Story Background
Early detection of any condition, chronic or not, gives us the ability to take necessary interventions. Once the diagnosis of diabetes is made, diabetes-related tissue damage occurs in nearly half of the patients. Even after diagnosis, the likelihood of leading to vascular complications is quite high. Evidence suggests that early detection of diabetes by appropriate screening methods, especially in subjects with high risk for diabetes will help to prevent or delay vascular complications and thus reduce the clinical, social, and economic burden of the disease. There is evidence to show that intervention at the prediabetic stage is superior to the diagnosis of diabetes.
Business Challenge
Providers and payors alike target diabetic patients with the intent to stem long-term impact on the health of the patient/member. There is a need to be able to use a predictive model to identify a cohort of patients/members. Early detection of diabetes can help prevent or delay complications and every co-morbidity associated with this chronic condition.
Business Outcomes
Implementation of this solution as a standalone or integrated as part of a bigger framework will lead to a significant lowering of the healthcare burden. For instance:
- Decrease in overall healthcare burden.
- Improvement in activities of daily living.
- Increase in the life span of members and quality of life.
Solution Overview
Use observational measures such as glucose, Urea Nitrogen, BMI, and BP at the point of encounter to predict prediabetes. Identify a population with greater than 50% propensity. Extract the latest PubMed topical articles on diabetes. Tag individual patients identified with a summary or abstract of PubMed articles in preparation for outreach. Currently, model accuracy is 96%.
Please follow the aka.ms/HLSBlog for all this great content.