EPIC working groups
The Statistical Working Group
The availability of data on a variety of exposures (dietary, lifestyle, anthropometric, genetic), assessed through questionnaire-based methods and laboratory analyses, renders the EPIC study an ideal setting to apply complex statistical modelling. The collection on a regular basis of information on individuals’ vital status and cancer end-points, and the ascertainment of incident cardiovascular diseases and diabetes allows a variety of etiological models to be developed and explored. Other than ensuring support to the analysis of epidemiological data with conventional statistical techniques, methodological work on statistics has focused primarily on the following areas:
- Measurement errors in dietary exposure. The measurement error structure of self-reported dietary assessments was investigated through comparison with biomarkers of exposure. Calibration models were developed to correct the diet—disease association in a multicentric context.
- Competing risks analysis. The availability of multiple disease end-points, and of cause-specific mortality data, constitutes the ideal setting for models accounting for multiple outcomes. This is a growing promising area for research, with the ultimate objective of providing estimates of absolute risk of disease exploiting the prospective nature of the EPIC data.
- Analysis of complex sets of data. The adaptation of long-existing multivariate techniques to the analysis of metabolomic profile data represents a valuable strategy. These techniques can be applied to similar large-scale sets of data, i.e. miRNAs, CpG islands (epigenetic), and genomic data.
Resources were also invested in the discovery of dietary patterns, multilevel modelling for the evaluation of between- and within-centre variability of exposure, and of aggregate- and individual-level components in disease models.
The EPIC Statistical Working Group meets regularly with the objectives of coordinating existing activities, stimulating collaborations among statisticians, promoting statistical expertise within the EPIC network, and possibly seeking funding opportunities. Collaborations with worldwide statistical experts are not only established but also continually sought.
- Ferrari P, Carroll RJ, Gustafson P, Riboli E. A Bayesian multilevel model for estimating the diet/disease relationship in a multicenter study with exposures measured with error: the EPIC study. Stat Med. 2008 Dec 20;27(29):6037-54. PMID: 18951369
- Fahey MT, Ferrari P, Slimani N, Vermunt JK, White IR, Hoffmann K et al. Identifying dietary patterns using a normal mixture model: application to the EPIC study. J Epidemiol Community Health. 2012 Jan;66 (1):89-94. PMID: 21875868
- Fages A*, Ferrari P*, Monni S, Dossus L, Floegel A, Mode N, et al. Investigating sources of variability in metabolomic data in the EPIC study: the Principal Component Partial R-square (PC-PR2) method. Metabolomics 2014 Dec 10(6):1074-83.
- Assi N, Moskal A, Slimani N, Viallon V, Chajes V, Freisling H, et al. A treelet transform analysis to relate nutrient patterns to the risk of hormonal receptor-defined breast cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC). Public Health Nutrition doi:10.1017/S1368980015000294. PMID: 25702596
- Agogo GO, van der Voet H, van't Veer P, Ferrari P, Leenders M, Muller DC, et al. Use of two-part regression calibration model to correct for measurement error in episodically consumed foods in a single-replicate study design: EPIC case study. PLoS One. 2014 Nov 17;9(11):e113160. PMID: 25402487
- Sera F & Ferrari P. A multilevel model to estimate the within- and the between-center components of the exposure/disease association in the EPIC study. PLoS One (in press).
Contact details/Working Group leaderPietro Ferrari, PhD
Nutritional Epidemiology Group (NEP)
International Agency for Research on Cancer (IARC/WHO)
150 cours Albert Thomas, 69008 Lyon, France