Brian Locke

Ethics of Clinical Research

Research Misconduct

OSTP 2000 definition of research misconduct: Falsification, Fabrication, or plagiarism (FFP) in proposing or reporting research

Category is meant to capture things that influence the reliability of research (as opposed to things like authorship disputes that may only affect the authors)

Many questionable practices fall outside of this: many behaviors which, if discovered, would lead to institutional or federal reprimand are more common than FFP. Examples: ignoring human-subject requirements, non-disclosure of conflicts, relationships, failing to present contradictory data or overlooking other's data, changing study design/methodology/results due to pressure from a funding source.

Approaches considering "bad-apples" targeted at only FFP are unlikely to make a difference. Behavior more likely driven by competition, difficult/unreasonable? Regulatory, social, and managerial demands.

Institutions have Research Integrity Officers (RIOs) who receive complaints. Their process is:

  1. Do an inquiry (1st pass with someone who understands the field to know if something is amiss)
  2. Investigation if suspicious - trying to prove that FFP was different from accepted norms, intentional, and at the level of preponderance of evidence (legal definition - less rigorous than beyond a reasonable doubt.. just more likely than not.)

Research of non-human animals

Biologically similar (makes them helpful) but ethically similar (makes the research fraught).

On the spectrum of beliefs about the permissibility of non-human animal research:

  1. No animal research is permissible (e.g. PETA, ALF)
  2. Animal research is always permissible (much of human history)

Middle ground (current paradigm) is that SOME research is permissible on SOME animals - but that we need to:

  • replace: use less human-like animals
  • reduce: use the least number of animals possible
  • refine: do the research in the way to inflict the least amount of suffering

Consider: what is it about animals that confers them ethical standing? Does a human in a persistent vegetative state lack these things?

Singer - speciesism: treating things based on a category that is irrelevant to the question (similar to sexism, racism, nationalism... etc.)

Historically: ALF -> tapes to PETA of baboon experiments at Penn (Unnecessary Fuss) led to facing out of primate research almost entirely.

Rights of animals generally - compare to the amount of suffering in food production. Comparison to farm animals? Much, much smaller scale.

Enforcement through two agencies: NIH Office of Laboratory Animal Welfare vs US Department of Agriculture (which doesn't have jurisdiction of mice). How to enforce improvements? NIH - University relationship is cozy. Perhaps welfare unit should be independent

What makes clinical research ethical?

History:

Nuremberg Code Declaration of Helsinki Belmont Report International Ethical Guidelines for Biomedical Research Involving Human Subjects

Social value of research

Traditionally only considered in terms of participant protection - ensuring favorable risk/benefit by some definition. However, neglect of possible value led to paternalistic undervaluation of benefits in situations such as the HIV/AIDS epidemic.

Re-evaluation also prompted by meta-research showing staggering inefficiency to the research enterprise.

Is it a necessary requirement for ethical research? Does it matter who the funder is? (e.g. can private companies spend their money how they see fit - and if not.. does the production of an expensive drug that is brought to market and may have negative consequence ethical? E.g.

How should it be specified?

How should it be implemented? Current evidence suggests that there is a gap here (e.g. IRBs do not share their processes, nor have ways to appeal judgements, etc.)

Who should be prioritized? Distributive justice: maximizing the health of the worst off and thus ensuring a baseline level of health for evyerone is one approach. Supported by practical consideration of increased efficiency in research for people with lower baseline (e.g. effective altruism, Singer)

Joseph Millum’s (2019) helpful overview of the ethical debate on how proven effective health interventions should be evaluated when priorities in health care are set [ ] Millum, J. 2019. “Introduction to Priority Setting.” Lecture, First-Year Seminar in Philosophy and Bioethics, Department of Bioethics, The Clinical Center, NIH, Bethesda, MD, March 7, 2019.

7 Requiremeents

JAMA. 2000;283:2701-2711

Potential for exploitation occurs whenever some people are placed at risk of harm for the benefit of others.

