Audit identifies the most read BMJ research papers

BMJ 2007; 334 doi: (Published 15 March 2007) Cite this as: BMJ 2007;334:554
  1. Susan Mayor
  1. London

    Research studies on the side effects of commonly prescribed drugs constituted the three most read papers published by the BMJ in 2005, according to an internal audit that assessed their use by readers.

    Studies that showed increased risk of myocardial infarction in patients taking cyclo-oxygenase-2 (COX 2) inhibitors and that explored the link between selective serotonin reuptake inhibitors and suicide came top. For the audit Sara Schroter, the BMJ's senior researcher, analysed research papers using three measures: the number of citations a paper received, the number of times it was accessed on the web, and the number of rapid responses it generated.

    The top scoring paper was a case-control study that showed greater risk of myocardial infarction in patients taking the COX 2 inhibitor rofecoxib (Vioxx) and in patients taking diclofenac and ibuprofen (2005;330:1366-9, doi: 10.1136/bmj.330.7504.1366).

    The online version of this paper was accessed 42 505 times in the first year after publication. It received 28 rapid responses and was cited 97 times in 2005 and 2006.

    At the time, there was widespread interest in the cardiovascular safety of COX 2 inhibitors after the withdrawal of Vioxx and the publication of several papers on the risk associated with these agents.

    A meta-analysis of drug company data that showed no evidence that selective serotonin reuptake inhibitors increase the risk of suicide but found weak evidence of increased risk of self harm was ranked second (2005;330:385-10, doi: 10.1136/bmj.330.7488.385). This was accessed almost 24 000 times and was cited 78 times in 2005 and 2006.

    In third place was a systematic review that showed an association between suicide attempts and the use of selective serotonin reuptake inhibitors and that highlighted limitations in the reporting of suicides in clinical trials (330:396-403, doi: 10.1136/bmj.330.7488.396).

    Trish Groves, deputy and senior research editor at the BMJ, said that the findings confirmed that the editorial policy of the journal fitted with what doctors want to read. “The BMJ gives priority to articles that will help doctors make better decisions—whether those doctors are practising clinical medicine, working in public health, developing and implementing health policy, or working mostly as researchers.

    “Most of the papers in the top 10 provided important new evidence on highly controversial questions that mattered then to patients—and still do.” She noted that several were about the safety and effectiveness of widely used treatments for common conditions.

    Three of the 10 most highly ranked studies were about treatments for depression, and two of these gave important evidence on the associations between selective serotonin reuptake inhibitors and self harm.

    Two other big questions alarming patients and the public during 2005 were the safety of COX 2 inhibitors and non-steroidal anti-inflammatory drugs, and the effectiveness of cholinesterase inhibitors in Alzheimer's disease. “Both debates were advanced by these highly read and cited BMJ research papers,” Dr Groves said.

    Other studies in the top 10 addressed practical issues, including a trial of calcium and cholecalciferol for prevention of fractures (2005;330:1003-6, doi: 10.1136/bmj.330.7498.1003), a prospective study of cannabis use and psychotic symptoms in young people (2005;330:11-4, doi: 10.1136/bmj.38267.664086.63), and a review of how well B-type natriuretic peptide predicts death and cardiac events in patients with heart failure (2005;330:625-7, doi: 10.1136/bmj.330.7492.625).

    Three of the top 10 papers were studies done in primary care, including the top scoring paper on the risk of myocardial infarction with COX 2 inhibitors and other non-steroidal anti-inflammatory drugs. Only a couple of the papers were randomised controlled trials.

    Dr Groves commented, “Randomised controlled trials can give robust and highly original evidence on treatments and other interventions, but they give only limited information, often about highly selected patients over a relatively short time. Eight of the top 10 studies provided a broader kind of evidence: by systematically reviewing and summarising the best data out there, or analysing what really happened to large numbers of people in real life over time.”


    • Competing interests: SM writes regularly for the BMJ and is paid for her contributions.

    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to to receive unlimited access to all content on for 14 days.
    Sign up for a free trial