False Positive—The Reality of Workplace Drugs in America
It isn’t what you recently heard. A special investigation by The Fix into the report that American employee drug use has declined 74% since 1988.
This past November 18th marked the 25th anniversary of President Reagan’s signing into law the Drug-Free Workplace Act. The bill required any institution receiving federal funds to establish and maintain an alcohol and drug-free workplace.
The anniversary might have passed without much comment except that one company was quite aware of the date. Quest Diagnostics, the most prominent corporation in the drug testing industry, saw it as an opportunity to grab headlines.
Quest that day issued a self-congratulatory report stating—unequivocally—that its own survey of 125 million drug tests from 1988 to 2012 proved that there was a 74% decline in drug use among American workers since the Drug-Free Workplace Act was signed. This figure was then heralded by a number of media outlets. The Wall Street Journal, for example, used exactly the same headline as the Quest press release, stating simply:
“Drug Use Among American Workers Declined 74% Over Past 25 Years”
The Journal then cited the Quest press release almost verbatim without undertaking its own closer analysis of the data.
“Overall,” the Journal reported straight from the press release, “3.5% of samples came back positive last year compared with 13.6% in 1988. The vast majority of tests, around 75% in recent years, were conducted for pre-employment screening. The rest were administered following accidents, after employers suspected drug use or as part of regular testing regimens.”
The Quest press release left no doubts that the company was making its bold claim for the entire U.S workforce and not just for Quest's sample of it. It also offered fairly minimal qualifications to its findings, not mentioning major factors that might skew even its own report on those workplaces for which it had some evidence. Instead, its most significant qualification to the news was: “. . .although the rate of positive test results for certain drugs, including amphetamine and opiates, continues to climb."
It then added: ". . .according to a landmark analysis of workplace drug test results released today by Quest Diagnostics (NYSE: DGX), the world’s leading provider of diagnostic information services.” (Quest conducts more than six million workplace drug tests annually, reported $7.4 billion in revenue last year and claims to service 30% of the adult American public with a wide variety of laboratory services, including drug testing.)
It would be a remarkable feat in workplace performance if 74% was an accurate figure. Unfortunately, on closer scrutiny, 74% turns out to be a “false positive.”
A careful inquiry into the details of the Quest study by The Fix reveals the following:
- Few workplaces even test for the widely-used new drugs defined by the DEA as currently threatening America’s health and safety, and so they are excluded from the Quest survey.
- Many employers do not even have a drug-testing program.
- The drugs actually being tested are not tested at uniform sensitivity levels.
- There is no complete uniformity in what drugs and how many different drug types are actually tested for by the employers who do test for drugs, and the Quest survey only includes the tests employers request. Most employers do not test for all drug types.
- Quest put itself forward as one company speaking for the entire American workplace based only upon its own large but hardly definitive sample.
AN EVEN CLOSER LOOK AT THE QUEST SURVEY REVEALS….
There are numerous and subtle factors that make up the world of drug testings in America. Many people applying for a job think you pee in a cup and cross your fingers that your last imbibe of marijuana moved out of your system; or they pray that the meds they are taking - over-the-counter and otherwise - don't create a false positive.
The varieties of contingencies in the world of drugs and drug tests are in fact extensive and complex - as is inevitably any attempt to measure a large chunk of the population's use of the scores of "Illicit" drugs out there along with the "licit" drugs that have gone black market.
Among the host of data influencing factors are where samples are collected, where they are "read" - it can be instantly like a pregnancy test or at the lab - and what drugs are being tested for and at what levels. Then there are the issues of the chain of custody of the samples, the ability of people to cheat on their tests, the facts that employers treat different drugs with different levels of leniency and may even throw away positive tests of the "lenient drugs."
Beyond this there's the reality that different employers might use tests for say, heroin, that are set at a different sensitivity level of detection than the company across the street uses. Some firms care to test for black market prescription drugs, the vast majority don't. Some want to screen for many drugs; most want to go with the basics, often for cost reasons. Some people fail tests, clean up, and then pass a test with a different employer. No data can track that. It goes on and on.
