InterventionsDeprecated

Deprecated Interventions

Mass drug deployment interventions (MDA/MSAT/T&T/IPT)

From version 33 onwards MDA or IPT is simulated using <treatSimple> (or <treatPKPD>) sub-elements of <component>. By including <screen> and "decisionTree" elements test and treat strategies can be simulated within this general framework. In earlier versions separate elements for MDA and IPT were used. The following is applicable to 5-day and 1-day time step models, up to schema version 32.

MDA (versions <= 32)

Prior to schema 32, MDA was simulated with the <MDA> element which would just contain deployment times and optionally a diagnostic:

<MDA>
  <description>
    <decisions>
      ...
    </decisions>
    <treatments>
      ...
    </treatments>
  </description>
  <timed coverage="0.6" time="160"/>
  <timed coverage="0.6" time="860"/>
</MDA>

e.g. MDA using DHP (note effective coverage of 75%){TODO, check this}

<effect id="DHP_MDA">
-<MDA>
-<drugEffect>
<compliance pCompliance="0.75" nonCompliersMultiplier="0"/>
<compliersEffective>
<timestep pClearance="0.8"/>
<timestep pClearance="0.5"/>
<timestep pClearance="0.3"/>
<timestep pClearance="0.1"/>
</compliersEffective>
</drugEffect>
</MDA>
</effect>

Primaquine acts by blocking transmission stages. This effect is equivalent to that of a transmission blocking vaccine (TBV) (though the duration may be very different) and can thus be parameterised as such.

-<effect id="Primaquine">
-<TBV>
<!-- Actual half-life of Primaquine is 8 hours. Use 2 days in order to accurately simulate gametocyte clearance -->
<decay function="exponential" L="0.005555556"/>
<efficacyB value="0.001"/>
<initialEfficacy value="1"/>
</TBV>
</effect>

Mass screen and treat and diagnostics

From schema version 29, a diagnostic can be used with the 5-day time-step model to simulate mass screen-and-treat. For example, the below uses a diagnostic simulating an RDT for Plasmodium falciparum in Africa source: Murray et al (Clinical Microbiological Reviews, Jan. 2008), e.g. in General:

<MDA>
  <!-- optionally a diagnostic here -->
  <timed coverage="1" maxAge="999" minAge="0" time="0"/>
  <timed coverage="1" maxAge="999" minAge="0" time="14"/>
  <timed coverage="1" maxAge="999" minAge="0" time="28"/>
</MDA>

<component id="msat" name="example with screening">
  <MDA>
    <diagnostic>
      <stochastic dens_50="50" specificity="0.942"/>
    </diagnostic>
    <effects .../>
  </MDA>
</component>

Some parameters for simulating microscopy (source: expert opinion): dens_50="20" specificity="0.75". Deterministic tests are also supported:

<diagnostic>
  <deterministic minDensity="40"/>
</diagnostic>

Here the test outcome is positive if the density is at least the minimum given (in parasites/µL, relative to malaria theropy data rather than densities estimated from Garki/non-Garki slides). If a minimum density of 0 is given, the test outcome is always positive.

e.g. for DHP

<effect id="DHP_MSAT">
<MDA>
<diagnostic>
<stochastic specificity="0.942" dens_50="50"/>
</diagnostic>
<drugEffect>
<compliance pCompliance="0.75" nonCompliersMultiplier="0"/>
<compliersEffective>
<timestep pClearance="0.8"/>
<timestep pClearance="0.5"/>
<timestep pClearance="0.3"/>
<timestep pClearance="0.1"/>
</compliersEffective>
</drugEffect>
</MDA>
</effect>

MSAT using Coartem (version 32):

<human>
  <effect id="Coartem">
  <MDA>
    <diagnostic>
      <stochastic specificity="0.942" dens_50="50"/>
    </diagnostic>
  <drugEffect>
    <compliance pCompliance="0.75" nonCompliersMultiplier="0"/>
    <compliersEffective>
    <timestep pClearance="0.8"/>
    <timestep pClearance="0.3"/>
    </compliersEffective>
  </drugEffect>
</MDA>
</effect>

MDA1D

The "MDA1D" component was introduced in Version 32. It uses 1-day health system code to simulate MDA, thus allowing decision trees and PK/PD deployment. The description here is for version 32, but the intervention was available before this (see note below).

With the 1-day time-step model MDA interventions are more flexible and can, for example, be configured as short-action anti-malarial deployment, deployment of SP-like drugs with a long prophylactic effect and mass screen and treat.

