Medical Research: Torture Data Until It Conforms to Desired Result

Torturing scientific data has exactly the same result as torturing humans. The victim or data gives up the desired answer, as demonstrated in a reanalysis of a salt study.

by Heidi Stevenson

1 August 2011

Mainstream medicine was aghast at a recent salt study showing that lower salt intake is associated with higher death rates. So, they tortured the data to claim it shows the opposite. Apparently, they couldn't find someone independent to take on the task, so a man who has made a career of advocating salt restriction acted as the inquisitor. And, of course, he claimed he had no conflict of interest.

As reported in Reducing Salt Does Not Reduce Heart Disease: Cochrane Study, a recent meta analysis demonstrated less salt is not beneficial. This follows another recent study, Heart Disease Deaths Four Times Higher With Low Salt Intake: JAMA Study, showing that deaths from heart disease are as much as four times likelier with low salt usage.

Conflict of Interest

The Cochrane study demonstrated that there is no benefit in taking less salt. They found no evidence that reduced salt intake reduces mortality. This, of course, upset Dr. Graham A. MacGregor, who is the chair of Consensus Action on Salt and Health and the World Action on Salt and Health (WASH), both of which exist to push lower salt intake. The Mission Statement of WASH reads:

World Action on Salt and Health's mission is to achieve a reduction in dietary salt intake around the world from the current intake of 10-15g/day to the World Health Organisation (WHO) target of 5g/day. This fall in salt intake and the resulting fall in blood pressure would lead to major reductions in both incidents of, and deaths from Cardiovascular Disease (CVD) i.e. stroke, heart failure and heart attacks, with a major reduction in the disability that results from CVD.⁽¹⁾

MacGregor has made a career of advocating reduced salt intake. He obviously has a lot to lose if the myth that salt is bad for health is refuted. Nonetheless, he claimed that he has no conflicts of interest on the topic!

So, he set out to reanalyze the Cochrane study data. Of course, the data really does show what its lead author claimed, which is a serious problem for MacGregor. But, when a great inquisitor wants his victim to say something, that victim will. That's as true when the victim is data as it is when the victim is human. Torture consistently gets the results the inquisitor wants.

How the Data Was Tortured

The Cochrane meta analysis correctly treated each of the included studies separately, thus avoiding mismatched comparisons. The original study used the Cochrane Handbook for Systematic Reviews of Interventions method of analysis, indicating that the authors made a point of following accepted methodologies.

MacGregor and his co-author, Dr. Feng J. He, decided that the study of heart failure patients wasn't appropriate because they wouldn't normally be placed on salt-restriction diets. Why that has any meaning isn't explained. Their salt intake was checked and its effect on their health recorded. In the context of the study, whether they would "normally" be placed on salt-restriction diets is not relevant to whether the amount of salt intake affects their health.

Then, they took the other six studies included in the meta study and lumped them all together, rather than treating them separately. Why this is not a legitimate approach may not be obvious. However, consider the number of subjects involved in each of the study types:

Normotension patients: 3,518
Hypertension patients: 758
Mixed normo- & hypertension: 1981

If the percentage of adverse events suffered by hypertensive patients is higher than those suffered by normotensives, they would carry less statistical weight when lumped together with normotensives because there were more than 4½ times as many normotensive patients. Therefore, if you properly compare hypertensives with hypertensives, you might find that the rate of death is about the same for low and high salt users. But, if you lump them all together, a false impression that low salt use is beneficial can emerge.

Relatively independent organizations like Cochrane set up guidelines for analysis to help assure honesty in reporting. When so-called researchers want to assure that a particular result is seen, they need to torture the data. And that's what MacGregor did. They tortured the data enough to assure that it would tell them what they wanted to hear.

Study Results

If a study is decently designed and the data is analyzed properly, it tells the truth. When tortured, it does what any torture victim will do: says whatever the inquisitor wants to hear. That's as true in salt studies as it was during historical inquisitions.

More and more, the most important thing one can do in determining the validity of a medical study is to follow the money trail—even when the researcher says there isn't one. As is obvious from this particular salt study reanalysis, an author's claim that there's no conflict of interest may be based in fantasy. Trust the claim at risk to your health.

*****************************************************************************

What do you think? Click here to comment!

*****************************************************************************

Search Gaia Health: