FundingThe US National Institutes of Health (NIH) supported the research and article publication.

AbbreviationsCI, confidence interval; FDA, Food and Drug Administration; ITT, intention to treat; LOCF, last observation carried forward; NDA, New Drug Application; NME, New Molecular Entity; NNT, number need to treat; OR, odds ratio.

CitationReporting bias in drug trials submitted to the Food and Drug Administration: A review of publication and presentation. Rising K, Bacchetti P, Bero L. PLoS Medicine 5(10): e217. doi: 10.1371 / journal.pmed.0050217
Download PDFAbstract

Background. Previous studies of drug trials submitted to regulatory authorities have documented selective reporting of both entire trials and favorable results. The objective of this study is to determine the publication rate of efficacy trials submitted to the Food and Drug Administration (FDA) in approved New Drug Applications (NDAs) and to compare the trial characteristics as reported by the FDA with those reported in publications.

Methods and Findings. This is an observational study of all efficacy trials found in approved NDAs for New Molecular Entities (NMEs) from 2001 to 2002 inclusive and all published clinical trials corresponding to the trials within the NDAs. For each trial included in the NDA, we assessed its publication status, primary outcome(s) reported and their statistical significance, and conclusions. Seventy-eight percent (128/164) of efficacy trials contained in FDA reviews of NDAs were published. In a multivariate model, trials with favorable primary outcomes (OR = 4.7, 95% confidence interval [CI] 1.33–17.1, p = 0.018) and active controls (OR = 3.4, 95% CI 1.02–11.2, p = 0.047) were more likely to be published. Forty-one primary outcomes from the NDAs were omitted from the papers. Papers included 155 outcomes that were in the NDAs, 15 additional outcomes that favored the test drug, and two other neutral or unknown additional outcomes. Excluding outcomes with unknown significance, there were 43 outcomes in the NDAs that did not favor the NDA drug. Of these, 20 (47%) were not included in the papers. The statistical significance of five of the remaining 23 outcomes (22%) changed between the NDA and the paper, with four changing to favor the test drug in the paper (p = 0.38). Excluding unknowns, 99 conclusions were provided in both NDAs and papers, nine conclusions (9%) changed from the FDA review of the NDA to the paper, and all nine did so to favor the test drug (100%, 95% CI 72%–100%, p = 0.0039).

Conclusions. Many trials were still not published 5 y after FDA approval. Discrepancies between the trial information reviewed by the FDA and information found in published trials tended to lead to more favorable presentations of the NDA drugs in the publications. Thus, the information that is readily available in the scientific literature to health care professionals is incomplete and potentially biased.

(PLoS) Editors' SummaryBackground. All health-care professionals want their patients to have the best available clinical care—but how can they identify the optimum drug or intervention? In the past, clinicians used their own experience or advice from colleagues to make treatment decisions. Nowadays, they rely on evidence-based medicine—the systematic review and appraisal of clinical research findings. So, for example, before a new drug is approved for the treatment of a specific disease in the United States and becomes available for doctors to prescribe, the drug's sponsors (usually a pharmaceutical company) must submit a “New Drug Application” (NDA) to the US Food and Drug Administration (FDA). The NDA tells the story of the drug's development from laboratory and animal studies through to clinical trials, including “efficacy” trials in which the efficacy and safety of the new drug and of a standard drug for the disease are compared by giving groups of patients the different drugs and measuring several key (primary) “outcomes.” FDA reviewers use this evidence to decide whether to approve a drug.

Why Was This Study Done? Although the information in NDAs is publicly available, clinicians and patients usually learn about new drugs from articles published in medical journals after drug approval. Unfortunately, drug sponsors sometimes publish the results only of the trials in which their drug performed well and in which statistical analyses indicate that the drug's improved performance was a real effect rather than a lucky coincidence. Trials in which a drug did not show a “statistically significant benefit” or where the drug was found to have unwanted side effects often remain unpublished. This “publication bias” means that the scientific literature can contain an inaccurate picture of a drug's efficacy and safety relative to other therapies. This may lead to clinicians preferentially prescribing newer, more expensive drugs that are not necessarily better than older drugs. In this study, the researchers test the hypothesis that not all the trial results in NDAs are published in medical journals. They also investigate whether there are any discrepancies between the trial data included in NDAs and in published articles.

What Did the Researchers Do and Find? The researchers identified all the efficacy trials included in NDAs for totally new drugs that were approved by the FDA in 2001 and 2002 and searched the scientific literature for publications between July 2006 and June 2007 relating to these trials. Only three-quarters of the efficacy trials in the NDAs were published; trials with favorable outcomes were nearly five times as likely to be published as those without favorable outcomes. Although 155 primary outcomes were in both the papers and the NDAs, 41 outcomes were only in the NDAs. Conversely, 17 outcomes were only in the papers; 15 of these favored the test drug. Of the 43 primary outcomes reported in the NDAs that showed no statistically significant benefit for the test drug, only half were included in the papers; for five of the reported primary outcomes, the statistical significance differed between the NDA and the paper and generally favored the test drug in the papers. Finally, nine out of 99 conclusions differed between the NDAs and the papers; each time, the published conclusion favored the test drug.

What Do These Findings Mean? These findings indicate that the results of many trials of new drugs are not published 5 years after FDA approval of the drug. Furthermore, unexplained discrepancies between the data and conclusions in NDAs and in medical journals are common and tend to paint a more favorable picture of the new drug in the scientific literature than in the NDAs. Overall, these findings suggest that the information on the efficacy of new drugs that is readily available to clinicians and patients through the published scientific literature is incomplete and potentially biased. The recent introduction in the US and elsewhere of mandatory registration of all clinical trials before they start and of mandatory publication in trial registers of the full results of all the predefined primary outcomes should reduce publication bias over the next few years and should allow clinicians and patients to make fully informed treatment decisions.