Smokescreen: Your Tax Dollars at Work

Abstract & Commentary

Synopsis: Direct-to-consumer marketing of helical computed tomography (CT) screening for older adult smokers is not advisable.

Source: Mahadevia PJ, et al. JAMA. 2003;289:313-322.

This study is a computer analysis of 100,000 hypothetical 60-year-old heavy smokers. (This population was used because they are at much higher risk for lung cancer than the overall population). The group was divided into 3 different cohorts: current smokers, quitters at the time of screening, and former smokers (those who quit 5 years prior to screening). The model allowed for annual screening from age 60 until age 80. Assumptions of lung cancer incidence, staging, and prognosis were taken from the SEER DevCan program.1 The hypothetical subjects were stratified by age, disease duration, and cell type. The model took into account such issues as overdiagnosis bias (picking up of tumors that do not cause disease), length bias (picking up of slowly growing tumors), and lead-time bias (advancement in the date of diagnosis). It also assumed a standard frequency of indeterminate nodules and of "patients" who are lost to follow-up. For non-small-cell lung cancer (NSCLC), Mahadevia and associates assumed a 50% "stage shift," that is, discovery of tumors in earlier stages (because it tends to be metatastic at diagnosis, they did not make this assumption for small-cell lung cancer). They assumed that the cost of clinical management, including diagnostic testing, would be the same for screened and nonscreened lung cancer cases and that the radiation exposure of the helical CT did not increase the risk of lung cancer. Quality-of-life measurements were taken from the EuroQoL multiattriubute utility scale,2 and cost estimates were taken from the AMA’s National Physician Fee Schedules Relative Value Scale. Outcome measures were the absolute and relative difference in lung cancer-specific deaths, the number of screening tests performed, harms from tests or surgery, and deaths from lung cancer. Costs were measured in dollars and incremental effectiveness in quality-adjusted life-years (QALYs) gained.

For current smokers, there was a 13% relative mortality reduction in lung cancer deaths for the screened group compared with the nonscreened group. However, there were 1186 invasive tests or surgeries for benign lesions in the screened group. For all groups, the cost of screening was between $4300 and $4600. For the smoking cohort, the cost of screening was $116,300 for each QALY gained, but the cost was $558,600 for each QALY for the quitting smokers and $2,322,700 for the former smokers. For current smokers, screening was most effective if begun between 55 and 65 years, and it lost effectiveness for all groups as they aged into the mid-70’s. Mahadevia et al were able to vary most factors in the analysis. For the "most favorable" scenario (low loss of follow-up, low estimates for length and overdiagnosis bias, better quality of life for localized-stage cancer, lower costs per case for CT screening, and no anxiety from indeterminate nodules), screening led to 900 fewer lung cancer deaths (16% mortality reduction), 1520 false-positive invasive procedures, with a price tag of $42,500 for each QALY gained. In the "worst case" scenario, there were 119 fewer lung cancer deaths (4% relative mortality reduction), and 993 individuals were harmed. Mahadevia et al concluded that direct-to-consumer marketing of helical CT is not advisable.

Comment by Barbara A. Phillips, MD, MSPH

Let me come out right now and admit a bias: I don’t think that you or I ought to pay for screening tests for lung cancer in every asymptomatic current smoker. Smokers who are worried about lung cancer ought to quit smoking! As to the notion that such individuals can and will pay for screening out of their own pockets, forget it. There is an inverse relationship between socioeconomic status and likelihood of smoking; those who generate the excess health care cost associated with smoking are least likely to be able to afford it, and taxpayers get stuck with the tab. (We seem to have forgotten that the Master Settlement agreement between states and the tobacco industry was the result of suits brought by states’ Attorneys Generals on behalf of taxpayers, originally intended to recoup the tax dollars spent on cigarette-related illness incurred by Medicaid recipients. Unfortunately, these dollars have been largely used to build roads and libraries and to replenish state coffers. But I digress). This model clearly demonstrated that quitting smokers and ex-smokers were at markedly reduced risk of lung cancer and that screening for cancer in this group was even more expensive than screening in the smokers.

But this is America. Commercially available cancer (and other disease) screening programs, fueled by direct advertising to consumers, are proliferating.3 In most instances so far, individuals are paying for these services out of pocket. US News and World Report has a remarkably balanced report on this phenomenon, noting that screening does not change behavior, physicians at screening centers are "delighted" to write orders for tests that require them, and that both the rate and cost of false-negative tests are very high.4 The article concludes, however, that, "Short of a bad economy or a skeptical public, scanning centers seem poised to continue to reap profits from the worried well."

So, how do you manage the patient who asks your advice about lung cancer (or other) screening, or who brings in results for you to interpret? The biggest headache appears to be the falsely positive scan, which can lead to a cascade of expensive, risky testing, including surgery. Patients need to know this up front, and physicians need somehow to avoid being drawn into the testing spiral as they practice defensive medicine. There is little guidance on this issue in the article at hand, or even in US News and World Report. The article did spawn a flurry of letters to the editor, which highlight the complexity of this issue and the likely ferocity of the impending debate about it. Perhaps the most useful letter was from Tom Petty, who once again urges us to go after the "low hanging fruit," those patients at highest risk.5 Those are current heavy smokers with airflow obstruction or an occupational risk or family history of lung cancer. In a study of 86 patients with this profile, 5 cancers were found at a cost of $12,900 per cancer, and 4 of these cancers were potentially curable.6 Petty urges us to consider this kind of "targeted case finding" for patients at highest risk.

The question is, who should pay for it? Do we really want to "enable" actively smoking COPD’ers by subsidizing their screening tests?

The accompanying editorial emphasizes that ". . . smoking cessation must remain the first and foremost priority in reducing the burden of lung cancer in the population,"7 and notes that the National Cancer Institute has recently initiated a large, randomized trial to test the efficacy of helical CT scanning as a lung cancer screen.8 Pending the results of this study, helical CT screening for lung cancer, ought to be considered a research tool. Patients who want it should probably be appraised of the risk of false positives up front, perhaps even with a documented disclaimer of liability obtained by the hapless MD who will wind up sorting it out. And they should pay for it themselves. And quit smoking!

Dr. Phillips is Professor of Medicine, University of Kentucky and Director, Sleep Disorders Center, Samaritan Hospital, Lexington, KY.


1. DevCan: Probability of developing and dying of cancer (computer program) Bethesda, MD: National Cancer Institutes: 2001.

2. Dolan P. Med Care. 1997;35:1095-1108.

3. Patz EF Jr, et al. N Engl J Med. 2000;343:1627-1633.

4. Comarow A. US News and World Report. May 19, 2003;48-52.

5. Petty TL. JAMA. 2003;289:2357.

6. Bechtel JJ, et al. Chest. 2002;122 (suppl):216S-228S.

7. Grann VR, Neugut AI. JAMA. 2003;289:357-358.

8. Broughton B. Lancet Oncol. 2002;3:647.