Configuration: an MDA Description element must be included, as, for example, below:

<component id="mda" name="example of MSAT with a 1-day health system model configuration">
  <MDA1D>
    <decisions>
      <decision depends="result" name="treatment" values="MQ,none">
        result(negative): none
        result(none): none
        result(positive): MQ
      </decision>
      <decision depends="" name="test" values="none,microscopy,RDT">
        RDT
      </decision>
    </decisions>
    <treatments>
      <treatment name="MQ">
        <schedule>
          <medicate drug="MQ" hour="0" mg="1200"/>
        </schedule>
      </treatment>
      <treatment name="none">
        <schedule/>
      </treatment>
    </treatments>
  </MDA1D>
</component>

This has the same format as part of the EventScheduler complicated/uncomplicated trees. The above example is fairly minimal: the test decision indicates that no malaria diagnostic is used, and the treatment decision indicates that the treatment schedule called 1 is used. Note that it may be preferable to specify dosing by age as with health-system treatment of uncomplicated malaria.

Intermittent Preventive Treatment (IPT)

From schema Version 32, this is implemented using the same approach as mass drug administration interventions and the separate IPT model is no longer available. Earlier Versions implemented explicity IPTi and IPTc: (Intermittent Preventive Treatment for infants or children). Antimalarial drugs are given at specified ages in infants according to the routine vaccination schedule (IPTi) or at regular intervals in children for whom delivery may be timed to coincide with the malaria season (IPTc). IPT is compatible with the 5-day timestep model. The drug action is similar to that of the mass drug administration but can have a longer lasting (prophylactic) effect and can affect individual infections with different levels of drug resistance differently.

<IPT>
  <description iptiEffect="18">
    <infGenotype ACR="1.0" atten="1" freq="0.8" name="wt" proph="10" tolPeriod="4"/>
    <infGenotype ACR="1.0" atten="1" freq="0.1" name="double" proph="2" tolPeriod="4"/>
    <infGenotype ACR="0.5" atten="1" freq="0.1" name="triple" proph="0" tolPeriod="4"/>
  </description>
  <continuous coverage="0.95" targetAgeYrs="0.25"/>
  <continuous coverage="0.95" targetAgeYrs="0.33"/>
  <continuous coverage="0.95" targetAgeYrs="0.75"/>
  <timed time="7" coverage=".8"/>
  <timed time="79" coverage=".8"/>
</IPT>

In this example, IPTi is given at 2, 3 and 9 months of age with a coverage of 95% for each dose. The drug in this example is sulphadoxine-pyrimethamine (SP) and new infections are randomly assigned one of three specified genotypes (wt,108,triples) which correspond to different degrees of SP resistance and have frequencies freq. Existing wildtype infections (wt) are cleared with a cure rate (ACR) of 100%, and incoming wildtype infections are cleared for a prophylactic period (proph) of 10*5 days. The intervention can also be timed to simulate seasonal delivery. iptieffect refers to whether the IPT or placebo group should be simulated (>=10 are IPT groups) and the choice of treatment drug in those who present with malaria fevers. atten and tolperiod are not used except with alternative models for IPT described in Ross et al, 2008 PLoS ONE e2661.

Vaccines

Three vaccine types can be simulated. Prior to schema 32, all configured vaccine types are administered simultaneously; from version 32, multiple descriptions for each vaccine type are allowed and each vaccine description may be administered independently. Types of vaccine:

• Pre-erythrocytic vaccine (vaccineType PEV): prevents a proportion of infections from commencing.
• Blood-stage vaccine (vaccineType BSV): acts as a killing factor on blood-stage parasites
• Transmission-blocking vaccine (vaccineType TBV): one minus this scales the probability of transmission to mosquitoes

Prior to schema version 25, the vaccine type codes were 1, 2, 3 instead of PEV, BSV, TBV.

Efficacy

Vaccine efficacy is calculated as initialEff * decayFactor, where initialEff is sampled from a Beta distribution, and decayFactor is described by the decay element and time since last deployment (see DecayFunction configuration).

The Beta distribution takes two parameters: β, set by efficacyB, and α, set as "β × μ / (1 - μ)". μ is set to one of the given initialEfficacy values: the first value the first time this vaccine component is administered, the second the second time, etc. (repeating the last value if necessary). Any number of initialEfficacy values can be given.

Note: the mean initial efficacy is μ (initialEfficacy), and the variance is "m × (1 - m)^2 / (1 - m + β)" (thus a large β=efficacyB can be used to minimise variance).

