The processes and benefits of sharing clinical data

The Respiratory Rate Estimation project aims to assess methods for automated respiratory rate (RR) monitoring of hospital patients. It consists of a series of studies of different algorithms for RR estimation from clinical data, complimented by the provision of publicly available datasets and resources.

Ethical procedures were followed to provide accountability for data collection, including obtaining approval from a Research Ethics Committee, written informed consent from each participant, and public dissemination of the protocol a priori.

Data were anonymised within the NHS to ensure confidentiality, before transfer to a university. The dataset was cleaned and formatted intuitively to allow external researchers to easily re-use it.A toolbox of algorithms was constructed to allow a systematic comparison. The dataset, toolbox, analysis code and results are publicly available. These resources facilitate transparency, reproducibility, and ongoing peer review of the research. Other publicly available datasets are now compatible with the toolbox, providing foundations for large-scale analyses across multiple datasets.

The accessibility of the dataset and resources will soon be widened by publication of an educational tutorial aimed at clinicians and data scientists, with exemplary analyses and code. Patients, students and researchers could all benefit from the sharing of this dataset.

Further details: http://peterhcharlton.github.io/RRest/