On 2016 Apr 07, Stephen L. Black commented:

Update: BMJ has now belatedly but effectively responded to my complaint, and a version of my original comment has been posted on the Open Diabetes website as I had initially intended.

On 2016 Jan 27, Valentine Njike commented:

The reviewer is right to suggest this point it getting excessive attention. It was a secondary outcome measure, and a significant within-group change only, when primary measures were between-group changes. This is indicated explicitly in the paper...

On 2016 Jan 22, Stephen L. Black commented:

This study claims to demonstrate that eating walnuts lowers cholesterol level, a claim enthusiastically and uncritically repeated in the media. Unfortunately, the claim is spurious, as the published data show no such effect. Instead the results do not show a statistically significant difference in cholesterol level between the experimental group which ate walnuts and the control group which did not. This is evident in Table 2 and is admitted by the authors in their discussion. In a puzzling comment, they attribute this negative outcome as “probably due to the placebo effect”. They instead focus on a significant decline in cholesterol level from baseline in the experimental group, ignoring the fact that a similar decline was evident in the control group which did not ingest walnuts. Thus, by the conventional logic of a placebo-controlled experiment, they failed to demonstrate that eating walnuts lowers cholesterol level. I find it inexplicable that a study with such an obviously-flawed conclusion should have been allowed to be published.

As the journal web page invites on-line comments, I submitted one, pointing this out. I was astonished to be told by the editor that my comment would not be published, not because it lacked merit, but because the journal does not publish comments. As a last resort, I protested to the BMJ itself this pointless practice of inviting on-line comments while refusing to publish them. Alas, the BMJ, which proudly states that it “welcomes complaints” and will acknowledge them “within three working days” instead ignored me. All-in-all, my failed attempt to merely correct a scientific error leaves me discouraged concerning the current practices of this once-respected journal.

On 2016 Jan 22, Stephen L. Black commented:

This study claims to demonstrate that eating walnuts lowers cholesterol level, a claim enthusiastically and uncritically repeated in the media. Unfortunately, the claim is spurious, as the published data show no such effect. Instead the results do not show a statistically significant difference in cholesterol level between the experimental group which ate walnuts and the control group which did not. This is evident in Table 2 and is admitted by the authors in their discussion. In a puzzling comment, they attribute this negative outcome as “probably due to the placebo effect”. They instead focus on a significant decline in cholesterol level from baseline in the experimental group, ignoring the fact that a similar decline was evident in the control group which did not ingest walnuts. Thus, by the conventional logic of a placebo-controlled experiment, they failed to demonstrate that eating walnuts lowers cholesterol level. I find it inexplicable that a study with such an obviously-flawed conclusion should have been allowed to be published.

As the journal web page invites on-line comments, I submitted one, pointing this out. I was astonished to be told by the editor that my comment would not be published, not because it lacked merit, but because the journal does not publish comments. As a last resort, I protested to the BMJ itself this pointless practice of inviting on-line comments while refusing to publish them. Alas, the BMJ, which proudly states that it “welcomes complaints” and will acknowledge them “within three working days” instead ignored me. All-in-all, my failed attempt to merely correct a scientific error leaves me discouraged concerning the current practices of this once-respected journal.

On 2016 Jan 27, Valentine Njike commented:

The reviewer is right to suggest this point it getting excessive attention. It was a secondary outcome measure, and a significant within-group change only, when primary measures were between-group changes. This is indicated explicitly in the paper...

On 2016 Apr 07, Stephen L. Black commented:

Update: BMJ has now belatedly but effectively responded to my complaint, and a version of my original comment has been posted on the Open Diabetes website as I had initially intended.