Jeffrey M Drazen’s fascinating keynote speech underlined the importance of accurate, unbiased, scientific data in the progress of health science, specifically with respect to the causes and management of diseases.
Over the course of human history, theories concerning the transmission of contagious diseases have ranged, depending on the prevailing wisdom of the day, from religious to scientific. However, it has been the scientific analysis of the causative factors involved in contagion that has led to successful identification and treatment. For example, in the case of the highly contagious and lethal disease tuberculosis, before it was possible for information regarding common symptoms (such as coughing and fever) to be shared among health practitioners and before new technologies—such as microscopy, staining, and x-rays—had been developed, all that medical practitioners could do was try to treat for the symptoms and hope for the best. Thousands of people died of tuberculosis every year; a popular cure was to send those showing blatant signs of the disease to a sanitarium, far away from cities, where they could find “clean air”. That “treatment” did not guarantee a cure; however, assembling groups of patients made it possible for researching physicians to try their own “cures” on them (which Drazen referred to as “the early days of drug trials”). In the 1920s, great claims were made for the efficacy of of a gold-based compound, and data were provided to support its use. However, the data were flawed, and unsound conclusions were based on selection of patients in a nonrandom manner, which introduced bias. A later, thorough, randomized investigation of the compound revealed that it was not a cure at all; in fact, it increased the likelihood of death! That revelation focused the scientific community on the importance of eliminating bias from such studies and served as a warning about accepting results at face value.
Much research has since been dedicated to the development of thorough and unbiased methods for evaluating the efficacy of drugs and other treatments. Although the randomized controlled trial itself did not come of age until the 1960s, its importance for public-health management and disease control is widely recognized.
High-quality data are needed to make informed decisions about health, disease, and treatment. The desire to perfect statistical test methods and statistical analysis has resulted in many groups’ dedication to the cause, and the last 50 years have seen marked development of techniques for removing bias. Such reports as the Greenburg Commission report in 1967 and such publications as Data Monitoring in Clinical Trials by DeMets and colleagues have contributed to the improvement of research results.
However, there are still numerous examples of how poor study design, changes in protocols, and misinformation result in faulty drug evaluations. In 2004, given that study design and protocol are of such monumental importance, 13 scientific journals refused to publish papers for which test protocols had not been registered in a central National Institutes of Health database. There was initially substantial push-back from the pharmaceutical companies, but compliance has been increasing since 2005.
The future of scientific breakthroughs depends on the application of sound test methods, the open discussion and analysis of results among the community, and the constant improvement of the technology that we use to disseminate our data. With respect to public health, only the most ethical protocols can be considered, and careful oversight is required. Accurate and clear communication of results and discoveries is key, and the community needs to be dedicated to this goal.