Trends and Variations in Associations Between Survey-Derived Individual Characteristics and Opioid-Related Adverse Events in Community-Dwelling Ontarians: 2013-2024

1. Project Goal

This project aims to develop a predictive algorithm for the risk of drug overdose following the prescription of narcotics. It utilizes a survival-analysis approach, drawing on administrative and survey-based predictors from the Canadian Community Health Survey (CCHS).

2. Project Organization

The project is organized into the following directories:

• Data/: Contains the raw CCHS data for different survey cycles (2013-2018 cycles).
• R/: Houses all the R scripts for data loading, processing, analysis, and utility functions.
• worksheets/: Includes supplementary files like variable lists and details in CSV and Excel formats.

/Users/karimhalal/Desktop/The worlds greatest thesis/Thesis/
├───.gitignore
├───config.yml
├───README.md
├───Thesis.Rproj
├───Data/
│   ├───cchs2013_2014.RData
│   ├───cchs2015_2016.RData
│   └───cchs2017_2018.RData
├───R/
│   ├───dependency_table.R
│   ├───harmonized.R
│   ├───load_dependencies.R
│   ├───loadData.R
│   ├───special_functions.R
│   ├───table-1-a.R
│   └───testing.R
└───worksheets/
    ├───cchsflow_variables_details1.csv
    ├───deptable.xlsx
    ├───masterfilesheet.xlsx
    └───od_variables.csv

3. File Descriptions

Root Directory

• .gitignore: Specifies files and directories to be ignored by Git.
• config.yml: The main configuration file that defines paths to data and variable sheets, ensuring a centralized and easily manageable setup.
• README.md: Provides a brief introduction to the project.
• Thesis.Rproj: An RStudio project file that helps in managing the project's context.

Data/ Directory

This directory stores the CCHS datasets for three different cycles:

• cchs2013_2014.RData
• cchs2015_2016.RData
• cchs2017_2018.RData

R/ Directory

This is where the core logic of the project resides.

• load_dependencies.R: Loads all the necessary R packages required for the project.
• loadData.R: The main script for data handling. It reads the configuration from config.yml, loads the CCHS data, and then harmonizes it using functions from the cchsflow and recodeflow packages to create a unified study dataset.
• harmonized.R: Contains scripts for creating and manipulating the harmonized dataset.
• special_functions.R: A collection of custom R functions tailored for specific data transformations and derivations needed in the analysis.
• dependency_table.R: A utility script that generates a table of all package dependencies for the project, which is useful for reproducibility.
• table-1-a.R: Generates a summary table (Table 1) of the dataset's characteristics.
• testing.R: Includes experimental or test code, such as a function for imputing single-year age from categorical age data.

worksheets/ Directory

This directory contains human-readable files that provide metadata for the analysis.

• od_variables.csv & cchsflow_variables_details1.csv: These files define the variables to be used in the analysis, their roles (e.g., predictor, outcome), and other details.
• deptable.xlsx: An Excel file containing the dependency table generated by dependency_table.R.
• masterfilesheet.xlsx: A master sheet for variables.

4. Architecture and Workflow

The project follows a modular, configuration-driven architecture that promotes clarity and reproducibility.

Workflow:

1. Configuration: The config.yml file acts as the single source of truth for file paths and parameters.
2. Data Loading & Harmonization: loadData.R reads the configuration and orchestrates the data loading and harmonization process. It iterates through the specified CCHS datasets, applies transformations using recodeflow and custom functions from special_functions.R, and combines them into a single harmonized_data frame.
3. Analysis: Once the data is prepared, other scripts like table-1-a.R are used to perform the actual analysis and generate results.
4. Utilities: Scripts like dependency_table.R provide helpful utilities for managing the project.

This structured approach ensures that the analysis is easy to understand, modify, and reproduce.