The recent article by Ismail-Beigi et al. in the Annals of Internal Medicine suggests that we are finally beginning to take a more reasonable approach to setting treatment goals by proposing reasonable and considered criteria for the individualization of blood glucose targets for people with type 2 diabetes. Using expert opinion to interpret and extend the findings from recent landmark clinical trials, the authors suggest important clinical features that would be expected to modify clinical risks and benefits for people with diabetes. The authors suggest how these modifiers might be used to adjust individual A1c goals. Although the proposal is clearly in the realm of clinical opinion, not clinical fact, the authors address issues that many clinical providers confront every day. The thoughtful discussion and direction provided is certainly welcome.

As the Annals article describes, important findings from recent trials have provided strong evidence for promoting more individualized A1c targets, and many organizations have published evidence-based recommendations for diabetes management, including the American Academy of Family Physicians (AAFP), the American Association of Clinical Endocrinologists (AACE), the American Diabetes Association (ADA), and the United States Preventive Services Task Force (USPSTF). However, there are substantial differences between these recommendations, with little consensus about exactly how individualized targets should be translated into practical strategies. Nonetheless, since clinical performance measures have built the attainment of specific therapeutic targets into measures of quality of care, these targets are integrated into provider reimbursement. Reimbursement provides a strong incentive for clinicians, and, as the authors argue, would be fine if attaining the target was actually the best thing to do for the patient.

Unfortunately, in my opinion the article stops short of where we need to be. Although the text of the article recognizes potential A1c modifiers, including comorbidity, age, existing vascular disease, hypoglycemic events, and “psychosocioeconomic” (is that even a word?) context, the crude algorithm offered by the authors in table 2 includes only age, duration of diabetes, and the presence of microvascular or macrovascular complications in determining an appropriate A1c ranging from <6.5 to ~8.0. What happened to all the other modifiers the authors talked about?

Clearly the Ismail-Beigi article provides a good start for considering modifications in A1c targets. However, as a primary care doctor, I find the proposed disease-specific algorithm doesn’t really fit my population. None of my patients only have a case of diabetes. They have family and social problems, mental health challenges, insurance problems, educational and learning issues, challenges with physical activity and eating, job problems, and hundreds of different clinical conditions. They are all individuals, and as they change over time so does their most appropriately targeted A1c. The authors acknowledge the importance of individualizing therapy, and that is a step forward, but moving from one target A1c to even nine target A1cs isn’t what I would call individualized treatment. Individualized therapy doesn’t really fit nicely into an algorithm. If we can get the needs of individual patients integrated into existing quality measures, however, we will be on the way to more accurately measuring clinical performance and helping to create a better health care system.