Advanced topics

Here we describe how the multi-state life tables (MSLT) components are implemented.

Note that when you run a simulation using the simulate command:

simulate run reduce_acmr.yaml

The following sequence of operations will be performed:

1. The model specification will be read (in this case, from the file reduce_acmr.yaml) and a simulation object will be created.

2. The simulation.setup() method will call the setup() method for each of the MSLT components defined in the model specification.

   Note

   This is where components will load data tables, register event handlers, etc.

3. The initial population is created (typically by the BasePopulation component).

4. The time-steps will be simulated, with each time-step triggering the following events in turn:

   1. "time_step__prepare": The DelayedRisk component uses this event to account for transitions between exposure categories (i.e., uptake, cessation, and transitions between tunnel states). The Disease component uses this event to update disease prevalence and mortality for both the BAU and intervention scenarios, so that mortality and morbidity adjustments can be calculated. The TobaccoEradication component uses this event to move current smokers to the 0 years post-cessation exposure category when tobacco is eradicated.

   2. "time_step": The BasePopulation component uses this event to remove cohorts once they’ve reached the maximum age (110 years). The Mortality component uses this event to calculate the number of deaths and survivors at each time-step. The Disability component uses this event to calculate the HALYs for each cohort for both the BAU and intervention scenarios.

   3. "time_step__cleanup": no MSLT components respond to this event.

   4. "collect_metrics": the observer components will record relevant population details at the end of each time-step.

5. The simulation will trigger the "simulation_end" event and finish. The observer components use this event to write output tables to disk.

Model analysis

• Uncertainty analyses
• Alternative BAU: Immediate recovery upon cessation
  • Intervention comparison
• Alternative BAU: Constant tobacco prevalence
  • Intervention comparison

Writing your own MSLT components

• Writing a custom intervention
  • Structure of an intervention component
  • Example of an intervention component
• Writing a custom observer
  • Structure of an observer component
  • Example of an observer component
• Writing a custom risk factor
  • Structure of a risk factor component
  • Example of an intervention component