JMIR Publications

ABSTRACT

Background:

Identification of people with HIV from electronic health record (EHR) data is an essential first step in the study of important HIV outcomes, such as risk assessment. The task has been historically performed via manual chart review, but the increased availability of large clinical datasets has led to the emergence of phenotyping algorithms to automate this process. Existing algorithms for identifying people with HIV rely on a combination of ICD codes and laboratory test or closely mimic clinical testing guidelines for HIV diagnosis. However, we found that existing algorithms in the literature missed a significant proportion of people with HIV in our data.

Objective:

In this study, we developed and evaluated an empiric HIV phenotyping algorithm.

Methods:

We developed an algorithm using HIV-specific laboratory tests and medications and compared it to two previously published algorithms in national and local datasets to identify cohorts of people with HIV. Cohort demographics were compared to those reported in national and local surveillance data. Chart review was performed on a subsample of patients from the local database to calculate sensitivity, specificity, and accuracy.

Results:

Our empiric algorithm identified substantially more people with HIV in both national (38% to 86% increase) and local (26% to 83% increase) EHR databases than the previously published algorithms. Demographic characteristics in people with HIV identified with the empiric algorithm were similar to those reported in national and local HIV surveillance data. Our empiric algorithm demonstrated improved sensitivity over existing algorithms (98% vs 56% and 90%) while maintaining similar overall accuracy (96% vs. 80% and 95%).

Conclusions:

We have developed and evaluated an empiric phenotyping algorithm for identifying people with HIV in EHR data that demonstrates improved sensitivity over existing algorithms.