Of Drowning Children and Doubtful Analogies
26 HASTINGS CENTER REPORT July-August 2019
In an article in this issue of the Hastings Center Report, James Sabin and his colleagues ask how we should we think about “bystander ethics” in a learning health
organization.1 At the outset, I must reject the label “by-
stander ethics.” As understood by Sabin et al., the “by-
standers” here are the investigators actively designing
and conducting interventional research within a learning
health organization, but they are not bystanders at all. To
conceive of them as passive parties creates problems in
the resulting ethical reasoning that I will unpack below.
Nomenclature aside, I agree with Sabin and his colleagues
that it is important to ask what responsibility investiga-
tors in a learning health organization have to patients
when research—particularly research of which patients
might be unaware—illuminates problematic aspects of
the patients’ care.
Sabin and his colleagues were confronted by this ques-
tion in the midst of designing a randomized controlled
trial, IMPACT-AFib (Implementation of a Random-
ized Controlled Trial to Improve Treatment with Oral
AntiCoagulants in Patients with Atrial Fi
illation).
IMPACT-AFib sought to determine if an educational in-
tervention targeted at patients with atrial fi
illation and
their clinicians reduces underuse of oral anticoagulants.
The researchers also considered whether the study was
Anticoagulant Therapy in Patients with
Atrial Fi
illation (Review),” Cochrane
Database of Systematic Reviews 4, no.
CD XXXXXXXXXX): 3.
11. T. M. Scanlon, What We Owe to Each
Other (Cam
idge, MA: Belknap Press of
Harvard University Press, 1998), 224-26.
12. P. Singer, “Famine, Affluence, and
Mortality,” Philosophy and Public Affairs 1,
no XXXXXXXXXX): 229-43, at 231.
13. T. A. Lieu and R. Platt, “Applied
Research and Development in Health
Care—Time for a Frameshift,” New Eng-
land Journal of Medicine 376, no XXXXXXXXXX):
710-13.
14. J. E. Sabin and D. Cochran, “Con-
fronting Trade-offs in Health Care: Harvard
Pilgrim Health Care’s Organizational Ethics
Program,” Health Affairs 26, no XXXXXXXXXX):
1129-34.
15. S. Y. H. Kim, “Clinical Trials without
Consent?,” Perspectives in Biology and Medi-
cine 59, no XXXXXXXXXX): 132-46.
16. M. P. Kerlin et al., “A Randomized
Trial of Nighttime Physician Staffing in an
Intensive Care Unit,” New England Journal
of Medicine XXXXXXXXXX): 2201-9.
17. Kim, “Clinical Trials without Con-
sent?,” 140.
18. L. Shepherd and R Macklin, “Ero-
sion of Informed Consent in U.S. Re-
search,” Bioethics XXXXXXXXXX): 4-12.
19. S. Y. H. Kim and F. G. Miller, “In-
formed Consent for Pragmatic Trials—the
Integrated Consent Model,” New England
Journal of Medicine 370, no XXXXXXXXXX):
769-72.
20. E. Kodish, J. D. Lantos, and M.
Siegler, “Ethical Considerations in Ran-
domized Controlled Clinical Trials,” Cancer
65, supplement (1990): XXXXXXXXXX.
21. Institute of Medicine, To E
is Hu-
man: Building a Safer Health System (Wash-
ington, DC: National Academies Press,
2000).
22. Institute of Medicine, Crossing the
Quality Chasm: A New Health System for the
21st Century (Washington, DC: National
Academies Press, 2001).
23. A. Gutmann, “The Bioethics Com-
mission on Incidental Findings,” Science
XXXXXXXXXX): XXXXXXXXXX.
24. H. S. Richardson and L. Belsky, “The
Ancillary-Care Responsibilities of Medical
Researchers,” Hastings Center Report 34, no.
1 (2004): 25-33, at 29.
25. Faden et al., “An Ethics Framework
for a Learning Health Care System.”
26. American Medical As-
sociation, “AMA Principles of
Medical Ethics,” 2001, https:
www.ama-
assn.org/about/publications-newsletters
ama-principles-medical-ethics.
27. E. A. Largent, S. Joffe, and F. G.
Miller, “Can Research and Care Be Ethi-
cally Integrated?,” Hastings Center Report
41, no XXXXXXXXXX): 37-46.
28. A. L. Simpkin and R. M. Schwartz-
stein, “Tolerating Uncertainty—the Next
Medical Revolution?,” New England Journal
of Medicine XXXXXXXXXX): XXXXXXXXXX.
29. Faden et al., “An Ethics Framework
for a Learning Health Care System,” S25.
