Article Text

Download PDFPDF

Report of the Tobacco Policy Research Study Group on Reimbursement and Insurance in the United States
  1. Michael D Parkinson,
  2. Helen H Schauffler,
  3. Thomas E Kottke,
  4. Susan J Curry,
  5. Leif I Solberg,
  6. Charles B Arnold,
  7. Roger H Butz,
  8. Roger Taylor,
  9. James B Holloway,
  10. Corinne Meltzer
  1. Health Resources and Services Administration, US Department of Health and Human Services, Rockville, Maryland
  2. University of California, Berkeley, California
  3. Mayo Clinic, Rochester, Minnesota
  4. Group Health Cooperative of Puget Sound, Seattle, Washington
  5. Blue Cross/Blue Shield of Minnesota, St Paul, Minnesota
  6. Metropolitan Life Insurance Company, New York, New York
  7. Safeco Life Insurance Company, Redmond, Washington
  8. The Wyatt Company, Washington, DC
  9. Kentucky Blue Cross and Blue Shield, Lexington, Kentucky
  10. Centers for Disease Control, Atlanta, Georgia
  1. Correspondence to Dr Michael D Parkinson, Division of Associated and Dental Health Professions, Bureau of Health Professions, Parklawn Building, Room 8–101, Rockville, Maryland 20857, USA.

Statistics from

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.

The United States health care system is a unique and uncoordinated combination of private and public programmes involving employers, government, insurance companies, and individual consumers. Although spending on health care comprised 12% of the US gross domestic product in 1990, it is estimated that 35–7 million Americans are uninsured, three quarters of whom are fulltime workers and their dependents.1 Also, it is estimated that as many as 40 million people have partial but inadequate health insurance.

Currently, employed persons and their families are most likely to obtain health insurance through a group policy that is purchased by the employer from a private insurance company. The federal government has been heavily involved in financing the provision of medical care since 1966 through enactment of Medicare, which provides health care for the elderly. Medicaid, which is a joint federal-state health insurance programme, funds the provision of care for the indigent.

Health insurance, whether purchased by the private employer or paid for by the public sector, has traditionally been designed to pay for high cost, unpredictable medical expenses associated with acute illness, such as stays in hospital and services of a physician. The health insurance policy stipulates the specific health care services that are covered. Hospitals, health care facilities, and individual providers submit claims to the insurance company in order to be reimbursed for the services provided to the patient. Because of recent dramatic rises in costs of health care, patients are increasingly required to contribute more to the cost of the premium paid to obtain the insurance, and to the cost of the specific services provided once they have been delivered. The amount of copayment required of the employee or health insurance enrollee for either the payment of the premium or the service provided has been used as an incentive for the control of costs and for the use of what is, in the eyes of the insurer and employer, appropriate medical care.

Disease prevention and health maintenance services, because they are relatively low cost and predictable, have not been included in health insurance coverage in the United States. In recent years, both private and public insurance plans have begun to offer coverage for screening tests such as Papanicolaou smears, mammography, and selected vaccinations, because of the growing evidence of the clinical and cost effectiveness of these preventive services. This trend has also been fuelled by the increased cost of medical care, and rising public demand for coverage of these services. Counselling services for disease risk factors, such as specific smoking cessation interventions, are not currently covered by the overwhelming majority of private or public insurance plans. A comprehensive review of the status of reimbursement for preventive services has outlined strategies for improving payment for screening, counselling, and immunisations.2,3

The US health care system is now at centre stage in the domestic political arena because of concerns about rising costs, questions of quality, and lack of access for uninsured and underinsured citizens. The pressure to reform the US health care system, whether incrementally or dramatically, is likely to continue through the 1990s. Advocates of tobacco control should have an understanding of the economic and political issues shaping the reform debate in order to promote access to mediated smoking cessation services for those patients who would benefit from such interventions.

This paper provides a review of three major issues that have an impact on the decision to reimburse smoking cessation efforts and provide insurance based incentives to smokers to stop. These include the effectiveness of smoking cessation interventions, cost effectiveness of mediated smoking cessation, and the cost effectiveness of insurance premium discounts for people who smoke.


It has been estimated from a survey on national tobacco use that 90% of adult smokers in the United States who successfully stop do so “on their own” without the aid of formal smoking cessation programmes or other methods. This comprehensive review of the methods used to stop smoking concluded, however, that structured cessation programmes served a small, but important population of heavier smokers who were most at risk of smoking related morbidity and mortality.4 It is not clear however, to what degree financial constraints prohibit access to organised programmes.

