I can’t help but believe that there is a considerable number of occasions in which physicians and other medical personnel give patients the wrong idea. It’s not that the information is wrong or delivered without good intentions. Nonetheless, not unlike other fields, the management of information, particularly the ascertainment of accuracy and source, is a difficult matter for those who work in medicine. First, it is difficult for physicians (and most people) to retain massive amounts of information, particularly information that isn’t used on a daily basis. Most physicians store the information they use most often in easily-accessible places in their brains while maintaining familiarity with other related subjects. Secondly, the knowledge base of medicine is constantly and rapidly growing. Our medical school clinical professors frequently tell us, “This information will likely be outdated in five years, so you will have to stay up to date with the most current literature.” Thirdly, at least partly in agreement with friend and fellow blogger Eric, I suspect that many physicians have a poor understanding of statistics and epidemiology, leading them to believe in partial falsehoods passed on to them through teachings of questionable accuracy or unaccompanied by proper disclosure of caveats. Lastly, the information we need is often not available because it is too difficult to collect.
One example of the passing of information without proper disclosure of its flaws was readily demonstrated in a problem-based learning session for my second year Pathology course. It seems that whenever a patient is presented as “gay” in this course, the implication is that the patient is HIV-positive. However, I wonder whether our PBL writers, unaware of the effects of our learning in this course on our fledgling clinical acumen, are confusing the distinctions between risk factors and prevalence. I suspect that it would be difficult to accurately estimate the number of men who have sex with men (MSM, the medical term to avoid ambiguities of the social terms “gay,” “homosexual,” or “queer”), the demographic considered most at risk for HIV transmission, given problems with self-reporting. In the absence of such numbers, it’s not possible to accurately estimate the prevalence of HIV in the MSM population (the proportion of people who are HIV-positive). Instead, epidemiologists can only discuss incidence (the estimated number of new diagnosed cases each year). With respect to incidence, the proportion of HIV transmission in the MSM population is decreasing relative to other risk factors (reference). Instead of passing on the implicit assumption that all MSM are HIV-positive, medical students should be taught about the caveats and flaws of the available information: not all men who have sex with men, probably not even most, are HIV-positive, but they are a group that we should work closely with to prevent new HIV transmission since they are at risk.
Another example closer to my interests is tobacco smoking. I hate tobacco, but I don’t believe that we should tell every smoker and would-be smoker “You’re going to die of lung cancer.” Many people believe that the statistics don’t support this statement. The more we seemingly exaggerate, intentionally or unintentionally, knowingly or unknowingly, the more our patients will think to themselves, I don’t believe you. Instead, a better message might focus on tobacco smoking as the major risk factor in various cancers and chronic obstructive pulmonary disease (COPD, affecting approximately 20% of smokers), and as one of the biggest contributors to coronary artery disease. Or, we can learn how to better interpret the information we have and find information from accurate resources.
For example, a smoking patient might wonder, “What is my risk for getting (and dying of) lung cancer?” First, one should note that lung cancers are generally diagnosed at late “stages,” meaning that when the cancer is detectable and symptomatic, it has already grown into an unmanageable state (e.g. local invasion of other organs and metastasis to distant sites in the body). Accordingly, the five-year survival rate for lung cancer (the proportion of patients treated for lung cancer still alive after five years from the diagnosis) is relatively small – only 16% (40% are alive after 1 year, 27% are alive after 2 years). That doesn’t sound good at all, but what’s the risk of developing lung cancer for a smoker? According to the American Cancer Society, the relative risk for male current smokers for developing lung cancer is 23.3 times greater than that for non-smokers. For male ex-smokers, the relative risk drops to 8.7 times the risk for non-smokers. For women current smokers, the relative risk is 12.7 times greater than for non-smokers, and for women non-smokers, the relative risk is 4.5 times greater. OK, but what if that absolute risk is quite small so that even 23 times a small number is virtually nothing? Absolute risk data is harder to find, but here is some more perspective: the ACS states that women have a 1 in 16 chance of developing lung cancer during their lifetime, an men have a 1 in 13 chance. One might think, “That’s not so bad. I’ll probably not be the unlucky 1 in 13 (7.7%) or 17 (5.9%).” Well, that is the absolute risk data for smokers, ex-smokers, and non-smokers – an average of all three categories. Only about 20% of American adults smoke, which means that the other 80% are non-smokers or ex-smokers. Accordingly, the 1 in 13 and 1 in 17 numbers are substantially underestimating the chances that a smoker would develop lung cancer. Considering that the relative risk of developing lung cancer for smokers is 22.3 times greater for men and 12.7 times greater for women than that for non-smokers, the chance of developing lung cancer doesn’t look so small anymore. Combine this with the information that for men, there are 147,000 deaths to cancer every year, and while lung cancer, the greatest cancer killer in both men and women, amounts to 89,900 deaths per year, 104,200 of all cancer deaths in men are attributed to smoking (in women, 54,300 of 104,600 cancer deaths per year are attributable to smoking).
If we want to talk about death, though, we should be knowledgeable enough to speak convincingly and rationally. I believe, though, that with respect to tobacco smoking, we should discuss more the cardiovascular problems and lifestyle difficulties that smokers suffer throughout their lifetimes, not just at the end of their days. For all patients, we should know how to find information, determine its accuracy and validity, and then know how to communicate it meaningfully to our patients.