Just finished looking at Matt Hagen’s 2014 “Biological and clinical data integration and its applications in healthcare.” PhD dissertation. This is a great piece of work … You can find it here.
While its around 5 years old, the insights and discussion are excellent. I like the detailed breakdown of how different ontologies and vocabularies align (and how things fall through the cracks). I liked the discussion of using Neo4j to analyze relationships and simplify searches and relationship mappings.
Particularly liked the discussion of using ontologies. to” facilitate improved prioritization of intensive care admissions and accurate clustering of multimorbidity conditions”. THIS IS BIG! with enormous potential.
Discussion of his BioSPIDA relational database translator and its contrast with the separate Entrez Gene, Pubmed, CDD, Refseq, MMDB, and Biosystems NCBI databases.
His Table 7.2: Descriptions of patient clusters is rather illuminating, as his discussion and analysis of ICU Electronic Health Records and findings associated with morbidity outcomes.
For example Cluster 1 contains the following Most Prevalent Conditions: Coronary arteriosclerosis, Hypercholesterolemia, Diabetes, Gastroesophageal reflux disease, Atrial fibrillation, Hyperlipidemia, Tobacco dependence. Which led to the following Most Prevalent Procedures: Catheterization of left heart, Cardiopulmonary bypass operation, Angiocardiography of left heart,.
I am surprised this work is not cited as much as it should be!. IMHO, this work definitely should be used as blueprint for additional investigations.