Principles of Virology, Volume 2. S. Jane Flint

(such as the 1793 epidemic of yellow fever in Philadelphia) may have a high incidence but a low prevalence, because many of the infected individuals either died or cleared the infection. In contrast, a virus that can persist in a host for decades is likely to have high prevalence. An example of high prevalence is provided by hepatitis B virus; of the 300 million to 400 million people infected globally, one-third live in China, with 130 million carriers. For this reason, incidence is an in formative measure for acute or highly lethal infections, whereas prevalence is often used to describe long-lasting or persistent infections.

BOX 1.4

      DISCUSSION

       Video games model infectious-disease epidemics

      The hugely popular online video game World of Warcraft became a model for the transmission of viral infections. In 2005, a dungeon was added to the fantasy world in which players could confront a powerful creature called Hakkar. In his death throes, Hakkar hit foes with “corrupted blood,” infected with a virus that killed the virtual player. The infection was meant to affect only those in the immediate vicinity of Hakkar’s corpse, but the virus spread as players and their virtual pets traveled to other cities in the game. Within hours after the software update, a full-blown virtual epidemic ensued as millions of characters became infected.

      Although such games are meant only for entertainment, they do model disease spread in a mostly realistic manner. For example, as in real life, the spread of the virus in Hakkar’s blood depended on the ease of travel within the game, zoonotic transmission by pets, and transmission via asymptomatic carriers. Moreover, such games have a large number of participants, at one point more than 10 million for World of Warcraft, creating an excellent community for experimental study of infectious diseases. The players’ responses to dangerous situations approximated real-world reactions. For example, during the “corrupted-blood” epidemic, players with healing ability were the first to rush to the aid of infected players. This action probably affected the dynamics of the epidemic because infected players survived longer and were able to travel and spread the infection. A more reality-based smart phone app called Plague Inc., downloaded more than 85 million times, asks: “Can you infect the world?” and gives players the opportunity to choose a pathogen and influence its evolution. Players compete against the clock, trying to destroy humanity before the world can develop a cure.

      Scientists themselves recognize the educational value of such games. A professor at Drexel University developed CD4 Hunter, in which players enter the bloodstream as a human immunodeficiency virus type 1 particle. The goal is to find and infect CD4⁺ T cells, white blood cells of the adaptive immune system that are the main targets in this infection. The game mimics virus binding and entry, and was created as a supplementary teaching tool for graduate students and undergraduates in advanced-level courses (http://bit.ly/Virology_Twiv489).

      With all of these games, successful players learn to integrate multiple variables simultaneously, including environment, time, and population density. These applications also demonstrate how the reproductive cycle of a virus may change over the course of an epidemic. However, the parallels to real-world epidemiology end there; a defeated player can begin again with the click of a button or the flick of a finger. Alas, real life does not come with “do-overs.”

       Lofgren ET, Fefferman NH. 2007. The untapped potential of virtual game worlds to shed light on real world epidemics. Lancet Infect Dis 7:625–629.

       Prospective and Retrospective Studies

      Although infections of natural populations differ from those under controlled conditions in the laboratory, it is possible to determine if one or more variables affect disease incidence and viral transmission in nature. Two general experimental approaches are used: prospective (also called cohort or longitudinal) and retrospective (or case-controlled) studies. In prospective studies, a population is randomly divided into two groups (cohorts). One group then gets the “treatment of interest,” such as a vaccine or a drug, and the other does not. The negative-control population often receives a placebo. Whether a person belongs to the treatment or placebo cohort is not known to either the recipient or the investigator until the data are collected and the code is broken (“double blind”). This strategy removes potential investigator bias and patient expectations that may otherwise influence data collection. Prospective studies require a large number of subjects, who often are followed for months or years. The number of subjects and time required depend on the incidence of the disease or side effect under consideration and the statistical power, the probability of detecting a difference that is sufficiently significant to draw conclusions.

BOX 1.5

      TERMINOLOGY

       Morbidity, mortality, incidence, and case fatality

      The terminology used to calculate the number of people who are infected and/or who become ill following a viral outbreak can be confusing. The following fictional example will be used to clarify these definitions.

      Imagine that, in a city of 200,000 residents, a virus causes infection of 50,000 persons (as determined by serology). Of these, 20,000 develop signs of illness and 10,000 die of the infection.

       The incidence of this infection is the number of people infected divided by the population (50,000/200,000, or 25%).

       Morbidity rate is the number of individuals who became ill divided by the number of individuals at risk (20,000/200,000, or 10%).

       Mortality rate is the number of deaths divided by the number of individuals who are at risk (10,000/200,000, or 5%).

       The case fatality ratio is the proportion of deaths within a population of infected individuals. This value is typically expressed as a percentage. Case fatality ratios are most often used for diseases with discrete, limited time courses, such as outbreaks of acute infections. In the above example, the case fatality ratio is the number of deaths divided by the number of individuals with illness (10,000/20,000, or 50%).

      Representation of incidence, morbidity, and mortality rates in a population. Each person represents 10,000 members of a community, as in the example above. Orange individuals are those who are infected; red are those who show symptoms of infection; the coffin indicates those who have died of the infection.

      As a real-world example, Nipah virus infection and resulting encephalitis in Southeast Asia in 2011–12 resulted in 280 cases and 211 deaths, a staggering case fatality ratio of 75%.

      In contrast, retrospective studies are not encumbered by the need for large numbers of subjects and long study times. Instead, some number of subjects with the disease or side effect under investigation is selected, as is an equal number of subjects who do not have the disease. The presence of the variable under study is then determined for each group. For example, in one retrospective study of measles virus vaccine safety and childhood autism, a cohort of vaccinated children and an equivalent cohort of age-matched, unvaccinated children were chosen randomly. The proportion of children with autism was then calculated for each group to determine if the rate of occurrence of autism in the vaccinated group was higher, lower, or the same as in the unvaccinated group. The incidence of the side effect in each group is then calculated; the ratio of these values between groups is the relative risk associated with vaccination. In this example, the rate