Viral Epidemiology

Epidemiology is the study of the distribution, dynamics and determinants of disease in populations, and its purpose in virology is practical: to provide a rational, quantitative basis for explaining why viral disease occurs where and when it does, and for directing control. Because a virus can only replicate inside a living cell, it survives in nature by maintaining an unbroken chain of transmission, so almost every epidemiological pattern reduces to a question about that chain: how fast the virus moves through susceptible hosts, and whether a continuous supply of them can be sustained. The measures and study designs below are the tools for answering it, distinct from the routes and mechanics of transmission themselves.

Measures of disease frequency

Epidemiological rates express the number of events in a standard population, for example per 1,000 or per 100,000, and often per unit of time. The denominator is the population at risk, which for many viruses is not the whole population but the susceptible fraction, usually taken as those lacking antibody to the virus. Two systematic gaps separate a measured rate from the true burden of infection. A rate based on the whole population understates the rate among susceptibles, and, because many infections are subclinical, the incidence of clinical disease is always lower than the incidence of infection. That gap can be extreme: measles is clinically apparent in almost everyone infected, whereas fewer than 1% of those infected with poliovirus or an encephalitis-causing arbovirus develop paralysis or encephalitis.

Measure	Definition	Best used for	Viral example
Incidence (attack rate)	New cases arising in a population over a defined period, per standard population and time	Acute infections of short duration	Influenza cases per season
Secondary attack rate	New cases among susceptible contacts of an index case within one incubation period, as a percentage of those exposed	Infectiousness in closed groups such as households	Chickenpox ~70%, zoster ~15%
Prevalence	Cases present in a population at one point in time, divided by the population size	Chronic or persistent infection	Chronic hepatitis B carriage
Seroprevalence	Proportion of a population with antibody to a virus; reflects cumulative past exposure	Mapping how far and how long ago a virus has spread	Cytomegalovirus, Epstein-Barr virus
Cause-specific mortality rate	Deaths from the disease in a year, divided by the mid-year population	Population burden of a lethal infection	AIDS deaths per 100,000
Case-fatality rate	Percentage of people with the disease who die of it	Severity, or virulence of the agent	Rabies approaching 100%

Incidence and prevalence answer different questions. Incidence counts new events over time and suits acute infections, where each case is a discrete, short episode. Prevalence is a snapshot of how many cases exist at one moment, and suits chronic or insidious infection, where onset is hard to date; it is a function of both the incidence and the duration of disease, so a rise in prevalence may mean more new infections or simply longer survival. Seroprevalence is a special case: because neutralising antibody often persists for life, a seroprevalence survey measures the cumulative experience of a population rather than current activity, and reading it by age group reveals how efficiently and how recently a virus has circulated.

Patterns of occurrence

The vocabulary of disease activity describes how a virus behaves against its own background rate. A sporadic infection occurs as scattered, unrelated cases. An endemic infection is maintained at a steady baseline in a population, the expected rate against which change is judged. An epidemic is a rise in incidence that exceeds that endemic baseline, spreading unusually widely and rapidly; the size of peak that counts as an epidemic is arbitrary and depends on the background rate and on how often infection is clinical, so a handful of encephalitis cases may be called an epidemic while the same number of colds would not. A pandemic is an epidemic spread across continents, the classic viral example being an influenza pandemic that follows the emergence of a subtype against which the population has little immunity.

A distinct pattern arises when a virus reaches a population with no prior experience of it. In such a virgin-soil epidemic the attack rate approaches 100% and, unusually, most deaths fall among adults rather than the very young or old. The severity owes less to any special genetic susceptibility than to the sheer proportion of the community infected at once, the collapse of care when carers themselves are ill, and the disruption of daily life. Peter Panum’s study of measles in the Faroe Islands in 1846 captured both the near-total attack rate and a second lesson about immunity: islanders who had survived the previous epidemic 65 years earlier remained solidly protected, despite no chance of reinforcement in between.

The shape of an outbreak over time reflects how it is spread. A point-source outbreak, in which many people are exposed to a single contaminated vehicle such as food or water, produces a sharp, compressed burst of cases. A propagated outbreak, spread person to person, climbs more gradually across successive generations of infection. The epidemic curve that distinguishes these two shapes is central to recognising and investigating an outbreak.

Epidemiological study designs

Establishing that a particular virus causes a particular disease, or that an intervention works, requires structured comparison. The main observational designs differ in their direction in time and in what they can measure.

