2 February 2003

Urine testing in drug treatments: Is it effective? Is it ethical? Is it necessary?

re: Goldstein A, Brown BW. Urine testing in methadone maintenance treatment: applications and limitations. J Substance Abuse Treatment 2003 25; 2: 61-63

Commentaries by Coppola; Dupont; Heaps and Lurigio; Parrino.

In addressing urine testing in methadone treatment, these authors seem to have forgotten that methadone prescription by doctors is a time-honoured practice in which compulsory urine testing is not only unnecessary, but also possibly unethical. There is no scientific evidence cited by them or anyone else to my knowledge to show that such testing reduces drug use or improves treatment outcomes in other respects. Indeed, the concept runs contrary to behavioural theory as well as to Hippocratic principles. That is not to say urine testing should not be done. It should. But it should be voluntary. Urine testing may be a valuable measure of outcome for practitioners in evaluating their own practices, rather than necessarily judging their patients. Further, it might be reasonable to suppose that testing improves outcomes if it is part of an environment that promotes communication with patients, even if there is no evidence currently. In this way, urine testing may be a better judge of treatment programs than of the patients in them.

The article by veteran researchers Goldstein and Brown and the series of commentaries published with it show the complete ‘divide’ which exists between good medical treatment and American drug treatment program protocols. Rather than a considered description of the subject, these authors write an ‘introduction’ followed immediately by ‘conclusions’. Neither the authors, nor their commentators give a clear description of how the use of such testing fits into the ethical framework of addiction treatment. We treat obesity and weigh people. We treat diabetes and measure the patients’ sugar levels. We treat tobacco addiction and alcoholism with a variety of means and outcome measures. A treatise on when and how to order these measures would be based on practical experience with the modalities available and their costs/effectiveness. Nobody would suggest withholding or altering treatment substantially if a patient declined to submit to testing on one or two occasions. At the same time, treatment would be fraught if there were no outcome measures.

Each of these authors appears to assume that urine testing reduces drug use and that to be effective, urine testing must be random and frequent, with an implication that it must be compulsory. None of them seriously broach the sensitive but important subject of ‘supervision’ or ‘witnessing’ of urine samples for accuracy and to prevent substitution. This should be implemented in close cooperation with consumer groups, needing scrupulous care in maximising privacy while maintaining the integrity of the process. In some American states it is considered illegal to ‘witness’ such testing. Patients who refuse testing should not be denied treatment although other means need to be sought to demonstrate their progress.

A senior colleague once said to me that the only unquestioned purposes of urine testing is in the course of research and practice evaluation, as well as to make money for pathologists. This should involve consent of the subjects. Most authorities would probably agree that a full dependency assessment also includes urine toxicology. Indeed, it is probably as fundamental as taking the blood pressure and urine test as components of a full cardiovascular assessment. But patients should never be forced to do such testing and the results should only ever be used in their direct best interests.

From the two misconceptions above, these learned American experts delve ever deeper into the realm of speculation, moving further away from science and ethics. How can we justify doctors ordering compulsory expensive and intrusive tests which have never been proven to be of any benefit? Indeed, in many programs the results may directly cause patients to be deprived of take-away doses … and this is lauded by one of the commentators as a ‘benefit’ to the long suffering patient! Such actions could lead to the loss of a job or prevent children getting to school on time. Take-home dosing should only be denied in cases where is it is causing greater risks than benefits, and it should never be withdrawn based on an individual urine test result. It is clear that with improved education that American treatment programs can respond to evidence and improve the quality of the treatment given (d’Aunno 1999). It is difficult to do so, however, when state mandated practice guidelines conflict with the evidence, as with take-home dose provisions, dose levels, counselling and urine testing frequency in some jurisdictions.
These authors seem to be learning things well known to every young medical student, yet they ignore some of the refinements of urine testing which can indeed be very useful for patient care. Results can be wrong for a variety of reasons such as labelling, storage, substitution, ‘inhibitors’, lab errors etc. Doctors’ usual response to unexpected results is to have them repeated. Such testing usually takes 24 hours. While some clinics now do them on site, few can provide a result and a medical opinion in the short time a patient is present for medication. Dupont seems to have realised that urine testing is not ‘delivering’ what we might like, has suggested the role for hair or saliva or sweat tests. These research tools are of interest, but are hardly near implementation, being without a solid research base nor current clinical guidelines.

