Can the analysis of health data predict which patients with diabetes may be at a high risk of developing chronic kidney disease?
Diabetes-related chronic kidney disease

Approximately 10% of people with diabetes can be affected by chronic kidney disease (CKD), a severe complication that may require dialysis or renal transplant.

A research team studied the data on 417,912 people who had been newly diagnosed with diabetes, without a previous history of kidney failure.

This analysis has shown that a new calculation method is much more accurate at predicting CKD. This can lead to preventive healthcare measures to prevent or to slow down this complication.

Read on to learn how sharing health data can save lives or have a look at our video

Click on image to start the video

Why was this work needed?

Chronic kidney disease (CKD) is one of the most severe complications related to diabetes. Approximately 10% of people with diabetes can be affected by this, within three years following their initial diabetes diagnosis. This severe complication may require dialysis or renal transplant.

An accurate way to predict which patients have a high risk of developing CKD, before it starts, can help clinical teams to try and prevent or slow down this complication.

An analysis was undertaken to assess the accuracy of a new calculation method for predicting CKD in a large patient population.

What did research find?

The analysis has shown that this new algorithm (calculation method) is much more accurate at predicting CKD.

What is the impact on patients?

Patients at risk of developing CKD will have more chances of being identified in time and receiving the appropriate preventive treatment.

With this algorithm being embedded in the software used by clinicians, doctors will have a tool that helps them identify the patients at risk more easily and accurately. Hence they can monitor their kidney function more intensively and take the necessary health measures to reduce the impact and severity of diabetes-related complications.

What data was used?

The research used robustly anonymised electronic health record information derived from multiple US healthcare providers. The research team studied the data on 417,912 people who had been newly diagnosed with diabetes (types 1 and 2), without a previous history of kidney failure. These data were already held in large health databases specifically established for research use. The reuse of this data has saved many years that might otherwise have been spent on a new clinical trial, at significant cost.

Who funded and collaborated on this work?

This research was a collaboration between Roche (Roche Diabetes Care) and IBM. The research was funded jointly by both companies.

Further information

This work uses data provided by patients and collected by the health professionals as part of the care and support provided.

i~HD