Access to the PPM Journal and newsletters is FREE for clinicians.
13 Articles in Volume 10, Issue #5
An Osteopathic Approach to Fibromyalgia
Co-Morbid Psychological Disorders in Interventional Pain Management
Compliance Monitoring and Effective Risk Mitigation Strategy
Cultural Differences and Pain Management
Electronic Prescription of Controlled Substances
Kinetic Chain from the Toes Influences the Craniofacial Region
Non-responsive Pain Patients with CYP-2D6 Defect
Platelet Rich Plasma for Hamstring Tears
The Iontophore
The Treatment of Achilles Tendonitis Using Therapeutic Laser
Thoracic Facet Injections
Urine Drug Testing as an Evaluation of Risk
Vitamin D Levels In Pain and Headache Patients

Urine Drug Testing as an Evaluation of Risk

Implementation of multiple risk assessment and monitoring strategies appears to lower the rate of inappropriate UDT results in a clinical setting as in the experience of this private pain specialty practice.

Editor’s note: Practical Pain Management continues to be the forum for opinions, strategies, precautions and background information concerning the creation of a standardized REMS program. In this article, Dr. Ted Jones et al use both controlled Urine Drug Testing and a literature search to propose a method of evaluating the effectiveness of a REMS system.

Robert Foery, PhD, DABCC/TC

Since the mid-1990s, chronic opioid therapy (COT) for the treatment of non-cancer-related chronic pain has become an essential aspect of modern medical practice in many parts of the United States. At the same time, the diversion of opioid substances has increased precipitously.1 This likely is not a coincidence. Hydrocodone, for example, is the most widely prescribed opioid medication, with over 131 million prescriptions dispensed in 2006. Hydrocodone is also the most commonly identified opioid in forensic laboratories in which cases of overdose or drug-related death are investigated.2

Practitioners who prescribe COT are responsible for performing due diligence in identifying, monitoring, and managing risk factors as they determine which patients to treat. Practitioners need objective data, in addition to “gut instinct,” to determine whether their risk assessment techniques are reliable and valid. They need to know they are treating patients who have the ability and willingness to handle medications appropriately. One recent editorial opined that physicians must acquire basic knowledge of sub-stance abuse and become proficient in screening and assessing risk.3

A number of clinical techniques may be used to assess and reduce risk. Some practitioners require patients to bring their remaining pills in the original bottles to follow-up visits, so that the pills may be counted. Such pill counts provide simple, rapid identification of which patients are unable to appropriately manage the supply of medications with which they have been entrusted, though it cannot identify individuals who may hoard pills.3 Additionally, state prescription-monitoring program reports are also readily available in many states, allowing providers to identify which patients are “doctor shop-ping” for controlled substances.2

Urine drug testing (UDT) has become “an essential feature of pain management, as physicians seek to verify adherence to prescribed opioid regimens and to detect the use of illicit or unauthorized licit drugs.”4 A recent position paper by the American Pain Society on COT calls for UDT on high risk patients and for clinicians to consider UDT on patients who are not considered high risk or who have shown medication aberrant behaviors.5 Other authors argue for UDT to be done on all patients on COT.6 Some type of monitoring is essential for patients on COT and UDT is an increasingly important tool in this regard.

While some studies seek solely to identify the presence of illicit or unauthorized prescription substances, the absence of a prescribed opioid medication—hereafter described as an “unexpected negative” finding—also raises a variety of clinically relevant possibilities:

  • patient never took the medication
  • patient took the medication but was for some reason unable to absorb the medication
  • patient took the medication but for some reason was unable to excrete the medication or its metabolites
  • patient last took the medication too many hours before the test for a detectable level to be present
  • patient lost the medication
  • patient sold or otherwise illicitly distributed the medication
  • medication was stolen
  • any combination of possibilities

While UDT does not capture all aberrant behavior, it appears to be a good overall indicator of the effectiveness of risk assessment techniques. However, reliance on UDT to confirm adherence “…can be problematic if the results are not interpreted correctly, and evidence suggests that many physicians lack an adequate understanding of the complexities of UDT and the factors that can affect test results.”5 One study found that only 30% of the physicians surveyed could answer half of the questions offered about interpreting UDT findings correctly.7

UDT is still under-utilized by healthcare providers. One study of primary care physicians found that only 15% of physicians performed UDT on patients after opioids had been prescribed.8 UDT is administered based on the assumption that substances consumed, applied, injected, or smoked by patients will be excreted either in raw form or as predictable metabolites in the urine. Tests may vary with regard to sensitivity and specificity, but are generally able to detect the relatively recent use of specified substances above certain minimum concentrations, sometimes referred to as “cut-off” levels.

