Business Ethics the topic of the ethics of drug testing in the employment setting.

workplace_drug_testing.pdf

Home >Business & Finance homework help >Management homework help >Business Ethics the topic of the ethics of drug testing in the employment setting.

Racial/Ethnic Differences in Report of Drug Testing Practices at the Workplace Level in the U.S.

William C. Becker, MD,1,2 Salimah Meghani, PhD,3 Jeanette M. Tetrault, MD,2

David A. Fiellin, MD2 1VA Connecticut Healthcare System, West Haven, Connecticut 2Yale University School of Medicine, New Haven, Connecticut 3University of Pennsylvania, Philadelphia, Pennsylvania

Background and Objectives: It is unknown whether racial/ethnic differences in report of workplace drug testing persist when analyzed within and across various occupations. We sought to examine the association between worker demographics, workplace characteristics, and report of employment in a workplace that performs drug testing. Methods: We performed a cross‐sectional study of the 2008–2010 National Survey on Drug Use and Health examining the relationship between race/ethnicity and report of workplace drug testing among employed, white, black, or Hispanic respondents �18 years old. In logistic regression analysis, we adjusted for demographic, occupa- tional, and other relevant variables and performed stratified analyses among three specific occupations. Results: Among 69,163 respondents, 48.2% reported employment in a workplace that performs drug testing. On multivariable analysis, younger age, male sex, black race, income greater than $20,000, completion of high school and non‐urban residence were associated with report of drug testing at one’s workplace among the full sample as were non‐white collar occupation, work in medium or large workplace, and absence of other substance abuse/dependence. In stratified analyses, black race was associated with report of workplace level drug testing among executive/administrative/managerial/financial workers and tech- nicians/related support occupations; Hispanic ethnicity was associated with the outcome among technicians/related support occupations. Conclusions: Racial/ethnic differences in report of workplace drug testing exist within and across various occupations. These differences have important public health implications deserving further study. Scientific Significance: Increased report of drug testing where racial/ ethnic minorities work highlights the potential bias that can be introduced when drug testing policies are not implemented in a universal fashion. (Am J Addict 2014;23:357–362)

INTRODUCTION

After and in spite of early controversy about the ethics of U.S. workplace drug testing,1 its prevalence has been increasing since the 1980s and is now commonly conducted in both prospective and current employees. A study from 2007 reported that 46% of workers and 90% of Fortune 200 companies report some form of drug testing.2 As employee drug use has been associated with accidents at work and elsewhere, absenteeism, lower productiv- ity, and decreased profits, many employers see drug testing as a cost‐saving measure.3,4 Additionally, some data demonstrate a deterrent effect of testing on worker drug use.5 Typical workplace drug testing uses urine samples, but blood and hair testing are also occasionally used. Urine tests usually include the so‐called “Federal 5” substances: marijuana, opiates, cocaine, PCP, and amphetamines but may be expanded to include other substances such as methadone, benzodiazepines, barbiturates, and oxycodone.

In terms of potential benefits and risks, much is at stake for prospective and current workers who undergo workplace drug testing. On the benefit side, a positive drug test could lead to a worker receiving treatment for a previously unrecognized substance use disorder through enrollment in an employee assistance or other available program. From a public health standpoint, workplaces may be made safer for both employees and the public if individuals with drug use are either not hired in the first place or suspended for ongoing drug use. On the other hand, employers may not hire applicants with positive tests that may have resulted from sporadic, low‐level use.6

Furthermore, workers with evidence of substance use may lose their jobs and health insurance for themselves and their families, with or without an opportunity for treatment. As employment and health insurance status are two of the most significant social determinants of health,7,8 it follows that workplace drug testing can have profound and long‐term effects on individuals’ health.

Since the results obtained from workplace drug testing may have major health‐related consequences for employees, an

Received December 11, 2012; revised June 4, 2013; accepted August 10, 2013.

Presented as an oral presentation at the Society of General Internal Medicine 33rd Annual Meeting, Minneapolis, MN, April 2010.

Address correspondence to William C. Becker, MD, 950 Campbell Avenue, West Haven VAMC, Mail Stop 151B, West Haven, CT 06516. E‐mail: [email protected].

