Validate the integrity of research data on COVID 19

Research integrity is one of the primary pillars of scientific research. Scientific research is the foundation for nearly all aspects of modern life, including public health, agriculture, education, technology, industry, and national security. Research is a multi-step endeavor, with each step requiring ethical vigilance. An alteration of any step could invalidate the results. Researchers must be rigorous and careful in designing and executing experiments and honest and transparent when reporting data, methods, and results.

Every now and then, scandals related to a lack of integrity in research arise. These scandals can have a significant impact on science and society when they involve the testing of drugs, biologics, or medical devices on human research subjects, since they may adversely affect the health of millions of people and waste billions of dollars in public or private investment of research. Scandals that have gained considerable attention in recent years have involved problems with data integrity (e.g. data fabrication or falsification, data suppression, or irreproducible results), or problems related to scientific integrity more generally, such as harassment, undisclosed conflicts of interest, or significant violations of human or animal research regulations.

Data auditing can play an essential role in preventing ethical problems related to data integrity. Data auditing involves carefully examining the research data to ensure that the output of research (e.g. the final data reported in a publication, results, figures, and diagrams) corresponds to the original data upon which it is based. To do this, one can examine all the data; or, if that is not practical, a statistically valid sample of the data. Data auditing can detect fabrication, falsification, errors, and inconsistencies. Data auditing takes around a week or two, depending on the size of the project. The cost of auditing research data has been estimated to be about 1% of the cost of the research project.

The primary goal of validation is to examine “the degree of correspondence between assertions of published data, with the original data (e. g, raw data). Validation is based upon comparing a statistically valid sample of the data to the final data.

Two papers that were recently retracted from The Lancet and the New England Journal of Medicine (NEJM) illustrate the potential role of auditing in preventing problems related to data integrity. The papers Lancet paper (Mehra et al. 2020) reported that the antimalarial drug hydroxychloroquine poses significant risks to patients with COVID-19 that outweigh its potential benefits, and the NEJM paper (Mehra et al. 2020a) reported that the drug does not prevent COVID-19 infection. The papers were retracted when concerns were raised about the validity of the data and the company that provided the data for the studies, Surgisphere, refused to allow auditors to have access to the original data, which consisted of electronic health records. While the authors have not been accused of fabricating or falsifying data, serious questions about the validity of data have not been answered and may not been until there is a thorough audit of the data. Data auditing would have also helped to validate the results prior to publication.

In the long run, science tends to be self-correcting, because fraudulent or invalid research may be retracted, and mistaken hypotheses or theories may be refuted by new data or experiments. However, this self-correcting process may take years or even decades, and the public often cannot wait for science to produce the correct solution to a problem, especially during a health crisis like the COVID-19 pandemic.

During this pandemic, rigor and auditing are important but so too, is access to data, openness, and transparency. The validation of research output will significantly enhance public trust in our research community.