An automated algorithm that analyzes patient EHR data may help to identify those at risk for HIV infection, according to research presented at IDWeek and reported in MedPage Today.
This algorithm aims to identify candidates for pre-exposure prophylaxis, or PrEp, which involves a high-risk patient taking daily HIV medication to lower their chance of contracting the infection. MedPage Today reports that the CDC estimates more than 1.2 million people would benefit from PrEp treatment. However, since sexual health and risk evaluations are not a standard part of patient care, these at-risk patients are often go unidentified.
The researchers used machine learning to analyze data patterns, investigating more than 100 variables from EHR data including demographics, diagnoses, prescriptions and lab tests. When identifying patterns in the data, the algorithm might, for example, associate previous sexually transmitted infection treatment with condomless sex, a risk factor for HIV. When applied to a medical practice in Boston, the algorithm indicated 1.1 percent of patients were potential PrEP candidates.
The researchers' next steps include validating the predictive algorithm at a community health center and conducting a pilot study to determine whether the predictive algorithm results in increased PrEP use.