Post on 30-May-2020
transcript
Robert D. Truog, MD Professor of Medical Ethics, Anaesthesia, & Pediatrics Director of Clinical Ethics, Harvard Medical School Executive Director, Institute for Professionalism & Ethical Practice Senior Associate in Critical Care Medicine, Children’s Hospital Boston
Ethical Issues in Adaptive Clinical
Trials
Adaptive designs: History of ECMO
• Extracorporeal membrane oxygenation is a form of heart-lung bypass for life-threatening cardiac or pulmonary failure
• 1980s: Surgeon Bob Bartlett began to use ECMO for neonatal PPHN with excellent anecdotal results
Adaptive designs: History of ECMO
• But – he realized practices would not change without validation from a randomized controlled trial
• Yet he was conflicted – proving the superiority of ECMO would require him to let some babies die that he believed he could save
Bartlett Article
Bartlett: Play-the-Winner Design
ECMO Survived
CMT Died
10 ECMO: survived
1 CMT: died
Bartlett Article
“The clinical indications for this new and complex treatment remain undefined. Further randomized controlled trials… will be difficult but remain necessary.”
O’Rourke Article
O’Rourke et al, 1989
• Also employed an adaptive design
• Phase I: 50/50 randomization until there are 4 deaths in one arm
• Phase II: Assign all patients to the more successful therapy, until there are 4 deaths in that arm or until statistical significance is achieved
Phase I
ECMO CMT
9 s, 0 d 6 s, 4 d
The Harvard Neonatal ECMO Trial: Results
Phase II 19 s, 1 d
Adaptive Randomization
• In the literature, the trial was criticized from both directions – No patients should have been assigned to CMT – Not enough patients were assigned to CMT
• Perhaps this approach was a good balance – 4 deaths was a psychological “breaking point,”
especially for the nursing staff
• But was it convincing?
The UK Neonatal ECMO Trial
• The existing “RCTs of neonatal ECMO… suggested reductions in mortality but were not conclusive.”
• Because they “used adaptive designs, which may have introduced bias…”
Field et al. UK collaborative randomised trial of neonatal extracorporeal membrane oxygenation. Lancet 1996;348:75-82
The UK Neonatal ECMO Trial
• 1993-1995: 185 neonates randomized to ECMO vs CMT
• Trial stopped early by DSMB, – ECMO survival 60/93 = 65% – CMT survival 38/92 = 41%, p<0.0005
• Were 22 babies unnecessarily “sacrificed”?
1) Social & Scientific Value
• Adaptive designs may influence the choice of the outcome variables – Must be one primary outcome of interest – Must be apparent within a short period of
time – May favor use of biomarkers or other
surrogate markers
2) Scientific Validity
• Frequent analysis of data may make confidentiality of data difficult
• Risk of accrual bias – Researchers may be able to guess study trends – May seek to enroll patients who they think will benefit – Subjects may prefer to enroll only in the later stages
of a trial
• Even if a trial is statistically valid, is it important that the validity be “transparent” to clinicians?
3) Fair subject selection
• Those who enroll later, benefit more – Subjects who do not understand this may be at a
disadvantage to those who do – Will those with superior understanding seek to
“game” the system? – Does this differ from current approach to Phase 1
trials?
• Investigators may find it more difficult to enroll patients to the arm they believe to be less favorable
4) Favorable risk-benefit ratio
• This is the primary advantage of adaptive designs
• For the investigator: Attempts to mitigate the conflict between clinician and investigator
• For the subject: Attempts to minimize number of subjects assigned to the less-successful therapy
5) Independent Review
• I don’t see any problem here as long as options are clearly defined up-front
– Study modifications made after the study begins could pose problems
6) Informed Consent
• How much do subjects need to understand about the study design to provide informed consent?
• Will they understand that modifications will be based solely on data and not on physician preference?
7) Respect for enrolled subjects
• Informing subjects of study results is increasingly encouraged
• How will subjects with early enrollment in the inferior arm perceive their involvement?
Meurer WJ, Lewis RJ, Berry DA. JAMA 2012;307:2377-8
The fundamental dilemma • A dilemma confronts physician-investigators…
• As physicians they are dedicated to caring for their patients…
• As investigators they are dedicated to caring for their research…
• These two commitments conflict whenever an individual physician/investigator comes face to face with an individual patient/subject.
Jay Katz, 1993
The ethical dilemma…
• The goal of research is to acquire generalizable knowledge to improve the health of future patients
• The goal of clinical care is to improve the health of our current patients
• A side-effect of research may be that subjects may be benefitted (or harmed) but it is never the purpose of the research
“Therapeutic Misconception”
• The false belief that the purpose of research is to benefit the patient
• Adaptive designs do not alter the primary goal of the research
• They may favorably impact the risk/benefit ratio, but may increase ethical concerns in other areas
Meurer WJ, Lewis RJ, Berry DA. JAMA 2012;307:2377-8
In my view, adaptive designs are not a “partial remedy for the therapeutic misconception,” but indeed make it even more problematic!