Other Voices Of Drowning Children and Doubtful
Analogies
by EMILY A. LARGENT
Emily A. Largent, “Of Drowning Children and Doubtful Analogies,” Hast-
ings Center Report 49, no XXXXXXXXXX): XXXXXXXXXXDOI: XXXXXXXXXX/hast.1032
July-August 2019 HASTINGS CENTER REPORT XXXXXXXXXX
eminiscent of the Tuskegee experiment, the U.S. Pub-
lic Health Service’s infamous study of untreated syphilis.2
(Given the contempt in which Tuskegee is held, the anal-
ogy to Tuskegee is striking; given the facts, it is rightly
ejected.)
As those overseeing the IMPACT-AFib trial saw it,
the ethical problem was that, whereas some patients with
atrial fi
illation would receive education about the im-
portance of anticoagulation, others would not. Though
Sabin and his colleagues repeatedly stress that the edu-
cational intervention under study in IMPACT-AFib was
in equipoise, they also seem to labor under a therapeutic
misestimation. If the study was in equipoise, then there
was no scientific justification for believing that patients
in the active group would be better off than those in the
control group. The very aim of the trial was to make that
determination.
Yet, wo
ied about harm that might befall patients in
the control group and fearing that they would be negli-
gent bystanders if they knew these patients were at risk
and did nothing, the IMPACT-AFib investigators ad-
opted a “workaround.” In the original IMPACT-AFib de-
sign, the investigators proposed to prospectively observe
a “usual care” control group. In the “workaround,” the
investigators instead identified a “delayed-intervention”
group without ascertaining the group members’ antico-
agulation status and with a plan to store the relevant data
and then retrospectively “trace” their care over the study
period.
The “workaround,” I suggest, was not a solution to
the negligent bystander problem identified by the IM-
PACT-AFib team. Nor was it a solution to the problem as
I would alternatively frame it—how to address instances
of suboptimal patient care identified through research
within learning health organizations. To see why, we must
consider two drowning children.
Sabin and his colleagues use Peter Singer’s classic es-
say, “Famine, Affluence, and Morality”3 to motivate the
moral problem at the core of their essay. They recount
Singer’s persuasive claim that if you (a bystander) see a
child drowning in a shallow pond, you should wade in
and save the child even if your clothes will get muddied.
Then, after describing the IMPACT-AFib “workaround,”
they claim that this revised trial design was “analogous to
choosing not to walk past or look at a shallow pond in
which a child might drown.”4
Let us accept as true the premise implicit in the au-
thors’ analogy: that we do not generally have an obliga-
tion to look for children who might be drowning. This is
at least superficially plausible. Most days, on my way to
work, I cross a
idge over the Schuylkill River, a body of
water in which a child might plausibly drown. Surely, few
would argue that I have a moral responsibility to scan the
water vigilantly as I walk to work, and none would argue
that I have a moral responsibility to always walk to work
over the
idge lest a child in the water need rescuing.
Even if we accept this premise as true, however, the anal-
ogy fails to support the argument being made in favor of
the “workaround.”
Sabin and his colleagues provide robust evidence at
the outset of their essay to demonstrate that oral antico-
agulants are underused in the treatment of patients with
atrial fi
illation. In light of their own evidence, they can-
not claim to be averting their eyes from a pond where
a child might drown. Through their research, they know
that some patients are receiving substandard care and
will come to harm (even if they do not know precisely
which ones). Though nameless to the investigators, these
patients are, of course, real individuals. And, as Singer
himself points out in “Famine, Affluence, and Morality,”
it “makes no moral difference whether the person I can
help is a neighbor’s child ten yards from me” or someone
“whose name I shall never know.” The IMPACT-AFib
“workaround” is more accurately analogized to knowing
that a child is drowning and averting your eyes to ensure
that your clothes do not get soiled. Unfortunately, what
saves one’s clothes does not always satisfy one’s moral
obligations.
This
ings us to another drowning child: a six-year-
old in his evening bath. This child inhabits another classic
essay, James Rachels’s “Active and Passive Euthanasia.”5
Rachels describes two men—Smith and Jones—both of
whom stand to gain a large inheritance if the boy in the
ath dies. Whereas Smith actively pushes the child under
the water, Jones passively watches as the child succumbs
to it. From a moral point of view, Rachels argues, Jones
is no better than Smith. Both men acted from the same
motive, and both had the same end in view.
Similarly, from a moral point of view, IMPACT-AFib
in its revised form is no better than it was as originally
proposed. The “workaround” was a legerdemain. It made
no material difference for the patients: the delayed-inter-
vention group continued to serve as the controls, and the
outcomes of interest continued to be anticoagulant use,
stroke, and bleeding. Same motive. Same end. The ma-
jor difference was that the investigators deliberately kept
themselves from knowing the anticoagulation status of
those in the delayed-intervention group, something that
was knowable by them.
If willful blindness is a successful strategy for avoiding
moral responsibility, one should conclude that the “work-
around” was preferable to the original IMPACT-AFib
design. But does one really want to say that? I think not.
Imagine I was crossing the Schuylkill River and knew a
child was drowning but