News & Events
Test Kits for Drugs of Abuse
D. Bruce Burlington, M.D.
Center for Devices and Radiological Health
Food and Drug Administration
Department of Health and Human Services
the Subcommittee on Oversight and Investigations
House Committee on Commerce
September 26, 1996
Mr. Chairman and Members of the Subcommittee, I am Bruce Burlington, Director of the Center for Devices and Radiological Health (CDRH), Food and Drug Administration (FDA or the Agency), and am here today to discuss how FDA reviews medical diagnostic tests and evaluates whether they should be marketed for their intended use. My testimony will cover the means by which FDA assures these products can perform safely, effectively and in accordance with claims made by those who develop and manufacture them. As you requested, I will specifically address the review of test systems for drugs of abuse and provide an update on the Diasensor 1000 glucometer device made by Biocontrol Technology, Inc.
Appearing with me today are two senior members of my staff, Ms. Lillian Gill and Dr. Steve Gutman. Ms. Gill is the Director of the Office of Compliance, CDRH, which oversees all regulatory compliance and enforcement activities. Ms. Gill's background is in clinical laboratory and regulatory laboratory science.
Dr. Gutman, a clinical pathologist, is the Director of the Division of Clinical Laboratory Devices, CDRH, which is responsible for evaluating diagnostic medical devices. Prior to joining FDA, Dr. Gutman was director of the clinical laboratory at the Veterans Administration Hospital in Buffalo, New York.
FDA REGULATION OF DIAGNOSTIC TEST SYSTEMS
CDRH evaluates approximately 1000 new and upgraded diagnostic medical devices produced each year. These include products used in laboratories, in medical offices and clinics, and in home settings to test for various types of cancer, diabetes, genetic defects, cholesterol and blood sugar levels, pregnancy, and many other conditions.
Diagnostic test kits or systems, such as those used to measure the level of glucose in a person's blood or to detect the presence of alcohol or drugs of abuse, belong to a category of devices referred to in vitrodiagnostics. An in vitrodiagnostic is a laboratory test which analyzes specimens taken from the body, such as blood, saliva or hair, and measures one specific chemical to see if it is present or measures how much is present. FDA regulates in vitro diagnostic test kits or systems as medical devices under the Federal Food, Drug and Cosmetic Act. Our statutory authority broadly encompasses the test systems, including the diagnostic test itself, as well as ancillary equipment and collection devices, regardless of the various sites at which these products are used to collect and/or process specimens. This includes tests used in point-of-care facilities, such as doctors' offices and emergency rooms; hospital-based laboratories; freestanding laboratories; and in the home. It also includes tests provided through intermediary service providers.
As a practical matter, we focus our regulatory resources on the products themselves and on those products that are manufactured as self-contained kits or test systems for commercial distribution. In certain venues where specimen collection or point-of-care testing occurs, such as the workplace or hospitals, and free-standing laboratories, we defer to other authorities, such as the Center for Substance Abuse, in the Substance Abuse and Mental Health Services Administration (SAMHSA). SAMHSA certifies drug testing laboratories and has provided specialized training for roughly 15,000 physicians on how properly to interpret drug testing results. We also rely upon agencies within the Department of Health and Human Services that are responsible for oversight of laboratories whose testing services are subject to the requirements of the Clinical Laboratory Improvement Amendments of 1988 (CLIA).
For all medical devices, our role is to assess whether a product proposed for marketing is safe and effective for its intended use. The determination of whether a product is safe and effective is based on a variety of considerations which provide the Agency the basis for assessing the risks and benefits of the device. The intended use of the device is always a key element in such an assessment, because the risks associated with a device may vary depending on how and by whom it is going to be used. As I discuss in more detail below, this is particularly the case with in vitro diagnostic tests.
What is the guiding principle that underlies our evaluation of diagnostic testing devices? Very simply, it is to ensure that diagnostic test systems produce test results that are reliable, clinically meaningful, and provide sufficient information to assure the results can be interpreted accurately by the intended users. Each test system has a number of components, including the specimen test itself, criteria for use of the test, specimen collection and handling, and reporting and interpretation of results. The analysis of these various components and their relationship to each other all contribute to the assessment of whether the system is safe and effective for its intended use.
