From blood disorders to data disorder. The thwarted path from onco-haematology clinic to occupational cancer epidemiology.

The promise of big data in public health has fueled expectations for leveraging electronic health records (EHRs) to advance epidemiological research, particularly in occupational health. Yet, hospitals-where patients' occupational histories are routinely collected-often do not transform this information into usable epidemiological data. This article examines the organisational and infrastructural barriers that hinder the use of hospital data for studying occupational cancers, focusing on blood cancers in France. Drawing on interviews, questionnaires, and ethnographic observations with onco-haematologists and Occupational Diseases Consultation Centers (ODCCs), we document two key mechanisms: (1) inscription practices, where clinicians collect but rarely record occupational histories in a structured, epidemiologically useful format; and (2) delegation practices, where referrals to specialised ODCCs are erratic and depend on local, ad-hoc relationships. Despite onco-hematologists' awareness of occupational risk factors, their documentation in medical records is fragmented, and referrals to ODCCs are inconsistent, resulting in non-representative data, unfit for epidemiological research. Considering the hospital as a whole as an information infrastructure (Bowker and Star, 2000) allows us to describe how it absorbs substantial information about patients' jobs and potential exposures, both through routine clinical work and specialised consultations. However, we also demonstrate that it converts only a small fraction of this information into standardised data that can be reliably incorporated into epidemiological research infrastructures. This study highlights the need for structural reforms aiming at stronger integration between clinical and occupational health services to bridge this gap. Our findings shed light on the social production of data disorder, which jeopardizes big data promises in the field of occupational health epidemiology. We call for greater attention to organisational and infrastructural constraints in hospital settings.