Emanuel claims these are necessary and sufficient (except 5 and 6, which are procedural)

  1. Social OR scientific value - evaluates diagnostic or therapeutic interventions that might improve health (promising, practical, might be disseminated). Based on wisely allocating finite resources and avoidance of exploitation.
  2. Validity - invalid research is ipso facto unethical; equipoise, good execution, proper method, correctly proposed question.
  3. Fair Subject selection - justice. Avoidance of convenience, concordance with goals, inclusion of those with potential to benefit. If found beneficial, will the population from which the sample is drawn have access to the therapy?
  4. Favorable risk-benefit ratio - non-malficensce, beneficence, non-exploitation. If there are bigger potential downsides to participation, there should be potential benefits that outweigh those. 'Proportionality' is non-quantifiable - 'risk-knowledge' trade-offs (e.g. phase 1 safety) are particularly hard, and utilitarianism is tough to apply.
  5. Independent Review - avoidance of conflicts of interest.
  6. Informed consent - allow for participation or not in, and to ensure research aligns with values. Autonomy. However, even patients without capacity to consent have interests and values (in general, surrogates using substituted judgement or, if impossible, their determination of best interests - though this must be utilized with great caution).
  7. Respect for potential and enrolled subjects - allowed to withdrawal without penalty, welfare monitored throughout the study, should be informed of what was learned from the research.

Independent review and informed consent are procedural requirements: "intended to minimize the possibility of conflict of interest, maximize the coincidence of the research with subjects’ interests, and respect their autonomy"

-> could advance directives regarding research substitute for informed consent?

Thoughts:

Science/Societal value and efficient use of limited resources [ ] this seems like a slippery slope to only studying biggest problems. Are rare diseases unethical to study? Scientific value seems like a slippery one.

1 person has to be torchered for 1 valuable piece of knowledge?

"If there exists a consensus about what is the better treatment, there is no null hypothesis, and the research is invalid." -> it seems like we have a poor track record of consensus being accurate, and in a way that's the whole justification for the research enterprise in the first place.

Individual to society tradeoffs

[ ] for our proposed project: get citation

6 Office of the Secretary. Protection of human subjects: informed consent and waiver of informed consent requirements in certain emergency research; final rules. 61 Federal Register 51498-51533 (1996).

  1. Truog RD, Robinson W, Randolph A, Morris A. Is informed consent always necessary for randomized, controlled trials? N Engl J Med. 1999;340:804-

In emergency settings that preclude time for identifying and eliciting the consent of a proxy decision maker, research can proceed without either informed consent or permission of proxy decision makers when conducted under strict guidelines.6 Most importantly, there should be clinical equipoise— the absence of a consensus regarding the comparative merits of the interventions to be tested.63 In such a case, the subject is not worse off by enrolling.

Need to demonstrate equipose, benefit to the patients who would be consented to replace the mechanism of consent "intended to minimize the possibility of conflict of interest, maximize the coincidence of the research with subjects’ interests, and respect their autonomy". Is individual consent needed in this case?

Need to also argue it is infeasble - time is survival.

Developing countries

What standard of care should be applied?

If proven useful, would the participants have access to the intervention? If not, then the population is being exploited for the benefit of others. Particularly if the reason for undertaking the research in the country is to avoid costs of conducting the research in the US (e.g. Pharma)

Quality of informed consent: issues around coercion, spheres of consent, and disclosure of information. Culture relevant to consent may differ greatly by locale.

e.g. reimbursement for participation: can be ethically justified if the societal value and thus risk/benefit suggests that everyone benefits from the arrangement.

Issue: is there anything special about countries? These same issues could apply within different segments of the population in 1 country.

Risks vs Benefits of Research

"The Common Rule" Requires:

Risks are reasonable in relation to anticipated benefits, if any, to subjects, and the importance of the knowledge that is expected to result. (Ethical principle of Beneficence)

Component analysis: different cconsideration of therapeutic and non therapeutic parts of an investigation (example: administration of the therapy under study -> possible therapeutic benefit to the patient; monitoring of drug levels not used in clinical practice -> not likely to benefit the individual, but will generate knowledge.)

For therapeutic components (re: the patient): should be clinical equipoise, control group consistent with competent, risks reasonable in relation to benefit. This can be considered separately as the protocol for the treatment group (ie risk:benefit of the experimental treatment) and the control group (ie. The risk benefit of no treatment, placebo, or usual care)

For nontherapeutic procedures (re; value to science/society): risks minimized as much as possible without compromising scientific design. Risk is worth possible knowledge gained... with a higher bar for vulnerable populations ("only minor increase over minimal risk encountered in daily life"). Involvement of community representatives should help in this determination (depends on social priorities)

Both components must be acceptable under this framework.

Incapacitated adults: included in vulnerable populations (those who are not able to protect their own interests) under the common rule. Interestingly, an addendum to the common rule prompted by determination that 'deferred consent' was not acceptable, now links the permissibility of research to severity of illness (sicker -> exposure to more risk for the benefit of science). Under the component approach, if there is true equipoise and non-therapeutic interventions are minimized to minimal risk criteria, then the research can be ethical.

What counts as human subjects research?

