The purpose of the this repository is to provide a method for determining the trajectory of Knee Surgery Outcomes for patients. The repository also contains a shiny-application for providers to utilize in practice.
The data is coming from patient information recorded in the UC Health system (REDCap).
/code folder contains the analysis code. The main prediction method is using the R package brokenstick, along with predictive mean matching and gamlss. Currently the code is under development to work within the caret and mlr packages.
For compiling gthe Scientific Reports .tex file (using the .Rnw extension with knitr()), I had trouble with one citation that had \textregister{} symbol in the bib file which causeed problems for bibtex compiling a .bbl file... remove the symbol and it works!
A HIPAA compliant web-application is currently under development.