Lexicon of Disease Spread Modelling terms
This document is based on a copy of the technical publication "Lexicon of disease spread modelling terms" originally published in Rev. sci. tech. Off. int. Epiz., 2011, 30(2), 547-554. It is provided here for convenient use in the Animal Disease Spread Model program under fair use. Formatting has been modified to fit the wiki format. ADSM specific entries have been added.
K. Patyk(1), C. Caraguel(2), C. Kristensen(1)& K. Forde-Folle(1)
(1) United States Department of Agriculture, Animal and Plant Health Inspection Service, Veterinary Services,
Centers for Epidemiology and Animal Health, 2150 Centre Avenue, Building B, Fort Collins, CO 80526-8117,
United States of America
(2) Canadian Food Inspection Agency, hosted at the Centre for Veterinary Epidemiological Research, Atlantic
Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, Prince Edward
Island, C1A 5H6 Canada
Over the past decade there has been a notable increase in the magnitude and variety of modelling work in the realm of animal health. Similarly, there has been an increase in the extent to which modelling is used as a component in the development of animal health policy. The increased dependency on modelling creates a need to enhance understanding and linkages between policy-makers (those that pose the policy or scientific questions, commission modelling work and use model outputs in the development of policy), intermediaries (those that are responsible for working with modellers and communicating model results to policy-makers), and modellers. Development of a lexicon of disease spread modelling terms can help support clear communication and collaboration between all players.
A contact between individuals that would lead to the transmission of disease if one individual is infectious and the other is susceptible (9). See also effective contact.
Synonyms: Animal Disease Spread Model
The current development fork of NAADSM being developed through the USDA
as a group effort between the United States and Canada. ADSM has an updated User
Interface based on web technology and can be used as a server to run many iterations
in parallel.
The distribution of microbial aerosols consisting partially or completely of microorganisms which can be drawn into lung alveoli. This type of transmission includes transmission by droplet nuclei and dust (6). See also direct transmission; indirect transmission; transmission.
A graph plotting the number of newly ‘detected’ cases per unit time (5). The number of ‘detected’ cases may be fewer than the number of ‘actual’ cases when detection is imperfect. See also actual epidemic curve; epidemic curve.
Unknown or uncharacterised information used in the model that is thought to be true. Models are typically based on assumptions that help simplify aspects of the system being modelled. It is recommended that all assumptions be explicitly stated and constantly re-assessed. ‘Hidden assumptions’ that could possibly impact the model outcome should also be described. If the assumptions used are incorrect, then output from the model is suspect (9).
The stochastic, spatial, state transition disease spread model that has been developed by the Australian Government Department of Agriculture, Fisheries, and Forestry to support exotic disease preparedness in Australia (3).
Synonyms: Trace back, Trace-in
To track or identify movements of animals, personnel,
vehicles, or equipment (i.e. direct and/or indirect contacts)
to a detected, infected premises for the purposes of
identifying a potential source of infection (13). See also
forward trace; tracing.
Synonyms: Ro
The number of secondary cases generated by the
introduction of an infectious individual into a completely
susceptible population (8). See also reproductive number.
A probability density function used to describe the uncertainty or variation around a probability or proportion. This distribution is defined by two parameters: alpha (), and beta () (15). See also probability density function.
A probability density function, derived from the Beta distribution, used to describe the uncertainty or variability of any continuous parameter estimate. This distribution is defined using three parameters: the minimum, maximum, and most likely (or mode) values. Often, subject matter experts are consulted to estimate the distribution’s parameters. Compared to a triangular distribution, a BetaPERT distribution provides less weight to extreme values (15). See also probability density function.
A disease state characterised by the presence of clinical signs and shedding of the disease agent (5).
A spatial and/or temporal concentration or group of individuals or observations. It is expected that individuals or observations within a cluster are more similar than those among clusters (10).
See face validity.
Disease is spread from one unit to another based on contact rate and the probability of infection transfer, which are set for each pair of production types. A contact rate is used to indicate the average number of contacts (shipments of animals in the case of direct contact, or movements of people, equipment, etc., for indirect contact) that are generated by each herd/flock on each day. For each unit that can infect others, the model simulates a number of outgoing shipments. A distance in chosen for each shipment from a probability density function of movement distances.
