Format

Send to

Choose Destination
See comment in PubMed Commons below

The future of epidemiology: methodological challenges and multilevel inference.

Author information

1
Norris Cotton Cancer Center, Section of Biostatistics and Epidemiology, Center for the Evaluative Clinical Sciences, Department of Community and Family Medicine, Dartmouth Medical School, Lebanon, New Hampshire, USA. Eric.Duell@Dartmouth.Edu

Erratum in

  • Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2006 Nov;49(11):1179-84..

Abstract

A decade ago there was considerable debate about the appropriate objectives and paradigms of modern epidemiologic research. One concern put forth in these debates was that "risk factor epidemiology" might be forcing our field to focus more on individuals and less on populations and public health. Today, most epidemiologists acknowledge that public health is influenced by both population-level and individual-level determinants. Ecologic studies are valuable tools for generating hypotheses and addressing group-level determinants of disease risk. Traditional risk factor studies and genomic studies have helped establish the multifactorial concept of disease causation. Individual-level studies also have provided the biomedical community with hypotheses that have stimulated research into disease mechanisms that have led to reductions in morbidity and mortality for diseases such as HIV/AIDS, cardiovascular disease, and cancer. Current debates about the role of genomic data in epidemiology and public health mirror the debates about risk factor epidemiology one decade ago. Genomic variation is measured at the individual level, but how this variation is maintained in human populations is a group-level (population) phenomenon that is worthy of epidemiologic investigation in its own right. Multilevel epidemiology seeks to understand multiple levels of inference, from genes to individuals to populations and could combine hypothesis-driven research with aspects of data mining. Multilevel epidemiology calls for the study of health and disease determinants defined at the population level and individual level for a more comprehensive strategy to understanding human disease etiology. With the continued development of multilevel statistical methods and the advent of data mining, the technical constraints of the past will become less relevant to the next generation of epidemiologists who wish to embrace a more multilevel epidemiology.

PMID:
16715181
DOI:
10.1007/s00103-006-1293-9
[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments

    Supplemental Content

    Full text links

    Icon for Springer
    Loading ...
    Support Center