20% LESS BABIES
IN INTENSIVE CARE
THE RIGHT PATIENT
Using TIMi on the data from the company, Sura developed predictive models to identify the patients with the highest risk of deadly or costly complication, such as heart attacks, dialysis, chronic liver disease or extreme cardio vascular risk.
BUILDING THE MODELS
In less than 3 months, they built 4 predictive models using 8 years of clinical history withwith over 23.000 variables and 4.5 million rows for the diabetic population, and integrated the results in analytics solution that helped articulate new patient care programs.
Sura obtained the following results.
Prediction of the last stage of diabetes :
over 80% AUC (i.e. ultra high accuracy)
Prediction of heart-related complications for diabetics:
ultra high accuracy
Early detection of pregrancy-related complications:
20% babies less in intensive care