Building a health outcomes database & Frontend Visualuzation
Sage Arbor Ph.D
Medicine and Health Sciences
This project created the first website capable of searching published biomedical knowledge and returning a rank ordered list of actions most effective at improving health. These actions have font scaled based on the odds ratios (O.R.), which is a statistical method to measure the relation between an incident occurring (as in a disease) and the variable of interest (the input/action). This website was designed with the intent of making data more easily consumable for the average user concerned with health outcomes. The database PubMed has over 1 million new articles added annually, and a single human can not read and/or compare the importance of these disparate outcomes. As part of the website, a user can enter information from published peer reviewed studies which will store the information in a SQL database. The data is then searchable via the search box. Data can be input into or searched from the database in lay terms; however, the information is codified using synonyms of Medical Subject Heading (MeSH) terms. Search capability allows for lifestyle input, interventions, or outcomes desired with the results visually scaled based on PubMed odds ratio effectiveness. This is the first time the magnitude of scientific importance, as published in PubMed, is translated in a fashion to draw people’s cognitive attention scaled to that O.R. importance. Future work may include providing feedback (e.g. educational credit) to incentivize submissions, or natural language processing (NLP) to allow real time automated incorporation of newly deposited PubMed articles.
Copyright 2019 all authors
Larson, Ashley OMS-2; Breeden, Scott; and Arbor, Sage Ph.D, "Building a health outcomes database & Frontend Visualuzation" (2019). MU-COM Research Day. 119.