INDUSTRY & RESEARCH
campusreview.com.au
Keep it clean
The simplest
data errors
can lead to
deadly flaws in
study results;
but most
mistakes can
be eliminated
once
academics
are aware of
the dangers.
Keith Baggerly
interviewed by
James Wells
M
islabelling, spreadsheet errors and
incomplete data are just some
of the basic mistakes in medical
research that can cost lives. But they can be
avoided if scientists are aware of how easily
they can happen, an expert says.
Professor Keith Baggerly, from the MD
Anderson Cancer Center, in the US, makes
a living as a forensic bioinformatician –
finding flaws in peer-reviewed, published
medical research papers that scientists and
journal editors may have overlooked. These
flaws are often caused by simple errors in
data compilation, interpretation of data,
mislabelling and computer errors.
These basic mistakes can have a human
cost. In 2010, Duke University researchers
conducted gene-based tests on a woman
with lung cancer, which they developed
from research they say could help pinpoint
the cancer’s weakness, bypassing trialand-error treatments. But the research was
wrong because of basic errors, and the
woman died.
Sitting down with Campus Review,
Baggerly recounted another incident, in
which medical researchers had become
wrongly convinced they’d developed a
serum that could