Inputs & Outputs

The HEP-CE model requires structured data in order to obtain meaningful outcomes. Here, we'll describe, in detail, how your data should be structured when using the model.

Directory Structure

The example, reiterated below, from the landing page illustrates some details about the necessary file structure to use the model:

data-source/
├── input1/
│   ├── inputs.db
│   └── sim.conf
└── input2/
    ├── inputs.db
    └── sim.conf

Let's break all this down. First, there's the data-source/ directory, containing individual input sets for the model. While the executable doesn't require a folder analogous to data-source, it can be useful to use such a folder as a record of what the inputs within it represent (e.g. basecase/, hcv-positive-cohort/, etc.) since you can name these folders arbitrarily. Regardless of whether or not you use such a directory, there must be at least one input folder, named input<X> where <X> is replaced with an integer, that contains the two required files sim.conf and inputs.db, which will be explained in detail in the following sections.

Note: There is no requirement that input folders start at 1. You can have any assortment of input numbers, though it is convenient for their numbers to be sequential, as using the executable with multiple input sets in one call expects them to be sequential. Consider the situation where you have the following input folders in data-source/:

input7, input8, input10, input11

Because they are not sequential, you would need to use two calls to the executable to simulate all of the inputs:

build/extras/executable/hepce_exe data-source 7 8
build/extras/executable/hepce_exe data-source 10 11

Our recommendation is simply that you structure your inputs in a way that makes them intuitive and interpretable, so that when someone looks at the data, they might understand the meaning without significant effort.

Input Files

Within each numbered input folder, there are two files, sim.conf and inputs.db, which are both required to use the HEP-CE executable. sim.conf is a standard, .ini-formatted configuration file and inputs.db is a SQLite database. In the next sections, we'll dive into the details of each of these files.

sim.conf

The sim.conf file defines the structure of a HEP-CE simulation run. In this file the discrete events, characteristics of HCV care, and single-value parameters are specified—there is no tabular data in this file. As is typical of .ini files, sim.conf is broken down into sections, each of which contain key-value pairs. HEP-CE has fourteen such sections:

• simulation
• mortality
• infection
• eligibility
• fibrosis
• fibrosis_staging
• screening
• screening_background_ab
• screening_background_rna
• screening_intervention_ab
• screening_intervention_rna
• linking
• treatment
• cost

An example sim.conf is provided within the GitHub repository for reference. All key-value pairs for each section are documented in detail in the example.

Input Database

All tabular data is stored in the database. In order to generate a database with the necessary structure, one simply needs to execute the script provided in the examples—the SQLite file is built accordingly. Due to data use agreements, we are not able to share the data we use with our model, only the schema. Users will have to provide their own data. The tables within inputs.db are named as follows:

• antibody_testing
• background_impacts
• background_mortality
• behavior_impacts
• behavior_transitions
• bool_lookup
• drug_behaviors
• fibrosis
• fibrosis_diagnosis_states
• fibrosis_states
• hcv_impacts
• hcv_states
• incidence
• init_cohort
• link_states
• lost_to_follow_up
• moud
• pregnancy
• pregnancy_states
• screening_and_linkage
• sex
• smr
• treatment_initiations
• treatments

Outputs

TODO

Previous	Next
Installation	Events