On November 15, 2012, the University of Alabama at Birmingham (UAB) School of Medicine and its health system became the first state academic health center in the United States to announce a policy of not hiring smokers. Beginning in the summer of 2013, candidates for employment were tested for nicotine. Those testing positive were not hired, save for those instances where it could be determined through additional testing that tobacco cessation therapy accounted for an initial positive test. In declining to hire employees who smoke, UAB joined a growing number of health care institutions that have taken this step1; a second major academic health center, the University of Pennsylvania Health System, announced its own hiring ban on smokers in the spring of 2013.1
These bans are the most recent signs of an emerging trend among health care institutions and other employers. Survey data from the late 1980s suggest that as of that time, 6% of all employers did not hire smokers (about 6,000 companies).2 Such policies, which appeared on a large scale in the early 1980s, generated a wave of legislation protecting smokers’ rights between 1989 and 1991.3 Health care institutions were not prominent among the first wave of antismoker workplace hiring policies, however. Present interest among such institutions in smoker hiring bans dates from 2007, when the Cleveland Clinic introduced its smoker hiring ban. As of mid-2013, Boston University’s Professor Michael Siegel estimated that 50 to 60 health care systems around the country had instituted such bans in recent years.1
In this essay, we take note of the competing ethical and policy claims regarding employee smoker bans and argue that adjudication of these claims (favorably or not) does not provide sufficient grounds for academic health centers and other health care institutions to exclude smokers from employment. We review arguments commonly made in support of employee smoker bans and suggest that these arguments are, at best, inconclusive. Most employers, in the 19 states where such bans are legal, could, however, legitimately choose to enact them as a matter of law and perceived organizational advantage. However, we suggest that leaders of health care institutions are not in the position of most employers. Those leaders ought to choose policy based not only on impersonal policy considerations but also on the unique character of the practices their institutions serve, most notably medicine and scholarship. We explain why we find that the normative standards of both of these practices are incompatible with employee smoker bans.
The Employee Smoker Ban Debate: Arguments For
Typical rationales offered for policies banning smokers from workplaces generally stress both their anticipated consequences and their expressive function. Businesses are said to gain financial benefit by not hiring smokers, who are less productive and who incur higher health care costs than nonsmokers do.4 Such policies also “send a message” about smoking to the larger community.
Institutions excluding smokers among new hires have not generally extended such exclusion to existing employees. Although having different stances toward new and existing employees opens employers to charges of inconsistency, it also avoids the upheaval that would likely accompany a demand that existing employees quit smoking or be terminated. Proponents of smoker bans would likely argue that compromise with existing employees is necessary to achieve the longer-term goal of a smoker-free workplace. And they would likely favor other measures calculated to achieve smoking cessation by current employees, such as wellness programs that include financial or other penalties for employees who fail to participate or to meet wellness targets.5
The financial advantage for employers of a smoker ban, in the form of lower health care costs, may be considerable. Berman et al4 offer the figure of $2,056 in excess health costs per smoking employee per year, and $5,088 in total yearly excess costs for such an employee. Health care institutions that are in the current wave of employers instituting smoker bans are not, however, emphasizing cost savings as their most important rationale. In public statements, more weight is typically placed on the expressive function of these bans. As a spokesman for the Cleveland Clinic put it, “As a health care institution, whose inherent mission is healing the sick and cultivating a healthier community, does it make sense to support a habit that leads to disease, disability, and death?”6 Proponents of smoker bans emphasize the bans’ societal impact even more than their advantages for individual institutions. Such bans, it is asserted, will eventually contribute to a lower prevalence of smoking in society, with accompanying benefits. One such benefit would be lower societal health care costs.