Design	Direction in time	Measures	Strength	Main limitation
Cross-sectional	Snapshot at one time point	Prevalence	Fast and inexpensive	Shows association, not time sequence or cause
Case-control	Retrospective, from disease back to exposure	Odds of prior exposure	Efficient for rare or slow diseases	Vulnerable to recall and selection bias
Cohort	Prospective, from exposure forward	Incidence, risk, vaccine efficacy	Strong evidence of cause and effect	Slow and expensive
Long-term population	Prospective over many years	Natural history, late and chronic effects	Reveals long-term outcomes of infection and intervention	Very slow; demanding to sustain

A cross-sectional study measures prevalence and can be done quickly, but it captures cause and effect together and cannot show which came first. Its results must be read with care, because populations are rarely uniform: a small high-risk subgroup may contribute most cases, and a pre-selected group such as blood donors is not representative of the whole. A case-control study is retrospective, starting from people who already have the disease and comparing their past exposures with those of matched controls; it is efficient for uncommon diseases but depends heavily on choosing controls that avoid bias. A cohort study is usually prospective, following an exposed population forward to see who develops disease, and it gives the strongest evidence of cause and effect, which is why it is essential for defining the safety and efficacy of a new vaccine. Its cost is time and expense.

The discovery of congenital rubella syndrome shows the designs working in sequence. The ophthalmologist Norman Gregg, noticing an unusual cluster of infants with cataracts and heart defects in Sydney in 1941, interviewed the mothers and found that most had contracted rubella early in pregnancy. That retrospective, case-control-style observation generated the hypothesis, which prospective cohort studies then confirmed by comparing malformation rates between women who did and did not have rubella in pregnancy. Human volunteer studies once filled the gap where no animal model existed, as for yellow fever, hepatitis and the common cold, and they remain part of the history of the field; the Willowbrook hepatitis studies of the 1960s, which distinguished hepatitis A from hepatitis B, are also the standard example of research that would be judged unethical today and would not pass a modern ethics committee.

Serological and molecular epidemiology

Two laboratory approaches turn stored samples and sequence data into epidemiological evidence. Sero-epidemiology examines sera for antibody, giving a more accurate measure of true infection than clinical reporting because it counts subclinical infections as well. Detecting antibody across age groups shows how effectively a virus has spread and, for a virus not currently circulating, how long since it last appeared; setting serology against clinical observation yields the clinical-to-subclinical ratio. Serological surveys can estimate the total burden of past infection, gauge the impact of an immunisation programme, identify occupational risk, and help decide whether a newly recognised infection is genuinely new or merely newly detected. Sera come from surveys, blood banks and diagnostic laboratories, and sentinel animals extend the same logic to arboviruses, with sentinel chickens used to detect the local circulation of encephalitis viruses.

Molecular epidemiology reads the viral genome to answer questions that serology cannot. Sequencing can distinguish a vaccine strain of poliovirus from a wild one and track its reversion toward virulence, assign a rabies isolate to a geographic region and reservoir species, and identify the origin of an outbreak strain: the West Nile virus that appeared in North America in 1999 matched a strain then circulating in the Middle East, pointing to its likely source. It can separate a common-source cluster from unrelated coincidental cases, detect a new influenza strain, and determine the antiviral susceptibility of an isolate. This rests on phylogenetic methods that reconstruct viral ancestry from sequence.

Both approaches feed surveillance, the ongoing collection of case data that allows unusual activity to be recognised. Surveillance may be passive, relying on clinicians to report notifiable conditions, or active, using dedicated reporting such as the international networks that monitor influenza for new antigenic variants and the acute-flaccid-paralysis system that underpins polio eradication. Surveillance is the routine measurement from which every other epidemiological signal is drawn, and its operational systems run from case detection through to coordinated outbreak response.

The reproduction number and herd immunity

The central quantity in the dynamics of an infection is the basic reproduction number, R₀, the average number of secondary infections produced by one infected individual in a wholly susceptible population. Its meaning is a threshold: if R₀ is under 1 the chain of transmission cannot be sustained and an outbreak dies out, whereas if R₀ is over 1 cases multiply until the pool of susceptibles is depleted. R₀ is not a fixed property of the virus alone but of the virus in a population, combining how transmissible the agent is, how long an infected person remains infectious, and how densely and often hosts contact one another. These components of R₀ each depend on the mechanics of spread. Once part of a population is immune, the relevant quantity becomes the effective reproduction number, R, equal to R₀ multiplied by the susceptible fraction, and control measures succeed precisely when they hold R below 1.

