Interest
in learning health care systems and in comparative-effectiveness
research (CER) is exploding. One major question is whether informed
consent should always be required for randomized
comparative-effectiveness studies, particularly studies conducted in a
learning health care system. Our answer to this question is no. It will
often be unethical to go forward with CER in which patients are randomly
assigned to different interventions without their written, prospective,
informed consent. However, in a mature learning health care system with
ethically robust oversight policies and practices, some randomized CER
studies may justifiably proceed with a streamlined consent process and
others may not require patient consent at all.
The current
oversight system, requiring informed consent for most clinical research,
grew out of a scandal-ridden period in which people were included in
research and exposed to considerable risk without their knowledge or
consent. In intervening decades, the clinical-research enterprise has
changed. Some research, including some CER, may pose only minimal risks,
yet the potential effect on patients' welfare of answering the core
question of CER — which standard interventions work best for whom — is
immense.
Elsewhere we have presented an ethical justification for
the transition to a learning health care system and for the
streamlining of both consent requirements and oversight practices within
the system.
1,2
A key premise in our justification is that current consent and
oversight practices too often overprotect patients from research that
has little effect on what matters to patients, whereas in other cases
oversight practices underprotect patients from medical errors and
inappropriate medical management because they make research to reduce
these problems unduly burdensome to conduct.
We also have put
forward an ethics framework for learning health care to serve as the
moral foundation for a learning health care system.
2
Our Common Purpose Framework builds on traditional principles of
clinical and research ethics, including the Belmont Report, but is
designed to provide guidance for activities in which research and
practice are integrated to enable rapid, systematic learning. The
Framework comprises seven moral obligations: first, respect the rights
and dignity of patients; second, respect the clinical judgments of
clinicians; third, provide optimal care to each patient; fourth, avoid
imposing nonclinical risks and burdens on patients; fifth, reduce health
inequalities among populations; sixth, conduct activities that foster
learning from clinical care and clinical information; and seventh,
contribute to the common purpose of improving the quality and value of
clinical care and health care systems. The first six obligations fall on
researchers, clinicians, health care administrators, institutions,
payers, and insurers. The seventh falls on patients to participate in
certain types of learning activities that will be integrated with their
clinical care.
Extensive consultation with patients and other
stakeholders is necessary for appropriate specification of the
institutional implications of the Framework. All involved must
appreciate that they are receiving care or working in an institution
committed to the shared mission of continuous learning that feeds
directly into improving patient care. An ethical learning health care
system must have core commitments to engagement, transparency, and
accountability in ways that are keenly sensitive to the rights and
interests of patients. Patients will be engaged in two respects: by
helping to set the CER priorities of the system and by serving on
ethics-oversight panels that will review proposed CER studies in light
of the obligations of the Common Purpose Framework and other ethical
requirements and determine the appropriate forms of consent and
authorization .
In this system, all patients will be told that
patients serve on ethics-oversight panels and how they operate. The
panels will determine whether particular CER (and quality-improvement)
activities fall above or below a threshold of negative effect on
expected clinical outcomes or other outcomes or values that matter
morally to patients. Research that falls below the threshold will be
integrated into clinical care without specific notification to or
consent from individual patients; however, public notification will be
provided to the community of the system, including patients. Other CER
studies, determined by panels to have minor but still meaningful effects
on patients' interests, will proceed with specific notification to
affected patients, who will have an option to decline participation.
Still other studies, determined to be clearly above the threshold, will
require prospective, written, informed consent before proceeding. The
system will thus aim to counteract problems of both underprotection and
overprotection.
Transparent mechanisms will ensure that patients
and other stakeholders can easily learn which CER studies are ongoing.
In addition, and critically, a learning health care system will be
accountable for rapid modifications of clinical practice that are
supported by CER findings and for providing public reasons when
modifications are not made.
In learning health care systems with
these ethically robust practices, it will be ethically acceptable for
some randomized CER studies, having no or only minor effects on
important patient interests, to proceed without informed consent from or
specific notification to individual patients. Consider, for example,
randomized studies that compare the effectiveness of sending medication
reminders by text or e-mail to patients who have previously given
permission to be contacted by either mechanism or the usefulness of
repeating a routine laboratory test once or twice during a patient
hospitalization when both are standard practice. In a mature learning
health care system, an ethics-oversight panel might justifiably approve
the integration of these studies into clinical care routines with only
public notification to the community of the system that the research is
being conducted.
