We thank Dr. Spitz (1) for his letter commenting on the Diabetes Physician Recognition Program (DPRP) criteria regarding HbA1c (A1C) levels. The DPRP criteria were changed in 2000 to coincide with those used in the Health Plan Employer Data and Information Set (HEDIS) program. More recently in 2004, a decision was made to include two measures for A1C, LDL, and blood pressure. In the case of LDL, the change reflected the HEDIS measure, National Cholesterol Education Program guidelines, and the American Diabetes Association recommendation. In the case of A1C and blood pressure, changes were based on current American Diabetes Association recommendations. Using two measures (which some refer to as good and poor control) allows a more comprehensive assessment of how well a group of patients is doing as this approach encourages both attention to persons in relatively poor control as well as allowing ongoing assessment of how the provider is doing in regard to meeting the stated guideline. For example, if only “% of patients with A1C >9%” were used, movement of patients from 9.1 to 8.9% would yield significant improvement, yet most would argue that little had changed. Using measures of “% >9%” and “% <7%”, however, would show that little had changed. If patients were moved from an A1C of 9.1 to 6.9%, using only the 9% measure would yield the same results as in the first case, but using both measures the rather significant change would be clearly indicated. Using both measures allows one to see continuing improvement over time as the “% >9%” should continue to decrease and the “% <7%” should continue to increase.
Dr. Spitz suggests that it would be useful (and more fair to those who are referred patients who are not doing well in regard to A1C) to add a measure based on improvement in A1C. The suggestion is well worth considering and has been reviewed in the past by experts in both diabetes as well as measurement. One obvious problem in having a change in A1C measure is that doctors caring for patients who are at goal would appear to not be doing well using this measure, as no improvement would be needed or likely seen. As well, the goal of using measures to document how a population of patients is doing over time would not be part of this metric. Simply awarding points for A1C improvement would create some potential unfairness as well, as it is generally much easier to get a patient doing poorly to reduce his/her A1C 1% (from 10 to 9%, for example) than a patient doing relatively well (to reduce the A1C from 8 to 7%). Secondly, all A1C improvements are not equal in regards to clinical benefit, as an improvement of 1% in A1C offers a different benefit if the change is from 7 to 6% vs. 12 to 11%, for example. Finally there is the problem of setting the time frame for the change and having to review charts for multiple values, not just the most recent.
Dr. Spitz is of course correct that any improvement in A1C is a positive change. The data he cites for his practice are very impressive in regards to the reduction in A1C levels he has achieved. We feel that the current measures, used accurately, fairly capture this aspect of diabetes care, and adding a new measure for A1C change is not likely to add substantial new information to the program. However, we feel it is worthwhile to bring this to the current DPRP advisory committee for discussion at their next meeting.