UDT and Risk Management

While healthcare providers who are treating patients with chronic pain should be administering UDT, questions are raised about what to expect on an aggregate basis. While each patient’s UDT can be judged and dealt with on a case-by-case basis, there is also the question of what the results are showing about the provider’s overall risk assessment and management methods. Risk assessment has become a standard of care for all providers of COT.5 If a provider is properly assessing risk, it stands to reason that the rate of inappropriate UDT results found in the overall practice patient population will be lower than it would be otherwise. While many factors are involved (patient population served, base rates of illicit drug use in the community, etc) the aggregate of UDT findings for a practice can be seen as a proxy for the success of one’s risk assessment and monitoring methods. Practices, however, rarely do a pre-post analysis when they implement new UDT processes, so interpretation of overall UDT results is problematic. One solution is to use pub-lished studies and their reported rates of inappropriate UDT results as a baseline for comparison. While practices vary widely, published rates of inappropriate UDT results offer one of the few available tools for interpretation and comparison of one’s own UDT findings.

The UDT Literature

A literature search was conducted through PubMed, PsychInfo and Medline on key terms including “urine drug testing” and “pain.” The following are summaries of studies that assessed the rate of abnormal UDT in a treated pain patient population (see Table 1). Katz et al followed 122 patients on chronic opiates for pain management for three years. They assessed patients over this period and found an overall 29% aberrant positive UDT rate (36 patients).9 More specifically, among these 36 patients, 26 of them were positive for an illicit drug (cocaine or marijuana), 17 were positive for an un-expected nonprescribed controlled drug (i.e., opioids, barbiturates, benzodiazapines) and three tested positive for ethanol. Unexpected negatives were not reported and apparently not assessed. Kell et al randomly tested 14,712 patients at 127 outpatient medical clinics.10 They found that 20% of patients tested were negative for prescribed opioids while 14% of patients showed positive for nonprescribed oxycodone. No testing for illicit drugs was reported. A study from physicians in an interventional pain management practice in Kentucky identified illicit drugs in the UDT of 16% of 500 patients.11 In that study, illicit use was defined as a UDT positive for THC, cocaine, and amphetamines or methamphetamines. “Opioid abuse” was assessed through records from physicians and pharmacies and not through UDT. Neither did they assess unexpected negative UDT results.

Table 1. Studies of Rates of Inappropriate UDT Results
Study Number of Patients Unexpected Positive % Unexpected Negative % Illicit Drugs % Total Inappropriate UDT results%
Cone et al (2008)14 10,922 Not tested Not tested 11% n/a
Fishbain et al (2008)17 review 20% combined 15% n/a
Fleming et al (2007)15 771 Not tested Not tested 24% n/a
Ives et al (2006)12 196 26% 8% repeated negatives 5% 32%
Katz et al (2003)9 122 14% Not tested 24% n/a
Kell (2005)10 14,712 14% (only oxycodone tested) 20% Not tested n/a
Manchikanti et al (2006)11 500 Not tested Not tested 16% n/a
Michna et al (2007)13 470 15% 10% 20% 45%
Schneider et al (2008)16 188 Not fully assessed Not fully assessed 10% n/a

Ives et al followed 196 patients for one year while on opiates for pain management.12 Results showed an overall opiate misuse rate of 32% or 62 patients. Of the total patients assessed 26% (50) evidenced cocaine or amphetamines in the urine toxicology screen (UTS) and 8% (15 patients) showed repeated negative urines for opiates despite being told after the first negative test that this should not happen. Five percent (9 patients) had inconsistent urines that were positive for opiates not prescribed by the physician. A study involving 470 patients tested with UDT in a pain clinic within an urban teaching hospital identified a 45% incidence of abnormal findings. This included 20% with an illicit substance and the remainder with either the absence of a prescribed opioid or with an adulterated specimen.13 It was felt that the relatively young age of those tested, as well as the presence of the practice in a large urban setting accounted for the disturbingly high incidence of abnormal test results. A very large study of 10,922 urine tests from COT patients identified illicit substances such as THC, amphetamines, cocaine, or gamma-hydroxybutyrate (Ecstasy) in 10.8% of the specimens.14 Unfortunately, that study did not specifically address the unexpected negative findings pertaining to the absence of prescribed medication and only addressed the presence of illicit drug use. A study of COT in a primary care setting was offered by Fleming et al.15 Twenty-four percent (177 patients) were found to be positive for illicit drug use. Nonprescribed opioid use or unexpected negative findings were not addressed.

