Maybe drug abuse doesn't turn people into criminals. But drug abuse drastically increases the amount of their criminal activity.

That is the principal finding of a new Justice Department study that may point the way toward doing something about both soaring street crime and overcrowded prisons.

Most criminals who wind up in prison commit maybe eight, nine, 10 crimes a year, James Stewart, director of the National Institute of Justice, explains. But criminals who are also drug abusers may commit hundreds of crimes a year.

The implication, while not made specific in the study, is that if we could do something about these "high-achieving" criminals, we could drastically reduce the amount of crime in our cities.

The 12-city study found that anywhere between 53 and 79 percent of the men arrested for serious offenses were also drug abusers. Not much surprise there; people who commit such crimes as burglary, grand larceny and assault are more likely to break the drug laws as well.

But add the finding confirmed by other NIJ studies -- that drug abusers, heroin addicts in particular, commit several times as many serious offenses as nonabusers -- and you have the basis for a policy choice: halt the addict's addiction, and you drastically reduce his criminal activity.

It is a choice Marion County (Indianapolis), Ind., already has made. The prosecutor's office there asks criminal suspects to undergo voluntary drug testing. Those who test "dirty" but who agree to enter a drug treatment program and submit to regular testing thereafter are recommended for reduced pretrial bond.

As Prosecutor Stephen Goldsmith explained in a recent ABC-TV interview: "We knew {even before the NIJ study} that drug use is pervasive and, more importantly, that the drug-using robber or burglar is far more prolific." His preference would be for mandatory testing for both suspects and applicants for parole, a policy certain to be resisted as amounting to unconstitutional self-incrimination.

"Mandatory testing postconviction is an excellent way to reduce prison and jail population, and preconviction {testing} is an excellent way to make an educated decision {as to who should be granted reduced bail}. . . . The great opportunity here is, for the first time, we have something that correlates with drug use and crime. A drug addict commits 200 to 400 times more crimes per year than a nonaddict. We ought to use that information."

If some way could be found to resolve the admittedly serious constitutional questions, the policy might make a good deal of sense.

Attorney General Edwin Meese, in releasing the NIJ study, said its purpose was to "track drug use trends among urban defendants suspected of serious crimes" on the assumption that "any increased drug use among offenders would represent an increased peril to the public."

"Drug abuse by criminal suspects far exceeds the estimated use in the general population, where it appears to be leveling off," Meese said. "Among criminal defendants, however, it seems to be increasing."

According to the study, cocaine use has nearly doubled in New York City over the past three years and more than tripled in the District of Columbia, where drug-related crime is at record levels. The question is how to use that information.

The present view is that crime (including drug abuse) is choice, not illness, and that punishment, not treatment, is the appropriate societal response. The problem is that the orthodox punishment -- imprisonment -- tends to make people worse, not better. And in any case, we are running out of prison space and the money to build more of it.

Given that state of affairs, it seems reasonable to try reducing crime by giving crime-prone drug abusers an incentive to get clean.