Soaring Demand Erodes Drug-Test Labs’ Efficiency

The rapid proliferation of employee drug-screening programs in America has created a gold-rush atmosphere in the nation’s testing laboratories, eroding the quality of lab work, according to many drug-testing experts and industry officials.

One consequence is that less-than-reliable urinalysis results are being used by many American employers to make hiring and firing decisions, they said.

As many as 30% of the nation’s testing laboratories “are not doing good work, and their results are open to question,” said chemist Robert E. Willette of Duo Research in Annapolis, Md., a former director of research for the National Institute on Drug Abuse.

Psychiatrist William H. Anderson of the Harvard Medical School, who counsels drug addicts undergoing therapy at Massachusetts General Hospital in Boston, added:

“We are always changing laboratories because we are not satisfied. Results on some tests are very often inconsistent with other data, so we have reason to believe that other tests are not being done as well as they could be.”

And John P. Morgan, a pharmacologist at the City University of New York who frequently testifies in court for employees accused of drug abuse, was even more outspoken. “These labs are no more ready to handle (the tests) than they are to fly,” he said.

These experts and others say that some labs are employing inexperienced technicians who are ill-prepared to interpret results and that many more lack the rigorous procedures necessary to prevent one urine sample from being mixed up with another or to prevent accidental contamination of a test specimen.

Such problems--coupled with the inherent limitations of the drug tests themselves--are spurring calls for government regulation on a fast-growing industry that markedly lacks uniform standards. Many labs are supporting such calls in order to weed out those with poor performance records.

In California, the Legislature passed a bill to require licensing of labs, but Gov. George Deukmejian in July vetoed the measure, saying it would raise the cost of testing and “inappropriately discourage” drug screening.

One measure of the uncertainty that surrounds the drug-testing business is that there is no authoritative estimate even on the number of laboratories performing drug tests.

There are about 5,500 independent clinical labs in the United States, about 6,000 clinical laboratories associated with group medical practices and health maintenance organizations, and about 7,000 labs associated with hospitals, according to Louis Frisina, vice president of Boston Biomedical Consultants, a marketing firm that specializes in biomedical fields.

In California, there are about 1,900 independent and hospital laboratories licensed by the state and another 1,500 in physicians’ offices, according to the state Department of Health Services.

Frisina estimates that about 30% of the nation’s independent labs and a much smaller fraction of the physician labs and hospital labs are doing drug screening. But at least half the hospital labs already have the necessary instruments to perform drug screening, he added, and most of them are “very interested” in getting into the field.

Sales of Supplies Soar

A more accurate reflection of the growth in the drug-testing industry can be seen in sales of the two leading tests and related supplies. In 1980, sales totaled $25 million, according to Frisina. By 1985, he said, the number had grown to $73 million. Sales are growing at a rate of 22% a year now, and can be expected to double again by 1990.

Experts say that cost of such supplies normally accounts for between 25% and 30% of a laboratory’s gross revenues. Based on that figure, total yearly revenues for drug-screening labs would be about $250 million per year.

“Because of the large amounts of money to be made, there are a lot of new laboratories in the business and they are not always qualified to do the tests,” said Northwestern University pharmacologist John Ambre, an outspoken critic of large scale drug testing.

The sharp increase in drug screening is straining the nation’s supply of toxicologists and qualified technicians, some caution.

“There just are not a lot of experts,” said Los Angeles toxicologist Gordon Hisayasu of Pacific Toxicological Laboratories. Hisayasu, who left the USC School of Pharmacy, where he headed the toxicology laboratory, in January to found Pacific Toxicological Labs, said he spends a large portion of his time training his own staff.

And expert technicians cannot be created overnight, noted toxicologist Ken Campbell of Kern Analytical Laboratories in Bakersfield.

“There are a lot of subtleties involved in testing and analysis that you can’t necessarily put your fingers on when you’re training a lay person,” he said. “And you never can tell when you are going to need it.”

‘Very Uncomfortable Feeling’

Peter Greaney, associate professor of community and environmental medicine at University of California, Irvine, and part-owner of a Tustin drug-testing consulting firm, cites the example of one inexperienced lab that was receiving samples from one of his clients.

“I was getting a very uncomfortable feeling with the results,” Greaney said, “the number of positives was very small. You just know that 9% to 12% of any workplace are using drugs, so you should detect at least 5%. But only about 0.2% of the client’s samples were coming back positive.

“Whenever I would call the lab, the phone went unanswered for a long time or was answered by people who couldn’t communicate in English or answer technical questions. They couldn’t distinguish between drugs of abuse and over-the-counter drugs. We switched labs, but that company is still in business.”

Even in the best of labs, many experts say, quality can vary from day to day and from one technician to the next. “You can never be sure that a good laboratory today will be a good laboratory tomorrow,” Anderson said.

Quality can suffer as labs try to retain or increase their share of the drug-testing market.

“Clinical laboratories have always been very highly cost competitive,” Hisayasu noted. “When more labs started doing toxicology, there was the same kind of competition on costs. The cost of reducing costs is quality.”

Experts agree that the best way to measure a lab’s quality is through proficiency tests in which the employer submits samples known to be contaminated with specific types of drugs, and then waits to see if the lab can identify which samples are drugged and which are not.

‘Blind Testing’

These tests can be done without warning to the lab, a process known in the industry as “blind testing.” More often, however, the lab is advised that it is working on a test sample, although the identity of the drug is not disclosed in advance. This is called “open testing.”

The two major testing agencies are the College of American Pathologists and the American Assn. of Bioanalysts, but only about half the drug-screening labs in the nation are enrolled.

