Most of the practices in the study community didn’t even have a computer, let a lone an EMR. Therefore, achart audit was conducted as part of clinical information systems. It was used to establish benchmarks of care in the community, and also as a comparison to out clinical trial population, so that we were able to tell if the results we found were generalizable to the community.

As part of the chart audit, everyone 18 years or older with a diabetes diagnosis during or prior to calendar year 1999 had their chart audited. Diagnoses were confirmed by: READ SLIDE

For the repeat chart review, a 28% random sample of the original 762 were audited at 12 month follow-up to see if changes occurred in practice patterns and clinical outcomes over the course of the study

219 charts provided sufficient power to detect differences in practice patterns and clinical outcomes from baseline to 12-month f/u if they truly existed.

Patients of designated physicians

The study’s PI trained two chart reviewers. Training was performed using a standard chart review protocol and consisted of both the trainer and trainee reviewing the same charts over a three day period (~20). After each chart was reviewed by both, discrepancies were noted and the chart was reviewed to adjudicate the discrepancies. If greater than 5% of responses were discrepant, chart audit procedures were reviewed.