Article Text

Download PDFPDF

Policy issues
  1. Kenneth E Warner
  1. Department of Public Health Policy and Administration, University of Michigan School of Public Health, Ann Arbor, Michigan, USA

Statistics from


I will open my remarks with an observation about the title of this conference, ‘Issues in Smoking Cessation: Who Quits? Who Pays?’ With the exception of a couple of presentations, I would suggest that we have focused relatively little on the questions in the subtitle. Dr Giovino opened the conference with an excellent discussion of who quits smoking, the first issue in the subtitle. Dr McGinnis addressed the second issue in his sobering, but realistic remarks on who is likely to pay in the near future. In between have come a series of presentations and panel discussions that have addressed a critically important issue that is not in the title explicitly, but has been the focal point: treatment efficacy, with particular concern with nicotine replacement and treatment delivery mechanisms. Other than Dr McGinnis’ remarks, the principal commentary on the issue of ‘who pays ’ has consisted simply of cessation experts lamenting the lack of insurance coverage.

I will address a very specific question: who should pay for smoking cessation, with special consideration of the issue of third-party coverage of cessation services. I will examine this matter in a policy context, but with more than a small dusting of personal philosophy thrown in. Note, incidentally, that when I say policy, I include insurance company policy and private business policy along with governmental policy (although the latter is clearly the principal concern of participants in this conference).

Prevention: the stepchild of health care

This conference is obviously, and probably not accidentally, very timely from a policy perspective. Insurance reimbursement of smoking cessation is one piece of a relatively small but visible dimension of the debate about health care reform, namely how to handle disease prevention within the health care reform framework.

Disease prevention (including smoking cessation) has long been the stepchild of American health care. Years ago, Dr William Foege, then director of the Centers for Disease Control, made a critically important observation that many conference speakers have recognised during this conference: to define a practice as medically acceptable, our health care system imposes very different standards of proof on surgical, medical, and prevention practices.1 For surgical procedures, the medical profession frequently requires merely a demonstration of safety; a medical procedure (eg, use of a new therapeutic pharmaceutical) has to be safe and effective; in contrast, prevention interventions seemingly have to be demonstrated to be safe, effective, and cost saving, a far higher standard of proof. Even that, Dr Foege noted, may not be enough.

In our more cost-conscious era, the standards of proof for surgical and medical practices may be increasing. Concern with the effectiveness (or at least efficacy) of surgical procedures is clearly more prominent today than a decade or two ago. Theoretically, the Health Care Financing Administration now requires demonstration of cost effectiveness to reimburse for use of medical and surgical innovations by Medicare patients, although implementation of this rule in practice is exceedingly difficult. Charged with developing appropriate practice guidelines, the Agency for Health Care Policy and Research simply cannot hope to do so in each of the areas in which such guidelines would be useful.

Regardless of the possibility of success in a few such endeavours, two critically important observations remain. First, theory and practice do not match in the case of medicine and surgery: cost effectiveness – the theoretically ideal criterion for choosing medical procedures2 – plays at best a rudimentary role in the actual determination of medical practice and reimbursement; in the case of prevention, in contrast, cost effectiveness is essential. Second, as Foege observed, the real criterion applied to prevention is not ‘merely’ cost effectiveness; rather, it is a demonstration of cost-saving potential. Prevention is asked to prove not merely that it accomplishes something important and does so less expensively than alternative interventions, but rather that it also makes money in the process. Many health insurers seem to want such an outcome; the wellness community clearly believes that the business world is concentrating narrowly on fiscal return in evaluating health promotion interventions.3

Cost effectiveness of smoking cessation

The best evidence indicates that many prevention measures are demonstrably cost effective, and smoking cessation ranks among the best of those. David Eddy refers to smoking cessation as ‘the gold standard’ of prevention cost effectiveness; he estimates that smoking cessation can be produced at $143 per year of life saved.4 He juxtaposes that number against figures ranging from $700 to $10 000 per year of life saved for common, widely accepted preventive screening measures, including cer-vical and colorectal cancer (which he considers the more cost-effective screening procedures), breast cancer and hypertension, and, at the poorer end, cholesterol.

In his luncheon address, Dr McGinnis said that smoking cessation costs $700 to $2000 per year of life saved, and that the addition of nicotine replacement raises the ‘ price5 of a life- year saved to about $4000. Much higher than Eddy’s estimate, these numbers come from studies of the cost effectiveness of clinical smoking cessation interventions,5,6 whereas Eddy’s figure most likely reflects ‘broad- spectrum ’ public health interventions (eg, mass media campaigns).