The first step in assessing whether a test system will be safe and effective for its intended use is having data that describes the accuracy of the specimen test itself. No in vitro diagnostic test yields perfect results. Sometimes the test is going to miss the presence of what it is supposed to be detecting. When this happens it is called a false negative. Sometimes it registers the presence of the substance even though it is not present. This result is called a false positive. Knowing the probability of false negatives and false positives, not just in abstract terms but why they occur and whether and how the rate varies among different populations, is absolutely essential in order to properly interpret the results and treat patients.
For example, let us look at the prostatic specific antigen test, which can be used to screen for prostate cancer. If this test detects the presence of cancer 95 percent of the time (sensitivity) when the disease is actually present, and detects the absence of the disease with the same high degree of accuracy (specificity), what then does a positive result really mean? On its face, this sounds relatively straightforward to interpret. But in reality, the result does not necessarily mean that any one individual with a positive result has a high probability of having cancer. If you were to test a healthy group of relatively young men, in which the expected cancer prevalence were 5 in 10,000, you would end up with 490 false positives for every one true positive result. Some men who received false positive results probably had prostatic infections; others would be unexplained but most would be disease free. This example illustrates two points: that tests are fully interpretable only in the context of the prevalence of the disease or condition for which you are testing and confounding conditions; and that interpreting them correctly presents complex issues.
Companies make diagnostic tests for use in a variety of settings: sophisticated research and clinical laboratories; labs located in community hospitals and nursing homes; at the point of care by health professionals; and at home by consumers. These tests can be available on a prescription basis from health professionals or sold directly to consumers. Testing processes can entail patients undergoing tests on-site at a laboratory or having a specimen collected elsewhere, including at home, and shipped to a lab for testing. Test results can be stated as raw data or as conclusions drawn from the data, and they can be provided with or without professional interpretation or counseling. When we move out of the sophisticated laboratory to point-of-care facilities and home settings, where testing is either performed on-site or test specimens are collected, there are a host of issues that are inextricably linked to the fundamental question of whether a product works or not.
After we establish the accuracy of a particular test, we must evaluate data that indicate whether a test will perform acceptably in the particular setting and under the conditions contemplated by the intended use. Will the test be performed by individuals with appropriate levels of training, knowledge and proficiency? Will the system provide sufficient information on how to use the results to support its intended use, or its use in a particular setting? When collection is performed in one place and testing in another, is there a chain of custody to preserve the identity of test sample? Does the test sample deteriorate over time? Is the collection method easily understood? Can the user proficiently conduct the collection? If a test yields a positive value, is confirmatory testing required? Is there a mechanism to provide access to confirmatory tests? Will the recipient of the test results know how to interpret them? If not, what assistance is available for such interpretation?
The full gamut of these questions does not apply to every test system or kit. The questions to which we seek answers depend heavily on how the product is to be used, for what purpose, and by whom. There are specific issues, for example, that pertain to all types of over-the-counter (OTC) diagnostic products. OTC products need adequate directions for use without the help or guidance of a health professional: directions that allow the lay public to understand the conditions under which use of the product is appropriate; directions on how to use the product properly; and directions that, for a diagnostic tool, enable the user to understand and act upon the information correctly. When we look at whether these principles have been met we need to consider the general state of knowledge in the community about the condition for which the test is being used, the likelihood of adverse events if someone uses the product without reading the directions or understanding the complexity involved in interpreting the results in various population groups.
Whether it is a new, breakthrough device like the Diasensor non-invasive glucose monitor or a test system to ascertain the presence of illegal drugs in humans, these are critical issues that must be resolved to ensure these test systems are safe and effective for their intended uses.
No parent wants to test their diabetic child's glucose level with a test that gives unreliable results, or results they are unable to interpret and apply rationally. Similarly, no parent wants to have their child's saliva or urine tested for the presence of drugs of abuse if the results of that test are not reliable, or they are unable to interpret and apply the results rationally. Being told there was evidence of opiate metabolite in a child's urine does a parent no good if she does not know whether it means the child had a poppy seed pastry for breakfast or used heroin on Saturday night.
Let me be clear: the question is not whether a non-invasive glucose test for glucose levels or a test collection system for drugs of abuse ought to be available, or available over the counter. The only issue is whether such a test is safe and effective for its intended use.