A series of actions, efforts, or measures taken to impede the progression of, reduce the frequency of, or eradicate, disease in a population. Control measures limit the probability of transmission between susceptible and infectious herds/flocks by either isolation (e.g. movement restrictions, quarantine) or decrease the production of the disease agent (e.g. treatment, culling), and/or increase the resistance of susceptible herds/flocks (i.e. vaccination) (10, 13).
Destruction, also known as stamping out or depopulation may be used in the simulation as a method of disease control.
A mathematical model in which the model input parameters are point estimates. This type of model does not incorporate random variation or uncertainty (13). See also stochastic models.
In ADSM, infected (i.e. latent, subclinical, or clinical) or naturally immune units identified by forward and/or backward tracing, can be detected by diagnostic testing. A diagnostic test in ADSM is characterized by its herd-level sensitivity and specificity.
The movement of animals within units (premises, section, pen) or from one unit to another unit with animals (5). See also adequate contact; effective contact; indirect contact.
The transfer of a disease agent by direct or close contact (14). See also airborne transmission; indirect transmission; transmission.
ADM is a state transition model. Each unit in a simulated outbreak exists in one of several disease states: susceptible, latent, subclinical, clinical, naturally immune, vaccine immune, or destroyed. The disease states that a unit moves through are described by the disease progression. During a simulation, units transition through these disease states according to user-defined parameters. For each production type, the user defines the length in days of these periods using probability distributions.
Models that incorporate epidemiological principles to simulate the spread of disease within a population. Epidemiological modelling provides a useful means of simulating disease spread when experiments and field observations are limited or impractical. Examples of disease spread models include: AusSpread, InterSpread Plus, and NAADSM.
A series of ordered stages which follow the progression of an infectious disease. The identification of the different disease states and the transition through them are fundamental to state transition modelling. Examples of various disease states include: latently infected, subclinically infectious, clinically infectious, and naturally immune.
A summary of the frequencies of each value or range of values out of all possible values (6). from detected, infected premises to other premises for the purposes of identifying the potential spread of infection (13). See also backward trace; tracing.
A procedure used to select a theoretical distribution which best describes empirical data within a sample dataset (15).
A contact between an infectious individual and a susceptible individual that leads to disease transmission (9). See also adequate contact.
A summary of the frequencies of each observed value or range of observed values within a sample dataset. Data points are directly applied to define an empirical distribution. Empirical distributions may be used in modelling when the available data are not amenable to fitting of theoretical probability density functions. See also probability density function; theoretical distribution.
The occurrence of disease in a population or region, in excess of normal expectancy (4).
A graph plotting the number of new cases per unit time (13). See also actual epidemic curve; apparent epidemic curve.
A herd exam is an inspection by trained experts of animals on a premises for clinical signs of disease. In ADSM, herds identified by forward and/or backward tracing can be detected as infected by a herd exam.
Information gathered through consultation with individuals recognised as knowledgeable regarding the subject at hand, to inform input parameter values for disease spread models. Expert opinion is based on personal experience, opinion, and assumptions and is an important source of information when data are needed but not readily available through other sources (12).
Synonyms: Conceptual validity
The process of determining if the interrelationships of the
real system are adequately represented by the model and if
the model outcomes make biological sense (11, 7). See also
validation.
Synonyms: Trace forward, Trace-out
To track or identify the movements of animals, personnel,
vehicles, or equipment (i.e. direct and/or indirect contacts)
Massey University EpiCentre. Further information is
available at: www.interspreadplus.com.
Synonyms: Recovered
A disease state characterised by the absence of clinical signs
and termination of agent shedding. See also natural immunity;
vaccine immunity.
The number of new cases of disease that occur in a population during a given time period (4).
An early disease state characterised by the presence of the agent in the herd/flock and by the absence of clinical signs. The state begins with initial exposure to the disease agent and ends at the first appearance of clinical signs. An incubating herd or flock may or may not be infectious (6). See also latent state.
