The COVID-19 pandemic has given the world a crash course in infectious disease epidemiology like no other. Concepts that were previously unfamiliar to most people, such as reproduction numbers, social distancing, and managed isolation, have now become part of everyday life. You may also have followed the Mycoplasma bovis outbreak in New Zealand since 2017, or the TbFree programme for bovine tuberculosis that has been operating for several decades, and wondered how policy-makers have devised the control strategies used in these programmes.
In the previous module, we introduced the process of investigating and controlling outbreaks at a more localised level, which is something you are likely to encounter regularly if you work in food animal or shelter medicine. In this section, we step up to a much larger population scale and explore the principles that underpin national infectious disease control programmes, so that you can better understand the rationale behind these large-scale responses.
A disease outbreak is broadly defined as the occurrence of cases of disease in excess of what would normally be expected in a defined community, geographical area, or time period. What constitutes “excess” depends on the disease, the population at risk, and the quality of baseline surveillance data available.
Outbreaks can vary widely in both scale and duration. Some are highly localised and short-lived, lasting only days or weeks, such as the 2019 Infectious Bursal Disease outbreak in the New Zealand commercial poultry industry. Others may persist for many years and affect large parts of the country, as seen with the Mycoplasma bovis outbreak that was first detected in New Zealand in 2017. In some situations, particularly where effective control measures are not feasible, a newly introduced disease may become established within the population. An example of this is Theileria orientalis in cattle following its introduction into New Zealand from Australia in 2014.
Epidemiologists use a number of specific terms to describe different patterns of disease occurrence:
Any one of these patterns may be sufficient for authorities to consider coordinated disease control measures at a regional or national level, depending on the severity of the disease and its consequences. In this section, however, the focus is on epidemic outbreak scenarios where a new disease has been introduced into a country or region, and where rapid investigation and intervention are critical to limit spread.
Once the first potential case of an emerging disease outbreak (also called the index case) has been reported to regulatory authorities, time is absolutely of the essence in confirming the diagnosis and identifying any individuals that may have been in contact with the index case (also called contact tracing) to try containing disease spread. This is also an important part of Step 4 surveillance activities in large outbreak investigations.
One of the reasons why the 2001 foot-and-mouth disease outbreak in the United Kingdom escalated so quickly is that a single batch of subclinically infected sheep were sent to market, sold to multiple different buyers, and disseminated across many different locations across the country. By the time officials diagnosed the index case, the wide geographic spread of cases posed significant challenges since there were insufficient numbers of trained personal that could be mobilised to cover all affected areas.
Mycoplasma bovis is another great of example of a situation where delays in identifying and confirming the outbreak resulted in a much wider spread of disease. Although the index case was found in a South Canterbury dairy herd in 2017, it is believed that the first occurrence of the disease in New Zealand (primary case) was actually from a farm in Southland that may have been infected up to two years earlier. Both of these premises had large numbers of cattle movements during the period from disease introduction to detection, which were difficult to trace due to poor data recording in the National Animal Identification and Tracing (NAIT) system.
Another good reason for tracking counts of confirmed cases over time is that we can generate epidemic curves, which provide a lot of key information about how the outbreak is progressing in the population. You will usually see the date on the x-axis and the case counts on the y-axis. Here are the three most common patterns*:
* Source: These examples were taken from the CDC epicurve training module
In a point source epidemic, persons are exposed over a brief time to the same source, such as a single meal or an event. The number of cases rises rapidly to a peak and falls gradually. The majority of cases occur within one incubation period of the disease.
The example here is from a cryptosporidiosis outbreak at a child day care centre.
In a continuous common source epidemic, persons are exposed to the same source but exposure is prolonged over a period of days, weeks, or longer. The epi curve rises gradually and might plateau.
The example here is from Salmonella cases associated with contaminated batches of salami from a meat processing plant that were distributed to many retail grocery chains in a geographic region.
In a propagated epidemic, there is no common source because the outbreak spreads from person-to-person. The graph will assume the classic epi curve shape of progressively taller peaks, each being one incubation period apart.
This example here is from a measles outbreak in a localised geographic region.
We can use the time between peaks to estimate the incubation period (approximate time from initial infection until the development of clinical disease). As we will see in the next section, we can also use the slope of the epidemic curve at the start of the outbreak to estimate the basic reproduction number (R0).
