Identifying avoidable waste in research at different research layers : analysis of methods in evidence synthesis, editorial policies for transparency, and definitions for predatory journals

Abstract

Transparency and reproducibility practices in research and science are a cornerstone for its progress. Adequate editorial and peer-review processes are adequately available, although authors sometimes do not adhere to them. Available resources and routines may contribute to the improvement of the aforementioned transparency in science. So, this thesis aimed to investigate potential problems and solutions in the biomedical research in different levels, as follows: the quality, transparency and apparently excessive redudancy of systematic reviews and meta-analysis (SRMAs) in high-intensity interval training (HIIT) on cardiorrespiratory fitness; how journals of Cardiology currently adhere to transparency and reproducibility practices in their policies; and how to handle the burden of predatory journals through evaluating available checklists to identify them. In summary, in the study 1 we concluded that SRMAs in HIIT have a low methodological quality and transparency practices, such as the very low proportion of registration records, modest comprehensive literature searches and limited assessment of probability of publication bias. There was also some redundancy and heterogeneity among the includedSRMAs in terms of population, intervention settings and sometimes the absence of a comparator group. In the study 2, the main message is that Cardiology journals adhere at low proportion to transparency and reproducibility practices in their policies, such the a moderate proportion on how to report the study or how to find a guideline (e.g., EQUATOR Network); the low proportion of raw data sharing statements the guidance on how to share the data. Importantly, only one journal has a mandatory policy for data sharing of randomized controlled trials. Finally, in the study 3, we found 93 unique checklists to identify potential predatory journals, which may difficult the choice of most useful resources, and that there is a low proportion of evidence-based checklists.