cleverly.ModifiedTreatmentPolicy#

class cleverly.ModifiedTreatmentPolicy(policies: Sequence[Policy], reference: str | None = None)[source]#

Request mean outcomes under modified treatment policies.

Parameters:

policiessequence of Policy: Named policies to evaluate: Shift, Scale, Piecewise or ModifiedPolicy. A RiskRatioTilt reports its own estimand, ey_rr_tilt, which has no typed request; this class refuses it.
referencestr or None: Policy label retained as the comparison reference.

See also

ModifiedTreatmentPolicyEffect: The contrast between these means.
cleverly.interventions.Shift: The additive dose shift, the simplest policy.
cleverly.interventions.check_policy_support: Whether the shifted dose is supported.

Examples

Raise every dose by one unit without imposing a cap:

>>> from cleverly import ModifiedTreatmentPolicy
>>> from cleverly.interventions import Shift
>>> estimand = ModifiedTreatmentPolicy(policies=[Shift(delta=1.0, cap=None, name="up1")])
>>> estimand.policies[0].cap is None
True

Methods#

Attributes#

ModifiedTreatmentPolicy.definition: str = 'mean outcome under each modified treatment policy'#

ModifiedTreatmentPolicy.name: str = 'ey_policy'#

ModifiedTreatmentPolicy.reference: str | None = None#

ModifiedTreatmentPolicy.policies: Sequence[Policy]#