First Virtual Meeting ICD10h Historic Strings (WG2)

The first meeting of the subgroup ‘ICD10h Historic Strings’ of Working Group 2: Methods and tools for standardization, coding and classifying causes of death takes place on Friday 29 November 14:00-15:00 CET.

TEAMS

The 2024 version of ICD10h comes with a set of Historic Strings in the English language. These include tidied phrases,[1] describing diseases, conditions, injuries and circumstances contributing to a death, which have been encountered in individual-level historic cause of death datasets, together with the ICD10h code and standardised description which appear in the ICD10h Masterlist. This is an extremely useful additional resource for those coding individual level deaths in the English language.

For example:

We would like to produce similar lists for other languages (as well as add to the English language list), and need the input of people coding their own datasets to ICD10h. This will involve more than just ingesting people’s coded datasets: the codes assigned will need to be checked and original cause of death strings tidied in order to make sure that strings which mean the same have been allocated the same ICD10 code, both within the dataset from which they were derived and across different datasets. We recognise the desirability of individual researchers maintaining primary use of their own dataset for analysis, so we ask that only unique strings with ICD10h codes are submitted. We also appreciate the need for the input and intellectual property of contributors to be recognised, so we propose that all contributors of data and members of the working group are listed as authors of the Historic Strings file when it is published.

We envisage the process taking place as follows:

• We start with a relatively small numbers of languages or language groups, and each language (group) forms a Sub Working Group.
• Members of the working group submit their unique strings with ICD10h codes (and source other coded data in that language if available).
• One or more members combine the different strings and run through a system of checks to ensure consistency of code allocation between and within datasets.
• Working group meets to discuss and resolve conflicts.
• HistoricStrings[Language] is deposited with authorship of the full sub working group.

[1] By ‘tidied’ we mean spellings corrected, excess white space removed, extraneous characters removed, and so on. It is also important to note that these are single causes which have sometimes been ‘parsed out’ of a longer list of causes assigned to one death.