Automated Identification of Acute Hepatitis B Using Electronic Medical Record Data
Author Information
Author(s): Michael Klompas, Gillian Haney, Daniel Church, Ross Lazarus, Xuanlin Hou, Richard Platt
Primary Institution: Harvard Medical School and Harvard Pilgrim Health Care
Hypothesis
Can electronic medical record data be used to accurately identify cases of acute hepatitis B for public health surveillance?
Conclusion
An algorithm using electronic medical record data can reliably detect acute hepatitis B and improve public health surveillance.
Supporting Evidence
- The algorithm identified 112 out of 113 patients with acute hepatitis B.
- The sensitivity of the final algorithm was 99% and specificity was 94%.
- Eight cases of acute hepatitis B were identified without any false positives.
Takeaway
Researchers created a computer program to find cases of a liver infection called acute hepatitis B using patient records, and it worked really well.
Methodology
The study developed and validated an algorithm using electronic medical record data to identify acute hepatitis B cases.
Potential Biases
Potential misclassification of acute versus chronic hepatitis B cases by the physician reviewer.
Limitations
The validation dataset was relatively small compared to the derivation set.
Participant Demographics
Patients from a large ambulatory medical practice in Eastern Massachusetts.
Statistical Information
P-Value
0.0001
Confidence Interval
95% CI, 94–100%
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website