Secondary Logo

Journal Logo

Economics, Education, and Policy: Special Article

Random Drug Testing to Reduce the Incidence of Addiction in Anesthesia Residents: Preliminary Results from One Program

Section Editor(s): Dexter, FranklinFitzsimons, Michael G. MD; Baker, Keith H. MD, PhD; Lowenstein, Edward MD; Zapol, Warren M. MD

Author Information
doi: 10.1213/ane.0b013e318176fefa

The incidence of substance abuse, including alcohol, among physicians is unknown.1 The incidence of substance abuse by anesthesiologists in training or in practice is also uncertain. Although the incidence of alcohol abuse among physicians appears to be no more prevalent than among other professionals, physicians may display a higher misuse of prescription opioids. Anesthesiology residents appear to have one of the highest known incidences of addiction to pharmaceutical substances of all groups of health care providers. The incidence of substance abuse is estimated as 1.6% of anesthesiology residents in the United States.2 This high incidence of substance abuse is believed due to a combination of workplace stress inherent in commencing this demanding profession (i.e., assuming responsibility for the safe induction, maintenance, and emergence of the anesthetized, paralyzed, often critically ill surgical patient), theorized second-hand occupational exposure and sensitization to the effect of opioids,3,4 and the ready availability of potent drugs used to anesthetize patients (particularly narcotics). Collins et al.’s5 survey of 111 training programs in 2005 reported that 80% of programs had experience with trainee impairment, primarily opioid abuse. Nineteen percent of programs reported at least one death due to overdose or suicide between 1991 and 2001. The highest risk of drug-related death for anesthesiologists is within the first 5 years after completion of medical school.6 Residents in anesthesia are over-represented in the Medical Association of Georgia’s Impaired Physicians Program.7 Additionally, anesthesiology residents and attending anesthesiologists have more years of life lost due to suicide and drug-related deaths than internists.6

For the past decade, anesthesiology residency programs have relied on education (lectures by recovered physicians, movies depicting the impact of physician drug addiction including loss of career or life, etc.) and strict control of substances (daily accounting, establishing operating room [OR] pharmacies, etc.), to detect and discourage substance abuse by anesthesiologists. These measures have not reduced the incidence of substance abuse.1 Recent technologic advances, including surveillance of drug transactions via anesthesia drug dispensing systems (Pyxis) along with analysis of anesthesia information managements systems and pharmacy information management systems may allow earlier detection of diversion by analysis of abnormal patterns of usage. These practices are not yet widely adopted.8

Other professions responsible for the lives of others (aviation, transportation, etc.) that have experienced problems with substance abuse are now required by the United States Congress to conduct random urine testing to attempt to reduce risk to the public. Illicit drug use decreased significantly after random testing was initiated in the US military in the early 1980s.9 Among impaired physicians, recovery is improved when random urine monitoring occurs because of the consequences of a positive substance screen.10,11 Presumably, physicians who understand the consequences of a positive urine screen would avoid use of illicit substances. Pre-employment drug testing of housestaff physicians at a teaching hospital has been reported.12 We are not aware of any civilian anesthesiology residency programs that require random drug testing. Mandatory random substance testing is common among physicians in recovery programs.

Over the past two decades, substance abuse within the residency of the Department of Anesthesia and Critical Care (DACC) at the Massachusetts General Hospital (MGH) has reflected the national incidence, despite education and strict accounting of drugs. We therefore decided to initiate preplacement and postemployment random urine testing of all anesthesia residents to attempt to deter and detect substance abuse.


A DACC Committee on Chemical Dependency was established in 2003 to investigate the feasibility, cost, and ramifications of random urine testing for occupationally available controlled substances. The committee was chaired by the department chairperson (W.M.Z.) and included the residency program director (K.H.B.), director of critical care, volunteer faculty members, chief residents, Office of General Counsel, and the medical director of the MGH Occupational Health Clinic (OHC). The senior hospital administration was informed and reviewed and approved our proposed program. The MGH OHC was invited to participate in the development of our program, because they had established protocols and instituted preplacement and “for-cause” urine testing for employees under the regulations of the Department of Transportation guidelines.

At the time our policy was approved (2003), mandatory testing had not been a requirement for becoming a resident in our program. Thus, current residents and those already committed to our program through the National Resident Matching Program were entered into the testing process on a voluntary and anonymous basis. The election of participation or nonparticipation of any resident was not known to our committee or the department leadership. However, beginning in 2004, we informed all resident applicants that random testing would be required if they matched to the MGH residency in anesthesiology. All new residents who began training from July 2005 onward have signed formal written contracts that acknowledge that they agree to our mandatory random urine testing policy.

We planned our urine testing frequency based on reported risk levels and our experience with substance abuse among our own residents. Residents in the first clinical anesthesia year (CA-1) are subject to at least two random tests per year with an additional 20% of the class subjected to a third test. Residents in their second and third clinical anesthesia year (CA-2, CA-3) are subject to at least one test a year with 30% of the trainees subjected to a second test. The additional testing is intended to eliminate any belief that once any individual completes a test, he/she would be exempt from screening for an entire year. We believe that more frequent testing would be disruptive to patient care.

The random urine testing protocol is in accordance with the previously established MGH OHC protocol for preplacement and for-cause testing. We decided not to observe collection of urine specimens to maintain privacy and dignity. Each urine sample is split to allow later confirmatory analysis if required.

The DACC committee determined the occupationally available substances that we would screen for. We focus on substances commonly available from the OR pharmacy. However, we omitted two substances, sufentanil and propofol, which are available. At the time we designed our panel, propofol was not considered a significant drug of abuse; however, propofol is an increasingly common substance of abuse13 and metabolites do appear in urine.14 Sufentanil was not added to our panel because it is rarely used in our department and would add substantial costs to testing. Distribution patterns for sufentanil are monitored by our pharmacy. The for-cause test includes a broader spectrum of substances of abuse. Our protocol allows screening for other substances as warranted by the circumstances (Table 1).

Table 1
Table 1:
Urine Testing Panel (2007)

Urine drug testing is performed at an outside facility. The sample is initially screened for substances by enzyme immunoassay. Confirmatory analysis of a positive immunoassay is via gas chromotograpy ± mass spectroscopy.

An independent certified medical review officer (MRO) receives, interprets, and reports all results of the workplace urine drug-testing program.15 When necessary, the MRO directly contacts the tested resident and determines whether the results were truly positive or due to confounding factors, such as taking a prescription medication. Results confirmed as positive by the MRO are reported directly to the Chairperson of the MGH DACC. Any positive immunoassay, which is ultimately deemed excusable by the MRO, is considered a negative result and is not reported to the department in any fashion.

Any resident suspected of illicit drug use because of their behavior is immediately removed from clinical duty and placed on a Medical Leave of Absence. The Hospital’s Professional Staff Health Status Committee is informed as required by Joint Commission on the Accreditation of Healthcare Organizations. The resident is confronted by at least two individuals, including the department chairperson and/or a designee. A psychiatrist experienced in dealing with drug abuse is in attendance whenever possible. The individual then undergoes mandatory urine drug testing. A positive urine test results in immediate referral for clinical evaluation or inpatient treatment. Outpatient treatment may occur through the Massachusetts Medical Society’s Physician’s Health Service.

Each state has its own requirements for reporting substance abuse among practitioners. Physicians are generally required to self-report their own dependence or abuse. Certain circumstances may dictate that a hospital directly report abuse to the state medical board. The American Board of Anesthesiology currently requires that a resident with documented abuse of alcohol or drugs receive an unsatisfactory evaluation for “Essential Character Attributes.” Other agencies such as the Department of Public Health or Federal Drug Enforcement Agency, and local police may need to be notified.

We have data on MGH DACC residents for the 6 yr before implementation of our policy (1998–2003) and for the 4 yr after testing began (January 2004 to December 2007) at MGH. The 6-yr period before testing was the time for which we could obtain reliable employment records. Resident drug abuse “events” were reported and analyzed for the year of occurrence and training level at the time of the event. Drug abuse events were defined as a positive urine test for illicit drugs, admitted diversion of a substance for personal use, direct observation of illicit use of a substance, or request for treatment for substance abuse or dependence. We calculated our incidence in events/resident year. The incidence before and after testing was compared using Fisher’s exact test.


Program Initiation

Forty-three percent (43%) of residents voluntarily elected to participate in random urine testing during the initial phase-in. The CA-1 class, which knowingly would be subjected to the most frequent testing, consented at the highest rate (65% of the class) (Table 2). Approximately one-third of the CA-2 and CA-3 classes consented. We found no evidence that introduction of our policy interfered with resident recruitment. The program matched every position offered between 2003 and 2006 and did so with our usual “ranked to match” ratio.

Table 2
Table 2:
Participation Rate for Voluntary and Anonymous Resident Urine Testing During Phase In

Since the inception of the urine testing program for residents, 236 urine tests (preplacement, random, and for-cause) have been performed (Table 3). Eighty-four preplacement, 150 random, and two for-cause tests have been performed.

