One step closer to better Electronic Health Records data
Real-World Data (RWD) like Electronic Health Records (EHR) is crucial for understanding drug usage and various treatments and generating Real-World Evidence (RWE). Risk prediction has been a major application where EHR is used, and there is now a shift towards causal inference, which requires data of even higher quality. Patients undergo treatments (drugs, procedures) at various times during their hospital stays, yet the data being recorded are messy and error-prone for various reasons. Analysts spend significant amount of time to sit together with clinicians to identify and understand abnormal records, and unfortunately this process is challenging to automate.
This talk will use an example on antibiotics prescription and use at a Nordic hospital to illustrate how some EHR systems can improve for better clinical decision-making and better data for research. I will also introduce a pilot R package (ggehr) that facilitates visual exploration of EHR data, and how it can help reconstruct patient journeys and enable analysts to perform effective quality control.
