Article by Judith Garber, May 9, 2020
Abstract Dr. Cécile Bour MD
Judith Garber is a scientist in political science and health policy at the Lown Institute. (The Lown Institute based in Massachusetts, United States, is a non-profit organization, a "non-partisan think tank that advocates innovative ideas for a fair and caring health system").
Screening procedures and routine medical visits have declined since the Covid-19 pandemic. According to the author, this trend is not necessarily deleterious, as there will likely be a decrease in unnecessary and harmful care.
A recent analysis conducted by the EPIC* Health Research Network used electronic health record data from 2.7 million patients in the U.S. and compared the rates of cervical, breast and colon cancer in the pre-Covid period with the current Covid period. They found that screening appointments in March 2020 decreased by 86-94% compared to the average number of screening appointments that occurred monthly from January 1, 2017 through January 19, 2020.
*EPIC: The Epic Community is a global community of healthcare organizations.
What effect would this drastic decrease in screening have on patient health? The authors of the EPIC report claim that it will be disastrous….
But is this grim prediction likely? According to J.Garber, several points need to be considered.
Mortality by disease vs. all-cause mortality
Although there is some evidence that cancer screening can reduce specific mortality, i.e. mortality from specific cancers, there is much less evidence of benefit on all-cause mortality (the risk of dying overall).
Editor's note: In the above cited study, the effect of screening with fecal occult-blood testing on colorectal-cancer mortality persists after 30 years but does not influence all-cause mortality.
( Note that the interest in all-cause mortality is that it includes both mortality from the disease and mortality from treatment of the disease.
If there is an impact with decreased all-cause mortality, it means that screening (and treatment) are effective. The overall mortality thus serves as a warning signal; e.g., if there is an abnormal offset between overall and disease-specific mortality, this should raise the question of whether an adverse effect of screening may have been insufficiently taken into account. Overall mortality is not generally used as a judgement criterion, which is regrettable because it implies a loss of information. Editor's note)
In 12 randomized screening trials, reviewed by Dr. Vinay Prasad and colleagues in an article published in the BMJ, seven failed to show a decrease in overall mortality, despite lower rate of disease-specific mortality. In some cases, overall mortality in the screening group was therefore higher than in the non-screened group. What is the reason for this?
The possible explanation is that the negative effects of screening, in these cases, may counteract the benefits. While cancer screening reduces mortality for some cancers, for others it leads to many more false positives with unnecessary biopsies, as well as over-diagnosis and over-treatment. Thus unnecessary surgery and complications of diagnostic and therapeutic tests and procedures have a negative effect on health in screened groups. See also .
The authors of the EPIC report should therefore have included in their alarming conclusion not only potential cancer deaths due to lack of screening, but also potential false positives and cascade of events avoided (over-diagnosis followed by over-treatment), and should also have taken into account the lack of evidence regarding the overall benefits of screening on overall mortality.
Not everyone benefits from screening
Not everyone is equally likely to benefit from cancer screening explains J.Garber. The likelihood of a young adult developing cancer is very low, making them more likely to be exposed to screening damage than actually helped.
At the same time, at a very advanced age, cancer screening also becomes less beneficial because older people no longer have the life expectancy to benefit from screening, but on the other hand are more susceptible to the harmful complications of testing and treatment that result from the detections. For example, the US Preventive Services Task Force (an independent US agency that reviews health devices) recommends colorectal cancer screening only between the ages of 50 and 75.
Yet many people are being screened for cancer outside the recommended ages, even though they are unlikely to benefit from it, the author explains.
In one 2014 study published in JAMA, among elderly patients with a very high mortality risk , 37.5% were screened for breast cancer, 30% were screened for cervical cancer and 40% were screened for colorectal cancer. Among women who had previously undergone a hysterectomy but did not have cancer, 34% to 56% still had been screened for cervical cancer within past three years.
Although cervical cancer screening is not recommended for young women aged 15 to 20, an estimated 1.6 million Pap tests (smears) are performed unnecessarily on women in this age group each year.
In the EPIC Network's analysis above, these are not considered. If most of the people who missed out on screening in March 2020 were those at lowest risk, then a reduction in screening participation may not be such a bad thing overall - especially if it was the number of obviously inappropriate screenings that was decreasing.
However, with the information we do have, we do not know which patients have avoided screening, making it impossible to estimate the real number of lives lost or damage reduced in this 'Covid pause' from screening.
An opportunity for research
The authors of the EPIC report explain the issues that seem clear to them, namely the 'lives lost' due to reduced screening. However, it is very likely that this number is overestimated, given the high rates of inappropriate screening in the real world, as well as the potential negative and unaddressed effects of screening on overall mortality.
Although the true impact of the sudden halt in screening due to Covid-19 remains unknown, according to J. Garber we have an opportunity to find out.
Such a drastic change in screening is unusual, so we must take advantage of this unique event and follow the results of this historical experiment to better understand the actual health effects - both beneficial and harmful - of cancer screening.
A few cautions: this monitoring of data will depend on the evolution of the outbreak and whether or not screening is resumed. Two months of interruption, as is the case in France, may be insufficient to draw these conclusions or to discern real variations in cancer incidence and mortality rates. This can be made even more complicated and the conclusions even more hazardous if the resumption of screening is not carried out in a uniform and homogenous manner in all regions of France, depending on the persistence here and there of Covid clusters that curb the public's willingness to attend the various screenings. Only if the interruption of screening would continue for a few more months, either because of the persistence or resurgence of the outbreak, or because of public resistance (fear of going to the doctor's office or of having to undergo a colonoscopy in the context of an epidemic), the situation would be quite different and results could emerge.