of newborns in the Austrian population is significantly positively associated with incidence.

CONCLUSION

The observation of higher incidence in rural areas in Austria during recent years may reflect a lower exposure to protective environmental factors, e.g. infections early in life. The observed association with the BMI in newborns indicates that neonatal or prenatal metabolic aspects may modulate the risk for childhood diabetes. We cannot explain the change in the spatial distribution of incidence.

Improved approximations to scaling relationships for species, populations, and ecosystems across latitudinal and elevational gradients

Historically, allometric equations relate organismal traits, such as metabolic rate, individual growth rate, and lifespan, to body mass. Similarly, Boltzmann or Q(10) factors are used to relate many organismal traits to body temperature. Allometric equations and Boltzmann factors are being applied increasingly to higher levels of biological organization in an attempt to describe aggregate properties of populations and ecosystems. They have been used previously for studies that analyse scaling relationships between populations and across latitudinal gradients. For these kinds of applications, it is crucial to be aware of the "fallacy of the averages", and it is often problematic or incorrect to simply substitute the average body mass or temperature for an entire population or ecosystem into allometric equations. We derive improved approximations to allometric equations and Boltzmann factors in terms of the central moments of body size and temperature, and we provide tests for the accuracy of these approximations. This framework is necessary for interpreting the predictions of scaling theories for large-scale systems and grants insight into which characteristics of a given distribution are important. These approximations and tests are applied to data for body size for several taxonomic groups, including groups with multiple species, and to data for temperature at locations of varying latitude, corresponding to ectothermic body temperatures. Based on these results, the accuracy and utility of these approximations as applied to biological systems are assessed. We conclude that approximations to allometric equations at the species level are extremely accurate. However, for systems with a large range in body size, evaluating the skewness and kurtosis is often necessary, so it may be advantageous to calculate the exact form for the averaged scaling relationships instead. Moreover, the improved approximation for the Boltzmann factor, which uses the average and standard deviation of temperature, is quite accurate and represents a significant improvement over previous approximations.

Regional distribution of risk for childhood diabetes in Austria and possible association with body mass index

This study was performed to investigate whether the body mass index (BMI) of diabetic children at manifestation is related to the non-random risk distribution in Austria and whether there is an association with the BMI of the background population. All newly diagnosed cases aged <15 years (n=1644) from 1989 to 2000 were allocated to districts using the postal code. BMI at the time of diagnosis was derived from the Austrian incidence data sheet in 99.3% of the registered cases. Data on BMI of the whole population >14 years of age for the year 1991 and birth weight for the years 1989-1999 were derived from Statistics Austria. Linear regression as well as ecological regression (Poisson) models with correction for non-linearity were used. BMI at diabetes manifestation was positively associated with year of diagnosis (P<0.05) and the age at diagnosis (P<0.01). An increase in BMI when moving from West to East (P<0.01) and a higher BMI at diagnosis in boys than girls (P<0.01) was observed. The ecological regression model showed a positive significant association of BMI in newborns and a positive trend of BMI in the background population with incidence rates.

CONCLUSION

our data support the hypothesis that obesity may be a risk factor not only for type 2 but for type 1 in children acting as an accelerator for the clinical manifestation.

[Methodological problems in health-environment ecological studies]

Different approaches are used to study the effects of the environment on health. We restrict our discussion to observational studies at the aggregated level, the so-called ecological studies. We discuss several sources of bias for group-level studies and consider questions relating to the link between individual-level and group-level dose-effect relationship, the difference between group exposure and environmental exposure, and the influence of measurement error and variability in the exposure. Taking into consideration confounding factors in the analyses is another important item that is discussed. A final item concerns the necessity of studying the temporal direction of the effect, as well as assessing the existence of a potential threshold in the effect. As a broad conclusion, we can say that realistic quantification of uncertainty in dose-effect relationships is a delicate task that requires the systematic consideration of all sources of variability, as well as a transparent sensitivity analysis of the choices and hypotheses made during the statistical analysis.

Biases in ecological studies: utility of including within-area distribution of confounders

This paper is centred on studies commonly known as ecological studies, where one seeks to estimate risks from data aggregated on a geographical basis. The aggregated nature of the data creates difficulties for interpreting, at an individual level, any ecological association found, difficulties which are generically referred to as 'the ecological bias or fallacy'. Here, we address an important component of this bias related to the problem of misspecification in ecological studies. We consider how aggregated level dose-response relationships are derived from integrating individual level ones over the group, and how their correct specification requires, in general, knowledge of the within-group joint distribution of the relevant risk factors, which is rarely available. We discuss in detail the common situation where data on the proportion of persons exposed in each area to several dichotomous risk factors are available. We show that ecological regression estimates of the relative risk for each factor will be improved by including in the regression, besides the linear terms, cross-product terms of the marginal prevalences. Results from a simulation study are discussed and an example concerning the geographical analysis of lung cancer mortality in France is presented.