Mapping sepsis of oral origin

Objectives: Sepsis of oral origin is defined as life threatening infection originating from the oral cavity leading to severe organ damage, dysfunction, and death. For sepsis of oral origin, little is known about the model of its occurrence, treatment, and outcomes. The objective of this study is to vest a graph database with data to predict clinical patterns and outcomes of sepsis of oral origin for mapping clinical practice guidelines. Graph databases are faster and more efficient than relational databased in storing, updating, and retrieving data for decision-making.
Methods: A total of 22 papers including 8 case studies were obtained using PubMed and Google scholar. The process to map patterns and outcomes used Neo4j graph database principles. A traversal graph of patterns and outcomes updated Subject Matter Expert knowledge level of node to node relationships with edge-probabilities reported in the literature to each reported outcome terminus. Data points were subjected to a meta-analysis (p=.05 significance level) when multiple data entries were reported. Averages or weighted average calculations were performed (Shapiro-Wilks, p=.05) when data did not meet the requirements for meta-analysis.
Results: As an example of tracking stored data, we have shown that the probability of an infection of oral origin becoming septic (SIRS diagnosis known or unknown) was 36.6%, not septic was 63.4%. If septic, the nidus of infection originated from oral origin 72.5% of the time, from non-oral tissue 26.7% of the time. If of oral origin nidus, 66.5% originated from soft tissue, 33.5% from hard tissue. Of the hard tissue cases, 60% had survival outcomes. Of the soft tissue cases, 60% had survival outcomes.
Conclusions: Mapping of clinical patterns and outcomes using graph database modeling produced the ability to map and store data on sepsis of oral origin and quantitate patient-centered outcomes. With this model, we will be able to increase predictive knowledge for decision-making.