HIV Comorbidity in TBSim

This notebook demonstrates how to use the HIV comorbidity module in the tbsim package.

Module Location: tbsim/comorbidities/hiv/

---

Overview

The HIV comorbidity module in TBSim allows users to simulate the impact of HIV infection on TB disease dynamics. Rather than modeling HIV progression in detail, it serves as a lightweight comorbidity that modifies TB risks based on HIV state.

The module is composed of three key components:

• HIV Model: Represents HIV infection as a comorbidity.

• HIV Intervention: Enables scenario testing via prevalence and ART coverage.

• TB-HIV Connector: Bridges HIV and TB models, applying comorbidity risk logic.

---

HIV Model

The HIV model simulates HIV infection status and its progression through simplified states:

• ACUTE

• LATENT

• AIDS

This implementation is intended as a modifier for TB outcomes, rather than a standalone disease model.

Key Parameters

Parameter	Description
init_prev	Probability of being HIV-positive (ACUTE stage) at simulation start
init_onart	Probability of being on ART if HIV-positive
acute_to_latent	Daily probability of transitioning from ACUTE to LATENT
LATENT_to_aids	Daily probability of transitioning from LATENT to AIDS
aids_to_dead	Daily probability of death after reaching AIDS (currently unused)
art_progression_factor	Multiplier reducing HIV progression for agents on ART

*** Important: ***

At step 0, the model looks for the existance of an HivIntervention handling the HIV infection and art coverage, if none is found then it uses the values defined at Disease level to set the initial values.

---

HIV Intervention

The HivIntervention class allows control over HIV prevalence and ART coverage in the population. This is useful for scenario testing and simulating public health campaigns.

Supported Modes

• 'infection': Adjust HIV infection prevalence

• 'art': Adjust ART coverage

• 'both': Adjust both infection and treatment

What It Does

• Applies changes to the HIV state of the population

• Operates within a specified time range (start to stop)

• Can be targeted to specific age groups or entire population

intervention = HivInterventions(pars=dict(
    mode='both',
    prevalence=0.2,
    percent_on_ART=0.5,
    minimum_age=15,
    max_age=49,
    start=ss.date('2000-01-01'),
    stop=ss.date('2010-12-31'),
))

Also, the parameters shown above are the same ones that you can modify in order to run a different scenario.

---

TB-HIV Connector

The TB_HIV_Connector class links the TB and HIV models, simulating the comorbidity effect of HIV on TB progression.

Purpose

Modifies TB progression parameters based on an individual’s current HIV state.

Risk Multiplier Table

HIV State	TB Progression Risk Multiplier
ACUTE	1.5
LATENT	2.0
AIDS	3.0

Behavior at Each Time Step

1. Identify all TB-infected individuals.

2. Look up their current HIV state.

3. Apply a multiplier to their TB rr_activation value (risk of progressing to active TB).

---

Process Flow Diagram

+-----------------+
|   TB Infection  |
+-----------------+
        |
        v
+-----------------+     +-----------------------+
|  HIV Comorbidity| --> | HIV State (ACUTE,...) |
+-----------------+     +-----------------------+
        |                          |
        |      TB_HIV_Connector    |
        +-------------------------> Applies RR Multiplier
                                   |
                                   v
                    +-------------------------------+
                    | TB rr_activation Adjusted     |
                    | Based on HIV State            |
                    +-------------------------------+

---

Example Code

Example

import tbsim as mtb
import starsim as ss
import sciris as sc
import numpy as np
import matplotlib.pyplot as plt# Rendering with Quatro

Intervention

Adding a function to simplify the generation of an intervention.

This intervention will try to simulate a steady prevalence and on ART values along the entire simulation

def make_hiv_intervention(pars=None):
    if pars is None:
        pars=dict(
                mode='both',
                prevalence=0.30,            # Maintain 30 percent of the alive population infected
                percent_on_ART=0.50,        # Maintain 50 percent of the % infected population on ART
                min_age=15, max_age=60,     # Min and Max age of agents that can be hit with the intervention
                start=ss.date('2000-01-01'), stop=ss.date('2035-12-31'),   # Intervention's start and stop dates
        )
    intervention = mtb.HivInterventions(pars=pars)
    return intervention

Creating the simulation

def build_tbhiv_sim(simpars=None, tbpars=None, hivinv_pars=None) -> ss.Sim:
    sim_pars = dict(
        dt=ss.days(7),
        start=ss.date('1980-01-01'),
        stop=ss.date('2035-12-31'),
        rand_seed=123,
        verbose=0,
    )
    tb_pars = dict(
        init_prev=ss.bernoulli(p=0.25),
        rel_sus_latentslow=0.1,
    )

    hivinv_pars = hivinv_pars or dict(
        mode='both',
        prevalence=0.20,
        percent_on_ART=0.20,
        min_age=15, max_age=49,
        start=ss.date('2000-01-01'),
        stop=ss.date('2010-12-31'),
    )
    hiv_intervention = make_hiv_intervention(hivinv_pars)
    hiv = mtb.HIV()
    connector = mtb.TB_HIV_Connector()

    tb = mtb.TB(pars=tb_pars)
    people = ss.People(n_agents=1_000)
    network = ss.RandomNet(pars=dict(n_contacts=ss.poisson(lam=2), dur=0))

    # --- Assemble Simulation ---
    sim = ss.Sim(
        people=people,
        diseases=[tb, hiv],
        interventions=[hiv_intervention],
        networks=network,
        connectors=[connector],
        pars=sim_pars,
    )
    return sim

Running the simulation

# Run the simulation:
sim = build_tbhiv_sim()
sim.run()

HIV intervention present, skipping initialization.

Sim(n=1000; 1980.01.01—2035.12.31; connectors=tb_hiv_connector; networks=randomnet; interventions=hivinterventions; diseases=tb, hiv)

Plot the results

sim.plot()
plt.show()

Figure(1200x900)