Table 3
Table 3:
Total (Random and Preplacement) Urine Drug Tests (January 2004 to December 2007)

Substance Abuse Events

Overall, the incidence of substance abuse was 1% in the 403 resident-years during the 6 yr before testing began. During this same time, in the most highly vulnerable CA-1 residents, the incidence of drug abuse in the 138 resident-years was 2.2% (three events). During this time period, one event occurred in a resident during the second year (CA-2). In contrast, no events have occurred during 330 resident-years since testing began in 2004. The data are associated with a P value of 0.13 by Fisher’s exact test. Thus, we cannot conclude from our data whether there has been a decrease in the incidence of abuse.

Cost Analysis

The total costs of the testing program are composed of two elements, test collection and analysis and administrative fees. We estimate collection and analysis expenses at less than $20,000/yr. The program requires approximately 20% of a full time equivalent nurse practitioner, which amounts to $20,000 per year. We estimate that at full implementation, a total cost of $50,000 per year would be associated with testing all residents in our program at the desired rate.


Substance abuse appears to be more frequent in anesthesiologists than in other medical specialties. Education and substance control measures have not significantly reduced this incidence. This report describes the development and initial experience of mandatory random drug testing in a civilian academic department of anesthesiology. Only 8% of anesthesia residencies use random urine testing, and all of them are military programs.2 More than 60% of all program directors would approve of random urine screening.2 This is the first report of random urine drug testing of civilian anesthesia residents.

Our preliminary experience does not have the statistical power to determine the effect of urine drug testing on the incidence of resident drug abuse. Assuming an incidence of 1%, an estimated decrease of 20% and a power of 80%, it would require a comparison of 800 resident-years, with and without drug testing within our own program, to establish a statistically significant effect at the P < 0.05 level. Large multicenter studies will be required to demonstrate efficacy.

Anesthesiologists are responsible for the care of the public and this carries an ethical obligation to ensure that medical care is delivered by unimpaired individuals. Though empirical data indicate that the greatest risk of substance abuse is early in the anesthesia residency and diminishes with time, the hazard remains thereafter.16

Any drug testing program is expensive. However, the costs must be weighed against the benefits. The United States Postal Service estimated a cost savings of more than $100 million dollars over a 10-yr period when they studied the effect of preplacement urine testing between September 1987 and May 1988. The savings were primarily due to lower rates of absenteeism, reduced involuntary turnover, fewer Employee Assistance Program referrals, fewer medical claims, and less frequent disciplinary action than would have been realized had those testing positive been employed.17 Zwerling et al.18 evaluated the cost-benefit relationship of pre-employment drug screening among United States Postal Service workers in Boston, MA. They were unable to find a definite benefit, because the costs and benefits of drug testing are based on many assumptions, such as the incidence of substance abuse within the population tested.

We estimate our annual cost at less than $50,000. The financial benefits of a drug testing program are more difficult to determine. Detoxification may involve 3–7 days of intensive medical and psychiatric care at a fee of up to $9000. Inpatient treatment for 30 days is approximately $25,000. Residential treatment of up to 90 days is common. Outpatient treatment for 4–8 wk approaches $8000. Our previous experience suggests that residents generally take 6 mo to return to duty after substance abuse events. We estimate the cost of diagnosis, initial management, and lost clinical revenue at more than $60,000–$70,000 for a single event. When the costs of psychiatric care, follow-up through physicians’ health services for 3–5 yr, and mandatory drug testing for a physician in recovery is considered, the total cost of returning a physician to unrestricted medical practice is likely in excess of $100,000. Thus, if we deter a single physician from illicit drug use we can save a significant amount of money.

Establishing a program of substance testing was not easy and we encountered several limitations and challenges. First, the OHC moved off-site shortly after our testing program was established. This required our residents to change clothes, leave the hospital, and walk two blocks. We have addressed this by configuring a newly constructed toilet adjacent to the OR suites for urine sampling. Second, most substances are only detectable in urine for about 2–4 days after use. Detection times are dependent on dose, sensitivity of the method of detection, route of administration, duration of substance use, and variability between individuals.19 Our initial protocol permitted a 36-h window to appear for testing. This time frame increases the admittedly small likelihood of a false negative result. Because anesthesiologists work in an environment where their presence is critical at nearly all times, residents cannot be called away at a moment’s notice to undergo a test. Thus, we believe that we should allow some time for residents to make appropriate arrangements for urine sampling after clinical obligations are complete. However, the establishment of our new sampling site will enable reduction of the time between notification and mandatory urine sampling. We are planning to reduce the interval.

