We were on high alert when reviewing this story about CT screening and lung cancer deaths, because of questions about the methods used in the study and the advocacy history of the lead researcher. The headline, lead sentence and concluding section… the most important sections of the story… tilt toward meek acceptance of the study claims. However, readers who take in the whole story are treated to a solid examination of many of the questions this study raises. The top and bottom of the story should have more closely reflected the meat of the report. Also, the story emphasizes the apparently large (36 to 64 percent) relative reductions in death rates without giving equal billing to the absolute reductions of less than one percent or the hundreds of smokers that would have to be screened in order to prevent one lung cancer death. See our primer on absolute versus relative risks. Also see risk communication expert Gerd Gigerenzer’s comments on the importance of communicating absolute risk data.
Lung cancer is the leading cancer killer of men and women in US. While the most effective response is reducing smoking (which also reduces the heart disease and other smoking-related ailments that kill and disable far more people than lung cancer), effective lung cancer screening could offer noteworthy benefits to those who have smoked. But screening presents numerous logical traps, including exposing people to costs and harms of treatment for tumors that turn out be either harmless or merely advancing the starting date of treatment that ultimately fails to make any difference in the outcome. Earlier published studies suggested earlier diagnosis of lung cancer in a group of heavier smokers, but the eventual outcomes of the diagnosis and the impact of CT screening remained unclear. This new observational study provides some additional insight.But, as independent experts emphasize in this story, there are still many issues to be resolved about lung cancer screening and the high false-positive rate is still an issue.
The story reports that this sort of scan costs about $300. The story also notes that if screening is effective, the scans would need to be repeated, though there is no consensus on what schedule would make the most sense. The CDC estimates that more than 45 million Americans currently smoke cigarettes and more than that many are former smokers.
The story should have more clearly pointed out that the cost of a screening program is not the cost of a single test, but the cumulative cost of repeatedly screening many millions of people as well as the cost follow-up tests and treatments… both the treatment of potentially dangerous lung cancers, but also the treatment of the many suspicious lesions that turn out not to be cancer. A separate study found a false positive rate of more than 20 percent, suggesting that follow-up tests (and the costs they incur) would be a frequent added expense. The story should have at least referred to the national costs and the additional testing and treatment costs triggered by false positive scans.
This story generally meets the specific requirements of this criterion, but the overall tone and structure of the report lead readers toward an exaggerated sense of the benefits reported in this study.
Putting aside fundamental questions about study design addressed earlier, there should have been more attention to the absolute differences in the actual versus expected outcomes for the smokers included in this study. The story highlights mortality reductions of 36 to 64 percent. While it does report deep into the story that 64 smokers died of lung cancer out of almost 8,000 in the main study group, readers would have to pull out their calculators to learn that based on these numbers… a 36 percent relative reduction in lung cancer mortality translates into an absolute reduction of less than half a percent. Specifically, 0.8 percent (64 out of 7995) of the smokers screened with CT died of lung cancer, while based on one comparison the researchers say the expected lung cancer death rate would be 1.2 percent (99.8 out of 7995). That is a difference of 0.4 percent. The story should have noted that based on these statistics, more than 200 smokers would have to be screened in order to make a difference for one of them.
The story also errs by highlighting that another report (based on an actual comparison of screening to not-screening) reported a 20 percent reduction in lung cancer deaths, without also noting that those researchers reported about 300 smokers would have to be screened in order to stave off one lung cancer death.
The story should have also clearly told readers that even if screening reduced lung cancer deaths by 20 percent in the most rigorous trial to date, that does not mean that screening would prevent 30,000 lung cancer deaths a year (20 percent of the 157,000 lung cancer deaths per year reported in the story).
While too many news reports take a “why not?” attitude toward medical screening tests, this story confronts some of the potential harms. It refers to a recent report that more than one in five “initial lung CT scans show suspicious lesions that turn out not to be cancer, but lead to needless invasive follow-up procedures and radiation exposure, as well as stress and anxiety for patients and their families.”
This criterion is of critical importance given the controversies over lead researcher Claudia Henscke’s long-standing advocacy of CT screening of smokers and criticism of some of her earlier publications on this topic. This latest study raises a number of questions about the methods used to compare lung cancer deaths in three different groups of people. The screened group was made up of heavy smokers living in New York who were recruited by Henschke and her team specifically for a study of CT screening. There was no control group, so in this study statistics from two other unrelated studies are used. One group (CPS-II) was made up of people who reported being smokers as part of a large nationwide study of cancer in general. The second group (CARET) was a subset of smokers who took part in the control group of a trial of drugs to prevent lung cancer.
The story takes note of earlier controversies involving Henschke and points out that this study attempts to compare groups of people with important differences and then calculate estimates of the number of lung cancer deaths would have been expected in the first group if they had not been screened.
Although the headline, lead and concluding lines of this story don’t incorporate these and other concerns, we give the story credit for attempting to address challenging questions of scientific and statistical methods.
The story reports credible numbers describing the toll of lung cancer.
The story quotes two independent experts who provide valuable perspectives on this study. We wish more stories provided similar attention to independent experts. It also notes that Dr. Henschke has been criticized in the past for accepting funding from tobacco companies and that this study was supported in part by manufacturers of CT scanners.
The story does not discuss alternatives. It would require only another sentence to at least note the reduction in the risk of lung cancer, heart disease and other ailments that come from quitting smoking.
Although this story does not directly address the availability of CT screening specifically designed to identify lung tumors, it does point out that such screening is generally not covered by insurance.
The story puts this study into context with other recent reports, thus making it clear that the evaluation of CT screening for lung cancer is a process that has been going on for a number of years and is far from finished.
The story clearly does not rely on a news release.