If clinical interventions are typically more costly per year of life saved than non-clinical interventions, are they therefore undesirable on grounds of inferior cost effectiveness ? Not necessarily. The real issue here, a subtle one, is who is quitting, in response to what, what is it costing, and what is it worth to get an additional person to quit? Logic certainly recommends that society shouldcbegin’ Smoking cessation with those measures that are most cost effective: this will maximise the number of quitters for the resources invested. But there will be a (large) group of ‘recalcitrants’, smokers who want to quit but do not (perhaps cannot) in response to those non-clinical measures. For them, such measures are not at all cost effective. The reasonable question to ask then becomes, is it worth it to spend up to $4000 to save additional years of life? In the context of health care treatment interventions in general (and even common sense), the answer is a resounding affirmative. The literature indicates a range of from $10000 to $70000 and even more than $100000 per year of life saved for many widely accepted secondary and tertiary care measures.7

Obstacles to prevention

If even the ‘ Cadillac ’ forms of clinical smoking cessation treatment are highly cost-effective medical care, why is there any doubt about reimbursing smoking cessation treatment services and products? There are a number of answers to this question, some reflecting undesirable barriers to reimbursement of preventive services in general, others possibly representing legitimate reasons for caution in moving toward the funding or reimbursement of services like smoking cessation.

The first answer relates to a fundamental fact of life: the current health care system discourages the utilisation of many preventive services, particularly those primarily involving behavioural counselling. All of the system’s major actors are culprits: providers find prevention uninteresting, of questionable efficacy, and unremunerative; insurers remain skeptical about its potential contribution to cost containment, especially to their own cost containment and especially in the short run; finally, consumers really don’t want it anyway, at least not if it takes time, effort, and/or money. I will examine each of these briefly in the specific context of smoking cessation.

A couple of minutes of informal counselling by a physician can roughly double the background rate of smoking cessation.8 As several speakers have emphasised, the addition of nicotine replacement therapies (NRT) can increase that rate even more. In a group of 100 counselled patients, two or three will quit smoking each year independent of counselling.9 Conservatively, another two or three will do so as a result of counselling (perhaps three or four with NRT). Among these extra treatment-induced quitters, perhaps one will avoid a smoking-related death that will translate into 20 additional years of life expectancy.10 There are few, if any, medical interventions that are so simple, so inexpensive, and so effective in terms of health outcome.

Consider the pieces of the puzzle from the perspective of each actor, however. From physicians’ perspective, the intervention was a complete failure. The physicians felt uncomfortable delivering the cessation message. They obviously received little or no training about behavioural counselling in medical school, and the message made patients squirm. The following year, 90 or more out of every 100 patients counselled to quit smoking returned to their doctors’ offices still smoking. The intervention was an abject failure, a waste of time. Worse yet, absent some questionable diagnostic accounting, the providers received no reimbursement for counselling time for insured patients or else the providers put the financial burden of this ineffective service (as they perceive it) directly onto the patient. The poor patient had to suffer through the physicians’ inept chastisement and pay for the privilege as well!

Why did the insurance company refuse to pay for the counselling (and prescribed products)? Clinical smoking cessation is almost uniquely cost effective. But cost effectiveness is not synonymous with cost saving. Physicians counselling their patients not to smoke may produce health benefits at a low cost, but that cost is greater than zero. Unless counselling- induced cessation reduces other health care costs in a meaningful manner, covering behavioural counselling is going to cost the insurance company additional monies, compelling the company to raise premiums to the insured population receiving the counselling benefit. The pressure today is to lower premiums, not to raise them.

The conventional wisdom is that smoking cessation will avoid health care costs down the road. The best evidence suggests that this is indeed possible,11 although the amount may be far less than many members of the public health community expect. While socially encouraging, however, this conclusion affords little reassurance to individual insurance companies. Cessation counselling costs accrue at present. The benefit of reduced utilisation attributable to smoking cessation is realised years into the future. The individual insurance company reimbursing for smoking cessation faces the likelihood that it will incur the costs of the service but not derive the dollar benefits of reduced health care spending in future years. Ironically, those will go to a competitor, to the insurance company that insures the former smoker years down the road when the health benefits are being realised. Is counselling cost-effective health care? Absolutely. Is it cost-saving health care? Socially, probably yes. To an individual provider, probably no. (The same logic applies to health maintenance organisations and businesses contemplating worksite smoking cessation programmes.)