The society-wide financial impact of lower smoking prevalence is actually unclear, however. Health care costs diminish in early years after smoking cessation but then increase because the former smoker lives longer. Barendregt et al7 suggested that with no discounting, smoking cessation lowered societal health care costs for up to 26 years of follow-up, but the beneficial effect on costs reversed if follow-up was longer. The break-even point at 26 years was extended to 32 to 35 years with the use of conventional discount rates. It is unclear that a financial analysis of decreased smoking prevalence should be limited to health care costs, however. Diminished collections of cigarette taxes are a cost to society of smoking cessation that must figure in striking a society-level balance sheet about the effects of smoking cessation. Of course, if one’s focus is population health rather than costs, the benefits of smoking cessation are clearer. Even if financial costs are higher, incurring these costs is a relatively efficient means to achieve diminished prevalence of fatal disease, and, hence, the costs are worth incurring.8
Employee smoker bans are more coercive than traditional smoking bans, as they control behavior outside of the workplace. The case for the additional coercion in smoker bans is twofold. First, proponents of employee smoker bans join proponents of more traditional smoking bans in pointing to harms inflicted by smokers on others as a reason for coercive antismoking measures. Of course, the spillover effects of smoking on nonsmokers are more difficult to invoke as a reason for smoker bans than for smoking bans, because the latter are sufficient to prevent exposure to secondhand smoke. But increasing concern about “thirdhand smoke”—the residue of nicotine and other toxins on the hair and clothing of smokers that persists after smoke has cleared—offers a similar rationale for smoker bans.9 Second, proponents of smoker bans contend that the putative benefit of more smokers quitting is worth the price of the additional coercion involved in smoker bans. Smoker ban proponents thus exemplify an increasing tendency in the public health community to dispense with any requirement for spillover effects before engaging in coercion. Public health policy makers have recently begun to heed calls for a low threshold for embracing “hard paternalism”—that is, coercive interference with purely self-harming behavior as a means to achieve the improved welfare of those coerced.10 And smoker bans are one example of such coercion.
The final element in the case for employee smoker bans is an appeal to the desirability of further “denormalizing” smoking in society. And, as above mentioned, in the case of health care institutions there is an accompanying claim that excluding smokers from employment furthers institutional coherence.
The Employee Smoker Ban Debate: Arguments Against
Opponents of employee smoker bans dispute both the public-welfare claims attributed to such bans and also the ethical case for subjecting smokers to hard paternalism. The claims for institutional health care costs related to employees who smoke are open to question. Berman and colleagues’4 published estimate for the costs of a smoking employee ($5,816 per year, including absenteeism, unproductive work hours, and smoking breaks) applies hourly compensation estimates ($26.49/hour), exceeding compensation for positions most typically held by smokers at academic health centers. The lion’s share of Berman and colleagues’4 estimated financial loss ($3,077) assumes that every smoker takes two unsanctioned smoking breaks of 15 minutes apiece (30 minutes per day). Whether licensed practical nurses or janitorial staff take such breaks, and thereby proportionately reduce the number of patients cared for or buildings cleaned, is unstudied. Empirical research shows that smokers outside office buildings consume a cigarette in an average time of 3.9 minutes.11 Finally, a significant component is based on estimates of unproductive work hours, a survey-derived measure that did not meet its developers’ criteria for predictive validity.12 In short, although employer cost calculations may help to justify efforts to assist employees who smoke to stop smoking, they do not readily justify categorical denial of employment to persons who smoke at home.
The claims for societal benefits from smoker bans are even more questionable than those for institutional benefit. Most important, the likelihood that smoker bans can have a causal influence on smoking prevalence is low. Reviews of comprehensive and partial laws to ban smoking did not find that they reduced the overall prevalence of smoking.13,14 If legislative bans do not reduce population-level active smoking, it seems excessively hopeful to presume that employee bans will do so.
In contrast to the likely lack of efficacy in diminishing smoking prevalence, employee smoker bans may be anticipated to cause tangible harms to the communities where the employers instituting such bans are located, particularly if those employers employ a high proportion of a community’s working citizens. Smokers are only 18% of the adult population in the United States, but they make up 26.1% of those with incomes less than $35,000/year and 25.5% of those without a high school diploma.15 U.S. health care workers reflect a similar pattern: Smoking is uncommon among physicians and pharmacists but moderately prevalent among persons in most other health care occupations, especially service staff and licensed practical nurses (see Figure 1 for statistics about employee categories common in U.S. hospitals).16,17 In many urban settings where large medical schools operate, it is likely to be poorer individuals, including those from minority groups, who, under a ban on employees who smoke, will lose the opportunity to work for an employer that offers health insurance, long-term advancement, and retirement benefits. In response to the incentives created by no-smoker hiring policies, some will, in fact, quit smoking. But with success most likely following fewer than 5% of quit attempts,18 most will not. In short, at the community level, employee smoker bans are more likely to be harmful than beneficial.
