Using disease-specific mortality in discussions with patients

I applaud Canadian Family Physician on its planned series of articles on prevention in primary care, beginning with the “Better decision making in preventive health screening” article in the July issue.¹

However, although I agree with most of the article, I disagree with the authors’ claim that disease-specific mortality is an appropriate outcome measure to evaluate cancer screening.

I suggest that the core of the issue is this: disease-specific mortality’s appropriateness is dependent on whether the reduction in disease-specific mortality is matched by the reduction in overall mortality. If overall and disease-specific mortalities are similarly reduced by the screening, then disease-specific mortality data are useful at the population level. However, if we are considering discussions with individual patients in daily practice, disease-specific mortality does not improve the data we bring to discussions with our patients about the likelihood of mortality.

More important, when disease-specific mortality for a cancer is reduced by screening but overall mortality is not, it means that we have simply traded death from that specific cancer for death from another illness. Further, it suggests that the screening and treatment process for the cancer with lower disease-specific mortality actually causes an increase in disease-specific mortality for other illnesses—something that we have suspected in prostate cancer.

Taking this to its logical conclusion, when considering interventions that reduce disease-specific mortality but do not also reduce overall mortality, we will find ourselves talking with patients about which disease they would prefer to die of. That is an unusually nuanced qualitative decision, one I have found that my patients are ill prepared to contemplate. My experience is that in such conversations patients fall prey to the cognitive error of “availability bias,” whereby they are most influenced by what they have seen in their personal lives. And that means that our attempt to collaboratively discuss options deteriorates into the patient choosing anecdote over evidence. Although I am willing to attempt such conversations, I doubt that the overall well-being of anyone is improved by trying to pick a mortality-causing disease.

As such, I would urge the Canadian Task Force on Preventive Health Care to distinguish between using disease-specific mortality for the purpose of establishing population-level guidance and its suitability for use by front-line family physicians in discussions with patients.

• Copyright© the College of Family Physicians of Canada