357

equitable system of implementation is critical to avoid burdening vulnerable populations with a disproportionate share of risk.9 Universal testing, exemplified by the U.S.’ Omnibus Transportation Employee Testing Act of 1991 requiring all transportation industry employees in “safety‐ sensitive” positions to have drug testing, is one strategy to promote equity. However, beyond these laws pertaining to a small subset of occupations, testing policies in the U.S. are subject to wide variation in state laws/regulations concerning when, where and how employers can implement workplace drug testing programs. Generally, employers must have a written policy requiring new hires to be drug‐free; written notice must be given before the applicant may be tested. The same testing program must be implemented for all applicants in a particular category though a company may require testing for pre‐specified positions only, provided this is stated and justified in the written policy.10

Regarding healthcare settings (as opposed to workplaces), studies have hypothesized that the absence of universal drug testing policies across or within practices promotes differential implementation by race, since, in the absence of protocols, prejudice and bias may influence decision‐making.11,12 Indeed, in clinical encounters, differential implementation in drug testing has been reported: non‐white individuals have been shown to be more likely to get tested.13 Similarly, the question of equity or fairness within U.S. workplace drug testing programs is of particular significance given the potential health‐related consequences noted above. Previous research using 2002–2004 data demonstrated increased reports of workplace drug testing by black workers, by individuals employed in certain occupations, such as transportation and material moving workers, and by individuals working in larger workplaces.14 However, the question of whether racial differences in reports of workplace drug testing persist both within and across occupations remains unanswered. As such, the objectives of our study are to determine whether race/ ethnicity is associated with differential report of being employed in a workplace that performs drug testing in an updated data set, and to determine whether differences persist within and across occupations.

METHODS

We performed a pooled cross‐sectional analysis of the U.S. National Survey on Drug use and Health (NSDUH) years 2008 through 2010.15–17 The NSDUH is an annual, self‐report survey of the civilian, non‐institutionalized U.S. population that has undergone multiple validations over the past 30 years, including a recent analysis of test‐retest reliability of key measures.18 Potentially sensitive topics, which constitute a large proportion of the survey, are queried using a computer‐ assisted, self‐administered platform that has also been extensively validated for accuracy. Youths are over‐sampled and larger states contribute more respondents.17 This study was considered exempt from full review by the VA Connecticut

Healthcare and Yale University Human Investigations Committees.

Analytic Sample Our analytic sample included individuals who self‐reported

past‐year, part‐ or full‐time employment; age greater than or equal to 18; white, black, or Hispanic race/ethnicity and an answer to the question: “Does your workplace ever test its employees for drug use?” Each survey year had approximately 3% missing data pertaining to this question. Response rate for the overall survey averaged 76% per year.

Independent Variables Our independent variable of interest was race/ethnicity, for

which we included self‐reported white, black and Hispanic respondents. Covariates in our analyses fit in the following three categories: (1) demographic, which included age, sex, marital status, income, education, and population density of residence; (2) occupational, which included occupation type, for which respondents could choose among 15 options that we classified dichotomously as white collar versus trade/craft/ labor based on the U.S. Office of Personnel Management’s system19; and workplace size, categorized as small (less than 25 employees), medium (25–99 employees) or large (100 or more employees); and (3) other potential confounders that might influence the odds of reporting workplace testing which included past‐year arrest, alcohol abuse/dependence and other substance abuse/dependence.

Dependent Variable The dependent variable in all analyses was the binary

response to the question “Does your workplace ever test its employees for drug use?”

Data Analysis Data analysis, using SAS 9.0 Cary, NC, proceeded in

several steps. First, we evaluated frequencies of the indepen- dent and dependent variables described above. Next, we assessed bivariate associations using chi‐square tests and multivariable associations using logistic regression between each independent variable listed above and report of workplace drug testing. Since occupation type was associated with both race/ethnicity and report of workplace drug testing, and thus could confound the relationship between the two, we then performed stratified multivariable models of drug testing using logistic regression in three different occupations, namely (1) executive/administrative/managerial/financial, which had the lowest proportion of non‐white respondents; (2) technicians/ related support occupations, which had the highest proportion of non‐white respondents, and (3) transportation & material moving, an occupation wherein federal law requires drug testing as noted above, using logistic regression. These models did not contain the occupation classification variable as it would have caused co‐linearity. We used sample weights to adjust for the complex sampling design of the survey using SAS‐callable SUDAAN.

358 Differences in U.S. Workplace Drug Testing July–August 2014

RESULTS

There were 69,163 respondents meeting our inclusion criteria (Table 1). This sample of working individuals was younger, had higher income and education attainment and lower prevalence of substance use disorders compared to the general U.S. population (data not shown).

