| July 16, 2002
'Underpowered' Clinical Trials Pose
Scientific and Ethical Problems
Studies with too few subjects may be 'scientifically
useless,' thus unethically exposing participants to the
risks of human research.
PA)-- In an article in the July 17th issue of the Journal
of the American Medical Association (JAMA), three
researchers at the University of Pennsylvania School
of Medicine claim that many clinical trials are both
scientifically deficient and unethical because they
enroll "too few participants to reliably determine
whether or not an intervention works."
The scientific and ethical problems of such "underpowered"
trials were first suggested more than 20 years ago.
"Because [underpowered trials] may not adequately
test the underlying hypotheses (of the study), they
have been considered 'scientifically useless,' and therefore
unethical in their exposure of participants to the risks
and burdens of human research," write the article's
authors, Scott D. Halpern, MSCE, Jason H.T.
Karlawish, MD, and Jesse A. Berlin, ScD.
Despite this earlier consensus that trials must enroll
adequate numbers of subjects, the authors found that
not only do many investigators continue to enroll too
few subjects, but that the practice has recently gained
In their article, entitled, The Continuing Unethical
Conduct of Underpowered Clinical Trials, the authors
review the many scientific problems associated with
conducting underpowered trials. And, as Halpern states,
"studies that are scientifically deficient are
also ethically deficient."
According to the researchers, scientific issues such
as statistical power become ethical problems because
most investigators do not tell people who enroll in
their studies that there will not be enough subjects
involved to make a reliable judgment on the study's
"We believe investigators who conduct small or
underpowered studies must tell patients that their participation
might not provide information that will be of immediate
clinical value, though it may provide information that
will guide future studies," explains Halpern.
"Because patients often choose to participate in
research because they want to help people with similar
diseases, or help advance the cause of medicine, it
is reasonable to assume that they would rather enroll
in studies that have a better ability to provide healthcare
benefits than in studies that have a lesser
ability to provide healthcare benefits," he added.
"The smaller the study is, the less it is of clinical
value. To deny patients the opportunity to know the
relative clinical value of a study that they're considering
enrolling in is unethical because it prevents them from
making a fully informed decision."
Halpern says that what constitutes an adequate number
of patients for research varies from study to study,
but that as a rule, clinical researchers determine this
number by using a specific formula which uses a number
of variables including: the effect size being sought,
the predicted variance in patients' outcomes, the level
of statistical significance being used and the level
of power sought. "The prominent factor that goes
into determining the number of patients you would need
is the magnitude of the effect of an intervention that
you're looking for," he explains. "Researchers
are generally guided by a clinical consensus among experts
in a particular field as to what a clinically important
effect is." In cases where such clinical consensus
on a meaningful effect does not exist, Halpern and colleagues
suggest that researchers use informed predictions of
an intervention's likely effect to calculate the number
of subjects that will be needed.
The authors say that conducting underpowered trials
is only justifiable in narrowly defined circumstances.
For instance, it might be difficult for investigators
to find enough patients with a rare disease to conduct
an adequately powered study. In such cases, Halpern
says underpowered trials may be acceptable as long as
the investigators can document plans at the trial's
outset to combine the results with those from similar
trials conducted elsewhere. Such investigators must
also tell potential research subjects that it is uncertain
whether the trial will yield clinically important information.
The authors write that failing to disclose this information
may be deceptive and that "if such deception were
publicized, it could undermine people's trust in science,
further curtailing future enrollment [in trials]."
In their article, the authors condemn trials in which
investigators should know, before commencing recruitment,
that too few subjects are planned for inclusion. They
do not specifically address trials in which patients
drop out or fail to comply with the research, which
can also decrease power. Trials suffering from these
problems should be considered valid if there were enough
subjects at the start, says Halpern. But he adds that,
ideally, investigators should plan for a certain number
of "drop-outs" when determining the number
of subjects needed for a study, and thus attempt to
enroll more patients than the minimum number needed.
Halpern is a seventh-year MD/PhD student, a fellow in
the University of Pennsylvania School of Medicine's
Center for Clinical Epidemiology and Biostatistics,
and a co-investigator with the school's Center for Education
and Research on Therapeutics; Karlawish is a faculty
member in the University's Center for Bioethics, the
Department of Medicine and the Leonard Davis Institute
for Health Economics; and Berlin is a professor of biostatistics,
a Senior Scholar in the Center for Epidemiology and
Biostatistics, and also a co-investigator with the Center
for Education and Research on Therapeutics.