Disease progression

Epidemiological MODeling software (EMOD) represents the progression of TB disease using a series of state changes as individuals progress from susceptible to latent, to active infection, and then to recovery or death. The latent states represents asymptomatic individuals who are positive for the tuberculin skin test(exposed) but not infectious. The active states represents infectious individuals who have signs or symptoms of TB. Therefore, you can simplify the TB model to replicate a simple SEIR model where latency corresponds to the exposed compartment and active TB corresponds to the infectious compartment as follows:

However, EMOD provides additional disease states to better capture the heterogeneity in disease progression and infectiousness. Within the broad categories of susceptible/latent/active, the model includes states for latent individuals whose progression to active disease is slow or fast and active individuals who are presymptomatic, who have pulmonary TB that is smear-negative or smear-positive, or who have extrapulmonary TB. The full TB model disease states are represented below:

Individuals progress from latent disease to an active presymptomatic state before progressing to the active symptomatic state, reflecting the spectrum of disease from latent to active disease. The presymptomatic state represents a period of reduced infectivity where patients may have objective signs of TB (for example, an increased cough) but lack subjective symptoms (they do not notice their increased cough).

To view reference information for all TB parameter related to disease progression, see Disease progression parameters.

The following sections walk through some example simulations. To run these simulations yourself, see the TB/Disease_Progression scenario in the downloadable EMOD scenarios zip file. Review the README files there for more information.

SEIR baseline

By default, the TB model is configured with heterogeneity in disease progression and presentation. However, to better understand how to configure this heterogeneity, we recommend first simplifying the TB model to replicate an SEIR model by disabling TB-specific heterogeneity by configuring a single latent progression speed, a single symptomatic state, and a single disease presentation. You can use this example as a baseline, adding different kinds of heterogeneity to see how the output changes.

The following graphs show the inset chart and charts for all channels in a TB outbreak where the model has been simplified to replicate an SEIR model.

Figure 1: TB model flattened to replicate an SEIR model

Figure 2: TB model flattened to replicate an SEIR model

Latency progression

Tuberculosis can progress at very different rates for different people. For example, it tends to progress to disease more quickly in children than in adults. In the EMOD TB model, individuals are divided into two groups: fast and slow progressors. TB_Fast_Progressor_Fraction_Adult and TB_Fast_Progressor_Fraction_Child set the fraction of the adults and children, respectively, that are fast progressors, while the remaining individuals are slow progressors. The rate at which latent infections progress to active disease is set separately for these two groups.

The following graphs show the inset chart and comparison to the baseline SEIR TB outbreak where both children and adults are evenly split between slow and fast progressors, and fast progressors transition into the presymptomatic active state 10 times faster than slow progressors. In this example, however, all individuals move out of the presymptomatic state after only a day, effectively eliminating it.

Figure 3: Course of outbreak slowed by heterogeneity in progression speed

Figure 4: Difference from baseline with heterogeneity in latency progression

Active disease presentation

Individuals progress from latent disease to an active presymptomatic state before progressing to the active symptomatic state, reflecting the spectrum of disease from latent to active disease. The presymptomatic state represents a period of reduced infectivity where patients may have objective signs of TB (for example, an increased cough) but lack subjective symptoms (they do not notice their increased cough). The addition of the active presymptomatic state increases the overall time from initial infection to the active symptomatic state, ultimately slowing the overall spread of disease amongst the cohort.

Individuals with active TB are classified as smear positive, smear negative or extrapulmonary. The fraction of individuals in each of these groups must add up to 1. The active, symptomatic compartment can be subdivided into smear-positive, smear-negative, and extrapulmonary disease presentations. You must set the infectiousness of smear-negative and smear-positive TB; extrapulmonary TB is non-infectious and therefore has no configurable infectiousness. As with disease progression, children and adults are configured independently.

The following graphs show the inset chart and comparison to the baseline SEIR TB outbreak where presymptomatic individuals are not infectious and transition to active disease in approximately 30 days. The population is evenly split among smear-positive, smear-negative, and extrapulmonary disease presentations. Extrapulmonary mortality is 40% of smear-positive, and the mortality and infectiousness are the same for both smear-negative and smear-positive TB. It does not include heterogeneity in the latent stage.

Figure 5: Course of outbreak with heterogeneity in active disease

Figure 6: Difference from baseline with heterogeneity in active disease states