Indemnity payments are compensation payments made to individuals or organizations to compensate for the loss or destruction of certain animals. Parameters to estimate the costs for indemnification can be included in ADSM.
USDA-APHIS-VS Available at: http://www.aphis.usda.gov/mrpbs/fmd/downloads/Veterinary_Services_Indemnity_Payments.pdf
The movement of people, vehicles, equipment, etc. from one premises to another premises with animals (5). See also adequate contact; direct contact; effective contact.
The transfer of disease agent via movement of personnel, vehicles, equipment, etc. (14). See also airborne transmission; direct transmission; transmission.
A disease state characterised by the shedding of disease agents with or without the presence of clinical signs. See also clinically infectious; subclinically infectious.
Synonyms: Model input, Parameter
Values, sometimes in the form of distributions, that are
used in models to represent a component of the processes
or events being modelled. Common examples of animal
disease spread model input parameters include herd and
flock locations, distributions representing the duration of
various disease states, and the number of contacts that
occur between premises.
A spatially explicit, stochastic simulation used to model disease among animal populations developed by the
Synonyms: Realisation, Run
The process of achieving results through the successive
repetition of a sequence of steps (1).
A disease state characterised by the period of time that elapses between exposure to a disease agent and the onset of infectiousness (shedding of disease agent) (8). See also incubation period.
A measure of central tendency. The average value for a set of data calculated by summing all of the sample values and dividing by the total number of samples (13).
A measure of central tendency. The middle value for a set of observations arranged from lowest to highest such that half of the observations fall below the value and half of the observations are above the value (13).
A measure of central tendency. The most common value in a set of observations (6).
A representation of a system or process that is intended to simulate reality. Models are commonly used to investigate how the system may respond to different interventions (12).
See stability.
A simulation model in which input values are sampled randomly from input distributions for each iteration of the model (10).
Synonyms: North American Animal Disease Spread Model
A stochastic, state transition model which incorporates
spatial and temporal information to simulate the spread
of highly contagious animal diseases. NAADSM provides a
framework for the construction of epidemiological
simulation models. NAADSM is available at:
www.naadsm.org (5).
A disease state in which herds/flocks are immune due to natural progression through the disease states (i.e. previous exposure to the pathogen) (5, 6). See also immune; vaccine immune.
See input parameter.
The population is the collection of all units of varying, user-defined production types that will be included in a simulation.
The range in which an observation is expected to exist with a specified probability based upon past observations (2).
An area of land and the animals on it. Disease spread is sometimes modelled between premises, herds, flocks or pens rather than between individual animals, consequently certain input parameters (e.g. direct and indirect contact frequencies and distances) are established to represent between-premises dynamics.
The proportion of the population that is affected by a disease or a condition at a specific time (4).
A distribution of values for a continuous variable that is representative of the range of possible values for a disease spread model input parameter (2, 5).
For a discrete random variable, a function (e.g. Poisson distribution) that gives the probability of each value of the variable. For a continuous random variable, a curve (i.e. probability density function) used to specify the probability of a value for the variable falling within the frequency distribution (6). See also probability density function.
Disease is spread from one unit to another based on contact rate and the probability of infection transfer, which are set for each pair of production types. The probability of infection transfer is the probability that, if a contact occurs, it will be adequate.
In ADSM, a production type is a user-defined collection of herds or flocks with similar disease transmission probabilities, disease manifestation, disease detection probabilities, and control strategies. Production types are typically defined by animal species and/or by the management practices that are applied to each herd/flock. Production types may be broad (e.g., cattle, swine, poultry) or narrow (e.g., beef sale yard, beef cow-calf operations with fewer than 50 animals, beef cow-calf operation with more than 50 animals, or beef feedlot), depending on the needs or preferences of the modeler.
See iteration.
See immune.
The process of comparing the model outcomes of a disease scenario applied to several independently developed models to identify similarities and differences in results (7). See also validation.
Synonyms: R
The number of secondary cases generated by an existing
case. R=R0(S/N) where S is the number of susceptible
individuals and N is the total number of individuals in the
population (8). See also basic reproductive number.
