A computer-based algorithm based on a survey of diabetes thought leaders could be used as a decision-aid tool to help in determining the appropriate HbA1c target for individual patients with type 2 diabetes, a new study suggests.
The survey findings and resulting algorithm were recently published in Diabetes Care by Avivit Cahn, MD, of the Diabetes Research Center, Hadassah Hebrew University Medical Center, Jerusalem, Israel, and colleagues.
In 2012, the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) issued a joint position statement calling for a "patient-centered" approach to type 2 diabetes management with individualization of HbA1c targets for glycemic control.
The statement was an acknowledgement that not all patients benefit from attempts at normalization of HbA1c levels, and some — particularly those who are older and/or have comorbidities — may suffer harm from excess hypoglycemia and other adverse effects if attempts are made to drive HbA1c too low.
However, shifting to a more "individualized" approach leaves clinicians in a "gray zone" with no "gold standard," Dr Cahn and colleagues note.
"Physicians are often assessed by the HbA1c of their patients, and therefore a physician whose patients include elderly individuals with increasing comorbidities may not be fairly judged," Dr Cahn told Medscape Medical News.
In Israel, however, glycemic targets have been modified based on age and disease duration, "though clearly these are not the only parameters that need to be considered," she said.
Diabetes Opinion Leaders Worldwide Surveyed
To address the issue, Dr Cahn and colleagues surveyed diabetes opinion leaders worldwide to determine what factors they take into account when determining a patient's appropriate glycemic target.
Assigning points to each factor based on low, medium, or high risk, the authors devised formulas by which a clinician can plug in the numbers from a patient's medical record to generate an HbA1c target between 6.5% and 8.5%.
The basic algorithm includes five objective parameters, such as comorbidities and hypoglycemia risk from treatment, and a second calculation adds three subjective parameters, including "adherence and motivation" and "resources and support system."
"The formula is intended for use in all type 2 diabetes patients, as it takes into account each patient's unique characteristics. It is meant to be a supportive decision-making tool for the physician, though it cannot replace the physician's own clinical intuition," Dr Cahn explained.
Asked to comment, Anne L Peters, MD, professor of clinical medicine and director of clinical diabetes programs, Keck School of Medicine, University of Southern California, Los Angeles, told Medscape Medical News that she sees the algorithm as a positive step.
"Physicians are feeling more and more trapped by numbers and metrics we have to reach, and I think there has to be a way to individualize those that makes sense based on individual disease characteristics."
But the algorithm is just a first step, she noted. "This is about getting the physician to understand that in their heads they need to have an idea of what they're working toward and then add in the patient perspective. This isn't an algorithm that needs to be rigidly applied, but it's to get the conversation moving."
Constructing and Validating the Formula
Dr Cahn and colleagues first sent online surveys to 244 key opinion leaders in diabetology from around the world. Half were from North America and Western Europe, and the rest were from the Far and Middle East, Latin America, and Eastern Europe. A total of 151 responded.
Survey respondents were asked to rank in order of importance 11 parameters affecting a patient's HbA1c goal based on the ADA/EASD guidelines. "Risk of hypoglycemia from treatment" came out on top, followed by "life expectancy."
Factors ranked the lowest included "disease duration" and "resources and support system."
The survey also presented six clinical cases and asked respondents to input the target HbA1c he or she would recommend for that patient. Examples ranged from a newly diagnosed patient with no complications, to a frail elderly patient with multiple complications and/or comorbidities.
The model was also validated with three new clinical cases in a new set of thought leaders, including nonresponders to the original survey and investigators involved in some of the recent major cardiovascular-outcomes trials.
After eliminating some of the parameters because of redundancy, investigators assigned each of five objective parameters a relative weight based on the survey responses: risk of hypoglycemia from treatment (29.7%), life expectancy (27.0%), important comorbidities (17.5%), macrovascular and advanced microvascular complications (15.7%), and disease duration (10.0%).
The parameters are easily obtainable from databases that doctors are already using, say the researchers.
The user "scores" each parameters as 1 for low risk, 2 for moderate, and 3 for high risk. For diabetes disease duration, those scores are 1 for <5 years, 2 for 5 to 20 years, and 3 for >20 years. For life expectancy, the scores are 1 for long, 2 for intermediate, and 3 for short.
The weight of each parameter is multiplied by the user score of 1, 2, or 3, and the values are added together. The formula for the glycemic target is 6.5 + (sum of products – 100)/100.
So, for example, the target HbA1c for a 70-year-old nursing-home resident with moderate dementia and currently treated with basal insulin would be 8.1%, and a 45-year-old teacher with no complications, not taking any diabetes medications, and following a lifestyle program would have a target of 6.5%.
The authors also derived a second formula that includes three additional "subjective" parameters: adherence and motivation, cognitive function, and resources/support system. Adding these in can shift the target HbA1c up or down by 0.5%, the authors note.
A First Step
With further refinement and validation, this approach may turn out to be helpful in discussions with payers, Dr Peters notes.
"We're all measured now by metrics....I really reject that, because I see patients who are more difficult or sicker, and I need adjusted targets....It's not appropriate to have a 'one-size-fits all' number that you're judging me by in terms of patient care."
Dr Peters said she could see the algorithm as a teaching tool.
"You need to take this for what it is, but honestly we have to be having this conversation. We have to be able to help providers with this....You have to start somewhere. I don’t think people will wholesale adopt it, but it's a step."
Indeed, Dr Cahn and colleagues acknowledge that this proposed algorithm "needs further study and validation."
But, she told Medscape Medical News, she hopes this tool will eventually "enable physicians to deliver better patient-tailored care, avoiding intensified therapy in the frail elderly, while aiming for tight targets in the young population with long life expectancy who may benefit from it."
"I hope the paper will evoke further discussion as to the refinement of glycemic targets and awareness of the risks associated with hypoglycemia,” she concluded.
The study authors have no relevant financial relationships. Dr Peters has served as director, officer, partner, employee, advisor, consultant, or trustee for Amylin, Eli Lilly, Novo Nordisk, AstraZeneca, Abbott, Boehringer Ingelheim, Bristol-Myers Squibb, Dexcom, Medtronic MiniMed, Merck, Roche, and Sanofi. She no longer serves on any speaker's bureaus.
Diabetes Care. Published online October 30, 2015. Abstract
Medscape Medical News © 2015 WebMD, LLC
Send comments and news tips to firstname.lastname@example.org.
Cite this: Algorithm Helps Individualize HbA1c Targets in Type 2 Diabetes - Medscape - Nov 27, 2015.