Treatment
The vivarium_public_health treatment package provides
components for modeling health interventions and
their effects within a simulation. While vivarium supplies the value
pipeline framework for combining modifiers,
the treatment package uses those pipelines to model specific styles of
intervention.
The package is organized around four components:
Absolute shift — a simple intervention that directly sets the value of a target epidemiological measure for simulants within a configured age range. See
AbsoluteShift.Linear scale-up — a time-varying intervention that linearly interpolates exposure parameters between configured start and end values over a date range. See Linear Scale-Up.
Therapeutic inertia — a population-level scalar representing the probability that treatment is not escalated during a healthcare visit, drawn from a triangular distribution. See Therapeutic Inertia.
Intervention and intervention effect — components that wrap the
causal_factorframework to model dichotomous intervention exposures and their relative-risk effects on target measures. SeeInterventionandInterventionEffect.
Absolute Shift
The AbsoluteShift component
is the simplest intervention model. It registers an
attribute modifier on a target pipeline that replaces the
current value with a configured absolute value for all simulants within a
specified age range. When the target_value is set to "baseline", no
modification is applied.
configuration:
intervention_on_my_cause:
target_value: 0.01
age_start: 15
age_end: 65
Intervention and Intervention Effect
The Intervention class
is a specialization of
CausalFactor restricted
to "intervention" entity types. It models a dichotomous coverage exposure
(exposed vs. unexposed) and can source data from the artifact or the
configuration.
The InterventionEffect
class models the effect of an Intervention on a target entity’s measure
using relative risk data. It is a specialization of
CausalFactorEffect.