Administration of a urine-testing program in a large academic anesthesia department where residents cover multiple services (intensive care unit, pain, obstetrics) in several hospitals, and spend time in distant locations (laboratories, simulator, etc.) presented additional problems. All randomization, notification, and record keeping in our program occurs through the OHC. The clinic contacts individuals directly via the hospital paging system. If a resident does not respond and notify the clinic of his/her time of availability after three pages, a member of the Committee on Chemical Dependency or the residency Program Director is notified. If the paged physician is not able to free himself from clinical responsibilities for urine sampling he/she is rerandomized to another day. This provides a potential breach in the random nature of the program.

The greatest problem to date is that the number of tests obtained (236) has fallen short of the number called for (Table 3). The rate of testing increased progressively each year on the CA-1 and CA-2 residents. In 2007 we achieved an average of about 1.5 tests per CA-1 resident and approximately one test per CA-2. This increase was not observed in the CA-3 residents, and only approximately 1/3 of CA-3 residents were tested in calendar year 2007. We speculate that this is due to the many rotations outside of the MGH ORs and this hospital. This unsatisfactory rate needs to be addressed. Establishment of a comparable testing program at other hospitals through which residents rotate would solve much of this problem. Better coordination of the Occupational Health Service with our scheduling system so residents on vacation are not selected for sampling and so residents at other hospitals can be paged through the paging system of the outside rotations will likely increase adherence. We expect that our proximate and new testing location and a more intensive follow-up via our paging system will increase the number of tests we successfully perform each year.

The risk of a false positive result is also a major concern. Indeed, false positive results have been reported with urine screening. Rifampin and fludroquinolones have been reported to cause false positive testing for opiates20,21 and oxaprozin for benzodiazepines.22 Nonsteroidal antiinflammatory drugs have been reported to cause false positive results for barbiturates and cannabinoids.23 We experienced one false positive urine screen for morphine, which was attributed by independent MRO review to the consumption of poppy seeds in a bagel, and was considered a negative result. This event prompted us, in conjunction with the Occupational Health Service to raise our confirmatory level from 300 to 2000 ng/mL, the level accepted by the federal government. Poppy seeds have traces of morphine and codeine and positive urine screens have been reported after consuming them.24

Education of physicians and their families, and stringent substance control have not reduced the incidence of substance abuse in anesthesiology residents. We envision substance testing as one component of a policy intended to prevent individuals from embarking on a course of behavior with potentially deadly results. In fact, nearly 20% of addicted individuals will die or nearly die (require resuscitation) before substance abuse is detected.2 A large and sufficiently powered study to determine the impact of preplacement and random urine testing on the incidence of resident substance abuse is very difficult within a single anesthesia department. We do not know if those residents who were not tested in this preliminary experience were those at highest risk and avoided testing. It is possible that our announcing a mandatory policy served to drive at-risk resident applicants to other training programs that do not test and shifted the burden to another academic department within the field of anesthesia. A multidepartment study would need to address all causes of false negative results, including delay in testing after notification, failure to appear for testing, and substances that are not included or difficult to detect by urinalysis.

We have demonstrated that it is feasible to establish and operate a random urine testing program for civilian residents in anesthesiology and define some of the impediments to complete adherence and suggest measures to improve (and approach complete) adherence. Although our data may suggest a decrease in the incidence of substance abuse, we cannot make this definite conclusion. Only larger, presumably multi-institutional, studies will have the power to determine whether random testing will decrease substance abuse in anesthesiology residents. If it were proven to reduce the incidence of abuse in anesthesia residents, a program could be expanded to other specialties and groups at risk for substance misuse.

Physicians are placed in a position of direct responsibility for their patient’s safety. A system-level effort to address the effectiveness of substance screening among all physicians has been suggested.25 Substance abuse by physicians endangers patients and jeopardizes the public’s trust in our profession. The profession of anesthesiology should be encouraged to conduct a large trial powered to determine if randomized urine testing will reduce the incidence of drug abuse in anesthesia residents.


The authors thank Hui Zheng, PhD, for statistical assistance, Dr. Debra Weinstein, Vice President for Education, Partner’s Healthcare, Dr. Jo Shapiro, Senior Associate Director of Partners GME, and Ann Prestipino, Senior Vice President for Surgery and Anesthesia Services and Clinical Business Development, Massachusetts General Hospital for their administrative support instituting the Random Urine Testing Program.


