New findings suggest that simply using HbA1clevels to assess the performance of individual physicians or healthcare systems in diabetes management may be misleading or inaccurate.
Endocrinologists typically see more complex patients who require more time to improve their glycemic control, which makes their performance look worse when judged solely by HbA1c levels.
But new data reported here at the American Diabetes Association 2013 Scientific Sessions show that when diabetes patients are grouped by medication use — a proxy for complexity and stage of disease — HbA1c levels for patients cared for by endocrinologists are the same as or better than those for individuals seen by general internists.
Lawrence S. Phillips, MD, professor of medicine in the division of endocrinology at Emory University, Atlanta, Georgia, who reported the findings in a poster at the meeting, said looking at patients by medication group shows there is very little difference between the performance of specialists and generalists.
“The message is really for the payers and the government… They need to do something like this. They need to have some conservative way to give the provider a chance to improve things, and then they need to compare apples to apples. Just looking at A1c is not sufficient,” Dr. Phillips told Medscape Medical News.
Poster session moderator Sanjeev Mehta, MD, MPH, director of quality at Joslin Diabetes Center, Boston, Massachusetts, agrees. “Dr. Phillips’s data demonstrated that endocrinologists, in the practice setting he evaluated, were seeing patients with higher HbA1c levels. While this suggests appropriate referrals by primary-care physicians to optimize glycemic control, it also supports Dr. Phillips’s conclusion that an outcome-based quality measure [such as HbA1c] may be inadequate when assessing the quality of diabetes care across all providers, especially endocrinologists,” he said.
Dr. Mehta noted that the Agency for Healthcare Research and Quality (AHRQ) has endorsed theadoption of more sophisticated quality metrics, including linked action measures such as appropriate medication use, which would assess outcomes in the context of the care provided.
“I strongly believe this is the direction that all stakeholders in the diabetes community need to be [following to evaluate] high-quality diabetes care,” he told Medscape Medical News.
Comparing Apples to Apples Is Best Approach
Dr. Phillips and colleagues obtained Emory Healthcare data for a total of 5880 diabetes patients cared for by 8 endocrinologists and 8 general internists over a 24-month period. The proportion of patients whose most recent HbA1c was 7% or above was higher for the endocrinologists than for the general internists, 51% vs 38%.
Subsequent analysis was restricted to the 3735 patients who had been seen 3 or more times in the past 24 months and at least once in the prior 12 months. This group was divided into 3 groups by medication use: Those using only oral medications and/or incretin-based drugs (1880), those using basal insulin (with or without oral medications/incretins, 324), and those also using mealtime insulin in addition to basal insulin, with or without other medications (1531). The latter group included patients with type 1 diabetes, Dr. Phillips told Medscape Medical News.
Overall control was poorer among the insulin-using patients, with HbA1c levels of 7% or higher in 66% of those using mealtime insulin and 55% of individuals using basal insulin, compared with just 21% of those not using insulin (P < .0001 for trend). And endocrinologists had more patients on insulin than did the general internists, with 53% vs 22% using mealtime insulin (P < .0001), 10% vs 7% using basal insulin (P = .02), and 37% vs 71% not using insulin (P < .0001), respectively.
When examined by treatment group, however, the non–insulin-using patients of the endocrinologists actually had better HbA1cs: 18.8% of their patients had levels at or above 7% vs 23.4% of the general internists’ patients.
For the 2 insulin treatment categories, there was no significant difference between the endocrinologists and the internists. In both groups, just over half of the patients had HbA1cs 7% or above (P = .6) as did about two thirds of those using mealtime insulin (P = .9).
New Models Needed for Evaluating Care
Dr. Mehta told Medscape Medical News:”I think this poster highlighted the importance of adopting more sophisticated quality metrics, such as linked action measures, and the importance of ongoing collaboration with specialists and specialty centers in the care of adults with diabetes.
“Specialists and specialty centers may have an opportunity to translate best practices to their referring primary-care physicians, who will continue to care for the majority of adults with diabetes in the United States,” he added.
And specialists should be rewarded, not penalized, for their particular patient mix. “Those providers and practices that care for more complex patients need to be recognized, even reimbursed, for their ability to make meaningful improvements in health outcomes in high-risk patients,” he observed.
Dr. Phillips told Medscape Medical News that “diabetes is a heavy-duty proxy for healthcare systems as a whole, because a lot of people have diabetes, and it’s an expensive disease.”
He believes his “apples-to-apples” comparison could have implications for other areas of medicine as well. “I think it’s an important concept. You would think it applies to blood pressure, cholesterol, all the things that doctors do. We think this is a model for how you evaluate care.”