As in many countries, in France, drug authorization is supposedly carried out by ‘independent’ agencies using ‘evidence-based medicine’. French Health Ministers, however, have recently destabilized this system by intervening directly to remove specific drugs from the market. Drawing upon a study of one such decision concerning antidiabetics, this article refutes an explanation of ministerial intervention in terms of the ‘heroic action’ of politicians. Instead, drawing on concepts from sociology and policy analysis, a generalizable analytical framework is proposed to capture deep shifts in power relations between representatives of four ‘fields’ (medical science, administration, party politics and journalism) which underlie ministerial intervention of this type. Specifically, the claim tested is that power struggles within and between those fields strongly structure actor behavior, but also create contingency. Indeed, in the case studied, these struggles coincided in such a way as to prompt and legitimate ministerial coercion. Rather than perpetuate the illusion of agency independence, social science should therefore shed light upon the intra and inter-field frictions which inevitably occur around the authorization of drugs.