The AI Journal

In January 1962, when the Artificial Kidney Center at Seattle’s Swedish Hospital became the first facility anywhere to offer chronic, repeated dialysis rather than a single emergency treatment, it had three machines and, by the founders’ own reckoning, roughly sixty regional candidates who would die without one. Belding Scribner’s arteriovenous shunt had solved the technical problem — a reusable access point that no longer sacrificed a vein and an artery with every treatment — and immediately created a rationing problem the technology offered no help in solving. The center’s physicians declined to make the selection themselves and asked the King County Medical Society to appoint a lay committee instead: a lawyer, a minister, a homemaker, a state official, a banker, a labour leader and one physician serving, pointedly, as “doctor-citizen” rather than clinician. Shana Alexander’s 1962 account for Life, “They Decide Who Lives, Who Dies,” made the seven anonymous members briefly the most discussed judges in the country, and gave the committee the name it has kept ever since: the God committee.

The scandal, as it has been told for six decades, is the criteria. Alongside medical suitability, the committee weighed marital status, number of dependants, church attendance, employment record and an assessment of the candidate’s prospective usefulness to the community — a “social worth” standard that predictably rewarded married, employed, churchgoing men and penalised everyone who was not. David Rothman’s history of the episode, Strangers at the Bedside (1991), treats it as the founding event of modern bioethics: the moment medicine, faced with a life-and-death allocation it could not defend on medical grounds alone, handed the defending to people with no medical training and no stake in the outcome. Every subsequent retelling repeats the moral: a committee should never have been asked to rank human worth, and the criteria it used were exactly as bad as they sound. Both claims are correct. What gets lost in repeating them is the question of what replaced the committee, and whether what replaced it is actually better.

The proximate answer is that Congress abolished it. Public outcry over the Seattle criteria fed a decade of advocacy — congressional testimony from dialysis patients, position papers from the young field of bioethics, lobbying from nephrologists who wanted the rationing decision taken out of their waiting rooms entirely — that culminated in Section 2991 of the Social Security Amendments of 1972, which made anyone diagnosed with end-stage renal disease automatically eligible for Medicare coverage regardless of age or income. Richard Rettig’s account of the entitlement’s passage (1991) describes it as the first, and so far only, disease-specific universal entitlement Congress has ever enacted, and it was understood at the time as a direct repudiation of Seattle: no committee would ever again decide who received dialysis, because everyone who needed it would receive it. Renée Fox and Judith Swazey’s contemporaneous fieldwork among transplant and dialysis physicians found that by 1978 the profession had absorbed the lesson completely — denying dialysis to any ESRD patient on grounds other than medical futility had become, in the prevailing view, morally indefensible.

That self-congratulation did not survive contact with the patients the entitlement actually served. Coverage removed the financial barrier but did nothing to remove the clinical judgment calls that scarcity, ability and expected benefit still forced on frail, elderly and multiply comorbid candidates, and physicians kept making those calls — quietly, individually, without a published criterion or a named committee to defend them. Leslie Rothenberg’s 1992 history of withholding and withdrawing dialysis from elderly patients documents exactly this: informal age- and prognosis-based rationing persisting through the 1970s and 1980s inside a system whose founding premise was that rationing had ended. No journalist could write Alexander’s exposé about it, because there was no committee roster to publish and no criteria document to quote. The decision simply migrated from a body that had to state its reasons to individual clinicians who did not.

It has kept migrating. Bjorg Thorsteinsdottir, Keith Swetz and Jon Tilburt’s 2013 analysis of dialysis in the frail elderly describes a system now governed less by any articulated ethical standard than by Medicare reimbursement rules — incentives that rewarded initiating and continuing treatment for years, producing what they call routine overtreatment near the end of life, with a reversal already visible on the horizon in which changed payment formulas would begin rewarding the opposite. Their diagnosis is not that today’s clinicians are worse people than Seattle’s committee; it is that the entire arrangement lacks the transparency and shared decision-making that would let a patient, a family or an outside observer see what standard is actually being applied, or by whom. Whichever direction the incentives point, no one voted on it in public and no one has to answer for it in a magazine article.

This is the comparison the standard telling never makes, because it requires crediting the God committee with something. Its members were wrong about what counted as worth, and the outcome for the patients they turned away was exactly as final as any outcome dialysis rationing produces. But they were wrong in a form that could be named, quoted and refuted, and was — within a decade of Alexander’s article, “social worth” as an explicit criterion for withholding a life-sustaining treatment had become professionally indefensible, a discrediting fast enough that no other rationing standard proposed since has dared state itself so plainly. That speed was possible only because the standard existed somewhere to be read.

The obvious reply is that this defence proves too little and risks proving something monstrous: a transparent panel that had, say, refused treatment to unmarried patients would still have killed unmarried patients, and no later correction restores what was already lost. Publishing a bad criterion does not make the harm it causes acceptable, and there is no guarantee that a legible standard is on average a fairer one — a documented committee could just as easily entrench a prejudice for decades under cover of due process. That objection is right on its own terms, and nothing here should be read as arguing that Seattle’s patients were treated justly, or that visibility is a substitute for a defensible criterion. The claim is narrower: visibility is what makes a bad criterion correctable by people outside the institution applying it, and that mechanism, not any specific rule Seattle used, is what disappeared in 1972. An implicit standard cannot be shown to be wrong in a court of public argument, because it was never stated as a standard to begin with; it can only be inferred, years later, from a pattern of outcomes that a historian eventually assembles from hospital records. The 1972 entitlement did not choose between rationing and no rationing. It chose between a rationing standard that could be attacked and one that could not, and it is the second kind, not the first, that has now been quietly rationing dialysis for longer than the God committee ever operated.