The story about a reporter’s decision to treat a carotid artery blockage exemplifies the powerful strengths and weaknesses of first-person health care reporting. It carries a powerful emotional punch and an apparently happy ending that overwhelms the evidence presented. Numerous psychology studies document the power of “outcome bias,” in which the results of a decision affect beliefs about whether the decision was correct or not. We are very happy that all went well, but it is reasonable to assume this story would have looked very different if had not.
There is no mention of the costs of the surgical procedure. Neither is carotid stenting mentioned as an alternative to endarterectomy surgery. Discussion of carotid artery screening of people who do not have symptoms is mixed up with the reporter’s experience with testing done in order to diagnose the cause of symptoms – a somewhat common flaw of conflating diagnostic testing with screening tests. The risk that the surgery could trigger a stroke, including a fatal one, is highlighted, but the risk statistics reported here are lower than those published in major summaries of the evidence.
The headline of the story proclaims that a carotid artery scan “Saved My Life.” That may be true, but it is also unknowable. What the best summaries of the evidence conclude is that in cases similar to those of the reporter, surgery to clear a nearly-blocked carotid artery lowers the long-term risk of stroke and death. However, stories on this topic should highlight the fact that the procedure does not actually “save the lives” of most patients. As a systematic review of the evidence by the Cochrane Collaboration concluded, using a statistical analysis known as the number needed to treat (NNT), six patients need to be treated to prevent one stroke or death within five years of the surgery. That means that for five out of six patients, the procedure probably does not alter the outcome: Either they would not have had a stroke or died even without surgery or they still had a stroke or died despite having the surgery.
Another important concern with this story is that the reporter uses his own case — where the scan was part of an attempt to track down the cause of his neurological symptoms — to argue for scanning of all older adults, including those who have no symptoms at all. This is a misleading argument since people with blockage severe enough to cause symptoms are likely different in important ways (and more likely to benefit from this scan) than those who don’t display symptoms. While it is certainly reasonable to choose a treatment shown to reduce the risk of stroke and death, it is imperative that stories clearly communicate the difference between screening and diagnosis, and the uncertainty about whether any particular patient will benefit.
While the story reports the cost of scanning carotid arteries to be about $70, it does not discuss the cost of surgery. One analysis of Medicare reimbursements done a few years ago reported hospitals were paid about $11,000 for each patient getting a carotid endarterectomy. An earlier report estimated that in 2008 hospitals charged about $35,000 for the surgery and associated treatment, though the ultimate reimbursements by Medicare were much lower.
The story does report on the results of a major trial from 20 years ago that concluded patients who were randomized to surgery had about a 5 percent risk of stroke or death within three years, “less than half that of people treated only with drugs.” But this statistic is overwhelmed by the headline (The Scan that Saved My Life) and the punchline, “mostly I feel lucky to have gotten a test that may have saved my life.”
A more recent review of the evidence by the international Cochrane Collaboration concludes that the “number needed to treat‘ (for the type of blockage and symptoms the reporter had) would be six patients. Meaning, six people have to be treated to prevent one stroke or death over a five-year period. In other words, the surgery likely does not alter the outcomes of five out of six patients. Either they wouldn’t have had a stroke or died even without treatment or they did have a stroke or died despite undergoing surgery.
We will rate the story as satisfactory because it does clearly point out that the surgery is hazardous, sometimes triggering a stroke, and that some patients die soon after the procedure. However, the story reports stroke and death rates at the lower end of the ranges reported in medical literature. The story says a major task force report estimated the risk of stroke and death within 30 days after surgery as less than 3 percent overall, but up to 5 percent at some hospitals. However, the lower figure in that task force report uses data only from clinical trials, which typically involve highly skilled surgeons and have better outcomes than typical clinical care. That report also said that some observational data indicates a risk of stroke and death of almost 7 percent at some hospitals. The evidence review mentioned above concludes that the 30-day risk of stroke and death is 7 percent in clinical trials of carotid endarterectomy.
The story does warn readers that few hospitals and surgeons publish their individual results.
The story would have been stronger had it mentioned that there are harms that can occur if we started screening anyone, regardless of whether they had symptoms. As the USPSTF explains, “For the general primary care population, the magnitude of benefit is small to none. Adequate evidence indicates that both the testing strategy for carotid artery stenosis and treatment with [surgery or stenting] can cause serious harms.”
The story does refer to certain major trials and guideline reports. However, the story seems to cherry-pick conclusions, favoring those that tilt toward screening and treatment while questioning those that recommend against widespread screening. It features a guideline in favor of screening from a professional society of surgeons. It does not mention the evidence review by the international Cochrane Collaboration. It also does not mention important limitations of studies that appear to favor screening — including that the medical therapy used in some of these older studies is no longer considered state of the art. The headline and punchline of the story portray a sunny and one-sided view that dismisses the complexity and uncertainty of the available evidence.
Although the story accurately reports statistics on the number of people who have strokes, it confuses readers by referring to “screening” (which is testing people who don’t have symptoms) with the testing the reporter underwent to diagnose the cause of specific symptoms. After noting that a U.S. task force (the US Preventive Services Task Force) recommends against screening people without symptoms, the story says, “My experience, along with some evidence from screening of thousands of individuals, raises questions about the task force’s conclusions.”
Since his experience was with diagnostic testing, it is not relevant to a discussion of screening tests.
It is an inherent weakness of this sort of first-person reporting that there is an absence of independent sources. All but one of the physicians quoted in the story were personally involved in the reporter’s case. The one exception is a vascular surgeon who strongly advocates for broader carotid artery screening. There are no quotes from experts who were members of the task force that recommended against widespread screening.
The story makes no mention of carotid artery stenting. This omission is surprising, since the reporter has covered this alternative on several occasions:
(Full text versions of some stories are behind a paywall.)
It is clear that carotid endarterectomy is widely available.
The story makes clear that the procedure has been in use for decades. What could have used more exploration–from independent sources–is whether it’s overused.
The story is based on the reporter’s personal experience.