Targeted Smallpox Vaccinations Could Be Effective Intervention Against Deliberate Attack

M. Elizabeth Halloran, M.D., D.Sc., and Ira M. Longini, Jr., Ph.D., professors of biostatistics at Emory University's Rollins School of Public Health, along with colleagues Azhar Nizam and Yang Yang, constructed a model that simulated the spread of smallpox deliberately introduced by infected individuals moving through a community. The model is based on the investigators' previous experience with modeling flu epidemics. It assumes that people interact primarily within known contact groups, including their own household, schools or daycare centers, their neighborhood and their community. The model differs from other recent models of smallpox epidemics that assume random mixing of individuals within a large, homogeneous population, with the conclusion that mass vaccination is the only way to sufficiently control an outbreak.

The Emory model simulated a range of scenarios for smallpox epidemics, based on different probabilities of how smallpox might spread, whether or not there was residual immunity from prior vaccinations, at what stage individuals might stay home during different stages of their disease, and the effectiveness of smallpox vaccines in different age groups.

The investigators conducted 200 simulations for each proposed intervention within their model, using a statistical community of 2000 people based on the age distribution and approximate household sizes reported in the U.S. Census 2000. Each simulated community consisted of four neighborhoods, one high school, one middle school, two elementary schools, small play groups, and day care centers. Households had between one and seven people, and 33% of households consisted of single adults. The model assumed that in a deliberate smallpox attack, either one or five individuals already infected with smallpox would circulate through the community at the beginning of the infectious stage of their disease.

Drs. Halloran and Longini believe their model is more realistic than "deterministic" models that assume all individuals within a large, homogenous population interact randomly with others, concluding that an epidemic always follows the same course. The Emory study considered many different variables, including where transmission was likely to be highest (first within households, then schools, then neighborhoods, and lastly the community at large) and which individuals were likely to have residual immunity. For individuals vaccinated prior to 1972, the investigators assumed different levels of immunity ranging from zero to half the immunity level induced by a fresh vaccination. They also considered the three different stages of smallpox –– incubation (10 to 14 days); prodromal phase, (flu-like symptoms); and pox phase –– and the likelihood of interaction and transmission within a community each day following infection. Based on historical data, they assumed that smallpox is more likely to be transmitted through close contact than through casual contact on the streets or in a subway.

The model investigated two different scenarios of mass vaccination, either before an introduction of smallpox, with vaccine coverage levels of 30, 50, or 80%, or after an epidemic begins, with 80% coverage within 10 days.

In the targeted vaccination scenario, people in close contact with either 80% or 100% of identified cases were vaccinated. Targeted vaccination included all household members, all day care and play group contacts, 80-100% of children in the same school, and 1.5% of neighborhood contacts. Vaccinations began either with the first known case or after the 15th or 25th case to account for possible delays in vaccination. In simulations of both mass and targeted vaccinations, a rapid response to identified smallpox cases made a significant difference in the ability to prevent or contain an epidemic.

"Our study demonstrates the importance of modeling the details of likely contact patterns when developing strategies to prevent or contain a deliberately planned smallpox epidemic," says Dr. Halloran. "We found, as have other investigators, that a rapidly implemented mass vaccination strategy following a smallpox outbreak would be more effective than targeted vaccination if there is no pre-existing immunity.

However, our model found a much smaller difference in effectiveness between the two strategies than did previous models that assumed homogeneous mixing within a large population. And, taking into account possible pre-existing immunity, the potential for increasing that pre-existing immunity, and the desirability of decreasing harmful vaccine reactions, we believe it is important to continue a serious exploration of the benefits of targeted vaccination. We will continue to refine our simulation to better reflect the variations and conditions within a large U.S. population. It also will be very important to include new information, as it becomes available, about the immune protection offered by prior immunizations."

"The daily contacts that people make within their communities could have a significant impact on the speed of smallpox spread and the reach of an epidemic," Dr. Longini adds. "In future models, we will explore additional possibilities, such as the interconnectedness of different communities, the possibility of the original infected individuals moving among more than one community, and different methods of how smallpox might be introduced. We also will explore whether targeted vaccination could be focused on areas where epidemics occur, rather than mass vaccination throughout the country."