What investigation does the "Common Rule" apply to? Research involving human subjects that either:

  1. receives federal funding through a department that has accepted the common rule (private research is not within the scope of the common rule. Neither is executive branch entities such as department of labor, or judicial/legislative branch entities such as the general accounting office) - this is the derivation of the name 'common rule'.
  2. Occurs at an institution with Common Rule agency for IRB review (most research institutions fit this bill due to negotiations with DHHS called 'multiple project assurances')
  3. Some agencies have additional restrictions: e.g. FDA

Note: there was an update to the common rule in 2019 that influences some of the details

There is debate about whether research outside the common rule should have similar protections, but legally that is not currently the case.

Research: systematic (=organized) investigation, including (=specifically these things, and others) research development, testing, and evaluation designed (=if something happened unintentionally, then generates knowledge it is not necessarily research... though usage of that information likely would be. Intentions are not knowable which creates a challenge) to contribute to generalizable knowledge.

Human subjects: a living individual that an investigator obtains either data through intervention or interaction OR identifiable private information.

Note: deceased subjects = not human subjects; secondary research with coded info = not human subjects

Some research on humans subjects is exempted (though many IRBs will review it to determine whether it is exempted):

  • observations/surveys of public behavior if the investigator is also participating in that behavior
  • all parts of the research are normal educational practices in the usual setting
  • analysis of existing data/documents/records/specimens from public sources of information or nonpublic sources that are anonymously recorded.
  • analysis of data collected for surveys/interviews/observation of public behavior that is either a.) absolutely confidential, b.) of public officials, c.) anonymized or benign
  • research projects examining the public benefit of service programs approved by agency heads (does not include programs approved by state or local governments)
  • research on human subjects food / food intake.

The 2019 revision adds some exemptions

  • public health surveillance, oral histories.
  • broad consent for submitted info that can then be used for future research endeavors
  • secondary research (ie research that is taken out on data that was collected for another purpose)

Problems with determine what the "Common Rule" applies to:

  • what counts as a 'category' that is not on the exempted list? (We don't have a complete list of the other 'categories' of human subject research to reference..e.g. does research into IRB administrators count as it's own category?)
  • distinction of whether the research is "about" human subjects, or "about something else, but involves people" (e.g. a survey on perceptions about something) is grey.
  • assessing intentions (e.g. to generate generalized knowledge) is complex, and many activities are carried out with multiple intentions (e.g. QI, education)

QI vs research: differentiation as to whether the intent is to generate systematic knowledge or not. (Publishable vs not is NOT a relevant distinction)

Minimal risk exemption (e.g. for vulnerable populations) - the risk must not be different in terms of likelihood or severity compared with daily life.

Checklist

There are 4 options in determination of a project's status viz-a-viz an IRB:

  • not research
  • research, but not human subjects
  • human subjects research, but exampt
  • or human subject research that is subject to IRB review.

Step 1: is it systematic investigation undertaken with intent to create generalizable knowledge? (Meaning, is it research?) Step 2: Does it involve living subjects with whom you collect either samples or identifiable PHI [ ] Is the study about people who are currently alive? [ ] If the study involves samples, does it require some procedure or modification of usual care to obtain? [ ] If the study involves data collection, is the type of information normally private? And if so, could the identity of the person be determined from the information? Step 3: is it not one of the exempted categories (listed above)?

If no to any of those 3, not human subjects research

Conflicts of Interest

Alignment/Conflict of interests:

  • usually, interests align (e.g. discovery and fame.)
  • example conflicts: review grant proposal of former student, owning stock in a company for which you do research.

U of U definition: personal, financial, professional, or political interests that undermine ability to meet obligations. Not inherently wrong or research misconduct,, but dangerous to the extent they compromise credibility/reputation/trust (perception based)

Separated into

  • Tangible: financial - significant = 5000$ or more worth (U of U def. But does not include honoraria, salary). Must be disclosed. Attention focuses here due to quantifiable, impactful.
  • Intangible: personal relationship, intellectual commitments. Widespread but gets less attention.

How to handle:

  • Reduction
  • Elimination
  • Management (disclosure, delegation tasks, monitoring)

Notes from: UNEASY ALLIANCE: Clinical Investigators and the Pharmaceutical Industry THOMAS BODENHEIMER, M.D.

Traditional model: academic medical center, NIH funded, investigator initiated research

Prevention, chronic disease has made trials more expensive and logistically challenging to complete. Often require multiple sites.

Costs: As of 2000, New drug estimate: 300-600 million. Total funding from industry, 6 billion (3.3 billion in the US). 70% of funding comes from industry, 30% from NIH.

