COVID-19 Spatial Clustering I

Mar-03 : In this lesson, we will practice reproducing a prior study.

Goals and Expectations

Before Class

Based on reading alone, draw a workflow diagram representing your best guess at Chakraborty’s research design. You will need a digital version of the workflow for inclusion in your repository, whether drawn digitally or scanned.
I suggest printing the article or using Zotero to highlight three types of clues with three colors:
- Data sources and characteristics thereof
- Methods and analysis
- Intermediary and final results that could be compared with a reproduction study
It is unlikley that you will be familiar with all of the methods. At this stage, it is OK to document a method as one “step” in analysis and take notes on context clues about how that method works.

During Class

Create a reproducible compendium for reproducing the study and add data provided by the author.
Start by forking opengisci/RPr-Chakraborty-2021, which already contains some data
Draft an analysis plan as an orgininal study for implementing the study up to the point of defining clusters to be used in GEE (generalized estimating equation) models. Use the blank .Rmd markdown file template for this.
As part of planning, I suggest updating metadata from xml format into the newer markdown format.
Convert the analysis plan of the original study into a reproduction study, including the purpose of the reproduction study, any planned devations from the prior study and methods for comparing results
Knit the analysis plan to html
Update the project readme.md and LICENSE files, and documentation of procedure_index.csv, data_index.csv, and data/metadata.
Create a version release of your compendium on GitHub prior to analyzing any data and commit a version of the repository prior to analyzing data.

Chakraborty, J. 2021. Social inequities in the distribution of COVID-19: An intra-categorical analysis of people with disabilities in the U.S. Disability and Health Journal 14 (1):101007. DOI: 10.1016/j.dhjo.2020.101007.

Use Draw.io to draw a workflow diagram representing your understanding of the research design, or draw neatly by hand or with a tablet.

Abstract: one paragraph summary of original study; one paragraph motivation to reproduce it; full reference for original study
Study Design
- short summary of plan for reproduction study (what will you compare / test from the original study?)
- subsection for “Original study design” with summary of original methods
Fill out Data, Methods, and Results sections as if repeating the original study. Flag any planned deviation for … with bold label at the beginning of the paragraph and specify in … whether the deviation aims to reproduce, replicate, or reanalyze.
In Discussion section, ruminate on the conditional implications of results: how will you know if the reproduction study “succeeded”, and what will it mean if it succeeds? Fails?

SaTScan software for spatial scan statistics, including a user guide
SpatialEpi Package R package for spatial epidemiology, and its GitHub Repository and reference manual
R Data and Methods for Spatial Epidemiology: the SpatialEpi Package
Kulldorff, M. (1997). A spatial scan statistic. Communications in Statistics - Theory and Methods, 26(6), 1481–1496. https://doi.org/10.1080/03610929708831995
The resources above are collected here

CDC DSEPD Principles of Epidemiology in Public Health Lesson 3: Measures of Risk

University of Virginia Library GEE Introduction
gee Package and reference manual
The R Package geepack for Generalized Estimating Equations academic paper by the package authors, Halekoh, Højsgaard, and Yan