Prediabetes to Diabetes (Pre2D™) model study

Get a copy of the study:

Pre2D Algomarker™ study

  • We share our peer-reviewed studies via email.
    Enter yours and we'll send it to you.
  • This field is for validation purposes and should be left unchanged.

Share:

Share on twitter
Share on linkedin
Share on facebook
Share on email

In this study led by Hadassah University Hospital endocrinologist Dr. Avivit Cahn, an algorithm developed by Medial EarlySign data scientists was used to identify patients with prediabetes at highest risk of progressing to diabetes within one (1) year.

Objectives:  This study sought to determine whether a machine-learning model could improve the prediction of incident diabetes utilizing patient data from electronic medical records, compared to a logistic regression model.

Background: Identifying patients at high risk of progression from prediabetes to diabetes can help prioritize and target delivery of limited diabetes prevention program resources, while avoiding the treatment of patients identified as being at low risk.

Methods: A machine-learning model predicting the progression from prediabetes to diabetes was developed using a gradient boosted trees model. The model was trained on data from the UK (THIN) database cohort, and validated with internal and external (Canadian AppleTree and Israeli Maccabi Health Services) data sets that were not used for training. The model’s predictive ability was compared with that of a logistic regression model within each data set.

This peer-reviewed study in Diabetes Metabolism Research and Reviews gives healthcare systems, endocrinologists, primary care physicians, and population health managers insights on how prioritizing diabetes prevention programs for prediabetic patients on high-risk trajectories can be more cost effective than offering similar resources to those at lower risk.

Highlights include:

  • How ML models for prediabetes can help clinicians identify patients at highest risk for progressing to diabetes;
  • How such a model enables physicians to prioritize and target prediabetic patients based on risk, given limited resources for lifestyle intervention programs. while avoiding the burden of prevention and treatment for patients at low risk;
  • How random forest regression models were more predictive and discriminative than standard regression methods at patients with prediabetes who would become diabetic within one (1) year
 

Prediction of progression from prediabetes to diabetes: Development and validation of a machine learning model, Avivit Cahn, Avi Shoshan, Tal Sagiv, Rachel Yesharim, Ran Goshen, Varda Shalev, Itamar Raz,  2020 Jan 14:e3252. doi: 10.1002/dmrr.3252; PubMed id: 31943669



Share:

Share on twitter
Share on linkedin
Share on facebook
Share on email

Prediabetes to Diabetes (Pre2D™) model study

Are you looking to solve high-burden disease challenges?