What should the appropriate policy be towards promoting changes in financing for mediated cessation efforts? Would expanded coverage for such services undermine the “public health approach” to tobacco control and unnecessarily stigmatise smokers, making them somehow “disabled” and in need of professional help? Should expansion of reimbursement or insurance incentives for smoking cessation be targeted at only selected populations of smokers – for example, those most heavily addicted to nicotine?

Health care providers and researchers have shown that successful smoking cessation represents a dynamic process of change over an extended period rather than a discrete, single act. Social learning theory has often been useful to explain the process of smoking cessation5. Progressing from smoker to former smoker is most often a recurrent cycle of motivation and committment, behaviour change, maintenance of non-smoking behaviour, relapse to smoking, subsequent recommitment to cessation and so on.6 The average smoker cycles three to four times through these stages before attaining long term abstinence.7 Studies have shown that less than 5% of smokers shift from being current smokers to former smokers in a single attempt without experiencing any relapse.8 Although many smokers are motivated to stop by concern about their physical health,9 the acquisition of specific strategies and skills can facilitate the cessation process.10

The effectiveness of preventive interventions generally and more specifically, for smoking cessation efforts, is an important issue in decisions regarding payment for services. The US Office of Technology Assessment recently reviewed the so called “double standard” applied to preventive as opposed to “curative” services with respect to Medicare insurance cover.11 Two arguments were advanced to defend a higher standard. Firstly, preventive services, like all interventions, have risks as well as benefits.

Unlike diagnostic and therapeutic services, however, preventive services are offered to ostensibly healthy persons and, therefore, imply a promise of improving the patient’s health, rather than simply curing an acute complaint. Secondly, the challenge, particularly given the growing concern about cost and quality of care, is to raise the level of evidence required for payment for diagnostic and therapeutic services, not to lower that for preventive ones.

Effectiveness of smoking cessation interventions

Because of the multifactorial nature of patient motivation and the proved effectiveness of using multiple modalities in bringing about and reinforcing smoking cessation, researchers should explore reimbursement options for various effective interventions. Physician and other health care provider-based counselling and organised cessation programmes are two of the most common interventions currently studied.

Two comprehensive reviews and evaluations of smoking cessation trials summarised follow up cessation rates for a wide variety of interventions.12,13 Self reported rates were given for several of interventions, from physician-based counselling to organised smoking cessation classes. In general, most treatments produced one year cessation rates in the range of 10%–40%, based on all original participants. For a brief clinical intervention in the office setting, a realistic smoking cessation rate was estimated at 5%–10%.14 Cessation efforts performed by nurses on patients after myocardial infarction produced a 60% rate compared with 30% in controls who received “usual care”.15 A meta-analysis of a variety of successful smoking cessation interventions16 showed that intervention effects decayed over time. Factors predictive of success included multiple contacts, a long duration of intervention, the use of both physicians and nonphysicians, the use of both individual and group contacts, and offering multiple interventions.

Although physicians believe that they know which of their patients smoke and that they have told all their patients to stop smoking, less than 50% of smokers report that they have ever been advised by a physician to stop.17,18 Several National Cancer Institute sponsored studies show unequivocally that physicians can be trained to give advice on smoking cessation that results in an increased probability that the patient will stop smoking.1925 Training programmes have also been developed for use by dentists and other health care professionals.26

Methods to determine which smokers would most likely benefit from different modalities of cessation interventions require further study. Most patients do not accept interventions that are more intensive than brief advice or counselling. In one randomised trial of smoking intervention in a primary care practice, referral to smoking cessation classes was “strikingly ineffective.” Of 369 subjects referred to classes by the study, only 14% investigated the classes.27 In a study of a small primary care practice, 11% of those who had stopped had used nicotine gum, and 18% visited a nurse for assistance on one or more occasions.28

Nevertheless, there are patients for whom group classes, or nicotine replacement therapy or both, are likely to be appropriate and effective methods of smoking cessation.2931 About 15% of those successfully stopping report using assisted methods of smoking cessation (programme, psychiatrist or psychologist, nicotine gum, hypnosis, acupuncture). In general, smokers using assisted methods were more likely to be female, middle aged, college educated, and most importantly, persons who had tried to stop numerous times unsuccessfully on their own and who were heavy smokers (more than 25 cigarettes a day).4

