Loading…

48 Developing dummy patient data specific to Great Ormond Street Hospital

There are an increasing numbers of secondary use cases for patient data. However, access is tightly controlled and it is not possible for researchers to develop applications on patient data. These issues can potentially be addressed through anonymisation of data. However, it can be a laborious proce...

Full description

Saved in:
Bibliographic Details
Published in:BMJ paediatrics open 2023-12, Vol.7 (Suppl 2), p.A24-A24
Main Authors: Booth, John, Sheldon, Ewart J, Key, Daniel, Bryant, William
Format: Article
Language:English
Subjects:
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:There are an increasing numbers of secondary use cases for patient data. However, access is tightly controlled and it is not possible for researchers to develop applications on patient data. These issues can potentially be addressed through anonymisation of data. However, it can be a laborious process to remove all identifiable data and for many purposes, e.g. testing new systems, alternative datasets would be equally effective. There are anonymised datasets e.g. MIMIC-III, that could be used but they are disproportionately based on United States hospital data and unlikely to reflect the GOSH patient population or data structures. Also, there are no up-to-date anonymised datasets, e.g. MIMIC-III data was collected between 2001 and 2012. We developed an in-house tool to generate dummy patient data that mimicked the GOSH DRE database.Real data was analysed to provide a dataset level summary of each column, e.g. patient ages, and saved as a configuration file. A separate system was then used to generate data based on this configuration for an arbitrary number of fake patients. This resulted in dummy records that reflected GOSH data structures but were devoid of any clinically meaningful or real patient data.By generating dummy data that accurately reflects the structure of GOSH data it is possible to quickly make analytical tools without any information governance concerns or risk to patient data. Within the team it has been used to add new functionality to existing applications and has the potential of allowing collaborators to make tools for GOSH without the need to access patient data. This project has been completed by Great Ormond Street Hospital NHS Foundation Trust and Roche Products Ltd as part of a collaborative working agreement. Roche Products Ltd had no influence on the results or decision to publish regarding this work. M-GB-00013717 | August 2023Acknowledgements for funding or supportThis work is supported by the NIHR GOSH BRC. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. This work is support ed by the Great Ormond Street Hospital Children’s Charity. This project has been completed by Great Ormond Street Hospital NHS Foundation Trust and Roche Products Ltd as part of a collaborative working agreement. Roche Products Ltd had no influence on the results or decision to publish regarding this work.
ISSN:2399-9772
DOI:10.1136/bmjpo-2023-GOSH.48