Background and aims: While predictive tools are being developed to identify those at highest risk for developing diabetes, little is known whether these assays affect clinical care.
Methods and results: Thirty sites who used the PreDx(®) (Tethys BioScience, Emeryville, CA) abstracted clinical information from baseline clinic visits prior to a PreDx test and from the most recent visit at time of abstraction. All visits occurred between May 2008-April 2011 (median follow-up 198 days, IQR 124-334). The primary analysis was the influence of the PreDx test (5-year diabetes prediction) on subsequent care; descriptive statistics were used to summarize baseline and follow-up variables. Overall 913 patients with 2 abstracted visits were included. Relative to baseline, median SBP decreased 1.5 mmHg (p = 0.039), DBP decreased 2 mmHg (p < 0.001), LDL-C decreased 4 mg/dL (p = 0.009), and HDL-C increased 2 mg/dL (p < 0.001) at follow-up. Behavioral or lifestyle counseling was not significantly different from baseline to follow-up (71.2% vs. 68.1% (p = 0.077), but BMI was lower by 0.2 kg/m(2) at follow up (p = 0.013). At follow-up, more patients were prescribed metformin (13.7% vs. 9.7%, p < 0.001). A higher PreDx score was significantly associated with metformin prescription (p = 0.0003), lifestyle counseling (p = 0.0099), and a lower BMI at follow-up (p = 0.007).
Conclusion: The use of a prognostic test in patients perceived to be high risk for diabetes was associated with a modest but significant increase in the prescription of metformin and lifestyle interventions and a reduction in BMI.
Keywords: Diabetes mellitus; Quality improvement; Risk prediction.
Copyright © 2013 Elsevier B.V. All rights reserved.