The Reference Model describes chronic disease complications in a population.
It is:
• An ensemble model assembling multiple other models
• A league of disease models that compete and cooperate
• A validation model
• A medical knowledge accumulator

The Reference Model is now:
The Most Validated Cardiovascular (CVD) Diabetes Model known
United States Patent 9,858,390

The Reference Model can now:
• Determine CVD models that significantly behave better on several diabetic populations
• Deduce that CVD probability halves every 5 years due to medicine improving - according to information from the last 3 decades
• Calculate life tables for diabetics
• Interface with ClinicalTrials.Gov
• Create an interactive map of our cumulative computational knowledge gap

The interactive plot below shows our cumulative knowledge gap by showing the error in the vertical axis for clinical trial cohorts listed on the horizontal axis. Circles at the bottom have a better fit between trial results and model results. Hover over the circles to see additional details about the cohort. The slider determines the model optimization iteration, the best model is in the last iteration. User can explore the map by changing size and color attributes.


The Reference Model is a good way to cross reference information to find out pieces of information and assumptions that fit together, and allow competition against accumulated known data to guide our perception. High Performance Computing is a key to those capabilities and it provided using capabilities of the MIcro Simulation Tool (MIST) .

MIST also provides advance population generation techniques using Evolutionary computation. The Reference Model uses publicly available data such as clinical trial publications. This allows it to access more information since it allows accessing data that otherwise will be restricted from sharing. The Reference Model has an interface that allows it to read information from ClinicalTrials.Gov while maintaining tractability and reproducibility.

The Reference Model was created in 2012 and evolved since then. You can find key developments and publications by year in the news section .

Here are some videos describing the Model:

This video explains the model in a larger context as presented in AnacondaCon 2019:

This video summarizes a decade of work as presented in PyTexas 2017:

This describes the evolution of the model up to 2016 presented in PyTexas:

This describes the work presented in PyData in 2014:

Related Projects

The project owner recommends the following other projects:

MIcro Simulation Tool - MIST

Population Modeling Workgroup