Medicine Graph - sporedata/researchdesigneR GitHub Wiki

General description

Electronic Health Records (EHR) are a valuable yet underutilized resource for understanding the actual dynamics of clinical practice and for measuring how closely related various medical elements such as medications, illnesses, procedures, and instruments are.

The Medicine Graph is an extensive dataset that offers a method to quantitatively analyze patterns of comorbidity, drug interactions, and medication-disease relationships, which can be useful for various clinical, epidemiological, and financial purposes. It is a novel collection of co-occurrence matrices that measure the frequency of pairings between 3 million terms linked to 1 million clinical concepts. These measurements are derived from the unprocessed text of 20 million clinical documents, covering 19 years of data.

Clinical areas of interest

• Diagnosis
• Data Mining
• Epidemiology
• Outcomes research

Related publications

• Building the graph of medicine from millions of clinical narratives

Data access

More information about Medicine Graph can be found at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4322575/pdf/sdata201432.pdf

To access Medicine Graph data, visit https://datadryad.org/stash/dataset/doi:10.5061/dryad.jp917