ADSM specific definition - A ring is an area of action specifically for the control activities depopulation and vaccination. The ring is radial and the user defines the radius. A ring does not get named and cannot be distinctly identified in the results outputs. A ring is not the same as a Zone.
See iteration.
A collection of input parameters applied to an epidemiological simulation model to describe a biological system, event or process.
The process of assessing the influence of uncertain input parameters or assumptions on model outcomes (12).
A method for modelling the spread of infectious diseases which is similar to SIR models but adds an additional disease state, exposed (E). Members of the population exist within one of four disease states: susceptible (S), exposed (E), infectious (I), or recovered (R), corresponding to the letters of the acronym for which the model is named, and transition from one disease state to the next (8).
A method for modelling the spread of infectious diseases. Members of the population exist within one of three disease states: susceptible (S), infectious (I), or recovered (R), corresponding to the letters of the acronym for which the model is named, and transition from one disease state to the next (8, 12).
Synonyms: Model stability
The reliability of a model, i.e. the consistency with which
it is free of errors, especially those affecting model accuracy
and results (7).
A type of model that takes into account the variability and uncertainty of biological processes by using distributions as model input parameters. These models produce a range of possible outcomes and the probability of any particular outcome (13). See alsodeterministic models.
A disease state in which there is an absence of clinical signs but in which disease agent is being shed (5).
A disease state characterised by the capacity to become infected. Individuals in this disease state are neither infected, naturally immune, nor vaccine immune (8).
A distribution based upon mathematical formulas and calculations, with a defined shape and specific parameters. Statistical methods (i.e. goodness of fit tests) are often used to select theoretical distributions which best fit the observed data. See also empirical distribution.
See backward trace.
See backward trace.
See forward trace.
See forward trace.
To track or identify the movements of animals, personnel, vehicles, equipment, etc. (direct and/or indirect contacts) to and from an infected premises for the purposes of detecting the source of infection and to identify potential disease spread to other premises (13). See also backward trace; forward trace.
A process by which a disease agent is passed from one individual to another (6). See also airborne transmission; direct transmission; indirect transmission.
A probability density function defined by a minimum, a maximum, and a mode (most likely value). This distribution is often used to represent information collected from subject matter experts and can be used when information is scarce (15). See also expert opinion; probability density function.
Lack of knowledge regarding the biological system. Input parameters often include a degree of uncertainty due to a lack of thorough understanding or available data regarding the processes or events being modelled. Uncertainty in input parameter values is incorporated into stochastic models through the use of probabilistic methods (12).
A symmetrical probability density function defined by a minimum and a maximum. Uniform distributions are applied when there is an equal probability of values occurring between the minimum and maximum (15). See also probability density function.
In ADSM, a cluster of animals called a unit is the basis of simulation. A unit has a production type, number of animals, point location (expressed in terms of longitude and latitude), and a transition state.
Vaccination may also be simulated as part of a disease control program.
A state in which herds or flocks are immune due to vaccination (4). See also immune; natural immunity.
The process of assessing if a model is an acceptable representation of the biological system it was intended to emulate, if the model outcome makes biological sense, and if it mimics real life (11, 13). See also face validity; relative validity.
Differences between individuals (herds/flocks) or within an individual (herd/flock) under differing circumstances (e.g. environmental, climatic, regional, host susceptibility, agent virulence). Use of stochastic models is intended to account for the natural variability intrinsic to the model parameters through the use of probabilistic methods (12).
The process of determining if the mathematics and programming in a model are working correctly and as intended (7, 12).
A probability density function defined by shape () and scale () (where both and are greater than zero) used to model the time until occurrence of an event where the probability of occurrence changes over time (15). See also probability density function.
The average daily prevalence of infectious (subclinically and clinically infectious) individuals within infected herds, flocks or pens. See also clinically infectious; subclinically infectious.
Zones in ADSM are circular areas created around detected, infected premises. Inside zones, more stringent controls on movements may apply and more intensive surveillance activities may be carried out.
The authors gratefully acknowledge Barbara Corso, Emery Leger, Francesca Culver, Graeme Garner, Jane Rooney, Jonathan Happold, Katie Owen, Neil Harvey, Steve Weber, and Thomas Rawdon for their comments and helpful suggestions.