3 historical reasons for companies to rely on AMCs:

  1. Design expertise -> now many more employed directly through companies.
  2. Patient access -> increasing referrals in some sectors from private practice.
  3. Prestige of publication

Academic-Industry partnerships often require slow Sponsorship approvals and IRB approval. Thus, commercially oriented networks such as Contract-research Organizations (CRO - develops a network of sites; often the sponsor designs the trial and performs the analysis. May use both academic centers and community practices) and Site-Managemene Organizations (SMOs - often the CRO will contract with an SMO who organizes a network of community partners to generate patient referrals) have proliferated.

Around 2000 AMCs began to try to re-invigorate their partnership. Often, contracts submitted to AMCs must have publication clauses (about who does what, and whether the results will be publishable) that need the AMCs to stand firm on. Primarily because the company's 'output' is primarily an FDA new drug application, where the AMC is interested in peer-reviewed publication.

paradigm: incentivized to find positive 'indications'. OK if transparent data and strong norms around design? Who is incentivized to find negative (ie. Side effects, or improper analysis, poor data quality)? Would removal of those with industry ties from guidelines help?

Ways that drug companies will skew studies to their benefit:

  1. Enrollment of patients healthier, and thus at lower risk of side effects, than those who it is intended to be used on (e.g. NSAIDs primarily studied in those under 65 y/o).
  2. Use of surrogate endpoints that are not properly validated
  3. Suppression of negative results (perhaps not as much of any issue as it used to be with trial registration)
  4. Not allowing access to the data, often even to investigators

'Author non writer, non-author writer' aka guest-ghost syndrome - CRO employs a medical writer ('ghostwriter') who write the manuscript, and the investigator ('guest author') does not analyze the data. The ghostwriter usually receives a packet of materials that they write the article based on.

Counter argument: clinicaltrials gov publication rate is higher for industry funded studies that are registered.

Science and Society:

Dual use dilemma - create something for a good purpose, but might be used for an evil purpose. (e.g. sequencing pandemic influenza)

Approach? Often, some credential required to access the information.

Vannevar Bush model: government invests in basic research because it will ultimately become applied research that can be exploited for the nations benefit. Broad based funding was needed because it couldn't be predicted where discoveries would come from. Investment in researchers was also required to keep pipeline productive. Victory in WWII and Cold War need reinforced this model.

2 approaches:

  1. Conflicts of interests are bad, and thus should be avoided
  2. Many players encourage this: Federal govt (Bayh-Dohl Act - ownership of inventions from federal research can go to indivs/orgs, not govt. Incentivize research), UT (USTAR - state $ to give to university to invest in things that'll stimulate local economy), U of U (startups from academic research: Tech and Venture Commercialization office)

Researcher-Participant Relationship

Ethical research must be;

  • scientifically valid (in order to justify exposure of participants to risks)
  • socially valuable (same justification)

International Researcher

Exploitation: 1 party (predicts) realization of a disproportionate/unfair benefit.

10/90 gap: 90% of research dollars globally spent on diseases that effect 10% of the population. Yet, poverty and social deprivation make populations susceptible to exploitation.

Traditionally: access to fruits of investigation treated as primary criteria to ensure justice.

However, a minimalist approach (meaning, minimal requirements for a trial to be ethical) might weigh the local needs/vulnerabilities of the host population and the capacity for the research to benefit or to burden (essentially, requirement of nonmaleficence and beneficence to the population... positing that something that is nonmalifiscent, beneficent, and respects persons is by definition also just. ). "Fair benefits" approach - if no one is worse off than they would be had the trial not occurred (e.g. control group just has to be better than local standard of care), then it is just.

Critique: this makes the status quo in the host community the normative baseline (when in reality, they may be morally entitled to better conditions than they currently have, and someone may be obligated to provide those conditions ). Additionally, there is a.) not stipulation on the degree of benefit from different parties and b.) no reciprocal requirement that a trialist do a trial that would be mutually beneficial (ie. They can choose to or not - are people with less interesting conditions to the scientific community entitled to less help? Decision to offer come from the interests of the developed world)

Thus, research in a setting where, say, the government is failing to provide basic human rights - a research endeavor that is better than the 'status quo' may still be ethically problematic (if citizens have valid claims against their government, then collaboration with that government is compromised). Especially if there is some claim that the investigator has some responsibility (e.g. from a country that perpetuated the conditions - which might be considered the entire western world. Or, more specifically pharmaceutical companies that lobby against loosening patent protections)

Alternative: Social development criteria. Research is worthy of support (or at least, permissible) if it expands the capacities of the basically social structures of the community such that it can better achieve its own interests (such as community health needs, ability to respond to a current circumstance, etc.). Thus, if a community won't benefit from the intervention under study (ie. Too expensive or requires better healthcare system), it becomes the obligation of the investigator to contribute toward making them able to benefit, or do the research elsewhere..