THE DIASENSOR GLUCOSE MONITOR
A product that illustrates the importance of assuring that a device works is the Diasensor glucose monitor. I know, Mr. Chairman, that you and other members of the Subcommittee will remember from your hearing last year that this is a novel, but still unproven, device which its inventors claim can be used to measure blood glucose levels in diabetics through intact skin of the forearm. FDA shares your interest and the interest of the 14 million Americans who suffer with diabetes in seeing a non- invasive glucose monitoring system available as soon as possible as an alternative to conventional finger stick testing.
If this product works as intended, it will be a major advance in the clinical management of diabetics. It could, for the first time, offer vision-saving and even life-saving benefits through tight glucose control to millions of Americans who find the finger stick method of monitoring glucose unacceptable. Tight glucose control usually means one has to measure blood sugar approximately 4 times a day (before meals and at bedtime) and take 3 to 5 adjusted doses of insulin. The goal is to have average glucose levels close to that of nondiabetics.
Let me explain briefly a bit about the Diasensor system. The Diasensor machine is about the size of a typewriter. Hopefully, it can evolve in the future into a truly portable product. When patients use the current model, it must be individually calibrated and does not always give a reading. While this system would reduce the number of conventional finger sticks required on a daily basis, it would by no means eliminate them. Patients would need dozens of finger sticks during initial calibration of the machine and more as part of monthly recalibrations. There also would be the need for confirmatory testing if the machine's readings were unexpectedly high or low or it could not give a reading on any given try.
The ability to achieve tight glucose control is dependent upon careful calibration of multiple daily insulin doses to the exact blood sugar levels. If glucose levels are not accurately measured, the results can be catastrophic. Incorrect insulin doses can cause insulin shock, with the potential for coma and seizures, or if one under doses you can miss the whole intent and benefit of tight control -- close to normal blood sugar levels over the long term, which will result in fewer complications of diabetes. So it should be obvious that the stakes of our making the right decision are very high indeed.
As with any medical device, we have to be sure that products work as they should and produce clinically dependable and useful results before they move into the clinical area and patients come to rely on them. When the manufacturer of the Diasensor first came to us in 1994, very limited clinical testing had been performed. In fact, only 6 patients -- all adults -- had been studied. The testing did not include episodes of very high or very low blood sugar readings, readings which signal a dangerous situation for the patient. Factors such as age, and variability in skin color, thickness and condition, were not adequately explored, nor was the potential for a difference between patients being treated with insulin versus oral medication fully assessed. Finally, the testing was conducted using a prototype machine that collected but did not analyze patient data or produce actual readings. This analysis was done only at the company.
In mid-1995, after several meetings with the firm, we were told that the original study had been expanded to three clinical sites and a total of 85 patients. We carefully examined the data, but in the final analysis, the results were not encouraging. The company cited machine break-downs, calibration difficulties and a number of other problems that surfaced when the product was tested on a wider scale.
In the end, of the 85 enrolled patients the company was able to provide data on only 8 who had results the firm considered successful. And in these cases, the machine produced accurate readings less than half of the time. In those cases when readings were obtained, multiple differences were observed in glucose levels when compared to results from back-up conventional testing.
Despite these very limited results, we thought the product and study findings should be presented to an independent panel for review. The company had, in fact, brought a group of their consulting experts to the Agency for a discussion of the test (machine). The Agency, however, maintained we should follow established procedure of Advisory Committee review by a standing panel of independent experts (SGEs). The company wanted to delay this discussion first for a meeting with staff from our Center which occurred in mid-November; and second for a meeting which occurred this past January with me and senior Agency officials. At that January meeting, we informed the company that an open discussion of the testing results with our advisory panel was the obvious next step and best means for expeditiously reaching a decision on market approvability and the possible need for further data.
During review of the product by our advisory panel on February 26 of this year, there was strong agreement among representatives from the diabetic community and panel members on the need for a non-invasive method for monitoring glucose levels. With respect to the Diasensor product, however, the panel was concerned about the limited number of study cases. Although the company felt the results were acceptable, the panel concluded that the device's performance was totally unacceptable in at least one case and, even in the best cases, was too unreliable and unpredictable. Panel members also felt that the Diasensor product should be tested in patients with low blood sugar levels to determine if the machine could provide accurate detection.