Doses and schedules

Prior to schema 32, mass deployments select a different portion of the population each time according to the coverage given, as with any other intervention. Continuous deployments work in the same way, except that after selecting individuals according to the coverage described, any individuals who either missed a previous continuous dose (unless receiving a campaign dose to catch up) or have already received this scheduled dose (through campaign) will not receive their dose. (Thus true coverage may be less than the coverage requested.)

For schemas (?) up to 31, a description element is given for each enabled vaccine type within the vaccine element, so vaccines are specified as follows:

<vaccine>
  <description vaccineType="TBV">
    <decay L="0.5" function="exponential"/>
    <efficacyB value="10"/>
    <initialEfficacy value="0.4"/>
    <initialEfficacy value="0.46"/>
    <initialEfficacy value="0.52"/>
  </description>
  <description vaccineType="BSV">
    <decay L="0.5" function="exponential"/>
    <efficacyB value="10"/>
    <initialEfficacy value="0.4"/>
    <initialEfficacy value="0.46"/>
    <initialEfficacy value="0.52"/>
  </description>
  <description vaccineType="PEV">
    <decay L="0.5" function="exponential"/>
    <efficacyB value="10"/>
    <initialEfficacy value="0.4"/>
    <initialEfficacy value="0.46"/>
    <initialEfficacy value="0.52"/>
  </description>
  <continuous coverage="0.95" targetAgeYrs="0.0833"/>
  <continuous coverage="0.95" targetAgeYrs="0.17"/>
  <continuous coverage="0.99" targetAgeYrs="0.25"/>
  <continuous coverage="0.99" targetAgeYrs="1.25"/>
  <continuous coverage="0.99" targetAgeYrs="2.25"/>
  <continuous coverage="0.99" targetAgeYrs="3.25"/>
  <continuous coverage="0.99" targetAgeYrs="4.25"/>
</vaccine>

Vector control interventions

Vector control interventions are described in terms of:

• deterrency describes a parameter x such that the availability of a protected host to mosquitoes becomes (1 - x) × preInterventionAvailability
• preprandialKillingEffect describes a parameter y such that the probability of a mosquito successfully biting a chosen protected host (instead of dying) becomes (1 - y) × preInterventionProbability
• postprandialKillingEffect describes a parameter z such that the probability of a mosquito successfully escaping from a protected human after feeding (instead of dying) becomes (1 - z) × preInterventionProbability

IRS

Prior to version 32, IRS (v1) was simulated using a specific sub-element of <intervention>. vectorDeterrent implemented a a sub-set of the same functionality. Both are now replaced by GVI. Examples of the old configuration:

    <IRS name="DDT test">
      <description>
        <decay L="0.5" function="exponential"/>
        <anophelesParams mosquito="gambiae_ss" propActive="1">
          <deterrency value="0.56"/>
          <preprandialKillingEffect value="0"/>
          <postprandialKillingEffect value="0.24"/>
        </anophelesParams>
        <anophelesParams mosquito="funestus" propActive="1">
          <deterrency value="0.56"/>
          <preprandialKillingEffect value="0"/>
          <postprandialKillingEffect value="0.24"/>
        </anophelesParams>
        <anophelesParams mosquito="arabiensis" propActive="1">
          <deterrency value="0.56"/>
          <preprandialKillingEffect value="0"/>
          <postprandialKillingEffect value="0.24"/>
        </anophelesParams>
      </description>
      <timed>
        <deploy coverage="0.95" time="7"/>
      </timed>
    </IRS>

an alternative parameterisation was also available:

    <IRS name="DDT test v2">
      <description_v2>
        <!-- Initial insecticide: units are μg/cm²; no variance here -->
        <initialInsecticide mu="118" sigma="0"/>
        <insecticideDecay L="0.15" function="exponential"/>
        <anophelesParams mosquito="gambiae_ss">
          <deterrency insecticideFactor="0.12" insecticideScalingFactor="0.05"/>
          <preprandialKillingEffect baseFactor="0.01" insecticideFactor="0.25" insecticideScalingFactor="0.008"/>
          <postprandialKillingEffect baseFactor="0.01" insecticideFactor="0.5" insecticideScalingFactor="0.006"/>
        </anophelesParams>
      </description_v2>
      <timed>
        <deploy coverage="0.05775" cumulativeWithMaxAge="0.5" time="18"/>
        <deploy coverage="0.1155" cumulativeWithMaxAge="0.5" time="19"/>
        <deploy coverage="0.17325" cumulativeWithMaxAge="0.5" time="20"/>
        ...
      </timed>
    </IRS>

Deterrency

    <vectorDeterrent name="hypothetical deterrency">
      <decay L="4" function="exponential"/>
      <anophelesParams mosquito="gambiae_ss">
        <deterrency value="0.2"/>
      </anophelesParams>
      <timed coverage="0.7" time="7"/>
    </vectorDeterrent>

The preprandialKillingEffect is only available as of schema 30, the other two previously (but the XML has a slightly different structure).