“Open” tests usually yield an accuracy rate of 90% to 95% among the 350 or so voluntary participants in the bioanalysts program, according to Nicholas Serafy Jr. of the association. The College of American Pathologists reported similar results in its most recent quarterly survey, which found a 94.9% accuracy rate in the 465 participating labs.

Doubts About Accuracy

But Serafy and industry officials concede that open testing may not yield the most accurate assessment of a lab’s proficiency, since it puts the labs on notice.

“If a lab has a hotshot chemist or technician, he is the one who is going to do proficiency testing, so the test results are always satisfactory,” said toxicologist Emil deVera of California Department of Health Services.

Doubt about the accuracy of open testing was reinforced by a federal study published in 1985 that concluded there was a “crisis in drug testing.”

The report was based on a 1981 study in which chemist Joe Boone and his colleagues at the Centers for Disease Control tested the proficiency of 13 laboratories that were serving 262 methadone clinics throughout the country. (Methadone clinics, which prescribe methadone to heroin addicts as an opiate substitute, are required to conduct urine tests of their clients to ensure they are not using heroin and are taking their methadone.)

69% Average Accuracy

The CDC group sent the laboratories serving the clinics hundreds of urine samples that contained not only methadone and heroin, but also amphetamines, cocaine and codeine.

When Boone sent out open samples, the accuracy of the 13 labs ranged from 76% to 100%, with a 98% average.

But when blind testing was conducted, the accuracy rate dropped dramatically. It ranged from 11% to 100%, with an average of 69% accuracy.

The CDC report, published in the Journal of the American Medical Assn., concluded that only six of the labs performed acceptably in detecting methadone, while only one or two labs performed acceptably in detecting barbiturates, amphetamines, cocaine, codeine and morphine.

But not everyone agrees with the conclusions of Boone’s study.

Those results “are not an accurate reflection of what is happening today for many of the drug tests ordered by employers,” said Peter Bensinger, former head of the U.S. Drug Enforcement Administration and now co-owner of a Chicago drug-abuse consulting firm.

Conditions Have Improved

Bensinger said that labs are more certain to confirm positive tests with a second test, more sensitive to contamination of samples and more careful to protect specimens from being inadvertently switched.

Boone conceded that conditions in drug-screening labs have improved since 1981, but adds that results from blind proficiency tests at all types of laboratories have historically been less accurate than open tests and that he would be surprised if the situation were any different now. “But we don’t have any current data--just suspicions.”

The National Institute on Drug Abuse is now trying to set up a voluntary accreditation program for drug-screening labs that would involve blind testing, but Richard Hawks, NIDA’s director of research, said it could not begin for at least two years.

Ensure Accurate Results

Most experts thus agree that the best way for a company to ensure accurate results from a lab under contract is to submit its own blind test samples. But among 35 companies contacted by The Times, only two, the DuPont Co. of Wilmington, Del., and Southern California Edison Co. are doing so.

Blind testing, however, has its own difficulties.

“First you have to get controlled drugs. Then you have to obtain good urine samples,” said chemist W. Lee Hearn of Toxicological Testing Service in Miami. “You have to ensure that you know what the contents are.”

Two firms have recently begun marketing such samples, and some experts predict that the selling of these samples will be the hottest growth area in the drug-screening industry.

The lab-regulating bill vetoed by Deukemjian was sponsored by Assemblyman Johan Klehs (D-San Leandro).

It would have required all testing labs in the state to be licensed and monitored by the Department of Health Services. And it would have mandated a “chain of custody” agreement between employers and labs to ensure that every urine sample be properly identified and safeguarded from contamination.

And because there is a significant potential for error in testing results, the bill also would have mandated confirmation of a positive test “by at least one fundamentally different testing method” and required the preservation of any positive sample for at least 6 months to ensure a worker’s right to appeal.

Finally, the bill would have required employers to post written notice to employees of their testing policies, to establish disciplinary and appeals procedures and to offer rehabilitation programs.

Bill Opposed

Both the California Manufacturers Assn. and the California Chamber of Commerce opposed the Klehs bill because it would prohibit member companies from using out-of-state laboratories or performing the screening tests on site.

Frisina estimated that about 10,000 U.S. companies are large enough to have a medical department and, thus, the capacity to perform drug testing on site. He also estimates that fewer than 10% of these companies are actually doing so.

The Legislature in 1973 passed a similar bill to regulate methadone-testing labs, and that program has worked well, according to Daniel Morales, chief of the health department’s division of laboratories. A key feature of that law was a provision for blind proficiency testing, but the law also established other standards for proper maintenance, use and calibration of equipment and handling of samples.

Labs Certified

Twelve laboratories initially applied to be certified by the state, Morales said, and 10 passed. Three other labs were subsequently certified.

Of those 13, two have since lost their licenses because they failed to meet one or more of the state requirements, two lost their licenses because of poor performance on proficiency testing, two lost their licenses for reasons that do not appear in the records, and two voluntarily relinquished their licenses.

Since 1973, Morales added, 16 other labs have applied for certification and have been rejected, and another 155 have requested applications and not completed them.

Most of the companies that lost their licenses to perform methadone tests have since gone out of business, Morales said. But one of them--which was decertified by the state for failing to meet minimum standards of accuracy--is still at work. Among other things, it performs drug tests for employers.

A RELIABLE DRUG-SCREENING PROGRAM

These are what the National Institute on Drug Abuse describes as key features of a reliable drug-screening program. There is no estimate of how many programs include such features: Full procedure should be in writing and made readily available.

Question employee or job applicant about use of legal drugs that may cause

positive result.

Watch production of urine sample.

A strict chain of custody to ensure samples are not tampered with or mixed up.

Select a good lab and then periodically test it for quality control.

Confirm all positive results with a different test than the original.