Prof Colleen Aldous clarifies the outcome of the colon cancer study as positive

This week, BizNews republished an article by Bloomberg which indicated that the findings of a 10-year study published in the New England Journal of Medicine questioned the benefits of colonoscopy screening exams, as the common procedure failed to reduce the risk of death from colon cancer. Shortly thereafter, the American Gastroenterological Association (AGA) released a statement indicating that the conclusions of the study aren’t necessarily applicable to colorectal cancer screening in the US and have been misrepresented in the media. This article by Professor Colleen Aldous concurs with the AGA statement highlighting the error of interpretation by the global mainstream media “which could prove to be deadly for many”. In the study, the researchers used the ‘intention to screen’ number as a denominator which does not give efficacy. Aldous clarifies that the paper gives an idea only of what the public health outcome would be if you offered free colonoscopies, and not of true efficacy. The results are therefore misleading. – Nadya Swart

Colonoscopy as a screening tool is effective in catching colorectal cancer early in patients, but a screening program using colonoscopy needs more public health education – study shows

By Colleen Aldous

A large randomised controlled study looking at the implementation of a screening program for colorectal cancer, which included offering colonoscopy, was published in the NEJM on 9 October (1). It wasn’t long before the global mainstream media (MSM), including Bloomberg and CNN, carried the story questioning the efficacy of colonoscopy for screening for colorectal cancer (CrCa). This is an error of interpretation that could prove to be deadly for many.

Professor Colleen Aldous

It was an international trial gathering data from Poland, Norway, and Sweden over 10 years. People between the ages of 55 and 64 were randomised into the intervention arm, where they were offered colonoscopy, or into the control arm, where they were offered nothing.

Let us look at the original data and do some maths:

A total of 84,585 participants were included in the study. 56,365 were not invited to have a colonoscopy (control arm), and 28,220 were invited to have a colonoscopy (intervention arm). Less than half (11843 or 42%) took up the opportunity of those invited to have a colonoscopy. This is a huge randomised controlled trial, and most who have been through medical school classes on evidence-based medicine (EBM) would not question the outcomes as being true. However, we need to look at the interpretation of the data more closely.

After ten years, 622 people from the control arm were diagnosed CrCa after enrollment, and 259 of those in the intervention arm had a diagnosis of CrCa.

If we calculate the percentages for the control arm: 622/56,365 = 1,1%

If we calculate for the entire intervention arm: 259/28,220 =0,9%

But, only 11,843 (the per protocol number) actually had the treatment, thus 259/11,843 =2.2%

So what can we conclude?

This paper shows that if you invite the population for free colonoscopies, only a fraction (42% from all three countries) of the people will take up the opportunity and see the remarkable benefit they offer. The countries where this study was carried out had very different uptake rates, with only 33% in Poland and 60,7% in Norway. This may indicate that there may be a difference in public health education regarding the benefits of screening in different countries. Nowhere does the research question the paper is based on, i.e., what are the benefits of a public health program where people are invited for colonoscopy, show that colonoscopy is an ineffective screening tool, thus refuting all previous studies which indicate that it does.

The paper uses the intention to treat principle (ITT), meaning that the number of ALL participants randomised to the screening arm is used as a denominator to calculate efficacy. In their study 28,220 participants were invited to have colonoscopies. Of these, only 11,843 took up the opportunity and were followed up. From this smaller group, 259 people were found to develop CrCa. To see if a screening program, not to be confused with a screening tool, for the entire population is effective, one would use the total number of participants to calculate the incidence, i.e. 259/28,220. This shows only the effectiveness of a voluntary screening program, considering that more than half the population will not take up the intervention.

If you look at those who participated fully by having colonoscopies, you will use the per protocol number, i.e. the number of participants who did take the intervention, as a denominator and calculate incidence as 259/11,843. This number shows that colonoscopy does indeed catch more CrCa early than in the control group, mirroring the findings of countless previous studies showing the efficacy of colonoscopy as a good screening tool for individuals. The American Gastroenterologist Association’s statement is quite correct in its conclusions. They interpret the per protocol figures to show a mortality reduction of 50% and increased CrCa prevention for colonoscopy as a screening tool. This does, however, bias the data towards compliance.

The NEJM editorial by Dominitz et al., (2) which accompanied the original article, has set this skewed interpretation of the data into the MSM by concluding:

“If the trial truly represents the real-world performance of a population based screening colonoscopy, it might be hard to justify the risk and expense of this form of screening when simpler, less invasive strategies (eg sigmoidoscopy and FIT (faecal immunochemical testing)) are available.”

Interestingly, two of the study’s authors declared conflicts of interest as speakers or consultants to eight companies involved in manufacturing instruments or diagnostics related to colonoscopy. For several years, the NEJM has diminished the importance of conflicts of interest by authors. A cynical reader of this science might see that this paper could be used in future product marketing.

This is not the only time the ITT principle has been used to obfuscate possible truth. The TOGETHER Trial results (3) were published in the NEJM earlier this year. The results were released to the MSM weeks before publication. A clever ploy because any criticism of the study could only be voiced weeks later after the negative results had become global knowledge.

In this study, 679 participants received ivermectin, and another 679 were allocated to the control arm. In the end, 624 remained in the intervention arm and only 288 in the control sample. More than half the control arm dropped out? Why? Could it be because they were sick with COVID-19 and were not getting better on the placebo and therefore decided to get treatment elsewhere rather than risk dying?

Let’s look at the maths for mortality control:

For the group that received ivermectin per protocol: 21/624 = 3,3% died. The paper’s authors ignored that more than half of their control participants failed to complete the study and used the ITT number 24/679 = 3,5%. This is nonsensical, and information gleaned from this calculation is useless. When we look at the reported deaths from the control group with the per protocol number, we see that ivermectin clearly prevented death. I asked the lead author if the 24 deaths came from the 288 to be certain of my thinking, and he refused to answer my question. I worked out a mortality rate of 8,3% in the per protocol placebo arm.

Clearly, the MSM, which is used as a tool to get medical information to the public, and the public itself, need to become more savvy with scientific communication. Science is not at fault here, but it’s what people do with science that we have to keep questioning.


  1. Bretthauer M, Løberg M, Wieszczy P, Kalager M, Emilsson L, Garborg K, Rupinski M, Dekker E, Spaander M, Bugajski M, Holme Ø. Effect of Colonoscopy Screening on Risks of Colorectal Cancer and Related Death. New England Journal of Medicine. 2022 Oct 9.
  2. Dominitz JA, Robertson DJ. Understanding the Results of a Randomised Trial of Screening Colonoscopy. New England Journal of Medicine. 2022 Oct 9.

Reis G, Silva EA, Silva DC, Thabane L, Milagres AC, Ferreira TS, Dos Santos CV, Campos VH, Nogueira AM, de Almeida AP, Callegari ED. Effect of early treatment with ivermectin among patients with Covid-19. New England Journal of Medicine. 2022 May 5;386(18):1721-31.

*Colleen Aldous has a doctorate and is a full Professor and Health Care scientist at UKZN’s medical school where she runs the doctoral academy at the College of Health Sciences. She has published over 130 peer-reviewed articles in rated journals.

Read Also:

Visited 1,045 times, 2 visit(s) today