This problem would disappear, of course, if patients viewed behavioural counselling sufficiently favourably that they were willing to pay higher premiums to cover the added services. But patients view prevention counselling with ambivalence, at best. Such services are considered discretionary. You don’t ‘need’ them in the same sense you have to treat an infection or set a fractured bone. The timing is discretionary, as well; it can always be post-poned. Typically, counselling costs out-of-pocket dollars, since insurance generally fails to provide coverage. And counselling can be unpleasant. Who wants to be lectured about a bad habit that you already know you shouldn’t be engaging in but you find difficult to alter? Behaviour change, after all, is a lot more difficult than simply getting a flu shot. And who wants to do that ? It takes time out of your day, it can hurt, it may or may not work, and if it does, you’ll never know it. And then, indignity of indignities, it’s probably going to cost you some hard-earned money.

For all three parties – providers, insurers, and patients – the fundamental problem is the fundamental problem of prevention in general. The benefits are abstract and deferred. The costs are tangible and immediate. Thus, the dilemma of whether and, if so, how to incorporate smoking cessation into mainstream health care is, in part, simply a reflection of the dilemma of disease control and prevention, especially in the area of behavioural counselling.

Unlike behavioural counselling, some disease prevention services are reasonably well integrated into the health care system, but they are interventions that tend to fit the medical model. Dr Benowitz discussed this phenomenon in his presentation at this conference. Immunisations involve giving a patient an injection. Screening for and treatment of cholesterol, hypertension, cervical and breast cancer have technological tests and specific chemical or surgical treatments. In contrast, behavioural counselling is less well accepted because it lacks a technical test and a pharmacological solution. In this regard, smoking cessation has become more mainstream as it has entered the medical model, thanks to the availability of NRT. Furthermore, the future of behavioural counselling may improve simply by virtue of Health Care Financing Administration’s recent implementation of the resource-based relative value scale method of determining reimbursements for physician services provided to Medicare patients. This reform places greater weight on ‘cognitive’ services, such as counselling.

Reimbursing smoking cessation: fundamental facts and possible implications

But where does this leave us? Here are some fundamental facts about smoking cessation: (1) It works. No intervention works as well as we might like, but many interventions work for many different people, and virtually all do so in a cost-effective manner compared to other health care interventions. (2) Utilisation of smoking cessation counselling services would definitely increase, probably substantially, if the costs were covered by insurance, be it public or private. (3) Reimbursement of professional smoking cessation counselling services would increase quitting and, thereby, the public’s health. In the process, it might also put a small dent in health care costs in the future, although this impact can be, and regularly is, exaggerated.

Here are some more fundamental facts about smoking cessation, facts that may give pause in advocating the coverage of smoking cessation services. Smoking cessation services are not particularly expensive. They range from a few dollars for self-help materials to a few hundred dollars for ‘ Cadillac care, ’ comprehensive behavioural counselling with NRT. As such, why do we need reimbursement? If smokers really want to quit, and if they believe that available services and products are potentially useful, why (with the exception of the truly impoverished) would they not pay for such services directly, without asking their fellow taxpayers and insurance premium payers to do it for them? (It is interesting to contemplate that even impoverished smokers pay $700 or more each year for their cigarettes.)

Clarity is lent to this question by considering the meaning of the word, ‘insurance’. In its typical (non-medical) usage, insurance is protection against high-cost, low-probability events that are not affected by its existence. Homeowner’s and automobile insurance are good examples. The idea of ‘insurance’ reimbursement for smoking cessation services is exactly the opposite on all three characteristics. It is relatively low-cost and covers a reasonably high-probability event. As for its effect on utilisation, support for it in the smoking- control community derives specifically from the expectation that its availability will affect utilisation! The implication is that reimbursement of smoking cessation services, far from constituting true insurance, is really subsidisation. Specifically, it is subsidisation of smokers by non-smokers.

Suppose that a decision is reached to cover smoking cessation services (regardless of whether we call it insurance or subsidy or whatever). Exactly which services should be covered? The health education community has called for certification of smoking cessation services for purposes of reimbursement, a call that has been echoed frequently during this conference. We have learned at this meeting that AHCPR is going to develop guidelines for smoking cessation treatment. I wonder whether we know enough to define what constitutes effective treatment for which groups of smokers. Dr Slade, for example, suggested in his presentation that we ought to at least consider the notion of lifelong nicotine maintenance. Will this be evaluated by the AHCPR task force? On what evidence? Following authoritative evaluation, is there a risk of gravitating towards the high-cost end of cessation services if some are designated as effective (and possibly worthy of reimbursement), while others are not so designated? Note that this is relatively likely to occur if effectiveness is the major criterion.

