Back in the day when I was a database manager on a study of HIV and cognitive decline, we would use something called multivariate analysis to control confounding variables. Before your eyes glaze over, let me give a real world example.
If we compare something like cognitive decline (or in the example above, viral loads) among people with HIV, we'll see some folks perform just as well as negative controls, while most do much worse. Drilling down into the data, we'll see that persons more likely to have poor outcomes are also more likely to have a history of drug use, have an untreated mental illness, have unstable housing, or they'll be more likely to be from a racial minority. To reduce the impact of the influence of these confounding factors, researchers "control" for these variables in the analysis. Unfortunately, when you control for all these issues, you're more likely to get a positive population made up of well-educated, middle-class, white gay men--which is not representative of the HIV-positive population in its diversity and complexity.
Things like injection drug use, gender, and race influence clinical outcomes. In a study published last year, "Improvement in the health of HIV-infected persons in care: reducing disparities," we can see in the graph reproduced above, that over the past fifteen years, differences in viral load (a good measure of successful HIV treatment) has been virtually eliminated between men and women, whites and blacks, and MSM and IDU populations. Part of these impressive outcomes are due to the fact that programs like Ryan White have had input from the local community, with funds targeting many of these subpopulations who have worse outcomes.
Current funding is now focused on the "test and treat" model, looking to diagnose as many positive people as possible, and getting them into medical care, on medications, and with an undetectable viral load, be less likely to pass along the virus. The good news is that with clinical outcomes being monitored, persons who miss appointments or have other risk factors for falling out of care are being caught when they fall through the cracks. The unfortunte flip side of these new efforts is that there is less funding that targets vulnerable populations. Let's hope that as we move forward, we're able to maintain these impressive outcomes like we've seen in the Baltimore HIV clinic presented in the Johns Hopkins data.
