Five key tips for using AI-based simulations

Mandates in education provision often give us challenges. Recently, the General Pharmaceutical Council (GPhC) mandated the incorporation of prescribing learning outcomes into the undergraduate pharmacy degree (MPharm). As a result, professional experiences within the MPharm programme have increased, to make sure that students are equipped with patient care and consultation skills.

This poses a dual challenge: the pharmacy profession lacks the workforce capacity to accommodate the large number of students, and universities struggle to ensure a quality-assured student experience across multiple providers. The NHS placement tariff for pharmacy students is approximately £3.40 per student per hour, not enough to cover the level of supervision needed for one-on-one training. This financial constraint also complicates the provision of traditional campus-based simulations and the scaling needed to build student confidence in practical tasks.

Is technology the solution?

Artificial intelligence-based simulations have been able to provide part of the solution to these challenges. By creating simulation scenarios based within the pharmacy workplace, students can receive training before their placement. These AI simulations provide immediate, personalised, high-quality feedback and allow for the students to repeat them, consolidating their learning. This approach reduces the burden on placement providers, enabling them to focus on providing the exposure and involvement in delivering patient care.

• A guide to making automation work in higher education
• AI as a catalyst for assessment innovation
• The two key steps to promoting responsible use of LLMs

For students, the depth and personalisation of feedback surpasses what is typically achievable in traditional simulations, which are often constrained by timetables and reliance on teacher recall. Additionally, AI simulations offer flexibility in terms of when, where and how long students engage with tasks, personalising their learning.

Putting it into practice

This approach was used at Aston University for its blended approach to professional experiences in the MPharm degree. The bespoke AI simulation scaffolds blocks of placement activity and provides students with experiences that could not be guaranteed in practice. As a result, we were able to increase our number of professional experience hours by 700 per cent from baseline.

Partner placement providers were involved in the design of the AI simulations and overall approach, so it was reflective of a broad range of needs. AI simulations contained simulated patient information such as patient charts; the student has everything they would in the workplace when they complete tasks. It has reduced the carbon footprint of training and enhanced employability, with one survey showing about 33 per cent of students receiving a job offer after their community pharmacy block. Providers have been positive about students and the preparatory training they receive.

Potential scope of the solution

This project used SimConverse and required authorship of brand-new scenarios suitable for pharmacy training, but other platforms are available, and scenarios for many healthcare professions already exist. With these existing libraries, it may be simpler for others wanting to take this approach. But the flexibility to author any form of communication-based scenario and involve partners opens scope for this technology – it could be used to replicate lawyer-client consultations, de-escalation training, marketing discussions and many others that have a basis in communication.

Advice for implementation of AI-based simulations:

1. Understand your product and the student experience it will give: AI platforms are not a solution on their own – learners will need some wrap-around support for their use and to discuss issues arising from scenarios.
2. Consider the quality of the AI scenario: Someone with significant experience of the task will need to write the scenarios, and they must be tested by an expert. One of the main reasons for authoring scenarios is to ensure that they fit the needs of your learning outcomes and discipline. Without this, scenarios are more likely to cause confusion and more wrap-around support and explanation will be needed.
3. Orientate students to AI simulation: Use a low-risk practice task to build familiarity, after discussing the theory, benefits and limitations of the method with students. Enabling students to see how this fits in the learning journey increases engagement.
4. Sequence appropriately: AI simulations are only part of the solution – students need baseline knowledge before they attempt tasks. Make sure that there is a logical sequence to the deployment of simulations. Whether the simulations come before or after placement experience will be determined by whether you are building competence at a new skill or refining existing as in stage-based learning.
5. Think about the quality assurance process: If you are authoring your own scenarios, you can build your own QA process to ensure that they remain contemporary and fit for use over time. If you are buying pre-built scenarios, check what the processes the supplier has for updating for new guidance.

Natalie Lewis is a senior academic at Aston University. The university’s work with SimConverse AI has been shortlisted for Technological or Digital Innovation of the Year in the 2024 THE Awards. A full list of nominees can be found here. The awards will be presented at a ceremony in Birmingham on 28 November 2024.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.