Opponents of no-smoking hiring policies also question the legitimacy of the hard paternalism involved in such policies. Employers, the argument goes, ought not to reach into the private lives of employees when private activities do not affect workplace performance.19 If employee smoker bans are acceptable, what is to restrain employers from prohibiting other employee behaviors similarly risky—such as overeating, bad driving, or sexual promiscuity? External constraints on self-harming activities are perhaps justifiable when the gains are great and the costs are low—as in the case of seat belt or motorcycle helmet laws. When the costs are high, as they are to smokers who either wish to continue smoking or cannot quit, such constraints should be avoided.20
How one adjudicates the policy and ethical debates over smoker bans will depend on the answers given to questions very much open in contemporary discussion. Is allowing smokers the unfettered choice to smoke outside the workplace more or less important than possible gains in smokers’ welfare that would follow from quitting under coercion? What constitutes autonomy, and how far is it legitimate to coerce those making self-harming decisions that are imperfectly autonomous? How bad is smoking compared with other risky behaviors, and how does one weigh the cost to smokers of denied employment against the putative gain of less societal or community future smoking? How one responds to these and other such difficult questions will condition one’s ethical conclusions regarding employee smoker bans. We do not propose to attempt a resolution of the ethical issues. It seems to us that, at the very least, institutional smoker bans are ethically permissible (in general), even if some think them to be ethically objectionable. The legal regimes in states without laws that prohibit discrimination against smokers clearly permit smoker bans by employers. Our preference for freedom of association suggests that most employers (but not health care institutions, as we explain below) should hire as they please within the constraints of applicable law.
The Relevance of Institutional Character to Institutional Conduct When Deciding About Employee Smoker Bans
The debate over smoker bans has a dimension for health care institutions that is absent in the wider societal context. While it is one thing for businesses to ban the hiring of smokers, it is another for health care institutions to do so in the name of health. Such an antismoker hiring stance implies a position on both the content of the professional norms of medicine and on how far those norms may properly govern health care institutional activities only indirectly related to the provision of health care, such as the hiring of nurses, doctors, janitors or clerks.
Using hospitals as an example, we will argue that it is indeed fair to extend the norms of medicine to a hospital’s hiring practices, but that proponents of smoker bans have mistaken what the norms of medicine demand when deciding whom hospitals should hire. Our argument will be decisive only for those who are both sufficiently part of the medical community to be governed by its norms and who also recognize our explication of those norms to be resonant with their own understanding of who they are as health care professionals. We thus proceed by appealing to presumptions about practical ends that we believe we share with those who favor banning smokers from employment in health care institutions: that is, an adherence to medical norms and a sense of what these are and how they bear in professional life. Such an appeal is an inevitable feature of ethical argument, which must at some level be a seeking of common ground with those of differing views.21
Although it is an open question how far professional norms should extend in determining the conduct of institutional activities subsidiary to professional work, proponents of antismoker hiring policies for health care institutions clearly connect these policies to the health care institutions’ mission of promoting health. They presume, rather than argue for, the importance of this connection, asserting that the congruence of smoker bans with the health-promotion mission of health care institutions is an important virtue of these bans. We shall take them at their word, agreeing that institutional norms can and should insinuate themselves into institutional activities subsidiary to the institution’s primary task—at least for the hiring of support as well as clinical personnel in the case of health care institutions. The question at issue then becomes, What do medical norms demand regarding the hiring practices of hospitals and other institutions whose identities are defined by the work of medical professionals?
