Here is an online news story from What Doctors Don’t Tell You… and here [PDF] is their source. I’ve previously written about WDDTY’s reporting of a paper by DeStefano et al. This isn’t quite as bad, but there are a couple of mistakes in what is a very short article.
Here’s what the assessment their article is based on says, under ‘implications for practice':
At a population level, aspirin for primary prevention of CVD is associated with net harm due to increased potential for bleeding, while the results for benefits are not persuasive. For the primary prevention of cancer we consider that more information is desirable.
The assessment looked at aspirin for primary prevention of CVD and cancer. For CVD, the authors concluded that aspirin was associated with net harm. In the WDDTY article, this leads to the headline ‘Just-in-case’ aspirin may be doing more harm than good, the statement that for every potential life saved there is an increased risk of internal bleeding and stroke, and an odd assertion. WDDTY would have us believe that while the authors think it’s a fine line between benefit and risk* “the figures they uncovered may not bear out that conclusion”. So. The balance between benefit and risk is, to WDDTY, clearly weighted in favour of either benefit or risk. But which?
Overall, they reckon that regular aspirin use has saved between 33 and 46 deaths from heart disease and up to 34 deaths from colorectal cancer in 100,000 people over a 10-year period. However, over the same period, the drugs have caused up to 46 major bleeds and 117 gastrointestinal bleeds in 100,000 people.
Perhaps WDDTY think the benefits of aspirin outweigh the risks? After all, death is a bit more serious than major bleeds or gastrointestinal bleeds. I think though, given their antipathy to conventional medicine, that they might be arguing that 46 major bleeds and 117 gastrointestinal bleeds is a worse outcome than up to 80 deaths from heart disease and colorectal cancer (it’s not actually up to 80 deaths from heart disease and CRC but I’ll ignore that for now). The headline gave me a bit of a clue too, mind you. As did the scary reference to aspirin causing strokes (figures for which are missing from the WDDTY article, presumably due to being too low to scare readers – “8 or 10 haemorrhagic strokes”).
What I think is really interesting is that WDDTY appear to think that they are better able than Sutcliffe et al to draw conclusions from the available data. Particularly given the errors in their very brief article. The estimates in the assessment don’t refer to 100,000 people over a ten year period – they refer to 10,000 persons over a ten year period, or 100,000 person-years (see, for example, p86 of the HTA – or use ctrl+f to search the entire document if you prefer). The authors don’t “reckon that regular aspirin use has saved between 33 and 46 deaths from heart disease”, they estimate 33-46 deaths for any cause (CV deaths are actually bundled together with MI and stroke as major cardiovascular events, or MCE). WDDTY claim “up to 46″ major bleeds (it’s actually “up to” 49).
If you’re going to tell the authors of an HTA that the results don’t bear out their conclusions, it might help if you are able to accurately report the results in your article.
*I happened to notice that the WDDTY tl;dr summary of the findings (“In short, it’s a fine line between benefit and risk, they say”) is quite similar to what the authors wrote in the conclusions section of the abstract (on page 8 of the PDF I linked to at the start) and the scientific summary (p15), and what one of the authors was quoted as saying in this press release. The press release, unlike the full assessment, also refers to 100,000 patients. I did briefly wonder if WDDTY had read the abstract and/or the press release but not the full HTA. I think there are a couple of things in the article that don’t appear in the abstract or press release so perhaps they’ve read the full HTA and used two or more of the above as sources?
I’ve written a brief article in which I’ve criticised WDDTY for making errors in a brief article. While it would be embarrassing to find out there were errors in my own brief article, I would prefer it if any errors were pointed out. If you have any criticisms, corrections or clarifications to the above post, do please comment below.