Blog Series - from Deathbed to Database: capturing causes of death in Scotland, 1855-1973.

Introduction - 'consistency' is key!

Understanding how difficult it has been during Covid19 to accurately compare information from neighbouring nations, makes it easier to understand how hard it must have been to compare and contrast causes of death spanning across several decades or a century. Researchers involved with the ‘Digitising Scotland’ (DS) project, had the difficult task of comparing and contrasting causes of death across a century, from 1855 up until 1973. During this period there were many changes in scientific understanding, the evolution of medical training/traditions and the development of recording practices/priorities which made it even more challenging to reach a meaningful interpretation of death certificates.

To gain a full understanding of the reason for a death; researchers had to become aware of the legislation surrounding the registering of a death and the certification of the cause of death, and what different interpretations can affect what is reported, such as:

• Who was expected to register a death or report how a person had died?
• Was the register kept by the Church or by the State?
• Was registration compulsory?
• Were there guidelines on how causes of death were to be recorded and, if so, who laid them out?
• What categories of death were given greatest prominence and why?
• Do particular terms mean the same across different times or in different languages?

Digitising Scotland

The ‘Digitising Scotland’ (DS) project, ran from 2012 to the start of 2020 and was funded primarily by ESRC, and involved members from the :

- Digitising Scotland team (made up from SCADR//SHPD colleagues at the University of Edinburgh), and the

- Britain’s First Demographic Transition team: an integrated Geography (GB Atlas) project based at the University of Cambridge and the

- Study of Health in Port Cities Project (SHiP) team based at Radboud University in the Netherlands

The first part of the DS project was to transcribe information from the indexed images (e.g. certificates stored as PDF scanned images) to create a research database containing machine-readable, individual-level, information on approximately 18 million people who had an birth, marriage or death 'event'  in Scotland from 1855. As part of this undertaking, and in order to enable analysis of the information on the death certificates, they have developed a coding system ‘ICD10h’ based on the WHO’s International Classification of Diseases, Version 10 (ICD10). 

While ICD10 contains many thousands of diseases, conditions and events which may cause ill health or death, but it did not contain historical disease descriptions, such as ‘childbed fever’, ‘teething’, ‘strangled by a peat creel’ or ‘found dead on privy’.  Therefore researchers had to add in these descriptions to the ICD10h system, which created codes, allowing researchers to assess how they might have evolved or if they have been superseded with new medical knowledge. 

A dictionary of historic causes in a variety of languages is also being built up to encourage international comparison of mortality conditions wherever individual cause of death is available. (See forthcoming edition of Historical Life Course Studies).

Our Blog Series

On further investigation, it becomes apparent that there are many reasons why death certificates maybe didn't state the precise reason for death, and in order to highlight the amazing work that has taken place by the DS project, we are developing a Blog Series, which highlights:

1. Changing causes of death over time: Register General's classification of causes of death, and how causes alter

2. Doctors and the Medical cause of death certificates - no consistency (told different things at different times) the MCDCs destroyed

3. Entries in the Civil Registers, medical/legal, different registration forms and instructions over the years - explore the registration processes - creating a death certificate

4. Transcribing the data and how to the need to 'parse' the causes (divide up the causes)

5. Coding the data - our use of ICD10h - coding the words; international comparisons (- explain the need for legislators, historians, computer scientists and linguists to help analyse the deaths between 1855 and 1973)

6. Assigning a 'primary cause', what 'additional causes' can tell us ( how doctors, lawyers and registrars, all played a part in understanding the evolving picture of how the people of Scotland met their deaths between 1855 and 1973)

7. Classifying the causes - some preliminary results (how the team managed to digitise deaths and create a standard classification of deaths)

Interesting read for Eilidh - https://www.newyorker.com/magazine/2014/04/07/final-forms