Session number: 1069
Abstract title: The Need for More Intensive Hypoglycemic Medication Therapy in Patients With Diabetes
Author(s):
DR Berlowitz - CHQOER Bedford VA Hospital
M Glickman - Department of Mathematics, Boston University
A Ash - Section of General Internal Medicine, Boston Medical Center
A Wong - CHQOER Bedford VA Hospital
Objectives: Most patients with diabetes have inadequate glycemic control. Understanding and improving how clinicians manage hypoglycemic medications is likely to be critical in improving outcomes. We now develop a measure of the intensity of medication therapy and examine its association with glycemic control.
Methods: We used existing databases to examine diabetes patients in VISNs 1 and 8 receiving regular VA care in 1998 and 1999. We used recursive partitioning to develop a model identifying factors associated with the decision to “increase” hypoglycemic medications (start a new medication or increase dosage of an existing one). This model assigned an expected probability of an increase to each visit. We next calculated a patient treatment intensity score over all medical visits during 1-1/2 years by comparing the actual number of increases to that expected. Regression models were used to examine the association with two glycosylated hemoglobin (HgbA1c) outcomes: change over time and whether it was >8.0%.
Results: We studied 266,309 visits made by 23,291 patients. An increase in medications occurred at 9.8% of visits. The strongest predictor of an increase was having a recent HgbA1c greater than 8.0%. Other significant factors included most recent glucose, time since last visit, whether on insulin, and recency of HgbA1c determination. Predicted probabilities of an increase at a visit ranged from 1.5% to 32.0%. Patient intensity of therapy varied considerably. Among the 12,347 patients with initial and final HgbA1c levels, intensity of therapy was significantly (p<.001) associated with improved outcomes. Among the quartile receiving most intensive therapy, HgbA1c declined over 1-1/2 years by 0.13% and 44% had a final value >8.0%, as compared to an increase of 0.08% and 54%, respectively, among those in the least intensive quartile.
Conclusions: We have developed a process measure describing the intensity of clinicians’ diabetes therapy and linked it to glycemic control. Those patients receiving more intensive therapy achieved better outcomes. Clinicians should be more aggressive in increasing medications when confronted with elevated glucose levels.
Impact statement: Existing databases may be used to profile clinicians’ management of diabetes. Interventions to change how clinicians decide to increase medications are likely to be successful in improving glycemic control.