Social value of knowledge: in order to generate value, there has to be some instrumental use of the knowledge that will benefit society. On the basis of calculated expected utilities of different actions, we can decide whether the expected value of outcomes with the information gained from the trial is likely to improve the outcome in a host community. These estimates have increasing uncertainty with time, thus immediate benefits should be prioritized as ways to justify a trial.

Research that departs from therapeutic norms is not inherently unethical

Uncertainty about treatments themselves vs what to do in each patient: then, it is a myth that trials are run at the cost to current patients for the benefit of future patients. Instead, current patients may benefit as there can be no personalized care in a case where treatment itself is not known to be beneficial. In the face of epistemic uncertainty, randomization offers a hedge.

Therapeutic misconception = patients confuse research participation with clinical attention.

Ethics of clinical research and clinical care differ (though they are conflated in the Declaration of Helsinki): consider that if the interests of the individual patient ALWAYS need to be prioritized over science and knowledge, how would any inconvenience that does not lead to improved individual patient outcome be justified? For example, phase 1 trials? Nontherapeutic interventions for data collection. Indeed, randomization is seeding our normal "decide which treatment we think is best for the patient" for the benefit of generating robust data. Fundamentally, the purpose of clinical research is not to benefit each individual patient.

Thus, maleficence and beneficence are both problematic to directly apply to research. Consider:

  • placebo controls when effective treatments are known. Always prohibited? Consider allergic rhinitis treatment.
  • sham surgery
  • challenge studies. (e.g. Co2 inhalation to provoke panic attacks)

A better framework than prohibition: is exposure to the risk justified in terms of the expected benefit to society (which requires robust assessment of the study design and ability to answer pressing questions to society)

  1. What risks are posted to subjects?
  2. Have the risks been minimized?
  3. Are the risks balanced by medical benefits to the subject? (If not..)
  4. Are these risks justified by potential benefits of scientific knowledge gained by the research?

Our moral intuition is that physicians, who are solely devoted to patient health, do not have this same moral obligation to research subjects (only that on balance, the benefits outweigh the harms and persons are respected).

Critiques:

  1. Without clinical ethics, what is to stop exploitation of individuals for scientific aims under utilitarian principles? The 7 principles above provide a framework to protect against this (and are not synonymous with clinical ethics)
  2. Does withholding a beneficial treatment for possible placebo breach standard of care? Yes (it seems to me, problematic for the referring clinician) - but that the patient can be free to refuse it on their own accord if they have capacity.

Public health vs clinical research vs individual clinical care - public health also shares some of these same principles (some individual harms undertaken for the benefit of the whole)

Fiduciary Relationship

Clinical relationships: fundamental obligation is beneficence, nonmalficense Research relationships: fundamental obligations are avoidance of exploitation, minimization of harm c/w scientific principles.

(Is medical malpractice based on fiduciary relationship? If so, suggests research malpractice may be possible)

Physician-researcher and patient-subject relationship is (possibly, not established by case law) fiduciary: where a person acts on behalf of another, putting their clients interests ahead of their own with a duty to preserve good faith and trust. Connotes both an ethical and a legal obligation.

Note: case law has not definitively weighed in on whether it IS fiduciary, but normatively it OUGHT to be. Criteria:

  1. Fiduciary has scope to exercise discretion or power
  2. Fiduciary can unilaterally exercise that power or discretion to affect the beneficiary's interests
  3. The beneficiary is peculiarly vulnerable to the fiduciary's power or discretion.

Informed consent = then establishes a fiduciary duty, as opposed to just allowing interaction that is not battery.

If a fiduciary relationship is established, ethical obligation to not abuse/misuse power (=exploitive, by definition), duty to manage conflicts between fiduciary and beneficiary's personal interests, protection against carelessness and neglect (though not necessarily assurance of a desired outcome). Main duties: loyalty, discretion, and care.

Equipoise: traditional needed as justification for a physician (clinical obligations) to be morally permitted to enroll a patient in a trial. However, can also be seen as an obligation of the trustee (researcher) in a fiduciary relationship.

IRBs/DSMBs: researcher cannot abdicate patient-protection to these entities if they have fiduciary obligations.