Abstract: Results summary tables are one of the most important outputs of a clinical data science project, but statistical summary tables can be very time consuming and tedious to generate, format, and manage for a large clinical research project, since they require composing many heterogeneous components and footnotes, and may need to be generated many times or for many different versions of a dataset. We have streamlined this process using a combination of tools for both R and Python and a Jupyter/Sagemaker-to-Google Sheets integration that automates our table generation pipeline, saves us time, reduces risk of human error, improves reproducibility, and allows us to share the results with collaborators in their tools of choice.
Bio: Katie Shakman is a Senior Data Scientist at Lyra Health, specializing in clinical research into Lyra’s best-in-class employee mental health programs. She earned her PhD in neuroscience from Columbia University in 2018, and prior to Lyra she worked on health outcomes research related to wearable devices as well as assisting data science and machine learning teams with tools and processes that enable more efficient data science workflows. She brings her enthusiasm for reproducible science, efficient and easy to use data science pipelines, and quality mental healthcare to her work at Lyra. More on her team’s work can be found at: https://www.lyrahealth.com/lyra-clinical-research/