AI tool spots human factors shaping maternity care

When an ‘adverse maternity incident’ occurs in England, investigation reports are produced to identify opportunities to improve safety. These detailed reports provide insights into clinical aspects that impacted care, such as health conditions, procedures, and tests.

Identifying where human factors played a role in these incidents can be difficult, as these are generally complex and nuanced. Currently, this labour-intensive process is carried out manually and is, by its nature, subjective. 

By contrast, the AI tool identifies and categorises human factors in reports rapidly. According to the researchers, the AI can analyse multiple reports and spot recurring factors, flagging areas likely to benefit most from additional support.

The AI model was trained and tested on data from 188 real maternity incident reports, successfully identifying human factors in each report. The work is published in the International Journal of Population Data Science (IJPDS).

"AI has transformed our analysis of maternity safety reports. We've uncovered crucial insights far quicker than manual methods," said Georgina Cosma, Professor of AI and Data Science and one of the research leads.