Harnessing machine learning to help patients with ALS

Research Lives: Anna Markella Antoniadi, PhD candidate, FutureNeuro centre for chronic and rare neurological diseases

Anna Markella Antoniadi’s project has focused on patients with amyotrophic lateral sclerosis (ALS). ‘It has been quite rewarding to see everything fitting together.’

Anna Markella Antoniadi’s project has focused on patients with amyotrophic lateral sclerosis (ALS). ‘It has been quite rewarding to see everything fitting together.’

 

What inspired your interest in using machine learning in healthcare?

I studied computer science as an undergraduate in Athens, where I grew up, and I went on to do a master’s degree in biostatistics in Glasgow. I liked that biostatistics applies to real-world problems, and my research there used machine learning to look at data from patients who had heart failure.

What prompted you to move to Ireland?

My partner and I moved to Dublin, and I got a PhD position at University College Dublin and FutureNeuro with Dr Catherine Mooney, to work more on how machine learning can analyse healthcare records.

The idea is that machine learning might be able to find less linear links between patient data and their needs, and this could help to support clinicians when they are planning care for the patient.

Tell us about the project you have been working on.

My project has been looking at patients with ALS, or motor neurone disease. Over the years, Prof Orla Hardiman and her team at Trinity College Dublin have worked with groups across Europe, and have gathered data about ALS patients with their consent.

With funding from the Health Research Board and other agencies I was able to interrogate these anonymised data, and additional information that the team was able to provide from consenting caregivers and patients, to explore what factors could be likely to affect their quality of life.

What did you find, using this machine learning approach?

There were some aspects for the patients – like the timing of when the disease symptoms started and whether they have issues with breathing when lying down – that could reduce their quality of life. Also for primary caregivers, how they view their role and purpose seemed to be linked to their quality of life.

How might the technology be used to help people with ALS?

The models that we made can be used as part of a clinical decision support system, which could automatically flag up to a nurse or doctor a pattern of patient or caregiver characteristics that suggests the patient or caregiver might be at risk of greater psychological stress or a lower quality of life. This would help them to build a personalised plan to support the patient and caregiver.?

What has kept you going through the research?

The human side of it. I was able to visit an MND clinic and observe some of the sessions with the consent of those attending, which gave me an important context – these data aren’t just numbers I was working with on the computer, we are talking about real-world conditions and interactions.

Also we did a user study on a prototype clinical decision support system with clinicians, to see whether and how clinicians would use such a system, and it was encouraging to see our research being translated into a real-world context.

You recently wrote up your thesis, how did you find that?

It has been quite rewarding to see everything fitting together. I was also able to move back to Athens and I will defend my thesis online, which is easier for all the examiners than travelling.

And finally, how do you like to take a break?

I like to do creative things and work with my hands, to get a break from the computer. During the lockdown in Ireland I made and decorated cakes and I also did embroidery. I find it’s a good balance to sitting looking at a computer screen.”

Futureneurocentre.ie