Two Ventilators, Four Patients
You chair the ethics committee at a 220-bed community hospital. A substation fire has cut the region's power, and the hospital has two working ventilators but four patients who could survive with one. Patient A is 34, a single mother of two, admitted with bacterial pneumonia; the clinical team estimates 70% survival to discharge. Patient B is 71, a retired judge whose living will explicitly refuses 'prolonged mechanical ventilation in the setting of advanced COPD'; clinicians say she has perhaps a 25% chance and would likely need a tracheostomy and long-term care. Patient C is 45 and unvaccinated against the pneumococcal disease that caused her ARDS; her survival estimate is 55%. Patient D is 19, severely asthmatic, 80% survival. Two more patients will probably arrive within 12 hours. The hospital has no crisis-standards policy; the state has published only non-binding guidance, and the chief medical officer wants a decision in 90 minutes.
1. Compare first-come-first-served, a lottery, and a 'save the most life-years' rule for allocating the two ventilators. Which is most defensible here, and why?
Read the worked answer
First-come-first-served is procedurally simple and is defensible in ordinary times, but here it rewards whoever deteriorated first, which is arbitrary with respect to everything clinically relevant, and 'arrival time' is a judgment some clinician records, so it invites quiet gaming. A straight lottery treats persons as equals and refuses to rank lives, but it ignores prognosis; applied here it might assign a ventilator to a patient with 25% survival who has documented that she does not want prolonged ventilation, and it discards information that could save lives. A life-years or 'save the most lives' rule, scored by short-term survival probability and secondarily by expected benefit, is what most crisis-standards frameworks adopt and is the most defensible here - but with two corrections. Do not apply expected-years in a way that discounts older or disabled lives as such; score prognosis and patient preference, not age. And compute scores blind to race, insurance, and disability status. A hybrid - lottery among patients above a prognostic threshold - is a defensible compromise.
2. How should Patient C's decision not to be vaccinated affect her priority, if at all?
Read the worked answer
There are arguments on both sides and they should be weighed explicitly. For discounting her: she declined a cheap, safe preventive measure, and luck-egalitarian theories permit people to bear the costs of their own choices. Against, two points. First, voluntariness is an empirical question, not an assumption - access barriers, cost, a contraindication, or medical misinformation may all have driven the decision, so ask before judging. Second, and more decisively, medicine does not ration by deservingness. The professional commitment is to treat the patient in front of you; a triage rule that punishes past behavior turns clinicians into moral investigators and predictably penalizes the poor, the distrustful of medicine, and people with mental illness, all of whom refuse vaccination at higher rates for reasons unrelated to culpability. A rule stating that we score current prognosis and not life history is impartial, transparent, and can be applied without adjudicating each patient's character. Do not adjust her score; her 55% survival speaks for itself.
3. What procedural safeguards should the committee adopt in the ninety minutes available so that the decision is legitimate even if it is contested?
Read the worked answer
Legitimacy in a 90-minute window comes from procedure rather than from getting the ranking exactly right. Four safeguards. Publicity: write the criteria down before applying them, and be prepared to make them public; a rule you would not defend in the newspaper is probably not defensible. Relevance: state the reasons in advance - short-term survival probability, expected clinical benefit, and patient preference as expressed in an advance directive - and explicitly exclude ability to pay, insurance status, disability, citizenship, and social worth. Appeals: designate a clinician not involved in the original scoring to hear a bedside challenge and record the outcome; the committee should not be judge and appellate court at once. Revision: schedule reassessment at 48 hours under a stated rule for reallocating a ventilator that is not helping, because the next surge should not be improvised again. Finally, document the decision in real time, including dissent, and have the treating team apply the rule to its own patient rather than the committee doing it for them.