Schneider and Miller published data from a pain management practice. Of the 188 patients tested, 15.4% showed unexpected findings.16 Those showing illicit drug use were 10% while those showing any unexpected result (including negative UDT) was 15%. However, the lab used in that study only tested for the opioids that were prescribed for the patient rather than all opioids. The authors noted that this omission may have contributed to an underestimate for those taking nonprescribed opioid medications. A follow-up survey was performed but this was only reported for those who had an expected UDT the first time and several patients could not, or were not, retested for various reasons. A UDT survey that examined all aspects of inappropriate UDT results was not conducted in that practice and the findings are likely an underestimate of the total inappropriate UDT results that would have been found with a more comprehensive approach.

Fishbain et al offer a review of the literature in a study of the development of addiction or aberrant drug-related behaviors.17 As a part of this review, they mention rates of inappropriate UDT results. They found a combined rate of unexpected negative UDT and presence of non-prescribed opioids in their reviewed studies of 13-40% with an average rate of 20%. The rate of UDT with illicit drugs was 4-57% with an average of 15%.

Taken as a whole, these data yield an overall rough range of total inappropriate UDT results of 32-45%. The range of unexpected positive UDT is 9-26% while the range of unexpected negative UDT is 8-20%. The range of illicit drug use found is 5-24%. While certainly not comprehensive or conclusive, these studies hint at estimated base rates of aberrant UDT findings for pain practices. The purpose of this study is to compare our practice’s rates with these. This comparison offers one possible way to assess the impact of a practice’s risk assessment and monitoring techniques.

Chou et al have reviewed the rationale and importance of assessing the risk of medication misuse in patients being considered for COT.5 However, particular strategies of risk assessment are still being developed. Written screening tools are recommended as helpful. These include the Screener and Opioid Assessment for Patients with Pain (SOAPP), the revised Screener and Opioid Assessment for Patients with Pain (SOAPP-R), the Opioid Risk Tool (ORT), and the Diagnosis, Intractability, Risk, Efficacy (DIRE) instrument. Urine drug testing, frequent visits, and pill counts are also advocated for use with some, if not all, patients. What is unknown is what the impact of these measures has on reducing adverse events and aberrant behaviors. This study was designed to investigate whether the risk-management strategies utilized in this particular private, multidisciplinary, pain clinic might result in a lower percentage of patients with aberrant UDT findings compared to rates identified elsewhere in the medical literature.

Materials and Methods

In this study, UDT specimens were collected in a sequential basis over a three-month period in 2008. Aberrant findings included the presence of illicit substances or ones from which the patient had been specifically instructed to abstain (including cocaine, THC, amphetamine/methamphetamines, and alcohol), the presence of a controlled substance not prescribed for a given patient, or the absence of a prescribed medication when the patient should have had access to and was supposed to have been taking that particular medication.

Data Collection

Between September 1st and November 30th, 2008, a total of 338 separate UDT specimens were collected from 332 existing patients who were receiving COT as a part of their pain treatment for varying lengths of time. The tests were collected by medical assistants on orders from nurse practitioners who manage the day-to-day prescribing of COT in this practice. Although the nurse practitioners and medical assistants varied from patient to patient and there was no forensic-quality chain of custody, the basic collection method was deemed simple enough to provide reliable results for the purposes of this study.

Patients completed forms indicating the medications and other substances they had taken or used in the preceding 30 days—whether prescribed or non-prescribed. Each patient produced a urine specimen into a new, wrapped specimen cup provided by National testing laboratory. Specimens were given unobserved but in a bathroom with no running water and no bags were allowed in the bathroom with the patient. Each specimen had to register between 90 and 100 degrees Fahrenheit on a cup-mounted liquid crystal thermometer to be considered a valid specimen. Patients were asked to hold their specimen until ready to be sealed. Each patient personally placed a serially-coded seal over the lid of the specimen cup. If the patient was unable to seal their specimen, then they watched the medical assistant do it so there would be no question regarding a mix up of specimens. Each patient would then personally place the sealed cup and the current medication list in the bag before watching the medical assistant seal the bag. Specimens were transported on a daily basis at the end of each business day via parcel service to Rhode Island for testing. Patients who refused, for whatever reason, to provide a UDT when asked were discharged from COT. The rate of refusal to do a UDT is very, very low.

