When societies experience health crises today, eyes often turn to epidemiology for explanations. Epidemiological knowledge is used as a basis for decision-making; data created with epidemiological methods are seen as representing truths about reality that are usefulin public health policies. Nonetheless, epidemiology is constituted by scientific practices shaped in specific social and cultural contexts, behind which lie complex processes andassumptions about causation and risk. This becomes particularly visible when scrutinizingthe categories employed in epidemiological research. Since the 1990s, categories of gender, class, and race/ethnicity are used extensively to examine and explain patterns of healthand disease in populations. On the one hand, such categories may be helpful in identifying inequities in health. On the other hand, as history shows in the case of race/ethnicity, they offer alarming possibilities of misuse.The 1945 census was the last to record ethnicity in Sweden, and the 1973 Law on datasecurity made ethnic registering virtually illegal. Swedish population statistics and official health and social registers, comprehensive as they might otherwise be, therefore do not contain information on the ethnic background of the people registered. Still, numerous reports indicate health gaps between majority populations and underprivileged minorities defined in ethnic or racial terms – in itself a crisis of society and modern medicine. This paper explores the epidemiological practices, classifications, and categorisations that examine and purport to explain these gaps in Sweden, a country where ethnic identities are neither counted nor registered.