denotes emergency department, and IQR interquartile range. † Race was determined by the clinical team. ‡ Obesity was defined as a body-mass index (the weight in kilograms divided by the square of the height in meters) of 30 or higher.

Endorsed by the American Association of Critical-Care Nurses and the Society of Critical Care Medicine Public health emergencies have the potential to place enormous strain on health systems. The current pandemic of the novel 2019 coronavirus disease has required hospitals in numerous countries to expand their surge capacity to meet the needs of patients with critical illness. When even surge capacity is exceeded, however, principles of critical care triage may be needed as a means to allocate scarce resources, such as mechanical ventilators or key medications. The goal of a triage system is to direct limited resources towards patients most likely to benefit from them. Implementing a triage system requires careful coordination between clinicians, health systems, local and regional governments, and the public, with a goal of transparency to maintain trust. We discuss the principles of tertiary triage and methods for implementing such a system, emphasizing that these systems should serve only as a last resort. Even under triage, we must uphold our obligation to care for all patients as best possible under difficult circumstances.

Background  Between 5 and 25 April 2009, pandemic (H1N1) 2009 caused a substantial, severe outbreak in Mexico, and subsequently developed into the first global pandemic in 41 years. We determined the reproduction number of pandemic (H1N1) 2009 by analyzing the dynamics of the complete case series in Mexico City during this early period. Methods  We analyzed three mutually exclusive datasets from Mexico City Distrito Federal which constituted all suspect cases from 15 March to 25 April: confirmed pandemic (H1N1) 2009 infections, non‐pandemic influenza A infections and patients who tested negative for influenza. We estimated the initial reproduction number from 497 suspect cases identified prior to 20 April, using a novel contact network methodology incorporating dates of symptom onset and hospitalization, variation in contact rates, extrinsic sociological factors, and uncertainties in underreporting and disease progression. We tested the robustness of this estimate using both the subset of laboratory‐confirmed pandemic (H1N1) 2009 infections and an extended case series through 25 April, adjusted for suspected ascertainment bias. Results  The initial reproduction number (95% confidence interval range) for this novel virus is 1·51 (1·32–1·71) based on suspected cases and 1·43 (1·29–1·57) based on confirmed cases before 20 April. The longer time series (through 25 April) yielded a higher estimate of 2·04 (1·84–2·25), which reduced to 1·44 (1·38–1·51) after correction for ascertainment bias. Conclusions  The estimated transmission characteristics of pandemic (H1N1) 2009 suggest that pharmaceutical and non‐pharmaceutical mitigation measures may appreciably limit its spread prior the development of an effective vaccine.