DHIN Member Spotlight: Malcolm Pradhan

Home/DHIN Member Spotlight: Malcolm Pradhan
DHIN Member Spotlight: Malcolm Pradhan 2023-07-31T14:13:04+10:00

Please tell us a little about yourself 

As a medical student at the University of Adelaide, I discovered my passion for medical informatics, then an emerging field in the USA. After gaining my medical degree, I further explored this interest through a PhD in Medical Informatics at Stanford University, studying Bayesian networks and decision theory for clinical decision support (CDS). After returning to Australia, I launched a Health Informatics program at Adelaide University and consulted on patient safety. In the early 2000s, I co-founded Alcidion, a company dedicated to improving patient safety through CDS. After leading Alcidion to a public listing in 2016 and a busy 20+ -year tenure, I stepped away in 2022. I now consult in Health AI, informatics, and go-to-market strategies for health IT start-ups.

How do you define digital health?

Image of Malcolm Pradhan
Headshot of Malcolm Pradhan

I think it’s pretty broad. Digital health is about using technology to improve health, wellness and healthcare delivery. Health Informatics is a subset of this, focusing on the content of the data; how to represent it and use it. 

What do you think will enable digital health projects and innovations to succeed?

We’re at the point the technology itself is not the barrier. The main hindrances to success are interoperability (getting data out of proprietary IT systems) and clinical workflow integration so we can help clinicians save time and improve care; in other words, how to make the right thing to do the easiest thing to do, rather than requiring of them everyday heroics to get their job done.

What do you think are the biggest challenges facing digital health at the moment? 

Interoperability is the main one. It’s not a technical issue but a purposeful strategy by companies to block data access. I believe all Health IT systems should be mandated for high degrees of interoperability. The other one is the workforce — we need cross-trained people who deeply understand the technology, clinical environment and patient engagement. 

Do you have any resources or links you would like to share?

I think the Australian Institute of Digital Health (AIDH) is a good start to get into the field. We are also trying to build a community around Evidence-based AI so it would be great participation in the LinkedIn group.

How do you see Digital Health and Artificial Intelligence intersecting?

Current forms of digital health aim to replace paper systems with a digital equivalent. Sadly, health IT hasn’t moved the needle on patient safety, and there’s good evidence that it’s worsened clinician productivity and cognitive burden. It is sobering to remember when COVID was peaking in the US, one of the first things governors did was exempt doctors from using EMRs because the systems were slowing down healthcare delivery! AI is now forcing us to rethink how to integrate technology to better support clinicians and patients. In fact, I think it will lead to us rethinking the roles of clinicians as technology becomes more capable. I believe clinicians have to make too many decisions; healthcare needs more automation so we can spend more time engaging with patients.

What do you feel is the most exciting thing happing in Health Artificial Intelligence right now?

Generative AI is in the limelight, but I think it’s too risky to use in a clinical setting in its current form because it hallucinates and can randomly omit important information. The underlying technology of ChatGPT is the transformer deep learning model (that’s what the ’T’ in GPT stands for). Transformers can also be used for prediction rather than generating text. This allows us to incorporate the contents of clinical notes into risk prediction, risk stratification, and candidate selection for pathways and clinical trials. Also, these models can incorporate multiple modalities such as images and signal data; I think this will lead to significant improvement in picking up changes in patient health trajectory earlier and allow us to personalise care better than we can today. I’m also excited about the resurgence of interest in causal models, which are one way of protecting against generative AI hallucinations.

How would you like to see Artificial Intelligence technologies transforming healthcare in future?

We will see the AI augmented clinician. This will invite us to redefine the role of the human in healthcare. Today we expect clinicians to deliver safe and efficient healthcare when we all know it’s not humanly possible without assistance, which we do not provide them. For example, how can a junior doctor optimally manage 25 sick patients on their ward when even one very sick person consumes almost all their time? Consider that 80% of healthcare is routine and could be managed through automation and enhanced monitoring. Clinicians should manage the exceptions, allowing them to focus on complex, human-centred care.
From a consumer perspective, AI promises personalised care, catering to individual healthcare literacy, culture, and engagement styles. A sustainable healthcare system must deliver care in lower-acuity settings like homes, nursing homes and clinics. This necessitates more intelligent systems for monitoring patient progress and providing engagement choices. Scaling healthcare isn’t possible if every piece of data from every patient requires manual human review.


Skip to toolbar