The paper by Bright and Nelson1 makes several important contributions to the fast-growing literature on medical researchers' ancillary-care obligations. First, by focusing on investigators' day-to-day decision-making needs, it helpfully highlights the need for simple algorithms. The authors' decision tree is the kind of tool that researchers in the trenches will need when dealing with the great variety of ancillary-care needs that can arise in any study, both foreseeable and unexpected. Second, the authors have enriched the discussion by putting on the table a new position on the scope of ancillary-care obligations. Since I think everyone agrees that researchers are not obligated to do things they lack the capacity to do (that ‘ought’ implies ‘can,’ as moral philosophers put it), I would, despite their article's title, call this new position an ‘urgency-based approach.’ Their novel suggestion is that, within the limits of capability, ancillary-care obligations apply when and only when the needs addressed are urgent.2 This differs from an approach based on the duty of rescue, such as that of Merritt and coworkers, which holds that ancillary-care obligations also extend to easily provided interventions, such as disinfecting a wound or supplying de-worming tablets, even when the need for these cannot be counted as urgent.1 It also differs from my partial-entrustment approach, which supports this kind of appeal to easy rescue but holds that, beyond that, researchers have special obligations to provide urgently needed care if the need is …