Diabetes following SARS-CoV-2 infection: Incidence, persistence and implications of COVID-19 vaccination: A cohort study of fifteen million people.

Abstract:

You can run this project via Gitpod in a web browser by clicking on this badge:

The preprint can be found
The project protocol which contains in depth details of the analysis plan can be found here (currently restricted access).
Developers and epidemiologists interested in the framework should review the OpenSAFELY documentation

Repository navigation

If you are interested in how we defined our code lists, look in the codelists folder.
Analyses scripts are in the analysis directory:
- If you are interested in how we defined our variables, we use study definition scripts to define three cohorts: pre-vaccination, vaccinated and unvaccinated. Study start dates (i.e., index) and end dates differ by cohort and are all described in the protocol. Hence, we have a study definition for each; these are written in python. Extracted data is then combined to create our final cohorts, in the preprocess data script.
- This directory also contains all the R scripts that process, describe, and analyse the extracted data.
The lib/ directory contains a list of active analyses.
The project.yaml defines run-order and dependencies for all the analysis scripts. This file should not be edited directly. To make changes to the yaml, edit and run the create_project.R script which generates all the actions.
Descriptive and Model outputs, including figures and tables are in the released_outputs directory.

Manuscript

A first manuscript is being drafted.

project.yaml

The project.yaml defines project actions, run-order and dependencies for all analysis scripts. This file should not be edited directly. To make changes to the yaml, edit and run the create_project.R script instead. Project actions are then run securely using OpenSAFELY Jobs. Details of the purpose and any published outputs from this project can be found at this link as well.

Below is a description of each action in the project.yaml. Arguments are denoted by {arg} in the action name.

vax_eligibility_inputs
- Runs vax_eligibility_inputs.R which creates metadata for aspects of the study design which are required for the generate_study_population actions.
- Creates dataframes that contain dates for each phase of vaccination and conditions for defining JCVI vacciantion groups
generate_study_population_{cohort}
- There are three generate_study_population scripts that are run to create the two study populations: vaccinated and unvaccinated. These are study_definition_vaccinated.py, study_definition_electively_unvaccinated.py and study_definition_index.py.
- These scripts are used to define JCVI groups, vaccine variables, variables used to apply eligibility criteria, outcome variables and covariate variables. Common variables used in all three scripts can be found in common_variables.py.
preprocess_data_{cohort}
- Runs preprocess_data.R to apply dataframe tidying to input_{cohort}.feather (generated by generate_study_population_{cohort})
- Tidies vaccine variables, determines patient study index date, creates addtional variables (e.g. covid phenotype), tidies dataset and ensures all variables are in the correct format e.g. numeric, character etc.
stage1_data_cleaning_both
- Runs Stage1_data_cleaning.R
- Applies quality assurance rule and inclusion/exclusion criteria
- Outputted dataset is analysis ready
stage1_end_date_table_{cohort}
- Runs create_follow_up_end_date.R which creates a dataframe of patient study end dates for each outcome of interest.
stage2_missing_table1_both
- Runs Stage2_missing_table1.R
- Check for missing data within variables
- Creates the summary statistics for Table 1 of the manuscript which can then be outputted
stage4_table_2_{cohort}
- Runs table_2.R which calculates pre- and post-exposure event counts and person days of follow-up for all outcomes and subgroups.
- Used for Table 2 in the manuscript
stage4_venn_diagram_both
- Runs venn_diagram.R
- Creates venn diagram data for all outcomes reporting where outcomes are sourced from i.e primary care, secondary care or deaths data
Analysis_cox_{outcome}_{cohort}
- Runs 01_cox_pipeline.R
- Each action runs all subgroups for the outcome and cohort of interest
- Detailed descriptions of each script used to fit the cox models can be in the model README file
stata_cox_model_{outcome}_{subgroup}_{cohort}_{time_periods}
- Runs cox_model.do and cox_model_day0.do
- When models do not fit successfully in R they are run in stata instead
format_stata_output
- Runs format_stata_output.R which combines all the stata results in one formatted .csv file
format_R_output
- Runs 07_combine_HRs_to_one_file.R which combines all the R results in one formatted .csv file

Creating the study population

In OpenSAFELY a study definition is a formal specification of the data that you want to extract from the OpenSAFELY database. This includes:

the patient population (dataset rows)
the variables (dataset columns)
the expected distributions of these variables for use in dummy data

Further details on creating the study population can be found in the OpenSAFELY documentation.

About the OpenSAFELY framework

The OpenSAFELY framework is a Trusted Research Environment (TRE) for electronic health records research in the NHS, with a focus on public accountability and research quality. Read more at OpenSAFELY.org.

Licences

As standard, research projects have a MIT license.