xiaodi hu

Graduate Trainee, Massachusetts Institute of Technology

1 active project

First Workspace on Diabetes

It’s estimated 5 million of the 18 million americans in the US with diabetes do not know they have it. Early diagnosis of diabetes and pre-diabetes is important so that patients can begin to manage the disease early and potentially…

Scientific Questions Being Studied

It’s estimated 5 million of the 18 million americans in the US with diabetes do not know they have it. Early diagnosis of diabetes and pre-diabetes is important so that patients can begin to manage the disease early and potentially prevent or delay the serious disease complications that can decrease quality of life. Some of these complications include premature heart disease and stroke, blindness, limb amputations, and kidney failure. Our project will review clinical data to estimate each patient’s diabetes risk, which will help identify high risk individuals who may need to seek medical attention.

Project Purpose(s)

  • Educational

Scientific Approaches

Given the importance of being able to correctly give early diagnosis, this team places a stronger importance for our model on predictive performance, which leads the team to initially lean towards using random forest or boosting (particularly XGBoost) models. These more complex models might not be as interpretable, and take longer to run, but they generally yield a much higher out-of-sample performance, which is the main goal of this team’s project. The team will also look into how well a CART model performs with the dataset given that a CART tree, if it is not too complex, would offer that interpretability for the end users of this research project. It is important to note that a CART model would be used in lieu of a more complex model if it performs on par with the more complex models.

Anticipated Findings

The overall goal of the project is to assist doctors with treating patients. Ideally, the data analysis would provide doctors with a model of what age range a patient is likely to be diagnosed with diabetes based on his/her risk factors. If we can provide doctors with this information, then they can have a conversation with patients about preventative care, such as lifestyle changes, in advance of that age to try to prevent patients from developing diabetes. It also allows doctors to know when to start testing patients’ A1C levels to diagnose diabetes so the disease can be treated as quickly as possible. This prevents the untreated disease from wreaking havoc on the patients’ bodies, preventing complications from diabetes down the line.

Demographic Categories of Interest

This study will not center on underrepresented populations.

Research Team

Owner:

  • xiaodi hu - Graduate Trainee, Massachusetts Institute of Technology
  • Alexander Warner - Graduate Trainee, Massachusetts Institute of Technology
  • Kenneth Fan - Graduate Trainee, Massachusetts Institute of Technology
1 - 1 of 1
<
>
Request a Review of this Research Project

You can request that the All of Us Resource Access Board (RAB) review a research purpose description if you have concerns that this research project may stigmatize All of Us participants or violate the Data User Code of Conduct in some other way. To request a review, you must fill in a form, which you can access by selecting ‘request a review’ below.