Opioid Overdose Regression Analysis – South Carolina

This repository supports the manuscript:

A Typology Based on Naloxone Availability, Opioid Overdose Mortality, and Emergency Department Visits in South Carolina

The study examines geographic variation in opioid overdose burden and naloxone access in South Carolina using county-level and ZIP-level analyses. The repository includes shareable county-level data, data documentation, selected derived outputs, and reproducibility notes. ZIP-level data are not publicly shared because the ZIP-level analytic dataset includes restricted health outcome data and non-public demographic, socioeconomic, community, and naloxone-access data that are not suitable for open posting under the project’s data-use and confidentiality constraints.

Repository Purpose

This repository is intended to support transparency and reproducibility for the manuscript by providing:

• Shareable county-level data and documentation
• Data provenance and variable definitions
• Selected derived outputs such as tables, figures, maps, and typology summaries
• Documentation for the restricted ZIP-level analysis workflow without releasing non-public or restricted ZIP-level data values

Study Overview

The analysis was conducted at two geographic scales:

1. County level

   • Outcome: age-adjusted opioid overdose death rate per 10,000 population
   • Main modeling approach: spatial regression, with emphasis on the Spatial Error Model (SEM) after testing for spatial dependence
   • Access measure: naloxone distribution center density per 1,000 square miles
   • Additional covariates: socioeconomic, demographic, healthcare access, and treatment access variables

2. ZIP code level

   • Outcome: average annual opioid-related emergency department visits
   • Main modeling approach: Negative Binomial regression due to overdispersed count data
   • Access measures: naloxone center counts and proximity-based measures
   • Additional covariates: socioeconomic, demographic, rurality, healthcare access, and treatment access variables
   • Data availability: ZIP-level analytic data are restricted or non-public and are not included in this repository

The study also develops burden–access typologies to identify areas with relatively high opioid burden and relatively low naloxone access.

Repository Structure

opioid-overdose-regression-analysis/
│
├── county_level/
│   ├── data/
│       ├── raw/              # Shareable county-level input data
│       ├── processed/        # Processed county-level analysis data
│       └── dictionary/       # County-level variable dictionary
│
├── zip_level/
│   ├── data/
│       └── README.md         # Restricted-data note and expected scheme
│
├── docs/
│   ├── figures/              # Selected manuscript figures and map outputs
│   └── tables/               # Shareable result tables
│
├── data-provenance/
│   └── PROVENANCE.md         # Data sources and processing notes
│
├── CITATION.cff              # Citation metadata for referencing this repository
├── REPRODUCIBILITY.md        # Notes on reproducibility and restricted ZIP-level data access
├── DATA_LICENSE.md           # License for shareable data
├── LICENSE                   # License for repository materials
└── README.md