Contact Rate

Contact Rate: Concept Architecture

Orientation and learning roadmap

This page follows contacts between people into an infectious-disease transmission model. It distinguishes the number of encounters from the chance of transmission per encounter, works through a one-day example, and shows why population mixing patterns matter for economic evaluation.

Define a contact that is relevant to transmission

State what qualifies as an encounter for the pathogen and intervention being studied, such as a close conversation, household exposure, or another specified interaction. Count those encounters per person over a stated time unit and identify whether repeated interactions with the same person are counted separately.

A social contact diary estimates a different object from a count of laboratory-confirmed transmissions. Contact frequency may vary by age, setting, day of week, behaviour, and policy, so a single population average can conceal important pathways. The Mossong and colleagues social-contact survey demonstrates how observed mixing patterns can inform infectious-disease models.

Combine contact frequency with infection prevalence

A susceptible person faces a risk only when an encounter is with someone infectious and transmission occurs under the relevant circumstances. A contact rate must therefore be combined with the proportion infectious among potential contacts and a per-contact transmission probability, with any other modifiers represented consistently.

For susceptible group $i$ meeting people in groups $j$, a simple frequency-dependent formulation is:

$$\lambda_i(t)=\sum_j c_{ij}(t)p_{ij}(t) \frac{I_j(t)}{N_j(t)},$$

where $c_{ij}$ is relevant contacts per susceptible person in group $i$ with group $j$ per unit time, $p_{ij}$ is transmission probability per infectious contact, and $I_j/N_j$ is the infectious fraction of contacts' source group. The result $\lambda_i$ is an infection hazard under the assumed mixing process, not the contact rate itself. The CDC explanation of infectious-disease transmission models distinguishes interactions at a contact rate from per-contact infection probability.

Worked one-day transmission example

Suppose a susceptible person has eight relevant contacts per day, 10% of possible contacts are infectious, and the probability of transmission from an infectious contact is 2%. Under homogeneous independent mixing, the approximate force of infection is $8(0.10)(0.02)=0.016$ per day.

If that hazard is constant over one day, the corresponding infection probability is $1-\exp(-0.016)\approx0.01587$, or about 1.59%. Eight contacts are not eight infection events, and 0.016 is a rate rather than exactly the one-day probability. The values are illustrative; repeated contacts, clustered exposure, susceptibility, and infectiousness differences can alter the result.

Represent who meets whom

An age- or setting-specific contact matrix records average contacts from each group to another, making its row and column definitions essential. Under reciprocal person-to-person contacts measured consistently, population totals can be checked so that reported contacts from group $i$ to $j$ approximately match those from $j$ to $i$ after weighting by group sizes.

The product of average contact frequency and per-contact transmission probability is sometimes called an effective transmission parameter, but definitions vary by model. Do not label that product as the observed contact rate or assume a contact matrix estimated in one setting transfers unchanged to another.

Evaluate interventions and costs

School closure, remote consultations, masking, or vaccination can affect different terms in the transmission model rather than all contacts in the same way. A reduction in contact frequency can change infections indirectly through later transmission chains, so a static per-person model may miss population feedback.

An economic evaluation should connect modelled infections and downstream care to the resources and health outcomes of interest. Interventions can also carry educational, work, or access consequences that depend on the chosen perspective. The CDC modelling study of respiratory-pathogen interventions separates effects on contact patterns, transmission probability, and infectious duration.

Validate the rate and its use

Check that contacts are measured per person per day or another clearly specified unit, that matrix entries are defined consistently, and that the prevalence multiplier is drawn from the right population. Compare modelled incidence with independent data where available, while recognising that matching incidence alone may not separately identify contact frequency and transmissibility.

Explore uncertainty in contact survey coverage, recall, temporal change, and the classification of exposure intensity. Avoid double-counting an intervention effect by reducing both contacts and per-contact probability for the same mechanism without justification. Report how contact rates were collected or calibrated and when those data remain relevant to the policy being modelled.