During the early stages of an infectious disease outbreak, regulatory authorities need to learn as much as possible about the pathogenesis of the disease and how the disease spreads between individuals so that they can make more informed recommendations around diagnostic testing as well as other control measures. For diseases that have been around for a very long time, we might already have a lot of this information available from previous outbreaks. For new diseases, however, we might need to collect more descriptive data (Step 5), develop hypotheses (Step 6), and perform additional analytical studies (Step 7) to get the information we need.
When we think about the pathogenesis of a disease (also called the natural history), we are really trying to identify key epidemiological states that an individual may pass through from the time they initially get infected until they either recover or die. We often divide these into:
This type of information is very useful for developing the compartmental models, which describe the set of mutually exclusive epidemiological states that individuals can occupy. Remember that an individual can only have one disease status at a time, but a population contains individuals that are distributed across the different disease states.
Here is an example of a compartmental diagram showing the dynamics of COVID-19 infections. Individuals start in the susceptible state and become infected after exposure to an infectious person. Following infection, there is an average incubation period of around 4 to 5 days before individuals begin shedding virus, although this can be up to 14 days in some cases. People are typically infectious for approximately 8 to 10 days and may remain asymptomatic or develop clinical signs partway through this period. During the infectious phase, individuals can transmit the virus to others. Most infected individuals then recover and develop immunity, while a smaller proportion progress to severe disease and death.
In order for a pathogen to take-off and persist in the population after the initial introduction, it makes intuitive sense that individuals must be getting infected at a greater rate than they are recovering from infection (i.e. allowing the I compartment to fill up).
In an outbreak situation, we usually describe the pathogen transmission dynamics in terms of the basic reproduction number (R0), which is defined as the average number of secondary cases produced by a single infected individual over the course of its infectious period in an otherwise susceptible population. The media has been referring to this as the “R-number” in the COVID-19 pandemic.
As we touched on briefly in previous modules, this depends on the number of direct contacts (like standing in close proximity to someone) and indirect contacts (like contaminating a door handle which is later touched by someone else) that individuals make during the infectious period as well as the probability that the contact will result in disease transmission. The latter depends on agent factors (like virulence of the pathogen), host factors (like immunity, health status, or hygiene practices), and environmental factors (like temperature and cleanliness govern the pathogen load that individuals may be exposed to in their surroundings).
Obviously, the larger the R0 value, the faster the rate of epidemic growth. We can estimate R0 by measuring the slope of the epidemic curve in that early exponential growth phase (upward rise of cases early in the outbreak). Many epidemics will naturally reach a peak before starting to decline, which usually corresponds to a depletion in the number of susceptible individuals in the population.
The main purpose of learning more about the disease pathogenesis and transmission is so that we can develop and implement appropriate strategies for bringing the outbreak under control (Step 8). I find the “bucket and tap” model shown below to be a useful framework for conceptualising how the various different disease control measures we have can help us to work towards eliminating the pathogen from the population. In essence, we are trying to reduce the flow of individuals into the I compartment by lowering the water level in the S compartment and closing the tap between them. We can also lower the water levels in the I compartment by either temporarily or permanently removing infected individuals from the population or speeding up their rate of recovery from infection to reduce the total time they spend infectious.
The “water levels” in each bucket correspond to the number of individuals in the population in that state and the “taps” represent the flow of individuals between those states. Our ultimate goal is to empty the I bucket completely so there are no more infected individuals in the population.
There are 4 main strategies we can use to achieve this:
There are four main strategies we can use to reduce the number of individuals in a population that are susceptible to the disease at any given point in time:
You may have heard of “chicken pox parties” where parents intentionally expose their children to another child with chicken pox since the virus causes a mild short-lived clinical episode if you get infected when you are young, but can much more severe disease like shingles if you get infected as an adult. In the livestock world, you may have heard of farmers intentionally leaving an animal that is persistently infected (PI) with BVD in the herd to trigger natural immunity, but this is not recommend because of the significant impacts this disease has on animal health, welfare, and production. Improving an individual’s general health status through better nutrition and appropriate housing conditions can also work to make the immune system stronger against resisting pathogens. Although there has been some success with breeding sheep that are more naturally resistant to certain parasite infections, genetic selection is definitely not a control strategy that you can implement overnight.
There are two formula we can to use to estimate the minimum % of individuals in the population that must be immune to the pathogen to effectively control the outbreak based on the R0 value. The only difference with the vaccine coverage one is that we also need to account for the fact that the vaccines are not 100% effective and so we need to vaccinate a larger % of individuals to ensure that enough will end up with adequate immunity.
