OBJECTIVES:

Glycemic targets (HbA1c) have been recommended to guide therapeutic

treatment for patients with type 2 diabetes mellitus (T2DM) and reduce the risk of primary and

secondary complications. In this study, we describe a methodology using predictive models to

create individualized glycemic target ranges that are associated with a greater reduction of the

risk for complications.

METHODS:

The study population includes adult members of Clalit

Health Services with three-seven years T2DM duration, without concurrent serious chronic

conditions (cancer, chronic infections, and cirrhosis). We built a predictive model to assess the

future risk of common T2DM complications (macro/microvascular diseases, hypoglycemic

events and all-cause mortality), based on the index HbA1c, while controlling for baseline

demographic and clinical information. Individualized HbA1c target ranges were simulated in

order to determine which specific range would minimize each individual’s risk of complications

as identified by the predictive models. The final sub-analyses compared rates of complications

associated with the model-based individualized HbA1c target range to rates of complications

among those individuals whose index HbA1c was or was not within the target range.

RESULTS:

We developed a new methodology for the calculation of an individualized glycemic target. The

obtained targets yielded 20% more individuals within the recommended range, compared to the

standard guidelines, while maintaining the same outcome rates, and have the potential to more

accurately identify those at risk for future outcomes.

CONCLUSIONS:

We successfully created a

tool to calculate individualized HbA1c target ranges. Target ranges can potentially reduce the

need for intensive intervention in some populations and highlight other populations at greatest

risk. Validation of the tool using an independent external dataset is required. This study is the

first attempt to generate an individualized glycemic control target tool based on predictive

modeling and establishes how precision medicine can be incorporated into diabetes care

management.