Research Waste

Research waste refers to the expenditure of scientific resources on studies that fail to produce usable, accessible, or relevant knowledge due to problems in design, conduct, analysis, reporting, or dissemination. Chalmers and Glasziou estimated in 2009 that about 85% of the investment in biomedical research is avoidably wasted, a figure that has catalyzed reform efforts across the research enterprise.

Research waste represents a massive misallocation of human and financial resources. Funding agencies, taxpayers, charitable donors, and research participants invest in research with the expectation that it will advance knowledge and improve health outcomes. When studies are poorly designed, redundant, selectively reported, or inaccessible behind paywalls, that investment is lost. Research waste also has ethical implications: participants who consent to research have a reasonable expectation that their contribution will be used to advance science. When studies produce unusable results, that trust is violated. Reducing research waste has become a priority for funding agencies, journal editors, and research institutions worldwide.

Research waste arises at multiple stages of the research lifecycle. At the question stage, research may address questions that are not relevant to patients, clinicians, or policymakers, or that have already been adequately answered by existing studies. Failure to conduct a systematic review before designing a new study is a major contributor to redundant research.

At the design stage, inadequate study design, insufficient sample size, and inappropriate outcome measures produce studies that cannot answer their own questions. Underpowered studies are a particularly common source of waste, as they often produce inconclusive results that neither confirm nor refute the hypothesis, and the statistically significant results they do produce tend to overestimate the effect.

At the conduct stage, protocol deviations, poor data quality, and high attrition rates reduce the usable information a study produces.

At the reporting stage, selective reporting of outcomes, publication bias, and inadequate description of methods prevent the scientific community from using or replicating the findings. Studies whose results never become available, in a journal, a registry or elsewhere, are the most complete form of waste.

At the dissemination stage, findings hidden behind paywalls, written in inaccessible jargon, or not translated into clinical practice guidelines fail to reach the people who could use them.

Losses compound from stage to stage. Chalmers and Glasziou's estimate combined losses of about half at each of three stages: if about half of studies are never published, about half of published reports are too poorly reported to be usable, and about half of the rest have serious, avoidable design flaws, only about one study in eight (12.5%) passes all three, so about 87.5% of the investment is wasted, which they rounded down to 85%.

Strategies for reducing research waste include requiring systematic reviews before new studies are funded, mandating sample size justifications, enforcing trial registration and results reporting, promoting open access publication, and involving patients and stakeholders in research priority setting.

Consider research waste implications at every stage of the research process. Before starting a new study, conduct a systematic review to check whether the question has already been answered. During design, ensure adequate power and appropriate methods. During reporting, follow reporting guidelines and publish all results regardless of direction. When evaluating the research landscape in a field, assess the proportion of studies that produced usable evidence and identify the stages where waste is greatest.

A common error is launching a new study without reviewing the existing evidence, producing redundant research that adds little to the evidence base. Another mistake is designing underpowered studies that are unlikely to produce interpretable results. Researchers sometimes complete studies but fail to publish the results, either because the findings are negative or because other priorities intervene. Finally, publishing results in a manner that prevents replication (such as omitting methodological details or refusing to share data) limits the value of the research.

A systematic review of registered trials in a therapeutic area finds that of 200 registered trials, 80 (40%) have not published results five years after completion. Of the 120 published trials, 30 (25%) changed their primary outcome between registration and publication. Of the remaining 90, 22 (24%) were underpowered for their stated primary outcome. Only 68 of 200 registered trials (34%) pass all three checks (results published, primary outcome unchanged, and adequate power for it), and even these have not been checked for other problems such as poor reporting or high risk of bias, illustrating the scale of research waste in the field.