This threshold is the basis of herd immunity: once enough of a population is immune, each case produces fewer than one further case on average, and infection cannot spread even though susceptible individuals remain. The proportion that must be immune, the herd-immunity threshold, follows directly from R₀ as 1 minus 1/R₀, so the more transmissible the virus, the higher the coverage required. A virus with an R₀ of 4 needs about three-quarters of the population immune; measles, with an R₀ around 12 to 18, needs roughly 92 to 95%, which is why measles control demands such high and sustained vaccine coverage and is the first infection to return when coverage slips.

Virus	R₀ (approximate)	Herd-immunity threshold
Measles	12 to 18	~92 to 95%
Smallpox	5 to 7	~80 to 85%
Poliomyelitis	5 to 7	~80 to 86%
Rubella	6 to 7	~83 to 85%
Mumps	4 to 7	~75 to 86%
SARS (2003)	~3	~65 to 70%

Persistence of viruses in populations

Whether a virus can maintain itself indefinitely depends on how it survives between hosts, and viruses fall into a few broad patterns. Some cause acute, self-limiting infection followed by lifelong immunity, with no reservoir beyond the chain of infection itself; measles, mumps, rubella and poliovirus are examples, and this pattern demands a large, continuously renewed pool of susceptibles. Others cause acute infection with shorter-lived or narrower immunity, such as the respiratory syncytial virus, rotavirus and influenza, where reinfection is possible and antigenic change or multiple serotypes keep susceptibles available, so the virus can persist in smaller populations. A third group establishes persistent infection with a human reservoir, either latent with intermittent reactivation, as the herpesviruses do, or continuously replicating, as with HIV and hepatitis B and C; an infected person can then be a source for years or for life.

For the acute, immunity-conferring viruses, persistence sets a population threshold. The critical community size is the minimum population needed to maintain an infection indefinitely, because once every susceptible has been infected and made immune the virus fades out unless new susceptibles, mainly newborns, arrive fast enough. Measles requires a population of around 500,000 to remain endemic, which implies it could not have persisted before human settlements reached that size, and epidemics in smaller communities burn out and depend on reintroduction from outside. The contrast with chickenpox is instructive: despite conferring lifelong immunity it persists in communities of under 1,000, because varicella-zoster virus establishes latency and reactivates decades later as zoster, reseeding chickenpox in a new generation and so bridging the gap between epidemics without needing a large susceptible pool.

For many viruses persistence depends not on the human population at all but on a non-human reservoir. Most arboviruses, the rabies virus and the hantaviruses are maintained in animal hosts and reach humans only incidentally, so their distribution and abundance track the reservoir rather than human immunity. The existence and extent of an animal reservoir is decisive for any plan to eliminate a virus, because a human disease with an animal reservoir cannot be eradicated by controlling human infection alone. These reservoir cycles, and the drivers that bring them into contact with people, determine where and how often such viruses reach humans.

Overlaid on all of this is seasonality. In temperate climates most respiratory viruses peak in winter and most arbovirus infections in summer, when their vectors are active, while herpesviruses and the sexually transmitted viruses show no seasonal pattern. The drivers are partly physical, since enveloped respiratory viruses survive better in cool, and for some in low-humidity, air, and partly social, since winter crowding indoors concentrates respiratory spread. In the tropics the wet and dry seasons replace summer and winter, with measles and chickenpox often peaking late in the dry season and falling sharply once the rains begin.

South African context

South Africa’s epidemiological signals are dominated by the interaction of common viral infections with a large HIV-affected population, which raises the incidence, severity and duration of many other infections and shifts their age distribution. The country’s surveillance and reference-laboratory functions sit with the National Institute for Communicable Diseases (NICD), which runs sentinel programmes for influenza, arboviruses and other pathogens and publishes the communicable-disease data that define local endemic baselines. A defined list of notifiable medical conditions must be reported to the Department of Health under the National Health Act, providing the passive-surveillance backbone for recognising outbreaks. South Africa’s contribution to global polio eradication runs through its acute-flaccid-paralysis surveillance, the sensitive case-finding system that must keep detecting and investigating paralysis to demonstrate the continued absence of wild poliovirus.

References and recommended reading

• Burrell CJ, Howard CR, Murphy FA. Epidemiology of Viral Infections. In: Fenner and White’s Medical Virology, 5th edition, Chapter 13. Academic Press / Elsevier; 2017. The foundational backbone for this article: measures of disease frequency, study designs, sero- and molecular epidemiology, and the dynamics and persistence of viral infection in populations.
• van Seventer JM, Hochberg NS. Principles of Infectious Diseases: Transmission, Diagnosis, Prevention, and Control. In: International Encyclopedia of Public Health, 2nd edition, Volume 6. Elsevier; 2017. p. 22–39. A public-health framework source for the reproduction number, herd-immunity thresholds and the R₀ values tabulated here.