Tutorial 3: Interventions

This tutorial adds two interventions that represent a standard malaria control package: treatment-seeking care and insecticide-treated nets (ITNs). It introduces the campaign file and shows how to compare scenarios with and without interventions.

File: tutorials/tutorial_3_interventions.py

The campaign file

The campaign file defines what interventions to distribute, to whom, and when. In emodpy-malaria a build_campaign() function constructs this file and is passed to EMODTask via campaign_builder=:

task = EMODTask.from_defaults(
    ...,
    campaign_builder=build_campaign,   # previously None
    ...
)

Interventions

build_campaign() builds the campaign using the class-based intervention API. Each intervention is instantiated as an object and distributed via add_intervention_triggered() or add_intervention_scheduled():

tutorial_3_interventions.py, lines 94–147

def build_campaign(campaign):
    campaign.set_schema(manifest.schema_path)

    if deploy_treatment:
        clinical_drug = AntimalarialDrug(campaign, drug_type="Artemether")
        add_intervention_triggered(
            campaign,
            intervention_list=[clinical_drug],
            triggers_list=["NewClinicalCase"],
            start_day=60,
            target_demographics_config=TargetDemographicsConfig(demographic_coverage=0.7)
        )

        severe_drug = [AntimalarialDrug(campaign, drug_type="Chloroquine"),
                       AntimalarialDrug(campaign, drug_type="Lumefantrine")]
        add_intervention_triggered(
            campaign,
            intervention_list=severe_drug,
            triggers_list=["NewSevereCase"],
            start_day=40,
            target_demographics_config=TargetDemographicsConfig(demographic_coverage=0.9,
                                                                target_age_max=40)
        )

    if deploy_itn:
        bednet = SimpleBednet(
            campaign,
            repelling_config=Exponential(initial_effect=0.3, decay_time_constant=400),
            blocking_config=Exponential(initial_effect=0.9, decay_time_constant=200),
            killing_config=Exponential(initial_effect=0.1, decay_time_constant=300),
        )
        add_intervention_scheduled(
            campaign,
            intervention_list=[bednet],
            start_day=5,
            repetition_config=RepetitionConfig(infinite_repetitions=True,
                                               timesteps_between_repetitions=361),
            target_demographics_config=TargetDemographicsConfig(demographic_coverage=0.5)
        )

    return campaign

Treatment seeking distributes AntimalarialDrug directly in response to clinical events. The NewClinicalCase trigger fires when a person develops clinical malaria symptoms — the drug is given to 70% of cases. A second trigger handles NewSevereCase with a two-drug combination at 90% coverage for people 40 years and younger.

ITN distribution uses SimpleBednet with waning effect configurations for repelling, blocking, and killing. Exponential waning effects decay over time — the decay_rate controls how quickly effectiveness drops. TargetDemographicsConfig sets the coverage to 50%.

Comparing scenarios

Two boolean flags at the top of the script control which interventions are active:

deploy_treatment = True
deploy_itn = True

Toggle these and re-run to compare each intervention separately against the no-intervention baseline. Each combination writes to its own output directory:

`deploy_treatment`	`deploy_itn`	Output directory
True	True	`tutorial_3_results`
True	False	`tutorial_3_results_ts`
False	True	`tutorial_3_results_itn`
False	False	`tutorial_3_results_no_interventions`

Plotting with a baseline reference

plot_results() looks for tutorial_2_results/ from the previous tutorial and, if found, uses it as the no-intervention reference (plotted in red) so the intervention impact is visible directly. If you are starting here without having run Tutorial 2, the plot will still work — the reference line simply will not appear.

reference = None
if os.path.exists("tutorial_2_results"):
    t2_files = get_filenames(dir_or_filename="tutorial_2_results",
                             file_prefix="InsetChart", file_extension="json")
    if t2_files:
        reference = t2_files[0]

plot_inset_chart(dir_name=output_path,
                 reference=reference,
                 title="Tutorial 3 - InsetChart",
                 output=output_path)

Treatment seeking reduces the fraction of people infected; ITNs reduce the daily biting rate and vector population. Running each combination separately lets you see each effect in isolation.