Excess_deaths

Code to demonstrate ONS's new method for estimating the number of expected and excess deaths, as described in the methodology article. Data published on the ONS websites were derived using a Reproducible Analytical Pipeline (RAP), starting from the raw mortality data heald in a SQL database. The code provided here is to demonstrate the method, rather than be used for production purposes. Weekly data for 2024 and beyond will be published in the weekly deaths publication.

All code was developed using R version 4.1.3.

1. Prerequisites

• RStudio installed locally
• The 'openxlsx' package installed in your R library

2. Process

1. Create a new local folder (e.g. 'D:\ons_excess_deaths') - this will be your working directory
2. Download 'ons_weekly_ed.R' and 'ons_monthly_ed.R' from this GitHub repo into your working directory
3. Download the dataset dataset20240220.xlsx accompanying the methodology article into your working directory
4. Amend the parameters at the top of 'ons_weekly_ed.R' and/or 'ons_monthly_ed.R' (see below)
5. Run 'ons_weekly_ed.R' (for weekly estimates) or 'ons_monthly_ed.R' (for monthly estimates) in RStudio

3. Parameters

• dir: the path of your working directory (where the input dataset 'dataset20240220.xlsx' is saved and your outputs will be directed); this should be entered as a character string, is case sensitive, and should be specified using single forward slashes or double back slashes, e.g. "D:/ons_excess_deaths" or "D:\ons_excess_deaths", not "D:\ons_excess_deaths"
• ref_period: the reference week (for 'ons_weekly_ed.R') or month (for 'ons_monthly_ed.R') for which you want to estimate the number of expected and excess deaths; this should be a single numeric value (only one period is allowed per run of the code), entered in the format yyyyww or yyyymm, e.g. 202301 for Week 1 or January 2023

4. Outputs

The code will output one CSV file per run, saved in your working directory. The file will be named 'weekly_ed_yyyyww.csv' (from 'ons_weekly_ed.R') or 'monthly_ed_yyyymm' (from 'ons_monthly_ed.R'), where yyyyww or yyyymm corresponds to the ref_period paramater you specified before running the code.

The output file contains estimates of the total number of observed, expected and excess deaths in ref_period (as well as 95% confidence limits for expected and excess deaths), and breakdowns of these quantities by:

• Age group and sex (aggregated across all geographies)
• Geography: the four UK countries and the nine English regions (aggregated across all age groups and both sexes)