We appreciate the feedback and thoughtful discussion offered by Dr Hai and colleagues. We are pleased our paper is generating meaningful conversations as we all move through this uncharted territory of ethical conundrums created by the COVID-19 pandemic. As Dr Hai and colleagues point out, we are certainly aware of the differences between the HIV/AIDS epidemic of the 1980s/1990s and the current COVID-19 pandemic. The COVID-19 pandemic exceeds what any of us have seen, and therefore requires learning from all similar healthcare crises in history. We chose to focus on the AIDS epidemic because it is timelier than the 1918 Spanish Flu pandemic. In addition, we wanted to reference an ethically suitable precedent rather than a clinically similar disease process, although both would have been ideal points of reference. The H1N1 flu did not disrupt clinical care to the same extent as COVID-19, and the Spanish flu was sufficiently long ago that current ethical standards would be quite different from those at the time. Neither public health nor medical ethics had been institutionalized in either medical practice or in residency training a century ago. From a public health perspective, however, both are instructive.Dr Hai and colleagues raise an important distinction when they point out different expectations of duty for medical students and trainees. We wholly agree that medical students are under no obligation to provide care to COVID-19 patients. Residents, however, have formally entered the learned profession of medicine, having implicitly and explicitly dedicated themselves to prioritizing patients' interests above their own. There are certainly limits to such expectations. The same argument that protects attending physicians from attempting “suicide missions” holds true for trainees. Trainees should not be expected to provide care without appropriate personal protective equipment (PPE) or without sufficient safety training. We agree that when the availability of PPE is limited, minimizing the number of encounters with a patient makes practical sense, and therefore might guide institutional measures. One more consideration with regard to the possibility of requiring less of residents than of attendings would be that residents do not yet have the financial security of attendings, and therefore represent a vulnerable sub-group among physicians. We counter that the federal government, the source of funding for residents in training, should take actions like the programs enacted to protect injured first responders after the September 11th attacks on the Twin Towers.We share the urgency expressed by Dr Hai and colleagues in defending a cautious introduction of off-label medications in caring for COVID-19+ patients. We did not intend to suggest any ulterior motive in doing so. However, we did intend to encourage researchers and caregivers to resist a “What have you got to lose?” promotion of therapy or prevention minus reliable scientific foundation. Analogous to keeping one foot braced on solid ice while testing the thinness of the ice ahead, we must not sacrifice our still fragile research standards in our efforts to help frightened patients. Expedited approval—which is established for research that is known not to put participants at risk—is inherently dangerous when the result is relaxed research accountability. One guard against serious therapeutic misconception would be to focus patient discussions on what is not known more than on what might result. We must not forget the failed polio vaccine and the thalidomide tragedies.