ACEMS PhD Student to Represent Australasia at International Conference

“In my first semester I failed business statistics, and I actually hated statistics when I first took it.”

ACEMS PhD student Tea Uggen (UTS) has come long way since then. In fact, the International Biometric Society (IBS) just picked her as one of only five students from around the world who will present at this year’s Young Statistician Student Showcase at the 2018 International Biometric Conference (IBC 2018) in Barcelona, Spain in July.

“I’m truly honoured to be representing the Australasia region. I really want to share my research and I think this showcase is a great opportunity to do just that,” said Tea.

Tea’s research involves using statistical methods and machine learning techniques to classify aphasia severity in stroke patients. Aphasia is a communication impairment that occurs in approximately 30% of people that have had a stroke or brain injury. Tea is developing a method to quickly and automatically diagnose the severity of the aphasia.

“Once aphasia severity level has been identified in patients, clinicians can administer the appropriate rehabilitation.” said Tea.

“Generally speaking, the longer a patient has to wait to receive rehabilitation, the lower the chances of a full recovery. It doesn’t mean you won’t fully recover. It just reduces your chances of a full recovery.”

There is a gold standard assessment used to diagnose the severity of a patient’s aphasia in Australia. This is an extensive test which takes a long time to administer and assess, and is often physically, mentally and emotionally difficult for a patient to go through.

‘Discourse Analysis’ is another common assessment used for patients with aphasia. It is made up of three sub-tasks, one of which involves a patient having to describe a picture that they’re shown. Much like the gold standard assessment, this assessment takes a substantial amount of time to score. However, it takes much less time to administer and is a much less strenuous assessment for patients to undertake. The main drawback of ‘Discourse Analysis’ however, is that it cannot be used to classify patients by level of aphasia severity. It is only used in conjunction with the gold standard test to provide more information about patients’ speech impairment.

The aim of Tea’s research is to use the information from the second, more manageable test to predict an aphasia severity level that would be in line with the gold standard test. This would reduce the burden on patients and on the healthcare system.

Tea has a three-step process for doing this. First, the patient’s dialogue from the ‘Discourse Analysis’ test is transcribed. She takes that text data and runs it through an algorithm she has spent the past two years developing in Python, which converts the text into a numerical dataset of 18 variables. Tea then imports this dataset into R and uses machine learning techniques to predict severity level for each patient. It only takes a matter of seconds.

Tea is in the third year of her PhD at the University of Technology, Sydney (UTS) and would like to keep working on this problem as she moves forward.

“I would like to keep working with aphasia after my PhD. I think the algorithm I’m developing can be improved and extended to investigate which specific area(s) of speech a patient is having difficulties with. This information could help inform clinicians of the type of therapy that might be most appropriate for the individual patient. This work will take a test that is very meaningful but clinically not used because it is time consuming and difficult, and transform it for use in everyday clinics across the country,” said Tea.

Tea is originally from Norway but came to Sydney when she was 20. She started studying business at UTS. That’s when she failed her business statistics core subject. At that point, Tea was required to meet with her UTS course advisor who told her she had to take statistics again, because it’s assumed a student will have some knowledge of statistics when they graduate with a business degree.

“I was pretty upset with that, but I made up mind that I’m not going to fail it again because once was enough,” said Tea.

Tea said she took advantage of all the help that was available to her and made sure she didn’t miss any classes. She was able to grasp the key concepts, and everything started to click for her.

“So I was able to keep up and get good grades. That’s when I realised I really enjoyed it. After that, I took as many statistics electives as I could,” said Tea.

Tea’s interests have since transitioned from business statistics to biostatistics. It’s clear she’s passionate about her research and is very excited to share it with the world at the IBC 2018.

“I am grateful and really excited about this opportunity.”

Talk about making the most of a second chance with statistics – congratulations Tea!

Tea’s primary supervisor is Associate Prof Tapan Rai from UTS. Her co-supervisors are Associate Prof Erin Godecke from Edith Cowan University and ACEMS Chief Investigator Prof Louise Ryan from UTS.

Media contact: Tim Macuga, ACEMS Communications and Media Officer, 07 3138 6741, Timothy.Macuga@qut.edu.au

Green Acorn