Can you tell us about your work at MPS?

I’ve been at MPS, of which Medical Protection is a part, for a while now - over 16 years. The reason I have stayed at MPS for so long is because I truly believe in the values of the organisation and the high quality of the service and expertise that we offer to our members. I have seen it first hand from my role as a Medicolegal Consultant handling cases through being our Underwriting Lead, to now being the Deputy Medical Director. I can see the difference that the high-quality service makes to the lives and careers of members, particularly when they are facing challenging situations, and it’s a standout element of Medical Protection for me. 

How do you see your role as Deputy Medical Director supporting members?

One of my main responsibilities is to represent the member voice within the organisation. MPS is a mutual, not-for-profit that is owned by our members, so it is key that the member voice is heard and at the heart of all the decisions we make. 

Therefore, the other key element of my role is to make sure that I am getting out and speaking to our members and any relevant stakeholders as it is key for me to understand what’s changing within healthcare and what member’s thoughts and concerns are so that I can represent this view within MPS to ensure our members’ needs are met.

Via meeting with members, I also want to help increase awareness that MPS is not just there for them when something goes wrong, we are there to assist them throughout their career by providing educational material, an anonymous counselling service, lobbying on behalf of our members regarding important medicolegal issues, funding research into clinician wellbeing, and so much more. I want to ensure that members are aware of and access the full benefits of MPS, in addition to the exceptional service when they need us most.

What role do you think AI should (and should not) play in healthcare?

I think it is inevitable that AI is going to play quite a big role in healtcare. There is a real push for efficiency and AI has the potential to assist this while also improving patient safety. However, I think it is important that if people are using AI, that they are fully aware of the risks of using it and that they are fully trained and competent on the platform they are using.

It’s important that you feel able to explain how the AI works, what specific task it is being used for and where any data is being stored or used to gain consent from patients if using it. You and your patients need to be aware of the risks that go alongside it. Essentially, the clinician needs to remain in control of the AI, rather than the other way around.

How do you feel about the increasing reliance on digital triage systems and AI?

I think it’s crucial to remember that AI is not one homogenous entity, but that many different types of AI tool exist from AI scribes to recommender AI systems used for diagnosis and treatment, and any use or reliance will vary depending on the type of tool used.

The AI scribe is more about assisting you with administration. Hopefully it is enabling doctors to concentrate more on the actual consultation with the patient and resulting in better engagement. It takes some pressure off due to not having to remember everything in that moment so you can document it later which can be very distracting. Whereas AI triage is moving into that recommender style where potentially it will assist you with clinical decisions and that’s a whole different risk profile to consider.

It makes sense to utilise AI for some tasks where it is beneficial to you and the patient but it’s key to understand the risks, ensure you have had appropriate training on the system you are using and that you can fully understand the specific role of that AI tool including any limitations. Ultimately, it is the clinician who should still be in control of any clinical decisions and ultimately where any liability is likely to fall. It is important that you, as the clinician, always have the final decision in relation to patient care.

Also, it's important to remember as we rely more on AI systems that you need to be able to revert to not using them too. You need to have a backup system in place if something goes wrong with the AI or the patient does not consent to it being used in their care. It is important to not erode the skills of documenting notes appropriately, making clinical decisions yourself or providing treatment without AI, and other core skills. It is vital you that if you need to go back to previous processes you can without any detriment to the patient 

With AI tools becoming more common in triage, documentation, and diagnostics, what do you think constitutes adequate patient consent?

Consent is going to vary depending on what type of AI tool you are using. For example, an AI system involved in documentation or note taking is likely to be a simpler consent process than when you are using AI for any clinical decisions, but the key message would be to have a clear policy in place for what consent process should take place. I saw an example just the other week on social media of a patient who checked their electronic record and saw a note that they had consented to AI being used. They couldn’t recall ever being asked this and did not object to the AI but rather that they had not been consented appropriately or even to their knowledge.

If you are going to get consent from a patient about AI you are going to need to have a good understanding of what the AI system does, the benefits and limitations of it, and also where the data is being stored, how it’s being used, and who has access to it. We are increasingly seeing via our cases that one of a patient's main concerns is the data use and storage, so it is vital that you have a clear policy and understanding regarding this element. 

Medical Protection launched the AI Safer in Practice framework to help members think through how best to approach these sorts of challenges which you can explore here: https://www.medicalprotection.org/ai-framework