6689 How do we protect training in an era of chronic understaffing?

Objectives The 'Good Rostering Guide'1 published by NHS Employers recommends that 'junior doctor' rotas should be structured around training needs and ensure sufficient in- built time for training. However, with only a quarter of NHS staff working rotas that are sufficiently staffed,2 training is often sacrificed to maintain service provision. Template rotas cannot adapt to the challenges of understaffing or increasing numbers of less than full time (LTFT) staff. Current automated rostering solutions, which have been shown to be effective in reducing the number of unfilled shifts,3 attempt to mimic the heuristic shift-centric approach to rostering. In this project we investigate whether moving from shift-centric to a clinician-centric rota-design approach can safeguard training needs while maintaining safe staffing levels. Methods A new algorithm was developed jointly with trainees as part of Lantum (workforce management platform) to create estimated generic work schedules that can adapt to the number of clinicians available. The output was compared to a traditional rolling rota template for rotas with different degrees of understaffing. Specific comparison was made for the number of: 'Training shifts' (defined as clinical shifts with direct consultant supervision), based on targets defined by each rota manager. Unfilled essential shifts (defined as shifts that would require temporary staffing cover). Results At the time of submission data is available for 1 pilot comparison of a rota with 6.6 full time equivalent (FTE) clinicians compared to 9 FTE clinicians required for the existing rolling rota template. Using the novel predictive algorithm all clinicians could take their full annual leave and study leave entitlements and were allocated at least 5 'training shifts' every 6 weeks, corrected for FTE. Using the rolling rota template this was only achieved for 4 of the 8 clinicians as 'training shifts' were sacrificed to cover unfilled shifts. This predominantly affected full-time clinicians. The rota produced with the novel algorithm also had a 7 fewer unfilled shifts (35 vs 42). Data collection from 15 additional comparisons is underway and will be ready to present by the time of the conference. Conclusion Novel approaches, utilising advanced predictive and optimisation algorithms have previously demonstrated improved efficiency of medical rotas in a time where understaffing is common.3 Thus far these algorithms have been implemented in a heuristic manner – imitating the processes used by humans to generate rotas. However the data presented suggests that moving from the current shift-centric approach to a clinician-centred model may yield additional training benefits. References Good rostering guide, NHS Employers, 2018. NHS in a nutshell: NHS workforce, The King's Fund, 2023. The Good, The Bad and the Rota – solving workforce challenges and promoting flexible working through clinician-led innovative rostering technology, Archives of Disease in Childhood, 2023.