Introduction This project critically examines the use of language within NHS Serious Incident Reports, focusing on the balance between technical and humanistic language types. We hypothesise that the overuse of technical language in the current reports is hindering improvements by contributing to repetitive recommendations. While technical language is defined as specialised words with focus on processes, the concept of humanism is a more progressive one, aiding and affirming our ability and responsibility to engage in values-based decision making, leading to greater personal and organisational fulfilment and continual aspirations for greater good. The research aims to explore how humanistic and technical language could be used harmoniously to create a new perspective on serious incident reports. A new perspective could potentially transform recommendations, producing difference and novelty.

Aims and objectives of the research project or activity The research aims to contribute to a deeper understanding of words used in SI Reports, expressed in the following aims and objectives:

1. Evaluate the current language style and type used in Investigations and Clinical Reviews using a closed-system generative AI developed by our research team

2. 2.Investigate the potential benefits of an increased focus on humanistic language, building on the Systems Based Approach outlined in the Patient Safety Strategy

3. Develop recommendations that incorporate a more humanistic language, capturing a new perspective.

4. Implement and evaluate the impact of increasing humanistic language in a real-world setting.

Method or approach In the initial phase, a comprehensive background research revealed a repetition in current approaches, prompting the exploration of language analysis as a key avenue for improving recommendations within the NHS. Consultations with practitioners and field experts led to the strategic decision to employ AI for language analysis within SI Reports, subsequently developing ProtoNLP, an NLP system that balances human and machine interaction to achieve an insightful and applicable understanding of language use. A group of individuals from diverse professional backgrounds curated dictionaries categorising humanistic and technical language. The NLP filters out words into categories and flags up words that are yet to be categorised, providing a factual figure of the ratio between humanistic and technical language. The process is then intended to develop methodologies in co-creation with clinical reviews and investigators nation-wide to explore how humanistic and technical language can be brought into harmony in the production of reports.

Findings To date, we are witnessing an overwhelming proportion of technical language in high reports (~80% median figure).

Key messages We hypothesise that an over reliance on technical perspective in carrying out investigations and producing reports leads to a drowning out of the subjective opinion of the people involved in the incident. The outcome of this approach leads to reports which essentially all look the same. When carrying out investigations and writing reports, if we develop methods of bringing subjective experience and objective analysis into harmony the idiosyncrasies of each incident would be captured in addition to technical merits and deficiencies. This will allow the novelty contained within each incident to emerge, thereby generating difference in the recommendations. This occurs because it provides the reader to access to perspective, instead of just technical summary.