Consider also a pragmatic, randomized clinical
trial that compares two widely used hypertension medications, perhaps
two diuretics, and in which there are no delineated clinical
characteristics that would favor one drug over another for many
patients. Although an algorithm identifies eligible patients, treating
physicians make the final enrollment determination. Physicians and
patients can override the randomized choice. Physicians may change
drugs, adjust dosages, or add therapies for any patient at any time.
This study is unlikely to negatively affect expected clinical outcomes
for patients, and respect for physician judgment is maintained. The
drugs are similar in administration and side-effect profiles, both drugs
have acceptable side-effect profiles, and adverse events are rare. It
is unlikely that patients would have personal preferences for one drug
over the other. This trial therefore accords well with the obligations
in the Common Purpose Framework requirements.
3
In a mature learning health care system of the sort that we envision,
simply telling patients about the study through a streamlined process
and giving them an opportunity to decline participation would be an
ethically acceptable, warranted mechanism of authorization. It may even
be acceptable for an ethics-oversight panel to permit the study to
proceed with broad notification to the community of the system, without
requiring that individual patients be told about the randomization.
However,
some randomized CER studies in learning health care systems cannot be
ethically authorized by either of these mechanisms. Explicit informed
consent will be required if risk, uncertainty, or informational need is
higher. Included would be studies in which the prospect of differential
clinical outcomes or considerable risk looms large as well as studies in
which interventions are different in terms of other considerations that
matter to patients. Consider a study that randomly assigns patients
with back pain to acupuncture or to a home exercise regimen or that
randomly assigns patients with scoliosis to surgery or to bracing. Even
if the alternative treatments were considered standard practice and even
if clinicians were uncertain and evidence was lacking about which is
more effective, the two options have such different implications for
patients' lives that informed consent is essential. Among the critical
functions of having substantial patient engagement in ethics oversight
of CER (and other research) in learning health care is to ensure that
patients' values, beyond their interest in securing the best possible
clinical outcomes, are respected.
Our position that informed
consent is not a morally necessary condition for the conduct of all
randomized CER assumes a learning health care system grounded in a set
of moral commitments against which specific studies have been vetted and
found to satisfy the conditions that permit authorization through
processes other than informed consent. The transformation to a learning
health care system is still in its infancy, and no system on the path to
this important goal has yet to adopt an ethical framework with
accompanying policies and practices of the sort we are proposing.
However, the Common Purpose Framework can provide helpful guidance in
current health care settings. Some randomized CER studies that would
assess favorably against the first four obligations of the Framework
could proceed ethically with a streamlined consent process. These
include studies that, in comparison with what patients would otherwise
encounter in their care, have no expected negative effects on clinical
outcomes or on other considerations that matter to patients.
Consider
now the previously mentioned randomized clinical trial comparing two
similar hypertension drugs to see what authorization approaches might be
justified in the current environment. We suggested that in an ethically
robust learning health care system, characterized by extensive patient
engagement, transparency, and accountability, it would be ethically
acceptable for the study to proceed with a streamlined consent process
and potentially even without specific notification to affected patients.
In the present context, in which morally relevant features of a mature
learning health care system are not in place, proceeding without
specific notification to patients would not be ethically acceptable.
However, it may still be ethically justifiable to use a streamlined
consent process, similar to that suggested by others,
4,5
because the study has no apparent effects on the risks or burdens that
patients otherwise face in clinical care (the third and fourth
obligations), clinician judgment is respected (the second obligation),
and the interventions do not differ on matters of importance to patients
(the first obligation). In the streamlined process, physicians would
inform their patients about the study and the use of randomization.
Their explanations would be brief, akin to the conversation that
physicians typically have with patients about a new prescription, and
accompanied by a short, written description. Patients would be given an
opportunity to opt out of the research and to learn more if they wish,
but patients would not be asked for written informed consent. This
approach could be designed to be respectful of patients and less
burdensome for them and for clinicians than the lengthier process
entailed by current informed-consent requirements, thereby increasing
the numbers of clinicians willing to take part and increasing the
numbers of important clinical questions that can be addressed .
Clinical
research varies widely in the risks to which patients are exposed and
the degree to which research alters the care that patients receive in
ways that matter to them. The importance of streamlining oversight and
consent requirements, so that higher-risk research gets the focused
attention it deserves and less consequential research can proceed more
rapidly, is increasingly being acknowledged. As more low-risk CER is
planned, it will be essential to identify additional, valid
authorization mechanisms, rather than using a one-size-fits-all approach
to informed consent. The transformation to ethically robust learning
health care systems is critical to this goal.
ليست هناك تعليقات:
إرسال تعليق