Radiation Detector Program Delayed
DHS May Have Misled Congress, GAO Audit Finds

By Robert O'Harrow Jr.
Washington Post Staff Writer
Friday, July 20, 2007

A $1.2 billion program to deploy new radiation monitors to screen trucks, cars and cargo containers for signs of nuclear devices has been delayed by questions over whether Department of Homeland Security officials misled Congress about the effectiveness of the detectors.

Homeland Security Secretary Michael Chertoff announced the contracts for monitors with cutting-edge technology a year ago. He said they would improve radiation scans at borders and ports, while sharply reducing the number of false alarms. Congress had allowed the five-year project to move ahead after Homeland Security assured appropriators that the $377,000 machines would detect highly enriched uranium 95 percent of the time.

"What this next generation of detection equipment is going to let us do is make those determinations much more precisely, much more easily and much more quickly," Chertoff said.

But the department's Domestic Nuclear Detection Office did not know whether the detectors would work effectively, according to documents and interviews.

Auditors from the Government Accountability Office later found that the detection rates of machines tested by the department were as low as 17 percent and no higher than about 50 percent. The auditors said the department's optimistic report to Congress on the cost and benefits of the machines was based on assumptions instead of facts -- a finding that prompted lawmakers to put the project on hold last year.

Last week, the GAO told Congress that Homeland Security officials did not follow their own guidelines for ensuring that the cost-benefit report was accurate and complete. The GAO also said the director of the nuclear detection office was incorrect when he testified in March that the office was not aware of any specifics about whether officials followed the guidelines. A GAO official said auditors would release a report about the monitors next month.

Senate Homeland Security and Government Affairs Committee Chairman Joseph I. Lieberman (I-Conn.) said Congress will continue pressing officials for more verifiable details about the monitors before they allow the project to proceed.

"As DHS develops costly new technology critical to the nation's security, Congress must be able to rely on DNDO's claims about the technology," Lieberman said in an e-mailed statement. "DNDO's estimates of costs and benefits must be based on facts, not assumptions. And, while taking into account the effects this technology will have on commerce, it must be based first and foremost on how best to prevent nuclear smuggling."

Vayl Oxford, director of the nuclear detection office, defended the high detection rate cited in the report to Congress last year as a "high-water goal" the agency hoped to achieve, not an assessment of the monitors' capabilities. Oxford said recent tests of the monitors in New York show a "dramatic decrease" in false alarms. Oxford said eight monitors will be deployed at four border crossings and ports for further performance tests this week.

The government has had difficulty getting independent, reliable technical assessments about the plausibility, cost and benefits of advanced technology before Congress and agencies commit to spending. It has always struggled when buying new technology, which is why Congress created the Office of Technology Assessment in 1972. For two decades, scientists and engineers in that office helped sort out technical truths from wishful thinking in project proposals. But the office was killed in 1995 in an effort to streamline federal programs.

Since then, as government spending on new technology rose to record levels, the primary technical advisers to federal officials often have been the contractors themselves. Billions of dollars have been wasted on failed, flawed or speculative projects.

A new computer system for tracking imports and exports was delayed by years because of technical problems, and the cost rose by $1 billion, to $3.1 billion. A computer system for the FBI to track criminal cases was abandoned after more than $100 million was spent. A system designed to track the entry and exit of foreign visitors featured a "prototype" network for recording visitor exits that cost $146 million but does not work.

The radiation portal monitors were envisioned as the nation's key bulwark against attacks with radioactive material. After the terrorist attacks of Sept. 11, 2001, the government spent more than $200 million on detection equipment that could not distinguish nuclear devices from more benign sources of radiation, such as ceramic tiles and cat litter.

President Bush directed the establishment of the nuclear detection office in spring 2005 to be the main resource for assessing and buying monitors. Its mission includes providing technical advice to other agencies.

The office immediately began testing machines that, according to GAO estimates, cost about six times as much as current monitors. The Advanced Spectroscopic Portal radiation monitors rely on sensitive detection technology that had not previously been used in the field in the way officials envisioned.

Homeland Security officials tested monitors made by 10 companies. But before the results of those tests were made available to Congress, auditors from the GAO, in March 2006, raised questions about the procurement process.

The auditors predicted cost overruns of as much as $596 million and said the "prototypes of this equipment have not yet been shown to be more effective than the portal monitors now in use." The auditors concluded that it "is not clear that the dramatically higher cost of this equipment would be worth the investment."

In response, Congress told Chertoff and officials at the nuclear detection office to produce a "cost-benefit" analysis, comparing the existing machines with the proposed replacements.

In June 2006, the department delivered a report that said that the new machines "can correctly detect and identify highly enriched uranium (HEU) 95 percent of the time," according to the GAO. Congressional appropriators then approved the spending.

On July 14, 2006, Chertoff and Oxford announced that they had ordered the first 80 of 1,400 new monitors. The monitors, manufactured by three companies, were to be deployed last fall under a deal that officials said involved up to a year of research and development and up to four years of full-scale deployment.

In the meantime, the GAO auditors examined the detection office's cost-benefit report. In a private meeting last August, the auditors told lawmakers that the report used optimistic assumptions and overstated the acquisition costs of the existing detection machines, distorting any cost comparison.

The auditors concluded that "DNDO's cost-benefit analysis does not justify its recent decision to purchase and deploy" the new machines and that the nuclear detection office should not spend more money buying the machines "until it conducts realistic testing," according to documents included with a GAO report last fall.

That finding prompted Congress to tell Chertoff that deployment of the new monitors should not occur until he vouched for a new round of tests his department conducted in January and February, the results of which have yet to be released.

In March, Oxford testified before a House homeland security subcommittee that the GAO misunderstood the cost-benefit report. He said "we stand behind the basic conclusions" of the report, which he said was done to justify research and development, not full-scale production. Oxford said his office followed department guidelines in drafting the report.

That assertion was contradicted last week by the GAO letter, which said that Oxford's nuclear detection office did not meet seven of eight department guidelines. The letter also said that DHS officials were briefed on the requirements just days before the cost-benefit report was delivered to Congress.

During that March hearing, the GAO's Eugene E. Aloise warned lawmakers that "the data used in the [cost-benefit] analysis was incomplete and unreliable, and as a result, we do not have any confidence in it."

Lawmakers in both parties were also openly skeptical of Oxford's testimony.

Rep. Dan Lungren (R-Calif.) told Oxford that Aloise's testimony suggested that "you screwed up big time. You did what you weren't supposed to do."

In an interview this week, Oxford said the contracts for the project were written in a way to give his office flexibility to continue studying the performance of the monitors before they were deployed. He defended last year's cost-benefit report as a "preliminary" document that did not mean his office was prepared to authorize full production.

Oxford stood by his assertion that his office was not told by the GAO precisely what to include in the cost-benefit report. "We were never given the specific details of what they thought was flawed in our methodology," he said.

View all comments that have been posted about this article.

© 2007 The Washington Post Company