Cost effectiveness of mediated smoking cessation

Few research studies on the cost effectiveness of smoking cessation interventions have been conducted. Although there have been many estimates made of the costs of smoking and the benefits of stopping,3235 there have only been a few studies of the cost effectiveness of smoking cessation programmes. Two studies have evaluated the cost effectiveness of physician counselling33,36 and four studies the cost effectiveness of smoking cessation classes, contests, or self help programmes.3740

Despite methodological differences, particularly with regard to estimating costs and the use of outcome measures of effectiveness (per year of life saved, per life saved, and per cessation), these studies have generally drawn the same conclusion. Smoking cessation, regardless of the method used, is at least as cost effective as other widely accepted medical practices. Physician counselling for smoking cessation was found to cost $748 to $2020 per year of life saved.36 The addition of nicotine gum to physician counselling was also found to be relatively cost effective at $4113 to $9473 per year of life saved.33 A comparison of the cost effectiveness (measured in terms of cost per cessation) of a smoking cessation class, a smoking cessation contest, and self help materials found that the self help materials were the most cost effective ($22 to $144 per cessation), followed by the contest ($129 to $239 per cessation), with the class being the least cost effective ($235 to $399 per cessation).40

Also, a few studies have estimated the cost effectiveness of smoking cessation in preventing infant mortality and morbidity. Pashos estimated that smoking cessation programmes cost $100 to $325 per low birth weight prevented and $20000 to $41528 per infant life saved.37 Marks et al estimated that the cost effectiveness of a smoking cessation programme for pregnant women is $4000 per low birth weight prevented and $69542 per perinatal death prevented. Taking into account the costs of neonatal intensive care for low birth weight infants, Marks et al estimated that more than three dollars could be saved for each dollar spent on smoking cessation programmes for pregnant women.38

Use of insurance incentives to promote cessation

Premium differentials based on a subject’s smoking state are nearly universal for life insurance, and to a lesser degree property and casualty insurance policies. In 1987, 39% of insurance companies responding to a national survey reported offering health and behaviour related discounts on individual life insurance policies, with almost all of these including premium reductions averaging 12%-22% for non-smokers.12

There are no studies that consider the relation between smoking-related insurance premium differentials or discounts (health, life or property/casualty) and their effect on smoking cessation rates. There are no data available to suggest that the average insurance premium differential of $10 per month is large enough to motivate a change in smoking behaviour.41,42

Whereas premium discounts based on smoking states for individual and family life insurance policies are widespread, the adoption of similar incentives for group health insurance has been much slower.12 Health insurance companies are generally unwilling to offer a new group health insurance product for which (a) it is difficult to predict utilisation and costs, (¿0 there is currently little marked demand, and (c) there is limited evidence on cost effectiveness.

Recently, however, several health insurance plans have begun to offer group policy discounts based on the smoking state of their enrollees. The Contra Costa County (California) health plan and the King County (Washington) Medical Blue Shield plan recently introduced such group insurance premium discounts based on population smoking rates and the presence of smoking bans at the workplace. For example, the Contra Costa plan adjusts premiums based on smoking prevalence for employers of few people. If a company bans smoking in the workplace, and is shown to have a 90% prevalence of nonsmoking employees, a 15% group health insurance premium discount is offered. In the King County plan, employees can also take advantage of a $500 lifetime benefit for any smoking cessation intervention offered in the community, with a 25% copayment required.

Research priorities

This section describes four broad research questions, the rationale for selecting these questions, and suggestions for study design and study variables:

  • Which patient and provider incentive/ reimbursement schemes are most effective in increasing the demand for and use of clinically effective smoking cessation interventions? Which produce the systematic, organisational practice changes that are associated with higher cessation rates?

Investigating the efficacy and effectiveness of reimbursement and incentives is essential if employers and insurers are to expand these mechanisms. If reimbursement and incentives cannot be shown to have a statistically significant impact on smoking cessation rates in a given cohort (employee, insurance plan participants, and others), then wider application of these incentives is unlikely.

The funding of small demonstration projects exploring the impact on provider practice and smoking cessation rates of different reimbursement and incentive programmes would assist policy development. Such projects might evaluate the use of multiple smoking cessation techniques, the use of various types of health care providers, changes in techniques of office practice, and individual and group contacts.

  • Can insurance and reimbursement policy be effective in changing smoking behaviour and in reducing tobacco-related disease and associated costs?