We suggest that smoker ban proponents are mistaken when they maintain that smoker bans in health care institutions uphold professional norms in those institutions. Health care workers are expected to foster their patients’ healthy behaviors through respectful alliances with them, a posture articulated in the study of motivational interviewing.22 But health does not stand above care among the norms of medical practice; in fact, the latter is more important. Health care workers exemplify an ethic of care, including care for those whose ill health might be their own doing. Although there may be a case for having patients “take responsibility” for their health at the level of purchasing insurance, we do not consider individuals’ responsibility for their illnesses as disqualifying them from receiving care. Nor should we. This stance, which does not allow a person’s bad choices to influence our responses to his or her needs, is utterly at odds with employee smoker bans, which assign the moral status of the activity to the actor and label both as unwelcome. These bans reflect a moralization of health, characteristic of late-20th-century middle and upper class life, in which virtuous health behaviors serve as a marker for a “secular state of grace.”23 Whatever the merits of this moralization, its manifestation in smoker bans seems incompatible with mercy, charity, and even—insofar as such bans diminish the employment prospects of the poor—with social justice.
As health care professionals, we stand for health, but our more important priority of care suggests that while the message we ought to be sending to smokers need not be approval, it ought to be one of compassion, inclusion, and, as Dr. Albert Schweitzer put it, fellowship. If allowing smokers to work in an institution conveys institutional support of smoking, how does allowing smokers to be cared for in the same institution not similarly convey such support? Employee smoker ban advocates who presume that the divide between hiring and banning smokers tracks a divide between expressing support and disapproval of smoking are denying the possibility of making a distinction that we, as physicians and health care professionals, not only commonly make but must make if we are to be true to our profession: the distinction between the patients whom we are charged to care for and the multifarious unhealthy habits, activities, decisions, and propensities that contributed to their illnesses. Support of the one need never be tantamount to support of the other. Medical norms enjoin us to take care of the sick—not the deserving sick, the virtuous sick, or the well-behaved sick—merely the sick. The messages we send in our hiring as in our other institutional practices should be affirmation of all prospective patients, an affirmation that need not be construed as approval of their unhealthy activities. In the case of smokers, health care institutions ought to hire them freely and then encourage them to quit smoking. The message actually conveyed by an employee smoker ban to smokers is unlikely to be one of an affirmation of health; it is far more likely to be received as a personal affront or rejection. Such a message is incompatible with who we are as physicians and health care professionals.
We believe that an institutional identity allied to the profession of medicine should be sufficient to rule out employee smoker bans for such institutions. But what of the other aspects of institutional identity assumed by, for example, academic health centers? What, in particular, does an academic identity imply as to hiring practices? Although the distinguishing character of academic inquiry is generally taken to be the search for knowledge, academic institutions have often sought to cultivate distinctive forms of community. Prohibitions on smoking or other forms of unhealthy behavior seem neither more nor less anomalous as conditions for membership in a given academic community than are the restrictions often placed on students or faculty by religious, military, or other academic institutions distinguished by the special form of community they seek to cultivate. Of course, the case is different for academic institutions eschewing a particular sectarian, cultural, or other communal identification. Although nonsectarian public and private universities often endorse various social and political ends as goods, they generally make a point of avoiding the requirement, for students or employees, of behaviors or beliefs related to such ends. Members of nonsectarian academic communities are expected to conform to the law and to standards of behavior that allow academic inquiry to proceed. They are not to be otherwise required to adhere to standards of belief or behavior derived from particular visions of the good, even visions valorizing the cultivation of health. The hiring practices of secular academic institutions uncommitted to exclusive forms of community should reflect academic values, which suggest the inclusion of all who can participate in the academic enterprise—including smokers.
Hospitals, academic and otherwise, are among the most important of our health care institutions. If our arguments concerning smoker bans and the norms of medicine and academia are correct, smoker bans will not further align hospitals with the health care professions, and they will not align academic hospitals with the norms of the university. They will, instead, alienate the health care professionals working within hospital walls. That we have come to such a pass is in part a function of the recent evolution of hospitals. As professional dominance in hospitals has receded, hospitals have increasingly become aggregates of loosely coupled subunits, among which are multiple and vigorously contending “institutional logics”—most notably the professional and the managerial.24
Although there is no reason to suppose that managerial values (e.g., efficiency, entrepreneurship, cost containment, marketing) are intrinsically antithetical to professional norms, the two perspectives clearly offer the potential for conflict. As an operational matter, bans on hiring smokers have issued from the managerial side of hospital leadership, naturally enough, as these bans fall under the purview of personnel management. The obvious managerial priority served by these bans is cost containment. However, those issuing the bans have implicitly recognized the inadequacy of that rationale by appealing to the norms of health care for additional support. As we argue, the appeal misfires, and these bans will not have the anticipated effect of inducing professional solidarity within the hospitals instituting them.
