Printer Friendly Page
CANCER RESEARCHERS BLAST VA FOR WITHHOLDING
DATA -- Stonewalling by the VA is putting U.S.
cancer
surveillance and research in jeopardy,
according to
many of the researchers involved in those
fields.
For more on VA research, use the VA Watchdog
search engine...click here...
http://www.yourvabenefits.org/ses
search.php?q=va+research&op=ph
For more on veterans and cancer, use the VA
Watchdog search engine...click here...
http://www.yourvabenefits.org/ses
search.php?q=cancer&op=and
Story here...
http://www.medpagetoday.
com/InfectiousDisease/Pu
blicHealth/tb/6563
Story below:
-------------------------
VA Cancer Data Blockade May Imperil
Surveillance
By Michael Smith
Senior Staff Writer, MedPage Today
LOS ANGELES, -- Stonewalling by the Veterans Administration is
putting U.S. cancer surveillance and research in jeopardy, according to
many of the researchers involved in those fields.
After decades of sharing data freely and allowing researchers to get in
touch with its patients, the agency has been blocking such activity for
the past several years, according to Dennis Deapen, Dr.PH., of the Los
Angeles Cancer Surveillance Program and the University of Southern
California.
The result, Dr. Deapen said, is that California state data on cancer
incidence rates are being skewed. And that, he said, is likely to have
serious effects on national data.
The California Cancer Surveillance Program has seen a sharp drop in the
agency's reporting of new cases to Californian cancer registries
beginning in late 2004 -- from 3,000 cases in 2003 to almost none by the
end of 2005, according to an article in the September issue of Lancet
Oncology.
But the problem is not restricted to California, according to Holly
Howe, Ph.D., of the North American Association of Central Cancer
Registries (NAACCR) in Springfield, Ill.
"California has been very energetic in evaluating the impact of the loss
of VA cases on completeness and the ability to produce accurate
incidence data," Dr. Howe said.
"But it's not just California -- it's nearly every state," she said.
However, California and Florida -- where VA reporting of cancer cases
has also been blocked -- have large populations of veterans and large VA
medical facilities, she said.
Missing data from those two states has the potential to warp national
estimates, she said.
Lancet Oncology quoted Raye Ann Dorn, the VA's national coordinator of
cancer programs, as saying that only California and Florida were
withholding data, mainly because of privacy concerns.
Dorn was not immediately available for comment on the eve of the Labor
Day weekend.
Lancet Oncology said other VA officials pointed out that of the 130
medical centers that collect cancer data, only 29 withheld cases from
state cancer registries in 2006.
But the journal also said that, according to CDC data, VA centers in
seven states are not reporting cancer cases and in six others, at least
one VA facility is not reporting.
All told, "40 000 to 70 000 cases are potentially missed nationally each
year," the journal quoted a CDC spokesman.
Dr. Howe said her organization and others have been trying to persuade
the VA to resume wholehearted data-sharing, but with little success.
"We've been trying to solve this for over five years," she said.
Asked if she knows what's behind the policy, Dr. Howe said flatly: "No."
Representatives of a "whole cadre of associations" - including NAACCR,
the CDC, the American Cancer Society, and the National Cancer Institute
- met in early August to discuss the issue, Dr. Deapen said.
He said the VA position has two main effects.
The skewing of national and state cancer incidence rates, he said, is
"correctable."
"The VA still has the data," he said. "They could hand it out and then
we could correct incidence rate data."
What is "incorrectable," he said, is the effect the data blockade could
have on research.
Dr. Deapen said, for example, that researchers investigating the causes
of a particular type of cancer might be misled if they were not aware of
a cluster of cases being treated in VA hospitals.
"Once that study is done, (the researcher) doesn't get to go back and do
it over," Dr. Deapen said. Research during this period "will forever
require an asterisk" to remind other researchers that it might not be
correct.
But even when states get VA data, some cases may slip through the cracks
under a related VA policy that forbids interstate data-sharing, he said.
For instance, he said, it's common for veterans in some eastern states
to seek treatment in neighboring states.
The host state doesn't count them, because they live next door. And the
VA refuses to notify the home state or let the host state do so, so that
some cases are simply never counted, Dr. Deapen said.
Several of the cancer registries that are being locked out of VA data
take part in the Surveillance Epidemiology and End Results (SEER)
program, according to Brenda Edwards, Ph.D., of the National Cancer
Institute, which operates the database, a valued resource for
epidemiological research.
"This will significantly impact reporting in SEER," Dr. Edwards told
Lancet Oncology.
The collection of disease incidence data is a state responsibility, Dr.
Deapen said, but the VA - as a federal agency - is under no obligation
to comply with state laws.
Nonetheless, for years the VA voluntarily shared its data and allowed
access to patients, he said.
"We had it right and we were doing a good job," Dr. Deapen said. "Now we
need to get back on track."
Primary source: Lancet Oncology
Source reference:
Bryant Furlow. "Accuracy of US cancer
surveillance under threat." Lancet Oncology 2007;8:762-63.
-------------------------
Larry Scott --
