NHS England’s Transformation Directorate have published a resource exploring whether an artificial intelligence tool or approach can be developed to automatically generate student nurse placement schedules that adhere to a variety of stakeholder needs and constraints, whilst providing a more diverse range of placements for the students.
The resource highlights that there are specific requirements and contexts which must be taken into account for the three key stakeholders; the trust coordinating and hosting the placements, the university providing classroom study, and the students themselves. Factors include the set dates when placements will take place, whether students are required to visit specific wards, and a hospital’s maximum capacity for hosting students.
It notes that the current manual process means that it is not feasible for placements to cover a range of disciplines and specialities, but this is something that universities would like to support to enhance skill development.
The Transformation Directorate describe how they worked with Imperial College NHS Trust and North West London CCGs to develop an AI-driven solution to this using anonymised data and randomly generated student profiles.
“The task posed here is one of optimisation, of which there are many different approaches. The method selected was a Genetic Algorithm, where the best version of something is found through a process which is like the evolutionary process seen in nature,” the resource states.
The algorithm works by creating a population of objects (in this case, a population of potential schedules for all students) and applying mutations by randomly changing the allocated wards, which produced “offspring” in the form of combined/hybrid schedules, built by putting sections from different mutated schedules together. The mutated and offspring objects were put into the population and scored them to dictate what good looks like. This process was then repeated hundreds of times until a schedule is found which meets all needs.
The resource notes that Genetic Algorithms will not produce a “perfect” solution due to the random manner in which the problem space is explored, but “a more varied selection of placements will be allocated than at the beginning, or if the process was carried out with the current, manual process.”
The scoring process means that different schedules can be compared to each other, and this score can also be broken down into components so that the strengths and weaknesses of each schedule can be seen. As such, placement coordinators can evaluate which schedule suits their trust and its students. The Transformation Directorate call the scoring process “arguably the most important part of a Genetic Algorithm” and explains how scoring components can be divided into two categories: ‘must-haves’ (such as placement settings not being allocated more than capacity allows, and COVID risk levels), and ‘nice-to-haves’ (the number of unique wards to avoid repetition for the student, and focus on specific specialisms).
Streamlit, an interactive interface for web browser-based applications, provided the user interface. Streamlit allowed the user to adjust elements of the placement optimisation tool and to view progress as the AI tool produced schedules. Once the schedules were created, a comparison table was displayed and saved.
The resource highlights how the electronic generation of the schedules means that they can be easily reformatted to provide different views of the information, improving the insights available. “This simplicity extends to being able to calculate hours on placement each week,” it adds, “which is currently a mandatory reporting ask for trusts.”
The partnership has released the resulting code as an open source on GitHub; it is available for anybody to re-use here. A pilot is planned to evaluate how well the tool works in practice, working with Imperial College Healthcare Trust during a future student intake.
Hai Lin Leung, Programme Manager at North West London CCG, commented: “We can’t fix the nursing shortage without training more nurses and for that, we need to have (the right) clinical placements available. This tool will not only help us to allocate placements more efficiently and effectively, but it will also free up valuable time for the practice learning facilitators to focus on teaching and professional development for students. Ultimately more students will be able to get the placements and training they need.”