In the full sample, 48.2% indicated that drug testing occurred at their workplace. On bivariate analysis comparing

demographics and workplace drug testing (Table 1), respon- dents ages 26–34 and 35–49, men, black respondents, individuals with incomes $20,000 and above, respondents who graduated high school and non‐urban respondents were more likely to report workplace drug testing. With regard to occupational variables, respondents in non‐white collar occupations and individuals working in medium or large workplaces were more likely to report workplace drug testing. Bivariate analysis revealed an increased prevalence of

TABLE 1. Sample characteristics, proportion with report of workplace drug testing (N ¼ 69,163)

Characteristic Overall n (%) �

Report of workplace drug testing

n ¼ 32,421 (48.2) No report

n ¼ 36,742 (51.8) Chi‐square (p‐value)

Worker demographics Age

18–25 33,377 (15.2) 14,684 (44.4) 18,693 (55.6) 93.2 (<.0001) 26–34 11,501 (19.0) 5,867 (51.9) 5,634 (48.1) 35–49 16,886 (34.6) 8,626 (51.9) 8,260 (48.1) �50 7,399 (31.2) 3,244 (43.8) 4,155 (56.2)

Women 34,637 (46.6) 14,679 (43.0) 19,958 (57.0) 216.4 (<.0001) Men 34,526 (53.4) 17,742 (52.8) 16,784 (47.2) White 49,118 (72.9) 21,952 (45.9) 27,166 (54.1) 186.3 (<.0001) Black 8,347 (12.0) 5,159 (63.6) 3,188 (36.4) Hispanic 11,698 (15.2) 5,310 (47.2) 6,388 (52.8) Household income, $

<20,000 12,802 (11.2) 4,954 (39.8) 7,848 (60.2) 44.6 (<.0001) 20,000–49,999 15,847 (19.8) 7,727 (48.9) 8,120 (51.1) 50,000–74,999 21,020 (31.0) 10,478 (50.9) 10,542 (49.1) �75,000 19,484 (38.1) 9,262 (48.2) 10,232 (51.8)

Education �11th grade 9,032 (11.2) 3,904 (45.8) 5,128 (54.2) 68.9 (<.0001) Graduated high school 22,057 (29.8) 11,427 (54.2) 10,630 (45.8) College 38,074 (59.1) 17,090 (45.7) 20,984 (54.3)

Married 26,099 (55.9) 12,961 (48.7) 13,138 (51.3) 2.9 (.09) Not married 43,064 (44.1) 19,460 (47.6) 23,604 (52.4) Urban 29,191 (51.9) 13,273 (46.2) 15,918 (53.8) 40.2 (<.0001) Non‐urban 39,972 (48.1) 19,148 (50.5) 20,824 (49.6)

Occupational variables White collar 35,680 (57.5) 15,735 (44.0) 19,945 (56.0) 206.1 (<.0001) Non‐white collar 33,411 (42.5) 16,654 (54.0) 16,757 (46.0) Workplace size

Small (<25) 33,962 (47.3) 10,344 (30.2) 23,618 (69.8) 914.8 (<.0001) Medium (25–99) 16,141 (22.2) 8,304 (54.5) 7,837 (45.5) Large (100 or more) 18,881 (30.5) 13,675 (71.7) 5,206 (28.3)

Other covariates Alcohol abuse/dependence 8,635 (9.0) 3,705 (45.6) 4,930 (54.5) 6.9 (.01) No alcohol abuse/dependence 60,528 (91.0) 28,716 (48.5) 31,812 (51.5) Other substance 16,338 (16.0) 6,237 (40.5) 10,101 (59.6) 136.3 (<.0001) No other substance 52,825 (84.1) 26,184 (49.7) 26,641 (50.3) Arrested 2,871 (2.7) 1,252 (46.6) 1,619 (53.4) .66 (.42) Not arrested 65,740 (97.3) 30,942 (48.3) 34,798 (51.7)

�Results are reported as unweighted n and weighted percentages using SUDAAN.

Becker et al. July–August 2014 359

reporting workplace drug testing among respondents with no alcohol abuse/dependence and no other substance abuse/ dependence.

