District police and prosecutors will not have a reliable and accurate breath-test program in place until March, according to testimony Wednesday at a D.C. Council committee.

Until then, officers will continue to rely on urine samples and roadside sobriety tests, such as walking and turning, to make their cases — a more expensive and cumbersome way to curb drunken driving.

The testimony came at a roundtable called by council member Phil Mendelson (D-At Large), who has pressed police, prosecutors and the Office of the Chief Medical Examiner to rebuild the battered breath-testing program.

The 2012 projection means the city will go nearly two years without one of the most basic policing tools. Errors and tainted court cases doomed the old system in early 2010, when officials acknowledged that inaccuracies from miscalibrated equipment had overstated drivers’ breath-test scores in about 400 convictions dating to 2008.

Deputy Attorney General Robert Hildum acknowledged at the hearing that the breath-test program needs an overhaul and not the series of lesser fixes that have been tried since the scandal.

A panel of outside national experts told city officials in October that top-to-bottom change was needed. But the procedures for bringing rigorous scientific review and management to the testing, maintenance and training processes are being designed, testimony showed, and funding is not yet secured.

The District is in the final stages of applying for a federal grant of about $130,000 to help with the overhaul and expects an answer “very soon,” Hildum testified.

The absence of a reliable breath-test program has contributed to a drop of about 42 percent in drunken driving arrests by D.C. police — from 572 in the first five months of 2010 to 328 in the first five months of 2011, according to statistics cited at the hearing by Kristopher Baumann, head of the police union.

The deterrent effect also has eroded because of the turmoil, Baumann said.

Assistant Police Chief Lamar Greene said he wanted to double-check internal statistics he had gathered before releasing them to Mendelson.

Mendelson said, “It’s clear that drunk driving prosecution has stalled in the last year.” The delays in building a program across agencies mean that “basically government wasted a year” because of “bureaucratic inertia,” Mendelson said. “But at least now they are making some progress putting together a defensible program.”

Mendelson oversees police matters as chairman of the Judiciary Committee.

Since last year, District officials have repeatedly said that they were working on a solution — and had purchased $90,000 in new testing equipment to salvage the department’s program. But the equipment has been idle for nearly a year.

The planned revamping would bring the District closer to the breath-testing programs of Maryland, Virginia and many other states that put the responsibility under a forensic science unit. The D.C. medical examiner’s office would take on that role from the police department and the grant would help hire a lab technician to assist with the added work-load.

Hildum said the District has won “most” of the approximately 40 cases in which drivers returned to court to challenge a conviction that included the erroneous scores. The retried cases relied on field tests and officers’ testimony, Hildum said.

The flawed testing was uncovered in February 2010 by an outside consultant on his second day on the job for the Metropolitan Police Department. The consultant had been hired to replace an officer who had run the breathalyzer program for 14 years and moved to a new assignment.

After a false start to bring back breath-testing machines, D.C. Police Chief Cathy L. Lanier in February ordered officers to take urine samples at drunken driving arrests. Urine samples cost about $75 each, compared with less than $10 for each breath test, which had been done about 120 times a month before February, police records show.

During the hearing, Lucas Zarwell, acting chief toxicologist at the medical examiner’s office, said that urine testing is “an old way of doing things” and that alcohol levels registered through those tests “have very little correlation to blood concentrations for alcohol.”

As a result, Zarwell said that when he testifies about the urine results in court, he has said that the score alone is not a measure of drunkenness and that it is better to also have those results supported by an officer’s observations about a driver’s behavior.