Disease Control by the Numbers

Veterinary scientist Yvette Rauff analyzes statistical data to find patterns of disease distribution in a population.

Efforts to combat disease outbreaks, such as the Ebola virus in Zaire and mad cow disease in Britain, are increasingly in the public eye. Public health crises have even been featured in recent bestsellers and movies like "Outbreak," in which medical personnel encased in gleaming plastic protection gear heroically protect the American people from a deadly plague. But in the real world, scientists who work behind the scenes tracking the movement of virulent pathogens usually labor in obscurity. That's because these sleuths are likely to be number crunchers who get their answers using questionnaires and computers.

Consider the work of veterinary scientist Yvette Rauff, a graduate assistant in pathobiology. As part of her doctoral program, Rauff had to pinpoint conditions that contributed to an outbreak in 1994 of a particularly virulent strain of bovine viral diarrhea (BVD) in Pennsylvania dairy herds. BVD causes high fevers, loss of appetite, decreased milk production, diarrhea, and respiratory problems in cattle. In severe cases, cattle can die within 48 hours. Several of the Pennsylvania dairy operations affected by the 1994 outbreak lost so many cattle they went out of business.

Penn State researchers helping to contain the outbreak needed more detailed explanations of the conditions that might have caused it. That gave Rauff a chance to take on some daunting disease detective work as part of her studies. Luckily, as a returning adult student, she had lots of experience to draw on. She had earned a bachelor's degree in veterinary science and a doctor of veterinary medicine degree from Michigan State University, graduating as one of five female veterinarians in the class. In 1974, she spent a year in the Peace Corps working in the Philippines with the Bureau of Animal Industry to upgrade beef cattle quality. Afterward she practiced veterinary medicine in Wisconsin and Pennsylvania until 1992, when she decided it was time to seek new challenges.

"As a veterinarian diagnosing and treating companion animals, I was helping people individually," Rauff says. "But I really wanted to make a more direct impact on public health issues. I started out by taking a job as a technician in Penn State's veterinary science department. My first project was to develop a diagnostic test kit for Johne's disease, which causes diarrhea and wasting in cattle." This work spurred Rauff's interest in epidemiology, the study of disease distribution and control within a population. However, a veterinarian can't just become an epidemiologist. So at age 43, with a young daughter, she decided to go back to school.

Rauff approached Penn State epidemiologist William Sischo, a member of the veterinary science field studies group, who agreed to be her academic adviser. Field investigator Dale Moore had just discovered that herds involved in the BVD outbreak had not been vaccinated according to veterinary recommendations, or had not been vaccinated at all. This finding raised the question of how many other Pennsylvania herds were at risk. "The basis of epidemiology is trying to get unbiased information by collecting data, which often means putting together a survey," Sischo says. "That may seem routine until you actually try it. To do a survey well, you need to understand the mathematics of it and be able to visualize patterns within the numbers."

In order to bring the BVD numbers into focus, Rauff had to find a way to organize as much information as she could gather on how Pennsylvania dairy farmers manage their herds. She also needed to design a survey that could provide very specific information on herd management and vaccination practices. "Writing good survey questions is really an art," Rauff says. "The questions must be straightforward and unambiguous. Since some people may respond in the way they think you want them to answer, double-check questions have to be included, too."

Rauff also had to make sure the survey was concise and easy to understand, so respondents wouldn't feel intimidated or toss it into the trash without a second look. She designed her survey in booklet form, keeping the questions short. Some 600 surveys were sent out statewide in January 1995, and 378 usable responses were returned. Of those usable responses, 199 gave specific vaccination information. Rauff points out that the response rate was higher than normal because dairy producers are very concerned about BVD.

With the responses in hand, Rauff then had to construct what she calls "a snapshot in time" reflecting the vaccination and quarantine practices for Pennsylvania dairy herds at the time of the outbreak. Her first step was to determine which producers were adequately vaccinating their animals and which were not. Using Epi Info, a computer software program developed by the Centers for Disease Control and Prevention, she separated the respondents into two groups by analyzing each producer's answers to detailed questions about their vaccination practices. The next step centered on how each group differed. She factored in the following questions: Where in the state is the farm located? How big is the herd? Does the producer quarantine animals bought as replacements or returning from shows? Does a veterinarian or the producer vaccinate the animals? Is a killed vaccine, which introduces viral proteins to stimulate an animal's immune response, used, or a modified live vaccine, which mimics the actual infection?

"Once you have all that information analyzed, it's like having puzzle pieces scattered all over," Rauff explains. "It was up to me to put them together to create a clear picture." While analyzing the data, she discovered that the majority of producers had consulted their veterinarians for vaccination information. "That surprised me," Rauff admits. "I had some preconceived notions about why the outbreak happened, probably due to my bias as a vet. I thought the survey would show that producers were not consulting their veterinarians. I also found that larger herds–those with more than 100 animals–were more likely to be adequately vaccinated."

The statistical bombshell in the survey was the revelation that just 27 percent of the herds met the survey's definition of "adequately vaccinated," whereas almost 80 percent of the respondents had said they routinely vaccinated. "I really didn't expect the percentage to be that low," Rauff says. "And in reality it may be even lower because the people who took time to answer the survey are likely to be much more concerned about bovine viral diarrhea."

Rauff links this low number of adequately vaccinated herds to the producers' choice of vaccine. The majority used a vaccine containing a killed virus, which requires a booster shot two to four weeks after the initial vaccination. Many producers failed to follow up with a booster injection, yet believed they were following proper procedures. Rauff considers this to be the most important finding and attributes it to a breakdown in communication between producers and veterinarians on how to maintain good vaccination and biosecurity programs. "Producers were right in turning to their veterinarians for advice, but something was lost in the translation," she explains. "I think some feedback and education are needed on both sides. In addition, the pharmaceutical company that manufactures the vaccine may choose to make its directions more emphatic as well."

Another discovery Rauff made that has implications on a personal level is her talent for finding patterns and answers in statistical data. She is currently using these skills in her thesis project, which centers on the development of a computer program to help immuno-compromised human patients assess their risk for exposure to the parasite cryptosporidium, a diarrheal affliction that can be fatal for those with weakened immune systems. "Finding the answer to a problem is what appealed to me as a veterinarian," she says. "Taking symptoms, history, and other information and putting it all together was exciting. I didn't think I could find anything I enjoyed as much as clinical practice, but I think I have in epidemiology."

–John Wall