Skip Navigation

Frequently Asked Questions


How long ago did the project begin? Who was the founder? What was the original idea of the project?

The original project was started in 1998 by William Herman, MD. His main objective was to create a model for the development and complications associated with diabetes. Deanna Isaman started her project at the University of Michigan in April of 2006. Her research team's main goal is to take Herman's original idea of modeling diabetes, to create software that will be able to model many different types of chronic diseases, and to develop better ways to estimate transition rates between stages of disease.

What is the long term goal for the project?

The overall goal of this project is to create software that can model many different chronic diseases. Then doctors could give patients predictions of what complications they might get from a disease. This could help doctors treat patients earlier, which could potentially lower long term costs of medication and other treatments.

Who will benefit or what applications can this research be used for?

Other researchers will benefit from this project because they will not have to recreate every disease model if software is developed that will be general enough to handle many types of diseases. Also, better statistics are now being used to estimate parameters in the models, so other researchers will have better statistical tools than those that are currently available. The public will also benefit from this research project. As researchers better understand the progression of chronic diseases, better treatments can be designed to reduce the risk for patients with the disease. This will improve the quality of life for many people in America.

What has been the greatest accomplishment of the project so far?

The biggest accomplishment of the project so far is the new statistics. Because more studies are being analyzed, and more complicated statistical analysis can be done, the statistics used to help simulate the models are much more accurate. We call our statistical estimation approach the Lemonade method.
Also, new computer framework was developed and released. This computer software is able to accommodate different disease processes, beyond diabetes alone.

What is the Lemonade Method?

The Lemonade Method is a statistical method that allows combining summary information from multiple studies to estimate parameters of a given model. The model is generally a Markov model representing states and transitions probabilities between these states with some extensions and the studies can be found in the literature and typically available publicly. This approach allows creating a model of a disease process by using secondary data and provides an alternative to basing a model on a single study. The method is implemented as software available in this web site in the software section.

What has been the greatest obstacle or setback?

Engineering software that is general enough, and that has the capabilities of accounting for many different situations, has proved to be extremely difficult. This involved dealing with complex mathematical, numerical and programming issues.

Where can the Michigan Model for Diabetes be found?

Recent development of the Michigan Model for Diabetes can be found at the model section of this web site. It is possible to download it as a data file that works with our software.

Can I use the modeling software?

Yes, the software is available for download under GPL license. You can find it in the software section in this web site.