1. Brewster JM. Prevalence of alcohol and other drug problems among physicians. JAMA 1986;255:1913–20
2. Booth JV, Grossman D, Moore J, Lineberger C, Reynolds JD, Reves JG, Sheffield D. Substance abuse among physicians: a survey of academic anesthesiology programs. Anesth Analg 2002;95:1024–30
3. Gold MS, Byasr JA, Frist-Pineda K. Occupational exposure and addictions for physicians: case studies and theoretical implications. Psychiatr Clin North Am 2004;27:745–53
4. McAuliffe PF, Gold MS, Bajpai L, Merves ML, Frost-Pineda K, Pomm RM, Goldberger BA, Melker RJ, Cendan JC. Second-hand exposure to aerosolized intravenous anesthetics propofol and fentanyl may cause sensitization and subsequent opiate addiction among anesthesiologists and surgeons. Med Hypotheses 2006;66:874–82
5. Collins GB, McAllister MS, Jenson M, Gooden TA. Chemical dependency treatment outcomes of residents in anesthesiology: results of a survey. Anesth Analg 2005;101:1457–62
6. Alexander BH, Checkoway H, Nagahama SI, Domino KB. Cause-specific mortality risks of anesthesiologists. Anesthesiology 2000;93:922–30
7. Talbott GD, Gallegos KV, Wilson PO, Porter TL. The Medical Association of Georgia’s Impaired Physician’s Program. Review of the first 1000 physicians: analysis of specialty. JAMA 1987;257: 2927–30
8. Epstein RH, Gratch DM, Grunwald Z. Development of a scheduled drug diversion surveillance system based on an analysis of atypical drug transactions. Anesth Analg 2007;105:1053–60
9. Bray RM, Marsden ME, Rachal JV, Peterson MR. Drugs in the military workplace: results of the 1988 worldwide survey. In: Gust SW, Walsh JM, Thomas LB, Crouch DJ, eds. Drugs in the workplace research and evaluation data. Vol. II. NIDA Monograph Series No. 100. Washington, DC: NIDA, 1990:25–44
10. Jacobs WS, Repetto M, Vinson S, Pomm R, Gold M. Random urine testing as an intervention for drug addiction. Psychiatric Ann 2004;34:781–4
11. Shore JH. The Oregon experience with impaired physicians on probation. An eight-year follow-up. JAMA 1987;257:2931–4
12. Lewy RM. Pre-employment drug testing of housestaff physicians at a large urban hospital. Acad Med 1991;66:618–9
13. Wischmeyer PE, Johnson BR, Wilson JE, Dingmann C, Bachman HM, Roller E, Tran ZV, Henthorn TK. A survey of propofol abuse in academic anesthesia programs. Anesth Analg 2007;105: 1066–71
14. Simons PJ, Cockshott ID, Douglas EJ, Gordon EA, Hopkins K, Rowland M. Disposition in male volunteers of a subanesthetic does of an oil in emulsion of 14C-propofol. Xenobiotica 1988;18: 429–40
15. Sgan SL, Hanzlick R. The medical review officer. A potential role for the medical examiner. Am J Forensic Med Pathol 2003;24: 346–50
16. Gravenstein JS, Kory WP, Marks RG. Drug abuse by anesthesia personnel. Anesth Analg 1983;62:467–72
17. Peat MA. Financial viability of screening for drugs of abuse. Clin Chem 1995;41:805–8
18. Zwerling C, Ryan J, Orlav EJ. Costs and benefits of preemployment drug screening. JAMA 1992;267:91–3
19. Verstraete AG. Detection times of drugs of abuse in blood, urine, and oral fluid. Ther Drug Monit 2004;26:200–5
20. Zacher JL, Givone DM. False-positive urine opiate screening associated with fludroquinolone use. Ann Pharmacother 2004;38:1525–8
21. Daher R, Haidar JH, Al-Amin H. Rifampin interference with opiate immunoassays. Clin Chem 2002;48:203–4
22. Fraser AD, Howell P. Oxaprozin cross-reactivity in three commercial immunoassays for benzodiazepines in urine. J Anal Toxicol 1998;22:50–4
23. Rollins DE, Jennison TA, Jones G. Investigation of interference by nonsteroidal anti-inflammatory drugs in urine tests for abused drugs. Clin Chem 1990;36:602–6
24. Struempler RE. Excretion of codeine and morphine following ingestion of poppy seeds. J Anal Toxicol 1987;11:97–9
25. Leape LL, Fromson JA. Problem doctors: is there is system-level solution. Ann Intern Med 2006;144:107–15
© 2008 International Anesthesia Research Society