On multivariable analysis (Table 2), we observed among the full sample that younger age, male sex, black race, income greater than $20,000, completion of high school and non‐urban residence were each associated with report of workplace drug testing. Additionally, non‐white collar occupation, work in

medium or large workplace, and absence of other substance abuse/dependence were associated with report of workplace drug testing.

In our stratified models (Table 2), we observed an independent association between black race and report of workplace drug testing among respondents in both executive/ administrative/managerial/financial and technicians/related support occupations; however, there was no association

TABLE 2. Adjusted odds of report of workplace drug testing, multivariable models (N ¼ 69,163)

Characteristic Full sample OR (95% CI)

Stratified sub‐samples

Executive, administrative,

managerial, financial (n ¼ 7,043)

OR (95% CI)

Technicians and related support occupations (n ¼ 4,047)

OR (95% CI)

Transportation and material moving

workers (n ¼ 4,041) OR (95% CI)

Worker demographics Age

18–25 1.2 (1.1–1.3) 1.1 (.9–1.5) 1.8 (1.2–2.6) .9 (.6–1.3) 26–34 1.4 (1.3–1.5) 1.3 (1.1–1.6) 1.5 (1.00–2.1) 1.5 (.9–2.3) 35–49 1.3 (1.2–1.4) 1.2 (1.03–1.5) 1.2 (.9–1.7) 1.7 (1.1–2.5) �50 ref ref ref ref

Women, n (%) ref ref ref ref Men 1.5 (1.4–1.6) 1.2 (1.01–1.4) 1.7 (1.2–2.2) 1.7 (1.1–2.5) White ref ref ref ref Black 2.1 (1.9–2.2) 1.8 (1.2–2.6) 2.4 (1.7–3.4) 1.4 (.9–2.1) Hispanic 1.0 (.9–1.2) .9 (.6–1.2) 1.4 (1.01–2.0) .8 (.5–1.2) Household income, $

<20,000 ref ref ref ref 20,000–49,999 1.3 (1.2–1.5) 1.7 (1.1–2.5) .9 (.6–1.2) 1.4 (.9–2.0) 50,000–74,999 1.5 (1.3–1.6) 1.7 (1.1–2.6) .9 (.7–1.3) 2.0 (1.4–2.8) �75,000 1.4 (1.2–1.5) 1.8 (1.1–2.7) .7 (.5–1.1) 1.9 (1.3–2.8)

Education �11th grade 1.1 (.98–1.2) .9 (.7–1.3) 1.1 (.6–2.0) .9 (.6–1.5) Graduated high school 1.5 (1.4–1.6) 1.3 (1.1–1.6) .9 (.7–1.1) 1.4 (.9–2.0) College ref ref ref ref

Married 1.0 (.9–1.1) 1.1 (.9–1.3) 1.2 (.9–1.5) 1.2 (.9–1.7) Not married 1.0 1.0 1.0 1.0 Urban .8 (.7–.8) .8 (.7–1.00) .8 (.6–.97) .7 (.6–.99) Non‐urban 1.0 1.0 1.0 1.0

Occupational variables White collar ref n/a n/a n/a Non‐white collar 1.6 (1.5–1.7) Workplace size

Small (<25) ref ref ref ref Medium (25–99) 2.9 (2.8–3.1) 3.0 (2.4–3.7) 3.5 (2.7–4.7) 3.5 (2.6–4.7) Large (100 or more) 6.5 (5.9–7.1) 5.5 (4.4–6.9) 6.5 (5.2–8.2) 7.3 (5.0–10.6)

Other covariates Alcohol .9 (.8–1.02) .8 (.7–1.1) 1.0 (.6–1.7) .6 (.4–.8) No alcohol 1.0 1.0 1.0 1.0 Other substance .7 (.7–.8) .8 (.7–1.1) .7 (.5–.96) .6 (.4–.8) No other substance 1.0 1.0 1.0 1.0 Arrested 1.0 (.8–1.1) 1.8 (.9–3.7) .6 (.2–1.5) .8 (.5–1.1) Not arrested 1.0 1.0 1.0 1.0

360 Differences in U.S. Workplace Drug Testing July–August 2014

between black race and report of workplace drug testing among transportation and material moving workers. Additionally, among respondents in executive/administration/managerial/ financial occupations, we found increased odds of reporting workplace drug testing among respondents ages 26–34 and 35–49, men, individuals with incomes $20,000 and above, respondents who graduated high school and individuals working in medium or large workplaces. Among respondents in technicians/related support occupations, we observed increased odds of reporting workplace drug testing among respondents ages 18–25, men, Hispanic respondents, non‐ urban respondents, individuals working in medium or large workplaces and respondents with no other substance abuse/ dependence. Among respondents in transportation and material moving occupations, we found increased odds of reporting workplace drug testing among respondents ages 35– 49, men, individuals with incomes $50,000 and above, non‐ urban respondents, individuals working in medium or large workplaces, respondents with no alcohol abuse/dependence and respondents with no other substance abuse/dependence.

