This is a well-known problem in clinical trials: researchers said at the beginning that they would look for a specific result, such as a heart attack, but when they released the result, they would report something else.
This will make drugs or treatments seem safer or more effective than they actually are.
Now, a study that systematically investigates whether large journals are keeping their promises to ensure that the results are reported correctly has found that this effort seems to have failed, and many journals and authors are looking for various excuses.
When journals and researchers were asked to correct research, their answers were both interesting and worrying.
” Editors and researchers always misunderstand what a correct test report looks like.” Ben Goldacre, head of the project, said. Goldacre is a writer and doctor at the university of oxford in the uk and a transparent supporter of drug research.
Since four years ago, the Evidence – based Medicine Results Monitoring Project Center ( COMPare ) has conducted six-week research on all papers published in five journals: the Annals of Internal Medicine ( AIM ), the British Journal of Medicine ( BMJ ), the Journal of the American Medical Association ( JAMA ), the Lancet and the New England Journal of Medicine ( NEJM ).
This study covers a wide range from the health effects of drinking in diabetic patients to the comparison of two renal cancer drugs.
All five journals have recognized the Uniform Standard for Clinical Trial Reports ( CONSORT ).
One of CONSORT’s rules is that authors should describe the results of their planned research before the test starts and stick to these goals when publishing the test.
However, out of 67 tests published in 5 journals, only 9 reported correct results.
The Comparative team recently reported in Trials that 1 / 4 of the people did not correctly report the initial planned results of the study, 45% did not correctly report all secondary results, and others added new results.
This varies from journal to journal: only 44% of the trials in the Annals of Internal Medicine correctly reported the main results, while 96% of the trials in the New England Journal of Medicine were correct.
When the COMPare team published these problematic papers to the journal, only 23 of the 58 were published.
The Annals of Internal Medicine and the British Medical Journal both published, The Lancet accepted 80%, The New England Journal of Medicine and the Journal of the American Medical Association both rejected.
The editors of the New England Journal of Medicine explained that their editors and peer reviewers would decide which results to report. They said that although some of CONSORT’s rules are ” useful”, the author does not need to follow them all.
The Comparative research team also found that when the authors of the experiment responded to these letters, their comments were full of ” inaccurate or questionable statements and misunderstandings”.
Researchers said that like editors, many authors misunderstood CONSORT rules and the role of public registries in sharing pilot programs.
Some people attacked the COMPare project as a ” layman” of the research group. Others shrugged off the criticism and instead complained about how difficult their work was, while others denied that they had missed any results.
The COMPare team wrote that the purpose of their research is to inspire journals to better implement CONSORT and to re-examine their criteria for publishing papers and newsletters.
” We hope that editors will respond positively, constructively and thoughtfully to our findings.”