Claims-Based Algorithm to Identify Pre-Exposure Prophylaxis Indications for Tenofovir Disoproxil Fumarate and Emtricitabine Prescriptions (2012-2014): Validation Study
2024

Algorithm to Identify PrEP Use for HIV Prevention

Sample size: 10645 publication 10 minutes Evidence: moderate

Author Information

Author(s): Mavragani Amaryllis, Wang Weijie, Gonzalez Charles J, Zucker Jason, Sullivan Patrick Sean DVM, PhD, Mera-Giler Robertino M MD, PhD, Bush Staci PA, Shvachko Valentina MS, Sarkodie Eleanor MPH, O'Farrell Daniel MSPH, Dubose Stephanie MPH, Magnuson David PhD

Primary Institution: Rollins School of Public Health Emory University

Hypothesis

Can an algorithm accurately distinguish TDF/FTC prescriptions for pre-exposure prophylaxis (PrEP) from other clinical indications?

Conclusion

The PrEP algorithm accurately classified most patients treated with TDF/FTC for HIV prevention.

Supporting Evidence

  • The algorithm demonstrated high sensitivity (99.6%) and moderate specificity (49.6%) in electronic medical records.
  • In a community health clinic, the algorithm showed high sensitivity (99.6%) and specificity (87%).
  • The algorithm misclassified 4.6% of TDF/FTC exposure periods as PrEP when they were truly for chronic hepatitis B.
  • 9.3% of TDF/FTC periods classified as PrEP were actually for postexposure prophylaxis.

Takeaway

Researchers created a computer program to help doctors figure out if patients are taking a medicine to prevent HIV or for other reasons.

Methodology

The study developed and validated an algorithm using prescription, diagnosis, and medical procedure data to identify TDF/FTC prescriptions for PrEP.

Potential Biases

Misclassification could occur due to incomplete or inaccurate medical records.

Limitations

The algorithm's specificity was variable, leading to potential misclassification of non-PrEP users as PrEP users.

Participant Demographics

Patients prescribed TDF/FTC for various indications, including HIV treatment and PrEP.

Statistical Information

P-Value

0.0001

Confidence Interval

95% CI 98.7%-100%

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.2196/55614

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication