Evaluation of indicators supporting reproducibility and transparency within cardiology literature

Abstract

OBJECTIVES: It has been suggested that biomedical research is facing a reproducibility issue, yet the extent of reproducible research within the cardiology literature remains unclear. Thus, our main objective was to assess the quality of research published in cardiology journals by assessing for the presence of eight indicators of reproducibility and transparency. METHODS: Using a cross-sectional study design, we conducted an advanced search of the National Library of Medicine catalogue for publications in cardiology journals. We included publications published between 1 January 2014 and 31 December 2019. After the initial list of eligible cardiology publications was generated, we searched for full-text PDF versions using Open Access, Google Scholar and PubMed. Using a pilot-tested Google Form, a random sample of 532 publications were assessed for the presence of eight indicators of reproducibility and transparency. RESULTS: A total of 232 eligible publications were included in our final analysis. The majority of publications (224/232, 96.6%) did not provide access to complete and unmodified data sets, all 229/232 (98.7%) failed to provide step-by-step analysis scripts and 228/232 (98.3%) did not provide access to complete study protocols. CONCLUSIONS: The presentation of studies published in cardiology journals would make reproducing study outcomes challenging, at best. Solutions to increase the reproducibility and transparency of publications in cardiology journals is needed. Moving forward, addressing inadequate sharing of materials, raw data and key methodological details might help to better the landscape of reproducible research within the field.

Ähnliche Arbeiten