Can numbers save lives? Not directly, but people working with solid statistical insights can. Chanelle Howe, for example, is looking for the numbers that can explain why the HIV epidemic continues to hit some minority communities, particularly blacks, disproportionately hard.
“What determines why African-Americans are more likely to become infected by HIV?” asked the new assistant professor of epidemiology. “Once they become infected, they are also not likely to do as well on [antiretroviral] drugs and are more likely to have poor HIV-related outcomes such as death,” she said. “I’m trying to figure out the potential reasons and make recommendations for how this health disparity could be lessened.”
The problem can be stated simply, but the mathematics of making headway are a scientific challenge. That entices, rather than deters, Howe. Increasingly rarefied math has been her stock and trade ever since she enrolled at MIT as an undergraduate.
Coming out of high school, Howe had strong interests in chemistry and math so she chose to major in chemical engineering.
“With time, as I took more courses, I realized it wasn’t the right fit, so I started thinking about other options,” she said. “I discovered public health and epidemiology.”
For her Master of Public Health, she went to Columbia University, then attended Johns Hopkins for a second master’s in biostatistics and her doctorate in epidemiology. In 2009, she joined the research group of Stephen Cole at the University of North Carolina as a postdoctoral fellow.
In Cole’s group, Howe’s interest grew in using and developing advanced methods to investigate the HIV epidemic in general, but particularly among minorities. That focus allows her to obtain a deep understanding of both the public health issue (i.e., what questions to ask) and the needed statistical methods (i.e., how to find the answers).
Earlier this year in the journal AIDS, for example, Howe was the lead author on a paper that used advanced methods to help reconcile the conflicting literature regarding the effect of alcohol consumption on HIV infection among black injection drug users in Baltimore. She found that the more people drank, the more likely they were to become infected, independent of other behaviors like risky sex or drug use.
The possibility that reducing drinking could reduce infections might seem straightforward, but doing studies like this isn’t always that way. Epidemiological studies, especially ones that try to follow participants over time, are rife with data pitfalls that must be overcome with cunning mathematical methods.
“A lot of things can happen over the study time period,” she said. “People can say, ‘I don’t want to participate anymore -- this is too much of a pain to get to a study visit.’ People are often lost to follow-up or will have gaps in time between when they attend visits. Missing data can arise from either of these sources.”
One of the contributions Howe hopes to make is to find better ways to make good interpolations to fill in such missing data.
Another analytic issue that piques Howe’s interest is called “effect decomposition.” This is the method of teasing out what intermediary mechanism or pathway allowed something to cause something else. It’s a vital issue in public health, because it tells people what can be fixed.
“It creates more points of intervention,” she said. “I may not be able to change X because X is someone’s race or gender, but I might be able to intervene on things that occur between X and Y.”
For example, Howe is keenly interested in why antiretroviral therapy hasn’t been as effective for blacks as for other groups. A lot of research suggests that black patients receiving HIV care are more likely to miss clinic visits. What Howe wants to know is why, given that visit attendance may be related to therapy use. Then perhaps policymakers could help.
“Let’s say patients are missing visits because they don’t have transportation. Then maybe HIV clinics could provide a free shuttle bus to the clinic.”
Whatever her investigations may yield, Howe is excited to be at Brown. Epidemiology became an official department only this summer so Howe hopes to help shape the curriculum, for instance by contributing expertise at the border with biostatistics.
Howe is also looking forward to joining a smaller public health program where she believes she can work closely with colleagues and students. But Brown also provides considerable focus to go with its intellectual intimacy. Brown’s public health program won a $7.5-million federal grant last year to study the intersection of alcohol and HIV, and also co-hosts a Center for AIDS Research, much like the one Howe worked with in North Carolina.
“When I came to visit it seemed like a perfect match,” she said.