Do we want to discourage innovation in the smoking cessation field by declaring only certain forms of cessation services as good, effective, perhaps worthy of reimbursement? Frequently, although certainly not always, certification or licensing is proposed by professional groups purporting to represent the public interest, when analysis suggests that the primary effect may be self-serving: certification/licensing narrows the market and creates monopoly control for the professional group. I do not mean to impugn the motives of the health education community, which I feel certain believes it represents the public’s interests. This general observation reflects experience in other areas of health care.12

At the predecessor to this meeting,13 individuals concerned with alcoholism treatment urged the conferees to consider what has happened in their field. Reimbursement for treatment is available, but for years it covered only multiple-day in-patient programmes, despite the fact that there was no good evidence that in-patient programmes worked more effectively than brief, dramatically less expensive, out-patient programmes. In the process of developing reimbursement practices, however, the treatment community built an economic monolith – a billion dollar business – with a tremendous financial interest in limiting reimbursement to the type of expensive inpatient service they provided. Is there a lesson herein for the smoking control community? Even the Cadillac forms of smoking cessation do not rival in-patient alcoholism treatment for cost per case. But consider that a billion- dollar-per-year smoking cessation medical interest could develop if merely 2% of all smokers participated in a $1000 smoking cessation therapy.

Public health policy context

Individual treatment approaches definitely work for some smokers, but individualised treatment is clearly less cost effective at the societal level than other interventions, such as mass media campaigns.14 Consider a contemporary example of another tobacco-control policy measure: excise taxation. If the federal cigarette excise tax were increased by $2 per pack and maintained in real value thereafter, as advocated by the major voluntary health associations, an estimated 7.6 million people would either stop smoking or not start smoking as a result.15 To achieve a comparable behav- ioual outcome through individual treatment, over 30 million or more smokers would have to be treated. At a cost of just $300 per case, this would cost close to $10 billion.16

As several participants have observed at this meeting, we do not confront an either/or situation, either policy with its greater aggregate effectiveness or individual counselling with its greater effectiveness per participant. I concur with Dr Orleans’ conclusion that we should want both; as Dr Manley noted, there may be a synergistic effect. But thinking about the emphasis in a nationwide smoking cessation effort, were Congress considering a panoply of tobacco-control measures, I would urge each of us to write our representatives a minimum of 10 letters, and maybe 100, supporting an increased cigarette excise tax and a ban on advertising before we write the first letter urging funding of smoking cessation treatment services for individuals. This is not to suggest that the latter are unimportant, but rather to reflect my perspective on relative importance (and perhaps to reflect the much greater need for political support for the policy measures; reimbursement of smoking cessation treatment services has a natural, potentially influential constituency of articulate professionals).


In the final analysis, where do I stand on the issue of reimbursing the costs of smoking cessation? I retain my basic ambivalence about mandating the funding of smoking cessation publicly or through some form of private insurance. I do not see this as a theoretically proper role of either government or the insurance industry; however, the identical criticism applies to the way much of medical care funding is handled, and not simply the issue of smoking cessation.

Pragmatically, in the context of a system that covers so many other services that are demonstrably less cost effective than smoking cessation, I have some sympathy for considering its inclusion within a health care reform framework. I am particularly sympathetic to funding it as a benefit for the medically indigent as opposed to the rest of us. This distinction has received too little attention at this meeting. I see no reason why average Americans should be asked to pay for the costs of cessation treatment for affluent or even middleclass smokers. Such smokers can afford cessation treatment; certainly, poorer non-smokers should not be subsidising their care!

Were smoking cessation to become a reimbursable service, a host of other behaviour- based treatments would warrant coverage as well. This would create the prospect of our nation’s beginning to devote serious attention to disease prevention in the context of the formal medical care system, a highly desirable development. But it would also raise the spectre of further increases in health care costs and of redistribution of those costs from individuals who derive the benefit to other members of society. And it would challenge the medical and health policy communities to identify precisely which counselling services ought to be covered, and which not. The US Preventive Services Task Force has tackled the immense chore of distinguishing effective from ineffective clinical prevention services.17 But as a sobering new report from the Congressional Office of Technology Assessment concludes,18 distinguishing cost-effective from cost-ineffective prevention services is an order of magnitude more difficult.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.