Study on Diabetes Diagnosis Using Electronic Medical Records
Author Information
Author(s): Harris Stewart B, Glazier Richard H, Tompkins Jordan W, Wilton Andrew S, Chevendra Vijaya, Stewart Moira A, Thind Amardeep
Primary Institution: Centre for Studies in Family Medicine, The University of Western Ontario
Hypothesis
Can electronic medical records provide a validated definition for diabetes diagnosis?
Conclusion
The study successfully establishes a new standard definition for diabetes diagnosis using electronic medical records.
Supporting Evidence
- The new definition identified 13% more patients with diabetes compared to the Ontario Diabetes Database.
- The study maintained good sensitivity (75%) and specificity (98%) for the new definition.
- The research supports the incorporation of HbA1c ≥6.5% as a diagnostic criterion for diabetes.
Takeaway
This study found a better way to identify diabetes using electronic health records, which can help doctors take better care of patients.
Methodology
A retrospective cohort study using linked health administrative data and electronic medical records from 1 April 2006 to 31 March 2008.
Potential Biases
The dataset may be biased towards individuals who use health services, potentially underreporting those who do not.
Limitations
The dataset does not include consult letters or visit notes, and is not population-based, which may affect the generalizability of the findings.
Participant Demographics
Patients included both rural and urban populations in Southwestern Ontario, with a mean age of 42.5 years.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website