The two main strategies we can use to prevent susceptible individuals from getting infected are:
Both of these strategies are particularly effective in outbreak situations because are they are generally not pathogen specific – i.e. washing your hands will stop you from getting a myriad of bacteria, viruses, and parasites that you may have pick up from surfaces whereas something like a flu vaccine will only work against flu.
The two main strategies we can use to reduce the total number of infected individuals in the population are:
Obviously culling is not an option for human populations (!), but it’s something that is used frequently to control animal disease outbreak either because the pathogen is incredibly infectious and there is is good chance that other farms will be infected before recovery takes place (i.e. foot-and-mouth disease) or because the diagnostic tests to identify individual infected animals are too unreliable or too expensive (i.e. Mycoplasma bovis in cattle or avian influenza in commercial poultry) and so it is logistically easier just to cull all the animals on an infected premises. Some disease like bovine tuberculosis or Johne’s disease are incurable so the only way to get rid of infected animals in the population is through culling.
There is one main strategy we can use to reduce the number of individuals in a population that are susceptible to the disease at any given point in time:
Our ability to rely on treatments as infectious disease control strategy is becoming increasingly limited in the age of drug resistance. A lot of medications, particularly the antivirals, are also really expensive and may not even be an option for use in veterinary medicine. The most successful example from human medicine was the use of antiviral agents to treat HIV. Although the drugs can’t completely cure the disease, they are able to reduce viral loads to an almost negligible level which significantly reduces the risk of onward transmission and significantly delays the progression to AIDS. For a while after the introduction of HIV drugs, the prevalence of HIV actually increased despite the greatly reduced incidence of HIV simply because infected individuals were now surviving for much longer periods.
In most outbreak situations, it often requires a combination of approaches to get the R0 value sufficiently below 1 to bring the disease under control. In the next section, we’ll very briefly discuss how policy-makers make decisions about which measures to implement.
Steps 3 through 7 of the outbreak investigation process are vital for helping regulatory authorities gather the evidence they need to decide on an initial action plan for tackling the outbreak (Step 8) and then monitoring the results to verify whether or not the control measures are having the desired effect (Step 9). It’s not always easy for policy-makers because there is often considerable tension between what the science shows will be the most effective control measures and what society will accept as a national disease control programme. Clear communication (Step 10) at all stages of the outbreak is key for building trust and compliance with policy recommendations.
Once the biological characteristics of a disease and the available control tools are understood, the next step is to define the overall objective of the control programme. This decision reflects a balance between what is theoretically possible from an epidemiological perspective and what is socially, politically, and economically acceptable.
In practice, disease control programmes tend to fall into one of five broad objective categories.
Different countries adopted different objectives for managing COVID-19, reflecting differences in population size, geography, healthcare capacity, and tolerance for social and economic disruption. Many countries focused on controlling the outbreak by flattening the epidemic curve to keep case numbers below healthcare system capacity. New Zealand, by contrast, pursued an elimination strategy. This was possible due to a combination of factors including geographic isolation, early border controls, low initial case numbers, and high levels of public compliance with control measures.
One of the most common tools used to support decision-making in infectious disease outbreaks is computer simulation models. These basically try to create a virtual representation of how an infectious disease spreads through a population based on what we know about the pathogen and how individuals in a population typically interact with each other over space and time. We can then test out many different potential combinations of disease control programmes to predict which ones are likely to be the most safe, effective, and practical for getting the outbreak under control. It’s another way of allowing us to test hypotheses without all the expense and impracticality of doing this kind of research in real world.
The easiest way to monitor the effects of interventions for an outbreak is to continue recording the daily case counts to generate our epidemic curve where you will hopefully start to see reductions in the numbers of individuals getting infected. Remember that we can sometimes see an initial increase in the number of cases as regulatory authorities ramp up testing and people become more aware of the need for testing if they develop compatible symptoms.
In this section, we have reviewed the process of conducting outbreak investigations at a national scale, using examples from the COVID-19 pandemic. Strong surveillance systems play a critical role in detecting outbreaks early, allowing epidemic curves to be constructed and data to be gathered to better understand disease dynamics within the population. By combining this information with knowledge of disease pathogenesis and available control measures, it is possible to identify strategies that are most likely to reduce R0 sufficiently to meet programme objectives. These objectives can range from taking no coordinated action through to full eradication, depending on biological, technical, and socioeconomic constraints. As with any clinical problem, ongoing monitoring and clear communication with stakeholders are essential to ensure that interventions remain effective and appropriate over time.
4. Herd Outbreak Management
Advancing Animal Care through Shared Learning