Deployment of interventions: interventions

The interventions element is organized informally into three sections, first descriptions of interventions (without associating times), second continuous interventions, and third timed interventions.

Descriptions

Intervention descriptions are, at this time, all direct sub-elements of interventions and not some other element.

The MDADescription element is only required when using Mass Drug Administration interventions on a one-day timestep. It can then take a schedule sub-element describing a treatment dose or schedule taken by individuals covered by the administration campaign. There are currently no adherence/dose size/quality effects varying this; i.e. individuals either are not covered by the campaign and receive nothing or are and receive the full schedule.

The vaccineDescription element describes vaccines.

The iptiDescription describes Intermittent Preventive Treatment for Infants . One important point: including this element also enables the IPT drug/within-host model, which adds a protective effect to every treatment. (This behaviour could be changed in the future to use a separate model-option switch.)

The anopheles element, if used, should be listed for each mosquito species (using the mosquito attribute to associate with one of the mosquito species specified in entomological data). Each may contain descriptions for Insecticide Treated Net (ITNDescription), Indoor Residual Spraying (IRSDescription) and a mosquito repellant (VADescription — Vector Availability) intervention. Theses have sub-elements describing a parameter as a Weibull decay curve. All three have a deterrency parameter; additionally ITNDescription has preprandialKillingEffect and postprandialKillingEffect (chance of killing a mosquito before or after feeding) and IRSDescription has killingEffect (chance of killing mosquito while resting) parameters.

Continuous: continuous

Interventions with continuous deployment are described by a continuous element. This accepts sub-elements vaccine, ITN and ipti (see descriptions above); these may occur more than once. (They must be listed in order of type; that is all vaccine elements before any ITN elements, etc.) Each has the same type with the following attributes:

Element	Action
targetAgeYrs	Intervention has a chance of being given to the human at this exact point of age.
coverage	Chance of human recieving intervention at given point of age.
begin	First timestep at which this item becomes active (defaults to 0, the start of the intervention period).
end	First timestep at which this item becomes deactive (defaults to the maximum value for an int, 231-1).

Multiple elements may be used for each intervention, e.g. to administer at several time points or make the intervention active for multiple independant time periods (if end-time = begin-time for two items, they meet with no overlap).

Timed: timed

Timed interventions correspond to mass deployment at a specific time point. The timed element accepts a list of intervention sub-elements, each of which describes interventions occuring at a time specified by its time attribute. The following interventions may occur (and should be listed in this order within each intervention):

Element	Action
changeHS	Replaces the current health-system description with a new one. Must be a complete description of the health system (contain an ImmediateOutcomes or EventScheduler element, and a CFR element, as with the usual healthSystem element)
changeEIR	Replaces EIR data with new, daily, data described here. Data must be sufficient to last until the end of the simulation.
importedInfectionsPerThousandHosts	Not an intervention, but deployed the same way: a floating-point number describing the chance each individual recieves an infection (hence number introduced is unlikely to be exactly the number specified)
MDA	Mass drug administration (clears infections directly for models with a 5-day timestep and administers drugs for models with a 1-day timestep)
vaccinate	Mass vaccination campaign
ITN	Mass ITN-deployment
IRS	IRS spray round
VectorAvailability (VectorDeterrent)	Mass mosquito-repellent deployment (e.g. house spraying)
ipti	Mass IPTi campaign
larviciding	Mosquito emergence-rate reduction intervention. Has an anopheles sub-element for each mosquito sub-species, containing the attributes: mosquito (name of sub-species), effectiveness (proportional reduction in mosquito emergence), duration (number of days the intervention is effective for).

The MDA through to ipti interventions above are all mass deployment. Each has its deployment described by the following attributes: maxAge and minAge describe the age-range applicable in years (optional; they default to 100 and 0 respectively), and coverage specifies the chance that each individual receives the intervention.