Data between environment and health : an epistemological study of the exposome

Abstract

This thesis is a philosophical analysis of the epistemic role of scientific data in biomedical research. It is comprised of an introduction (Chapter 1), three articles (Chapter 2, 3, 4) and a conclusion (Chapter 5). I use a case study approach and focus on the epidemiology of the ‘exposome’, a new line of research based on a reconceptualisation of exposure and the use of new and diverse datasets. I argue that data can sustain the subject matter of exposome research by shaping concepts, strategies, techniques and what counts as evidence. Yet, the epistemic role of data is enacted by the ways in which it is used by epistemic agents and thus constantly connected to and mediated by other artefacts, components and features of scientific inquiry. In Chapter 2, I discuss the innovations and changes of the exposome. I argue that these should be framed as the establishment of a repertoire, as opposed to a paradigm. The exposome repertoire consists in many components transferred from other areas of the life and health sciences: thus, scientific change is the result of the alignment of these components and it is not due to only one of these factors, such as data. In Chapter 3, I discuss data practices in exposome research. I argue that researchers use evidential claims to specify the evidential and representational value of datasets. Three strategies for evidential claims can be distinguished, differing in terms of level of abstraction, lines of work and type of evidential claim and leading to a picture of evidence production as epistemic-intensive labour. In Chapter 4, I discuss how data is classified as evidence in exposome research, in the context of philosophical discussions of the types of evidence used for causal claims in biomedical research. I argue that molecular data collected in exposome research is used to study differences and dependences, as opposed to mechanisms; more generally, the classification of a dataset as a type of evidence is dependent on the ways in which the data is used, rather than its intrinsic properties.