The panel review also produced an agreed upon course of action that entailed a broader study with a statistically valid sample. The need for such a study was perhaps best expressed by one of the advisory panel members. Dr. James Cooper, a physician and researcher with the Agency for Health Care Policy and Research and National Naval Medical Center, who has a daughter with diabetes, said:
"I do remain concerned about safety [of the device] and I don't think that the data that has been presented has established safety in the low blood sugar ranges, and I look forward to improvements and further data before the machine is marketed." (Clinical Chemistry and Clinical Toxicology Devices Panel, February 26, 1996)
Following the advisory panel meeting, members of my staff reiterated their commitment to working cooperatively with the firm to design a mutually acceptable study protocol in an effort to facilitate the clinical investigation. This intermediate step would allow the company to get the device into broader, but still limited, circulation in order to gain clinical experience, test public acceptance and recover the costs of development and production.
Regrettably, the company did not avail itself of this option. We have had no further substantive response from with the firm in the intervening months and have not received any new data or applications.
DRUGS OF ABUSE TESTING PRODUCTS
Let me say as both a father of three teenage children, and as a practicing emergency room physician who has seen first hand the ravages of illegal drug use, I understand very well the need to combat this public menace. The physical and psychological trauma it inflicts on casual users, addicts and their families can be devastating. Consequently, FDA is very sensitive to the need for readily available tests that can deliver accurate, meaningful results.
A critical first step to diagnosing and treating a person with a known or suspected drug problem is detection. FDA has granted market clearance for literally hundreds of screening products capable of detecting marijuana, cocaine, amphetamines, opiates (heroin, morphine and codeine) and phencyclidine (PCP). Of these, 60 are approved for use in point-of-care facilities. The broad availability of these products and the concomitant array of testing services afford parents, law enforcement authorities, and employers, among others, a broad range of options as they attempt to assure that home, school and workplace are drug free.
The Agency has some experience in assessing and applying the questions identified above to the data on a specific product intended for use in the home, the "Aware" home urine collection system. This test system consists of a specimen collection container, mailing kit and drug analysis for detection of use of drugs of abuse. In reviewing this product we also considered the applicability of questions that had been raised in the review of home collection systems for HIV tests to a proposed system for drugs of abuse test. On the Aware product, we concluded that home specimen collection made sense in the context of professional use labeling: that is, when there was a drug abuse counselor, psychologist or physician 1) involved in the decision to use home collection with direct reporting of results to the home, and 2) advising on interpretation of the results. We did not find sufficient evidence, however, to support direct to consumer OTC marketing of the test system. It was the Agency's view that the test accuracy and interpretability of the results required, the involvement of appropriate professionals.
As is the case for any in vitrodiagnostic, in considering whether a product is safe and effective for use as an OTC test for drugs of abuse, the Agency considers the risks associated with such use. The impact of errors related to tests for drugs of abuse -- what can happen if someone relies on an inaccurate result or does not accurately interpret a result -- extends beyond purely physical medical consequences. Errors can mean either that a parent falsely believes that a child is using drugs, or that a parent falsely believes that a child is not using drugs. Such errors could cause unnecessary and harmful distress to the parents and child. Regardless of whether one characterizes such impacts on people's lives and well-being as social and ethical issues or as mental health issues, they are clearly foreseeable adverse consequences. Drug abuse is far too serious a matter to allow either situation to be commonplace.
As I said above, the issue is not whether parents should have such products available: the issue is to ascertain that such products are safe and effective for their intended use. In the final analysis, our market approval decisions are based on the safety and effectiveness of the device and its impact on public health.
Mr. Chairman, we share the same goals of getting valuable new medical products into the hands of American consumers and health professionals as rapidly as possible. When the information is available on how properly to interpret medical test results, there are adequate directions for lay use, and we are satisfied that the tests can produce consistently accurate results, we believe the products should be available OTC. Unfortunately, to date such systems for drugs of abuse have not been shown to meet these criteria.
Our statutory and public health responsibilities compel us to make market approval decisions on the basis of evidence produced with high quality science. To settle for less would undermine the public trust and cause harm to scores of people throughout our nation.
This completes my prepared statement, Mr. Chairman. My colleagues and I would be pleased to respond to your questions.