Effectiveness of an HbA1c tracking tool on primary care management of diabetes mellitus: glycaemic control, clinical practice and usability

• Main Authors: Jane Howard, Ryan Sommers, Odette Gould, Michelina Mancuso
• Format: Article
• Language: English
• Published: BCS, The Chartered Institute for IT 2009-03-01
• Series: Journal of Innovation in Health Informatics
• Subjects: computerised medical records system; diabetes mellitus; glycosylated haemoglobin
• Online Access: http://hijournal.bcs.org/index.php/jhi/article/view/713

<strong>Objective</strong> To determine if a laboratory data report (the HbA1c Tracking Tool) could be used as an effective intervention to improve diabetes management. <strong>Design</strong> A longitudinal quasi-experimental cohort design was used to test the effectiveness of an HbA1c summary report sent to primary care physicians for all patients having HbA1c levels greater than 7%. Setting Moncton, New Brunswick, Canada. Sample selection Administrative data from all adult patients with diabetes who had had at least two HbA1c measurements within the year prior to the initiation of the HbA1c Tracking Tool, and who had had five years of HbA1c measurements (2002-2007) overall was included. Interventions In March 2006 all primary care physicians began receiving HbA1c summary reports (through the HbA1c Tracking Tool) as a means to improving the management of diabetes. <strong>Main outcome measures</strong> (a) patient glycaemic control as indicated by HbA1c levels, (b) physician adherence to practice guidelines as indicated by measuring the mean number of HbA1c tests ordered per patient per year, and (c) physician usage rates of the HbA1c Tracking Tool in clinical practice. <strong>Results</strong> The sample (n=955)was divided into three subgroups based on flagged HbA1c level (7_&lt;8%, 8_9%, &gt;9%). The strongest effect of the intervention was found in the two groups with the poorest glycaemic control. The effect was stronger in the &gt;9% group (from 10.1 to 9.3%), than in the 8_9% group (a drop of 8.5 to 8.3%). Longitudinal analyses over a five-year period indicated the same findings. Patients were also found to receive more tests across time (from 2.45 tests per year to 3.0 across five years). In terms of usage, 92.1% of the physicians surveyed used the tool in their practice. <strong>Conclusion</strong> Routinely collected hospital laboratory data can be used both as the basis for an information- based intervention and as a tool to monitor quality of diabetes care.