Landmark Med Diet Study Corrected After Methods Are Questioned

Back in 2013, a study on the effect of the Mediterranean diet on peo­ple at high risk of heart dis­ease and strokes shook the med­ical com­mu­nity. The study had con­cluded that a regime based on the Med diet could help reduce those inci­dents by 30 per­cent.

But last week, the study was retracted to be reviewed by its authors, and then it was again released with the same results but a dif­fer­ent tone in the lan­guage used.

The study, pub­lished in the New England Journal of Medicine, was huge both in terms of the time it took and the peo­ple involved; It went on for five years mon­i­tor­ing more than 7,000 peo­ple in 11 areas of Spain, who were fol­low­ing a spe­cific diet assigned to them. Smokers, over­weight, and dia­bet­ics were found to undergo fewer car­dio­vas­cu­lar events if they switched to a Mediterranean eat­ing style with four table­spoons of extra vir­gin olive oil a day, or the same diet but with an ounce of nuts a day.

Until the study was per­formed, evi­dence that the Mediterranean diet could help peo­ple suf­fer­ing from heart dis­ease was mostly based on obser­va­tions that those liv­ing in coun­tries of the Mediterranean basin had lower rates of inci­dents. Scientists already sus­pected that their eat­ing habits were behind it, but there was room for other fac­tors too, like envi­ron­men­tal con­di­tions or their spe­cific lifestyle.

The ini­tial report had gained a lot of praise because clin­i­cal tri­als in the field are dif­fi­cult to carry through, due to the fact that par­tic­i­pants have to stick to a cer­tain behav­ioral pat­tern and must be mon­i­tored con­tin­u­ously.

But it was the sta­tis­ti­cal analy­sis of one physi­cian, John Carlisle, which showed that some of the data used in the study were not prop­erly col­lected. The actual prob­lem was that the study pre­sup­posed that diets were ran­domly assigned to the par­tic­i­pants to ensure that a pos­si­ble improve­ment in their health con­di­tion stemmed from the diet and only that, but this was not always true.

In a few cases, the same diet was assigned to a whole fam­ily or even a whole vil­lage, pos­si­bly caus­ing faulty results since data was col­lected from a group of peo­ple shar­ing many com­mon habits and envi­ron­men­tal para­me­ters. The inves­ti­ga­tors could not be cer­tain whether the diet or other com­mon fac­tors had led to their improved con­di­tions.

The re-eval­u­a­tion of the find­ings of the orig­i­nal research took almost one year to com­plete. In their revised report, the researchers made sta­tis­ti­cal adjust­ments to com­pen­sate for the mat­ters raised. They also changed the vocab­u­lary and expres­sions used.

It is quite com­mon for sci­en­tific papers to go under the micro­scope of aca­d­e­mics after pub­li­ca­tion who look for dis­crep­an­cies and pos­si­ble rogue data, but it is highly uncom­mon for authors to retract and review their work.

Miguel A. Martínez-González, one of the researchers, told The Washington Post that after review­ing their ini­tial work, he and the rest of the team were more con­fi­dent of their con­clu­sions since they were actu­ally dou­ble-checked and reaf­firmed. The team ​“took more care to adjust for poten­tial imbal­ances,“ he added. ​“No pre­vi­ous trial has under­gone such intense scrutiny.”

But the event caused mixed reac­tions from other researchers. Donald Berry, a sta­tis­ti­cian at MD Anderson Cancer Center in Houston, told the New York Times that the inves­ti­ga­tors ​“were sloppy and didn’t know they were being sloppy.” He added that, while he uses olive oil for cook­ing, he is still not sure about the spe­cific study.

Bradley Efron, a pro­fes­sor at Stanford, said that the reviewed study was not enough to make him stick to a Mediterranean diet.

Cardiologist Steven Nissen of the Cleveland Clinic said he was thrilled to see ​“an impec­ca­ble trial” in the first paper, and the review­ing of it ​“reas­sured” him that the results were cor­rect.

The Mediterranean diet is per­haps the most cel­e­brated dietary regime on the planet, with hun­dreds of reports and stud­ies show­ing its ben­e­fi­cial effects. The study in ques­tion was reviewed and re-pub­lished with the out­come being the same, that a Med diet eat­ing pat­tern can help peo­ple at high risk suf­fer 30 per­cent less from heart attacks and strokes.

But regard­less of the study’s out­come, as John Carlisle said, ​“Whatever you believe, it shouldn’t depend upon just one paper. Try to have not only a bal­anced diet but also a bal­anced view.”

• The Washington Post


• The New York Times