Initial testing of each specimen consisted of enzyme-linked immunoassay (EIA) screening for qualitative identification of the presence of various com-pounds. The specimen is tested for all opioid classes, a variety of sedatives and a wide range of illicit drugs. Adjuvant medications are not assessed. Any positive finding on EIA screening was submitted to further testing using gas chromatography/mass spectroscopy (GC/MS) for quantitative measurement of that specific compound’s unique molecular pattern. If a certain chemical compound’s concentration fell below the EIA sensitivity level (i.e., the cut-off level) and the specimen was judged to be devoid of that compound, then GC/MS was not used to test below the cut-off level. The testing laboratory kept the specimens in cold storage for a certain period after receipt to allow for GC/MS testing below the EIA cut-off level, if requested by the ordering clinician. The EIA sensitivity levels for each drug category were lower than DOT cutoffs, making this testing more sensitive on the first pass analyses and triggering more GC/MS testing on samples for confirmation.

Risk Assessment and Monitoring

At our practice (Pain Consultants of East Tennessee in Knoxville, Tenn.) many risk assessment and monitoring measures are used. Patients are given a thorough physical examination and a detailed history is taken. Records from past providers are obtained and reviewed in detail including results from the Tennessee Prescription Monitoring Program patient profile (a statewide database of a patient’s history of filling scheduled medications). Opioid medications are rarely prescribed at the first visit and patients are so informed in an introductory letter before their first visit. In addition, all patients being considered for COT are sent for a 45-minute psychological interview with a psychologist who has expertise in addiction and chronic pain issues. Written screening tools are administered by the psychologist. During the time of this study, the screening tools used were the Screener and Opioid Assessment for Patients with Pain Revised (SOAPP-R) and the Opioid Risk Tool (ORT) and all patients were administered these as part of the evaluation. Rating categories by the psychologist are low, low-medium, medium, medium-high, and high. The interview and the screening tools result in an overall initial risk rating that medical staff use in deciding which opioids are to be used. Risk ratings, while never ruling out the use of opioids, have a direct impact on the types of opioids used as well as the fre-quency of UDT for each patient.

Along with a relatively standard medication agreement and consent form, patients are required to attend a 75-minute educational session, currently taught by the psychologist, which reviews the medication agreement as well as offering other information about treatment at our practice. Common treatment issues for pain patients and the importance of adherence to the medication agreement are stressed in an effort to reduce inadvertent violations of the medication agreement. This session is mandatory for a patient who is prescribed COT. UDT are administered with increased frequency depending the initial risk rating. The minimum frequency of UDT is annual with some patients who are assessed as high risk requiring more frequent UDT (e.g., weekly). Pill counts are done at each scheduled visit. Unscheduled called pill counts and UDT are done between visits if there is cause to suspect medication misuse, though this is generally rare. At each medical visit the patient’s situation (analgesia, adverse side effects, activity and aberrant behavior)18 are assessed and treatment changed accordingly. Prescription monitoring reports are also obtained periodically on existing patients with each UDT.

The Tennessee Board of Medical Examiner guidelines on prescribing addictive drugs are strictly adhered to.19 All appropriate non-narcotic treatment options are used before opioid medications are initiated. These include interventional treatments, referral to physical therapy, and referral to psychiatry for untreated comorbid psychiatric disorders. The practice has also set certain parameters on the opioid medications used. Potent short-acting opioid medications such as Actiq, Fentora and Roxi-codone are generally not used. As Oxy-contin is widely abused in this geographical area, it is not used for new patients and only for a select few existing patients with whom there have been no apparent aberrant behaviors. The practice emphasizes long-acting time-release opioid medications, use of long-acting medications alone or with no more than #60 short-acting medication doses per month, and monthly doses of more than #100 short-acting medications if short-acting medications alone are prescribed.

Thus, decisions on who is prescribed opioid medication, when they are pre-scribed it, what they are prescribed and how they are monitored for misuse is all assessed and made in deliberate fashion —with assessed risk being a major variable to be considered. Taken together, the above risk management processes provide a comprehensive system of assessing and managing risk that appears to us to be more than the norm in pain man-agement practices.

Some demographic data on the practice population were obtained in a concurrent study at the same practice.20 A sample of 244 patients at the practice for that study found that they were generally middle-aged (mean of 47 ± 13 years). The vast majority were white (96%) which is reflective of the general population of the geographical area. The majority were female (61%) and half (50%) were married. Most were unemployed or disabled (75%), with 18% working full-time. The most common primary pain complaint was low back pain with radicular symptoms (37%), with other common pain diagnoses being pervasive joint pain (18%) and specific joint pain (12%). These data appear generally comparable to the patient populations described in the other studies reviewed here.9,11,12,15


Of the 338 specimens collected between September 1, 2008 and November 30, 2008, a total of 330 of UDT were analyzed. The remaining eight specimens leaked in transit and were rejected by the testing facility. None were lost due to adulteration or problems at the point of collection. Thirty-six specimens (10.9%) contained unexpected positive findings for various non-prescribed medications. These com-pounds included hydrocodone, oxycodone, oxymorphone, fentanyl, morphine, methadone, butalbital, prop-oxyphene, diazepam, carisoprodol, or some combination of these compounds—none of which the patient had reported taking. Seven specimens (2.1%) had unexpected negative findings for the opioids pre-scribed (including either hydrocodone, oxycodone, oxymorphone, or morphine). No patient for whom methadone or fentanyl was prescribed had an unex-pected negative finding. Thirteen specimens (3.9%) revealed illicit or unauthorized licit substance use (including five with THC, three with alcohol, two with cocaine, two with amphetamines, and one with both cocaine and THC). Five patients were negative for prescribed opioid medication as well as positive for non-prescribed opioids. One was positive for illicit drugs as well as positive for non-prescribed opioids. In total, specimens from 50 patients (15.1%) revealed inappropriate medication use or illicit substance use or both.


The literature pertaining to UDT in pain management settings is growing but still relatively sparse. It must be remembered that COT for noncancer pain is still a relatively recent innovation in the medical community. Acknowledgement of the need for robust risk identification, management, and amelioration strategies within the COT community has been an even more recent development.

That being said, the data found in this study show a lower rate of aberrant UDT than studies previously reported in the literature (see Table 2). While general aberrant (inappropriate UDT results) rates in the literature vary from 32 to 45%, the data here find an overall rate of 15%. The rates of illicit drug use found in UDT in the literature on pain practices varied from 5% to 24% while the rate found here was 4%. Rates of unexpected positive UDT for opioids and rates of unexpected negative UDT were also lower than the literature rates. These data support the hypothesis that the extensive risk assessment and management techniques conducted by this pain practice do indeed appear to have lowered the rate of aberrant behavior below the rates previously reported in the literature. This is despite the fact that the region where this practice is located (southern Appalachia) has a documented rate of prescription drug abuse above the national average.21 This makes these findings even more encouraging regarding the risk management strategies employed at this practice.

Table 2. Summary of this Study’s Findings
  Unexpected Positive % Unexpected negative % Illicit Drugs % Total pts w/ inappropriate UDT results %
This study 11% 2% 4% 15%
Studies in Table 1 14-26% 3-20% 5-29% 32-45%

It is quite clear that these data are not definitive in their support of the hypothesis. There is no intra-practice comparison group for this population. Neither was a pre-post study done on the risk measures implemented here. This lack of comparison data makes the results here quite tentative. It may well be that there is some demographic characteristics in this patient population that leads to lower inappropriate UDT results that has nothing to so with the risk assessment processes used. While it appears from what demographic data we do have that our population is generally similar to others, this needs to be investigated further.

It may also be that the lack of truly random UDT could be lowering the rate of inappropriate UDT results. One can argue that patients know that UDT may be done at their visit and will work their illicit or inappropriate drug use around so as not to be caught and thus the rates of aberrant behavior found here are artificially low. First, in our experience, true random UDT are a significant burden to pain patients and penalize functional patients who have jobs and can not leave work at a moment’s notice to give a UDT at their pain practice (as failure to provide a UDT often means discharge). Second, the fact that patients know when their appointment is has not apparently yielded lowered rates in the literature. As illicit drug use often represents an addiction which is defined by an inability to control one’s use, addicted patients appear to be caught anyway as they typically do not have enough control to avoid using before a medical visit. An open question for further study is if true random UDT would produce a higher rate of aberrant findings.


Clearly more data is needed, both about this patient population and about other pain practices, before any definitive conclusions can be drawn about the impact of these risk assessment tech-niques on UDT findings. More studies are needed on UDT in pain practices to establish a general base rate of aberrant behaviors. However, these data do lean in support of the notion that good risk assessment and management strategies can be employed successfully and lower the rate of medication aberrant behavior and, hopefully, the rate of adverse events. We look forward to more studies and data from other practices to further investigate risk assessment techniques in the context of COT.

Last updated on: November 8, 2012
close X