DISCUSSION

In this large U.S. community‐based sample of workers representing a wide spectrum of demographics and occupa- tions, we found that black race was associated with report of being employed in a workplace that performs drug testing. This association persisted after adjustment for relevant demograph- ic, occupational, and other covariates and among stratified analyses in occupations with the lowest and highest proportions of black workers. Interestingly, among transporta- tion and material moving workers, where workplace drug testing is U.S. government‐mandated and therefore may occur in a more uniform manner, we found no association between black race and report of employment in a workplace that performs drug testing. Additionally, we found an association between Hispanic ethnicity and being employed in a workplace that performs drug testing among technicians/related support occupations. Consistent with a study of another large community sample,14 workers in larger workplaces had higher odds of being employed in a workplace that performs drug testing, possibly due to larger workplaces’ increased adminis- trative capacity to implement programs and larger budgets to pay for them. An unexpected finding was the association between the absence of alcohol and substance use disorders and report of employment in a workplace that performs drug testing. This may relate to one of the limitations of the study, discussed below.

Our discussion will focus on potential explanations for these findings, their implications and areas for future study and potential policy interventions. Key to this discussion is an acknowledgement of one of the limitations of the study: the lack of specificity of the outcome variable: “Does your workplace ever test its employees for drug use?” Some respondents may have answered “yes,” referring to their

knowledge or perception that that drug testing took place in their workplace, when they themselves had not undergone testing. In light of this, we hypothesize two explanations for racial differences in report of being employed in a workplace that performs drug testing that are not mutually exclusive and may interrelate. First is the structural hypothesis: blacks (and Hispanic technicians and other support workers) dispropor- tionately held jobs where workplace drug testing occurred. Second is the discrimination hypothesis: employers may have consciously or unconsciously discriminated against blacks at the individual or group level by differentially requiring workplace drug testing even within the same job. Interrelation of these two hypotheses could have occurred; for example, an employer could have enforced mandatory workplace drug testing in an entire division of its workforce in part because of the racial composition of that division. Empirical data on workplace testing among communication workers unions supports this hypothesis.20

Due to the lack of specificity of the dependent variable, we were unable to fully test these hypotheses. However, through the use of detailed analyses—logistic regression and stratifica- tion among certain occupations—we observed a persistence of the association between black race and report of being employed in a workplace that performs drug testing, suggesting that the structural hypothesis cannot entirely explain these differences. Further study is needed at the individual workplace level to understand the potential mechanisms of structural versus discriminatory workplace drug testing especially given the potential for significant health‐related consequences of workplace drug testing, including the inability to obtain employment or the loss of employment and health insurance.

Our study has other limitations. First, data were obtained via self‐report, which has an inherent degree of recall bias, possibly leading to inaccuracy and misclassification of information.21 It is possible that recall bias operated differently for different racial/ethnic groups; for example, blacks may have disproportionately been more likely to recall (and thus report) workplace drug testing due to heightened awareness of perceived or actual racial profiling and stigma. Additionally, individuals without alcohol and substance use disorders may have disproportionately answered that they were employed in a workplace that performs drug testing, assuming that it was occurring in others with alcohol and substance use disorders. Also, individuals with alcohol or substance use disorders may have been underrepresented in the study since they may have been more likely to be unemployed because of results of prior workplace drug testing. Second, as the data are cross‐sectional, we could not infer causal relationships between variables and thus avoided doing so. Though we adjusted for several potential confounds of the relationship between race/ethnicity and report of being employed in a workplace that performs drug testing, the cross‐sectional nature of the analysis allows for the possibility of unmeasured confounders.

Our study, the first to investigate the association between race and report of being employed in a workplace that performs

Becker et al. July–August 2014 361

drug testing within and across occupations, revealed significant differences in report of a practice that can have far‐reaching implications for workers, their livelihoods and their health. The hypotheses that we have presented on the precipitants of this disparity deserve further study at a more detailed level in order to guide policy‐level interventions.