All this a company like Quest needs to take into account in any data survey it releases to the country if it wants it to be credible.
Presumably aware of all these factors, Quest notably handles one troubling issue by omission. That is, it doesn’t count in its survey data an untold number of “instant results” urine specimens that employers then send Quest for legally required follow-up testing with more sophisticated lab technology. Quest omits these numbers from its survey of American drug use on the very reasonable grounds that it has no way of knowing by comparison how many workplace instant results tests per year came up negative or how the employer handled the "chain of custody" of the instant tests.
On the same grounds, Quest also excludes all test results submitted by employers who use a mix of instant results tests and standard urine collection devices.
Scientifically and statistically, this makes sense. Yet it leaves hanging the question: how much of the population's test results - positive and negative - are thereby not factored into the broad statement of 74% improvement?
Asked whether the complete instant results test information would skew the survey if known, one executive told The Fix that there were a negligible number of such uncounted tests sold by Quest each year. Further questioning of Quest’s chief testing scientist, Dr. Barry Sample, revealed that, in fact, there are “substantially less than one million” such uncounted instant results units sold by Quest (the exact number is “proprietary”). Those test cups that are returned are omitted from the database, Dr. Small said, and no count is kept of how many.
Obviously, a large number might alter the glowing survey results if not offset by a known number of negative tests. No number at all leaves the survey to be judged on how Quest treated most of the other complicating factors in the drug testing universe. And that's where the problems with the Quest pronouncement begins.
WHAT THE GOVERNMENT DOESN’T REQUIRE
One of those factors is who must obey government rules and who can do what they want re drug testing. Hugely relevant to the 74% claim is the fact that the federal government issues guidelines naming exactly the drugs to be tested for use by federal workers who are in “safety-sensitive” positions. The government does not otherwise require private sector employers who are not receiving federal funds to test for drugs. The number of employees not being tested therefore is a guesswork moving target.
Even so, one unverified survey claimed that while 84% of workplaces conduct pre-hire screening, only 39% did random follow-up screening of hired employers. Even assuming this survey was accurate, that would leave 16% of job seekers not tested, and huge numbers of people never tested once hired.
These tens of millions of regularly untested workers, if counted by Quest, would clearly skew any survey one way or the other. Quest doesn't account for them in its data base because it can't - but it gives the impression in its public relations announcements that it has.
Equally Alice in Wonderland upside down is the fact that a majority of private employers in the U.S, according to Quest itself, simply ignore the latest federal instructions as to the sensitivity levels they are to use in the first round of drug tests of would-be employees or of already hired employees. Instead, this majority relies on older, pre-2010 test standards with a much higher detection threshold. This would be the equivalent of a police breath test for alcohol set at, say, 1.5 rather than the .08 common in many states, including California.
How many additional people would be found to be using drugs if the stricter standards were applied across the board?
One clue comes from Quest’s report that amphetamine usage has tripled since 1988, with the largest jumps appearing after the federal government changed its guidelines in 2010 to require a sensitivity level for the first round of testing at 500 ng/ml vs. the earlier standard for amphetamines of 1000 ng/ml.
Morphine also showed an increase of 34% between 2005-2012 in the first round of testing.
Accordingly, Quest's statement that there was a 74% decline in drug usage really means that there was a 74% decline only in the tests that were included in the database. These tests, of course, only recorded "positives" at whatever level an employer customer of Quest's chose to use from the available range.
(Note here that the federal government sensitivity levels on the first round of tests are not zero but range from a low of 10 ng/ml for the heroin metabolite to 2000 ng/ml for codeine and morphine. The federal standard used to be 300 ng/ml on the initial test for codeine and morphine. These were dramatically changed when both employers and employees complained that standards were too stringent.)
Quest’s admission to The Fix that it includes in its database the less stringent sensitivity levels that many employers still use, adds to the mish-mash of conflicting and incomplete data that goes into its Drug Testing Index (DTI) on which the company bases its claim that workplace drug use has dramatically fallen.