Welcome to This Week’s Edition of the Lo Down!

This week I was still fairly under the weather from this lingering cold. And I likely gained a few pounds because of a lack of rugby training.

Let’s dive in. 🚀

👨🏻‍⚕️ AI and Healthcare

The following are my thoughts around healthcare. I am not a physician and this is just a summary of my own personal experiences associated with the industry. In no way should my thoughts be construed as a reflection of the experiences of real practitioners.

There are a few aspects that I come to understand AI and health. I interviewed Dr. NT Cheung, the Head of IT at the Hong Kong Hospital Authority, on my YouTube channel where we did discuss a few aspects in which AI could impact healthcare based on their implementation. Another is through my work with Dr. Hunaid Gurji, my YouTube show guest and frequent collaborator with me on AI and its uses. I also advise one medical startup, focused on the fertility space, as well as having a friend focused in the AI diagnostics space.

In terms of diagnosis, the problem in which my friend is solving is where you have an MRI scan at a scheduled time period, say annually. That MRI scan will be fed into a diagnostics AI, and the patient would await the prognosis. Assuming a clean bill of health, the patient would then scan again at the set scheduled time, and then the AI would have two records to analyse. And so on and so forth.

This is meant to detect any changes to the body, however slight. He had quoted that not only would patients be better served by earlier diagnosis, but there are certain diseases that are only solvable by a much earlier diagnosis.

In terms of the startup I have advised around fertility, the idea is to use AI to generate snapshots of sperm health, using only the front-facing camera on a home kit where the customer can test in the comfort of their own home. The AI innovation is being able to evaluate with the world-renowned CASA analysis, using just the front-facing camera and AI that’s been trained over thousands of samples.

Both these direct experiences lend me to skew towards healthcare startups that are super specific in their focus and diagnosis. But having said that, there are more use cases that are not just diagnostic focused, but have ancillary effects to the healthcare industry that can have massive cost savings.

For example, in my interview with Dr. NT Cheung, we discussed the use of generative AI to write patient reports. This helps save doctors time and frees up their efforts for other things.

My friend Dr. Hunaid Gurji says that he uses AI to do medical training; so instead of having physicians and RNs train once every two years, they could do once a quarter, making sure their skills are fresh.

Both these use cases from a real implementation perspective save time from an organisational purpose, but also ensures that physicians and medical practitioners are able to minimise errors. This ultimately saves both time and money on misdiagnosis and mistreatment.

There’s also a whole emerging space around drug discovery and development. Using AI to compact drug discovery could potentially shorten both the discovery and approval time, allowing drugs to go to market faster.

Having said all this, there are ethical considerations around AI and healthcare. Doctors do need to understand why an AI has made a recommendation, which AI isn’t well known to do. The datasets that the AI is trained upon is subject to bias, so it’s important that the data sources are attuned to the patient. Privacy and security is paramount as well since the data used to train the AI is personal.

Subscribe to Default to read the rest.