Lessons Learned from Lives Lost

UF researchers uncover surprising patterns in the spread of the great flu of 1918

In 1918, an unusually deadly flu swept the world, claiming 50 to 100 million lives in a pandemic often called the Spanish flu. Kyra Grantz, a research assistant in UF's Department of Biology and Emerging Pathogens Institute, hopes to help prevent such an outbreak from happening again. With Derek Cummings, professor of biology, she has studied how sociodemographic markers and urban infrastructure affected the spread of the flu in Chicago in that terrifying year.

The Spanish flu, so-named because of disproportionate press coverage of its incidence in Spain, is the ancestor of all influenza (H1N1) epidemics since. Its direct descendant is swine flu, which is less deadly than the 1918 strain. While typical flu mortality is 0.1 percent, the 1918 rate skyrocketed up to 20 percent in some areas. The phenomenon has been a topic of fascination for researchers who aim to discover how and why the Spanish flu was so devastating. "It's a pet project for a lot of us," said Grantz, whose examination of the 1918 pandemic revolves around potential health disparities that exacerbated the flu’s effects.

Grantz uses regression analysis on flu mortality and sociodemographic data by census tract to determine spatiotemporal clustering, which can elucidate why some areas in Chicago were worse affected. Using 100-year-old data collected by the US Census and the Chicago Department of Health that include maps of Chicago with point locations given for reported flu and pneumonia-related deaths, she’s found that mortality rates increased with illiteracy and unemployment and decreased with homeownership. Analyzing the coincidence of these variables with the point data allows her to identify the spatiotemporal clusters of flu cases; for example, the study found that within 200 meters, death from flu infection was 1.2 times more likely to be accompanied by a second death within the week, and within 100 meters, coinciding deaths were 1.3 times more likely. This subtle but importance difference suggests that neighborhood-level outbreaks are a vulnerable point in disease control.

Grantz received her BA in Biology at the University of Chicago, where she had focused on lab-based research in microbiology and bacterial genetics. Then, she took an epidemiology course that introduced her to virology at the population level, which she realized offers "more ability to make a large-scale impact." She explains that understanding patterns of transmission is essential to developing effective prevention and control techniques. “There's something fundamental — and more fun — in studying what you can directly observe. It’s more tangible,” she said. Before graduation, she cold-emailed a couple of experts in epidemiology and connected with Cummings. They both arrived at UF in August 2015 and began working on a series of research projects that use statistical and mathematical modeling to monitor and predict the spread of contemporary epidemics such as Zika and dengue, as well as historical events such as the 1918 pandemic.

"There is something really fascinating to bacteria and viruses in particular," said Grantz, explaining how she got hooked on epidemiology. For example, dengue is “this really small, really simple — not even alive, depending on who you talk to — molecule that affects almost 400 million people per year. That's a fascinating phenomenon. There’s not really a better word for it.” As a budding scientist, she was struck by “the idea of being able to build things up from the bottom and something larger from that.” Indeed, her paper, which develops an epidemiological model from 100-year-old public health data, well reflects this philosophy.

Their research was published in the Proceedings of the National Academy of Sciences on Nov. 21, 2016.