Developing a Predictive Model to Prioritize Human Immunodeficiency Virus Partner Notification in North Carolina

In North Carolina, disease intervention specialists (DIS) have less time to conduct partner notification activities because of competing responsibilities and growing case loads due to expanded HIV testing. In the current study, the researchers developed a model to predict undiagnosed HIV infection in sexual partners in order to prioritize DIS interviews.

Using DIS records of HIV-infected persons reported in two North Carolina surveillance regions from Jan. 1, 2003 to Dec. 31, 2007, the team abstracted demographic, behavioral and partnership data. A predictive model and risk scores among newly diagnosed persons and their partners were developed using multiple logistic regression with generalized estimating equations. Sensitivities and specificities of risk scores at varying cutoffs were used to examine algorithm performance.

The results showed that five factors predicted a partnership between a person newly diagnosed as HIV-positive and an undiagnosed partner: 1) a period of four weeks or less between HIV diagnosis and DIS interview; 2) no history of crack use; 3) no anonymous sex; 4) fewer total sexual partners reported to DIS; and 5) sexual partnerships between an older index case and younger partner. "Using this model, DIS could choose an appropriate cutoff for locating a particular partner by determining the weight of false negatives relative to false positives," the authors wrote.

"Although the overall predictive power of the model is low, it is possible to reduce the number of partners that needs to be located and interviewed while maintaining high sensitivity. If DIS continue to pursue all partners, the model would be useful in identifying partners in whom to invest more resources for locating," the team concluded.

Back to other news for February 2012