We maintain that physicians and other health care providers working in hospitals presently feel besieged by requirements imposed by hospital management—by the deluge of bureaucratic work imposed on them by electronic health records and order entry; by performance measures that seem either unrelated to quality care or likely to subvert such care; and by sudden bureaucratic imperatives to address the management goal of the moment by, seemingly, any means necessary. Bans on hiring smokers will not, of course, directly affect most physicians. But they will be perceived as one more sign that hospital management is out of touch with what is important for any institution that claims to exemplify the norms of medicine.
Hospitals, including academic hospitals, and other health care institutions should welcome all potential employees committed to caring for the sick, and they should seamlessly extend their posture of caring for patients to employees who engage in unhealthy practices, including smoking.
4. Berman M, Crane R, Seiber E, Munur M. Estimating the cost of a smoking employee Tob Control..
7. Barendregt JJ, Bonneux L, van der Maas PJ. The health care costs of smoking. N Engl J Med. 1997;337:1052–1057
8. Temple NJ. Why prevention can increase health-care spending. Eur J Public Health. 2012;22:618–619
9. Burton A. Does the smoke ever really clear? Thirdhand smoke exposure raises new concerns. Environ Health Perspect. 2011;119:A70–A74
10. Conly S Against Autonomy: Justifying Coercive Paternalism. 2012 New York, NY Cambridge University Press
11. Chapman S, Haddad S, Sindhusake D. Do work-place smoking bans cause smokers to smoke “harder”? Results from a naturalistic observational study. Addiction. 1997;92:607–610
12. Ozminkowski RJ, Goetzel RZ, Long SR. A validity analysis of the Work Productivity Short Inventory (WPSI) instrument measuring employee health and productivity. J Occup Environ Med. 2003;45:1183–1195
13. Callinan JE, Clarke A, Doherty K, Kelleher C. Legislative smoking bans for reducing secondhand smoke exposure, smoking prevalence and tobacco consumption. Cochrane Database Syst Rev. 2010;4:CD005992 14
14. Bajoga U, Lewis S, McNeill A, Szatkowski L. Does the introduction of comprehensive smoke-free legislation lead to a decrease in population smoking prevalence? Addiction. 2011;106:1346–1354
15. National Center for Health Statistics. Summary Health Statistics for U.S. Adults: National Health Interview Survey, 2011: Data From the National Health Interview Survey. 2012 Washington, DC U.S. Department of Health and Human Services
16. Sarna L, Bialous SA, Sinha K, Yang Q, Wewers ME. Are health care providers still smoking? Data from the 2003 and 2006/2007 Tobacco Use Supplement-Current Population Surveys. Nicotine Tob Res. 2010;12:1167–1171
17. Ham DC, Przybeck T, Strickland JR, Luke DA, Bierut LJ, Evanoff BA. Occupation and workplace policies predict smoking behaviors: Analysis of national data from the current population survey. J Occup Environ Med. 2011;53:1337–1345
18. Rafful C, García-Rodríguez O, Wang S, Secades-Villa R, Martínez-Ortega JM, Blanco C. Predictors of quit attempts and successful quit attempts in a nationally representative sample of smokers. Addict Behav. 2013;38:1920–1923
19. Maltby L. Whose life is it anyway? Employer control of off-duty smoking and individual autonomy. William Mitchell Law Rev. 2007;34:1639–1649
20. Chapman S. The smoker-free workplace: The case against. Tob Control.. 2005;14:144
21. Taylor CNussbaum MC, Sen A. Explanation and practical reason. The Quality of Life. 1993 Oxford, UK Clarendon Press
22. Miller WR, Rollnick S. Motivational Interviewing, Helping People Change. 20123rd ed. New York, NY Guilford Press
23. Leichter HM. “Evil habits” and “personal choices”: Assigning responsibility for health in the 20th century. Milbank Q. 2003;81:603–626
24. Mendel P, Scott WRBird CE. Institutional change and the organization of health care: The dynamics of “muddling through.” Handbook of Medical Sociology. 2010 Nashville, Tenn Vanderbilt University Press