https://diabetes.acponline.org/archives/2015/11/13/1.htm

New algorithm may help physicians set glycemic targets for patients with diabetes

Researchers created a new algorithm to help choose target HbA1cs for patients with diabetes and compared the results to experts' advice.


Researchers have created a new algorithm for physicians that may help when developing individual glycemic targets for patients with diabetes.

They constructed the algorithm using survey input from 151 international diabetologists and validated it by surveying additional diabetologists. Results were published online on Oct. 30 in Diabetes Care.

Surveyed physicians ranked 11 factors they take into consideration when setting a patient's glycemic target in terms of relative importance. “Risk of hypoglycemia from treatment” ranked the highest, with more than 50% of those surveyed ranking it in the top 3. “Life expectancy” was ranked among the top 3 by 48% of survey responders. “Disease duration” and “resources and support system” were both ranked the lowest. The physicians then suggested appropriate glycemic targets for 6 clinical vignettes, which represented a wide range of patients.

The study's authors used this information to create the algorithm, which computes an individualized HbA1c target according to the severity score of 8 clinical parameters (they excluded 3 parameters for redundancy). The researchers restricted target HbA1c numbers calculated by the algorithm to between 6.5% and 8.5% because this range encompassed 95.1% of the recommended values proposed by survey respondents.

The algorithm was validated by presenting 3 new cases to 57 diabetes experts that did not participate in the original survey, who suggested glycemic targets that were very similar to those calculated by the algorithm.

Three of the 8 parameters used in the algorithm could be considered subjective: “cognitive function,” “adherence to therapy,” and “resources and support system,” the authors noted. The combined weight of these parameters is nearly 25%, and they could shift target HbA1c by up to 0.5%. The authors also proposed an alternative, more minimalistic model using only the 5 objective parameters, which are easily extrapolated from electronic medical records: “risk of hypoglycemia from treatment,” “life expectancy,” “important comorbidities,” “macrovascular and advanced microvascular complications,” and “disease duration.” (See study for full algorithm. Subscription required.)

The authors noted several limitations of their work, such as that the selection of surveyed physicians was not based on a systematic scoring system and that it's probable that many experts around the world were not included. They also did not collect data regarding the physicians' age or years in practice, although they did aim to include those nationally and internationally recognized for their contributions in the field.

This algorithm is an attempt at standardizing individualized care, and it needs further study and validation, the authors acknowledged. “The aid of a validated algorithm would have great clinical importance and would enhance our ability to deliver better diabetes care for our patients while avoiding the hazards associated with both over- and undertreatment,” they wrote.