Fighting disease with data

Interdisciplinary team takes on HIV/AIDS with big data tools

South Carolina has one of the highest rates of new HIV infections in the country, and national estimates show that up to half of HIV-positive individuals don’t continue regular medical care after diagnosis.

Without consistent medical supervision, HIV patients remain infectious and often have dire health outcomes. But two Arnold School of Public Health professors and an interdisciplinary team from the University of South Carolina have a plan to help turn the tide in the ongoing campaign to reduce HIV infections in South Carolina and make medical care more responsive for those diagnosed with HIV/AIDS.

Xiaoming Li, a health promotion, education and behavior professor, and Bankole Olatosi, a clinical associate professor in health services policy and management, are co-principal investigators on a project that will use “big data science” and predictive analytics to better predict HIV medical care utilization and identify gaps in HIV medical treatment. The team will analyze health information data from Health Sciences South Carolina, DHEC and the S.C. Revenue and Fiscal Affairs Office.

The project is part of a five-year, $3.1 million grant from the National Institutes of Health.

“Using big data science techniques, we will help DHEC identify HIV-positive individuals who were never in care, moved in and out of care or are at risk of dropping out of care so that they can be enrolled in HIV care by DHEC,” says Li, the SmartState Endowed Chair for Clinical Translational Research. “This is important because there is a strong relationship between early diagnosis, HIV medical care and survival.”

The team, which includes faculty members from the School of Medicine, geography and computer science and engineering, will generate a profile and pattern of care-seeking behavior for different demographic groups, use several measures of retention in HIV care such as missed visits, appointment adherence and gaps in care. With data mining and predictive analytics, they will also try to identify missed opportunities for HIV testing and gaps in treatment. Lastly, the team will work to develop a predictive risk model for targeting HIV care interventions for HIV-positive South Carolinians who are not getting medical care or are at risk for dropping out of care.

Keeping HIV-positive individuals enrolled in a regular regimen of medical care helps to ensure they are receiving anti-viral medication necessary to control their viral load, which is important for their own health as well as the health of others with whom they are in intimate contact. The state Department of Health and Environmental Control has used a laborious manual program called Data-to-Care, using HIV surveillance data to identify and reengage HIV patients back into care.

“By using data mining and predictive analytics, this study can significantly improve linking patients back into HIV medical care by working in tandem with the Data-to-Care program in providing real-time, population-based information about health utilization patterns and predictors,” Li says. “That will allow DHEC to focus on deploying staff and resources to bring each patients back into HIV medical care. And by developing risk profiles for HIV/AIDS patients at risk of dropping out of care, health authorities might be able to better prevent that from happening.”

More than 15,000 S.C. citizens have been diagnosed with HIV, an infection that, untreated, progresses to the often fatal Acquired Immune Deficiency Syndrome.

Share this Story! Let friends in your social network know what you are reading about