Safe People
Lead applicant organisation name	University Hospital Birmingham

Safe Projects
Project title	Identifying and mitigating biases in perioperative prognostic models and clinical scoring systems
Lay summary	"Doctors use scoring systems to diagnose and treat patients by analyzing medical records. These scores ensure quicker, safer care but are only reliable if the data is accurate. A major issue arises when scores are based on specific age, gender, or ethnicity groups, as they may not work well for others. Scoring systems are widely used in healthcare, including surgery decisions. The process from seeing a GP to recovery is called perioperative medicine. A recent report highlighted that many perioperative scores may not work as expected, as they were created using small, non-representative groups. This could lead to inaccurate treatments, especially for minority ethnic populations. Research is crucial to understanding how well these scores perform across diverse groups. Poor advice could harm patients, particularly those from underrepresented backgrounds. The research will use anonymous healthcare data to test scoring accuracy and collaborate with statistical experts for external validation. Public contributors helped shape the research design. Findings will be shared in scientific papers, conferences, and accessible videos with subtitles in multiple languages. Additionally, a new dataset will be created to help improve scoring systems."
Public benefit statement	"Clinicians will know whether a particular PMCS is likely to be accurate for a patient they are treating. At the moment, they usually only know whether a PMCS works well across the entire population, rather than thinking about how well it works for subgroups. In particular, they will know if a particular PMCS should not be used, because it is likely to be inaccurate. Patients deciding whether to have surgery will benefit because PMCS will only be used if they work for them as individuals. This will help them work with clinicians to make the best decisions about their health. Scientists will be able to learn from the lessons we uncover about any problems with the way PMCS have been created or used in the past. This means that when they create PMCS in the future, they are likely to work better for everyone, not just a privileged few. These lessons may also be relevant when creating artificial intelligence tools for healthcare."
Latest approval date	11/07/2024

Safe Data
Dataset(s) name	PATHWAY Research Data Hub: PWY018 dataset

Safe Setting
Access type	Data released via Letter of Authorisation. All researchers have received training in the care, use and protection of personal data, enabling them to comply with their responsibilities under the Data Protection Act.

Safe Outputs
Link	Not yet Published