These authors all seem to believe that a positive methadone result is an important finding to confirm compliance with treatment. Yet these patients are nearly all receiving one or two doses per week under supervision, making such testing superfluous as methadone remains detectable in the body for several days. Indeed, a negative test result is far more likely to be from a tampered urine sample or a lab error than a case of a patient not taking their medicine in my experience. Another unusual but interesting finding is a specimen which is positive for methadone but negative for methadone metabolites. This usually indicates a substituted specimen from a non-methadone patient with a soupcon of medicinal methadone added to the urine. Such unusual but worrisome results should be used with the utmost sensitivity and in the patient’s best interests. Why would patients be driven to do such things in a caring treatment program? How can there be a productive outcome to such a seemingly distrusting environment? Obviously repeated episodes would indicate a change of treatment, but a single such event may lead to a watershed if it leads to improved communication and/or more frank discussion in the therapeutic milieu.

Goldstein and Brown’s response to poor outcomes with standard testing is to do it more frequently and introduce true randomisation so clients could be urine tested on ANY day (even two days in a row). Also recognising the limitations of urine testing, Dupont suggests that hair testing is ‘impervious to cheating’, although goes on to say it is not without problems, as well. We have heard similar claims for drugs in sport or the workplace, equally without any serious evidence of safety nor effectiveness of the intervention.

My own attitude is that urine testing can and should be used by patients to demonstrate their progress in treatment. This should be used, along with evidence of employment or education, a drug ‘diary’, rent receipts, healed veins and other indicators of stability in order to qualify for less supervision in treatment, and thus more ‘normalisation’ of their lives.

Patients should be able to volunteer for a supervised (witnessed) urine test at a point when they wish to have it known that they are moving away from illicit drug use. In my experience, a person who uses a significant quantity of street heroin has a positive ‘EMIT’ opiate test (Syva Corporation, California) for at least 5 and often 6 days. Such a positive test may, however, equally indicate the use of codeine, cough suppressants or even poppy seeds. Hence a positive test is less ‘useful’ than negative one or the negative predictive value of the test is higher than its positive predictive value for heroin use. A confirmed test for morphine is suggestive of, but not proof of illicit heroin use. And in some countries heroin is not always illicit. The use of buprenorphine is directly parallel to methadone maintenance and urine testing principles should be the same in nearly all respects. To date most tests cannot detect buprenorphine, but this does not reduce their utility significantly unless there is no supervision of doses. Even in America, doctors may request pharmacists to supervise some buprenorphine sublingual doses - although for unknown reasons this has not been widely recommended, despite nearly all the excellent American research on buprenorphine using dose supervision.

I have discussed the effect of urine testing on treatment outcomes with an Australian expert who has confirmed that there is no good evidence in his experience and reading either for or against it. He pointed out further, that ‘punishment’ rarely has a beneficial affect on behaviour. On the other hand, there is ample evidence in the literature that ‘reward’ mechanisms do have a positive effect on outcomes. It is a tragedy that so many modern experts appear to be convinced of the utility of unproven modalities and that many such practices which may be outmoded, expensive and even dangerous are persisted with, despite the evidence.

An enquiry of the corresponding author about these matters yielded a reminder that the main thrust of the article was to point out that the probability of detecting a "dirty" urine is ‘vanishingly small’ in most cases of illicit drug use. This rather simplistic piece of mathematics was based on the unreferenced premise that negative tests only indicate an absence of heroin use in the previous 24 hours. The one area we seem to agree on is that most patients, even long term, stable people, should have up to 8 urine tests per year as a part of their normal treatment. But in my view such testing should be patient-initiated and the results should only be used as a guide to the patients’ level of stability along with other clinical indicators.

There is also a political reality that in most countries where addiction treatment is state-funded, community expectations are that urine testing, along with other safeguards and supports, be used as part of a strict and effective system of ethical health care delivery for those addicted to drug and alcohol.

While it can sometimes be disheartening to read superficial and unscientific research in our field, it is always uplifting and reassuring to receive feedback from so many caring and intelligent colleagues who are working effectively with dependent patients in their practices, clinics, hospitals, etc in many different parts of the world. More strength to you all!

Comments by Andrew Byrne ..