Dr. Becker was supported by a Veterans Health Adminis- tration Health Services Research & Development Career Development Award (08‐276); Dr. Fiellin was supported by NIDA R01‐DA020576‐01A1 and NIAAA U01‐AA020795‐01.

Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.

Declaration of Interest The authors report no conflicts of interest. The authors alone

are responsible for the content and writing of this article.

REFERENCES

1. Holden C. Doctors square off on employee drug testing. Science. 1987;238:744–745.

2. Carpenter CS. Workplace drug testing and worker drug use. Health Serv Res. 2007;42:795–810.

3. Cashman CM, Ruotsalainen JH, Greiner BA, et al. Alcohol and drug screening of occupational drivers for preventing injury. Cochrane Database Syst Rev. 2009; 1–21.

4. Kraus JF. The effects of certain drug‐testing programs on injury reduction in the workplace: An evidence‐based review. Int J Occup Environ Health. 2001;7:103–108.

5. Committee on Drug Use in the Workplace, National Research Council/ Institute of Medicine. Under the Influence?: Drugs and the American Work Force. Washington, DC: The National Academies Press; 1994.

6. Macdonald S, Hall W, Roman P, et al. Testing for cannabis in the work‐ place: A review of the evidence. Addiction. 2010;105:408–416.

7. Marmot M. Social determinants of health inequalities. Lancet. 2005;365:1099–1104.

8. Puig‐Barrachina V, Malmusi D, Martenez JM, et al. Monitoring social determinants of health inequalities: The impact of unemployment among vulnerable groups. Int J Health Serv. 2011;41:459–482.

9. Braveman PA, Kumanyika S, Fielding J, et al. Health disparities and health equity: The issue is justice. Am J Public Health. 2011;101(Suppl 1):S149– S155.

10. Bush DM. The U.S. Mandatory Guidelines for Federal Workplace Drug Testing Programs: Current status and future considerations. Forensic Sci Int. 2008;174:111–119.

11. Burgess DJ, van Ryn M, Crowley‐Matoka M, et al. Understanding the provider contribution to race/ethnicity disparities in pain treatment: Insights from dual process models of stereotyping. Pain Med. 2006;7:119– 134.

12. Crowley‐Matoka M, Saha S, Dobscha SK, et al. Problems of quality and equity in pain management: Exploring the role of biomedical culture. Pain Med. 2009;10:1312–1324.

13. Becker WC, Starrels JL, Heo M, et al. Racial differences in primary care opioid risk reduction strategies. Ann Fam Med. 2011;9:219–225.

14. Larson SL, Eyerman J, Foster MS, et al. Worker Substance Use and Workplace Policies and Programs. Rockville, MD: Substance Abuse and Mental Health Services Administration, Office of Applied Studies; 2007.

15. United States Department of Health and Human Services. National survey on drug use and health, 2009 [Computer file]. In: Substance Abuse and Mental Health Services Administration, ed. Ann Arbor, MI: Inter‐ University Consortium for Political and Social Research.

16. United States Department of Health and Human Services, Substance Abuse and Mental Health Services Administration. National Survey on Drug Use and Health, 2010 [computer file]. Ann Arbor, MI: Inter‐ University Consortium for Political and Social Research; 2010.

17. United States Department of Health and Human Services, Substance Abuse and Mental Health Services Administration. National Survey on Drug Use and Health, 2008. Ann Arbor, MI: Inter‐University Consortium for Political and Social Research; 2008.

18. Kennet J, Gfroerer J. Introduction. In: Kennet J, Gfroerer J, eds. Evaluating and Improving Methods Used in the National Survey on Drug Use and Health. Rockville, MD: Substance Abuse and Mental Health Services Administration, Office of Applied Studies; 2005.

19. U.S. Office of Personnel Management. Federal Civilian Workforce Statistics: Occupations of Federal White‐Collar and Blue‐Collar Workers. Washington, DC: U.S. Office of Personnel Management; 2000.

20. Gee GC, Curbow B, Ensminger ME, et al. Are you positive? The relationship of minority composition to workplace drug and alcohol testing. J Drug Issues. 2005;35:755–778.

21. Raphael K. Recall bias: A proposal for assessment and control. Int J Epidemiol. 1987;16:167–170.

362 Differences in U.S. Workplace Drug Testing July–August 2014

Copyright of American Journal on Addictions is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.