epi-scenario-5

[djinnome]

Scenario 5: Stock and flow testing and diagnostics model (relevant for ASPR demo)

One of the major challenges that ASKEM aims to address is the limited reproducibility of models that are created using proprietary or specialized modeling software. Within the epidemiological modeling use case, the use of proprietary modeling software is particularly common within the system dynamics (“stock and flow”) component of the modeling ecosystem. System dynamics models are useful for exploring multifaceted scenarios, such as supply chain connections with disease dynamics.

The goal of this scenario is to demonstrate that ASKEM can ingest and simulate a model that is representative of the types of system dynamics approaches that were implemented in the early stage of the COVID-19 pandemic.

1. Ingest the file IndiaNonSubscriptedPulsed.mdl, which contains a system dynamics model that was used to assess the effectiveness of testing, isolation, and quarantine on the potential trajectory of COVID-19 in early 2020. The preprint for the model, Effectiveness of Testing, Tracing, Social Distancing and Hygiene in Tackling COVID-19 in India: a System Dynamics Model is available at: arXiv:2004.08859
2. Inspect the flow diagram to ensure that it is consistent with the model as implemented in its original environment (the modeling platform Vensim). The key stocks, and flows between them, should match the diagram in Figure 2:

3. Simulate the model using the base configurations in the .mdl file. The simulation output for the base case should indicate a peak around day 250, at 5 million cases/day.
4. Demonstrate that you can adjust rates in the simulation configuration, such as “default delay disease diagnosis” and “testing impact on delay”, which affect the flow into hospital demand.
5. In their original documentation, the developers of this model expressed an interest in introducing uncertainty into model components related to testing. The accuracy of tests in the base configuration of the model is assumed to be absolute. Demonstrate that you can introduce uncertainty into (1) the accuracy of test results and (2) the amount of time it takes to receive a test result. Show that you can set multiple model configurations to fully explore the uncertainty range in the relevant parameters, and show how simulating across this range affects the peak day and peak caseload.
6. Demonstrate that you can modify the model structure by adding an additional testing modality, rapid antigen tests. The base configuration assumes only one type of testing with one set of rates, and adding antigen-based testing will allow a modeler to explore the proportion of fast, but less reliable, antigen-based tests, compared with slower but more reliable nucleic acid amplification tests. Include the uncertainty added in step 5 above for both test types. Note: this step is meant to be exploratory; there is no “correct” answer other than demonstrating that the ingested SD model structure can be meaningfully modified in the workbench.
7. Consider extending the model to account for other structural features not included in the original model, such as vaccination and the potential for reinfection. Incorporate uncertainty in the new features using evidence drawn from literature review in the workbench. Note: this step is meant to be exploratory; there is no “correct” answer other than demonstrating that the ingested SD model structure can be meaningfully modified in the workbench.