Ideally, a large, comprehensive, randomised controlled trial would be used to evaluate this issue. It would employ independent variables such as type of reimbursement/insurance scheme, type of employer/insurance incentives, and whether or not a worksite ban is in place. Outcome measures would include: cessation rates measured at several time points; provider participation rates ; institutional practice changes; cost ; tobacco-related claims; and the change in demand for reimbursement. It would evaluate the effect by type of insurance plan; determine the threshold level for discount/premium desired effects; the relation to subsequent direct (medical) and indirect (productivity, absentee days, etc) costs for both continuing smokers and those recently stopped”; the effect of voluntary v mandatory plans ; and the additive effects of other control strategies such as worksite bans or clean indoor air laws. Such a comprehensive trial could be conducted with populations from one or a consortium of insurance companies, a large self insured corporation, a state Medicaid programme, or by using Medicare participants.

Alternatively, a less costly and shorter duration alternative for obtaining part of this information would involve the analysis of existing data bases for health care cost of employers and insurers relating to their preliminary experience in this area. Data on cohorts of documented stoppers could be used to determine subsequent health care and costs. Conducting surveys to determine the characteristics of smokers, as well as the characteristics of employers and insurers who request, implement, and participate in these policies, would also be useful. Determining the range of interventions, reimbursements, and incentives that are acceptable to employers and insurers would assist in making practical recommendations.

  • What steps can be taken to encourage industries and individual companies to focus on health-promotion? What are the barriers to implementation of health promotion measures?

It is important to understand the factors that have led companies within the same industry or within the same geographic area to adopt health promotion programmes. Case studies tracking the “natural history” of a benefits/ reimbursement package over time and location would be useful. Also, analysis of factors such as cost, productivity, and employee satisfaction that led to the development of a given benefit or package of health promotion services are required. It is also critical to analyse failed programmes.

  • What are the adverse consequences of instituting reimbursement/incentives for smoking cessation and non-smoking state generally?

Which method(s) of promoting payment for effective cessation services would most improve overall care to the beneficiaries of the health insurance plan, particularly with regard to the underserved? For example, would there be any negative impacts of expanding smoking cessation benefits in the Medicaid population, such as limiting access to care through changing eligibility? If smokers or other risk groups pay a surcharge for their smoking, either directly or indirectly, will they leave the health insurance plan completely? In Kansas, for example, numerous employees declined coverage after the state implemented a mandatory surcharge programme for smokers. Furthermore, “risk rating”, in which premium differentials are based in part on behavioural risk factors such as smoking or obesity, undermine the fundamental principle of insurance – namely, pooling of risk. Efforts should be made to ascertain the untoward effects of widespread adoption of insurance premium discounts/surcharges on eligibility, costs, and benefit packages for other health plan participants. Do such policies have any impact on personnel issues such as hiring, firing, and promotion?


The vast majority of those who have stopped smoking have done so without the use of a specific cessation intervention or programme. These people have stopped smoking without the benefit of any current specific health insurance incentive or payment to a health care provider or programme for smoking cessation services. It is equally true, however, that for many tobacco users, particularly those who are highly addicted to nicotine, “public health” approaches to cessation (mass media campaigns, tax incentives, workplace bans, etc) are by themselves insufficient to produce lasting abstinence. A wide variety of antismoking weapons, some targeted at the community, others more specifically at the needs of the individual subject, are necessary in order to continue to decrease the toll of smoking related morbidity and mortality.

Dr Louis Sullivan, Secretary of the US Department of Health and Human Services, has issued a policy challenge for tobacco control advocates and health professionals who care for smokers: Given the enormous benefits of smoking cessation, and the fact that good smoking cessation programs can achieve abstinence rates of 20 to 40 percent at one year follow-up, these programs are likely to be extremely cost-effective compared with other preventive or curative services. Therefore, I would encourage health insurers to provide payment for smoking cessation treatments that are shown to be effective. At a minimum, the treatment of nicotine addiction should be considered as favourably by third-party payors as treatment of alcoholism and illicit drug addiction.6

Much remains to be done if Sullivan’s wish is to become a reality. Unfortunately, in the eyes of insurers and employers, it is unlikely that redefining all tobacco use as nicotine addiction will readily lead to improved reimbursement for smoking cessation services. Clearly, however, many who are highly addicted to nicotine would benefit from the expanded access to effective smoking cessation interventions such changes in finance might provide.

Identifying which smokers are most likely to benefit from such interventions will be useful in designing insurance plans to cover such services. Defining and understanding the issues surrounding the decision of employers and employees to seek such coverage for cessation services, and for insurers to design and offer such incentives and reimbursement, are also critical. Focusing on how to best structure, utilise, and pay for smoking cessation interventions in either the individual or group clinical setting are important goals in our overall effort to decrease the use of tobacco.