AI In Healthcare

An Interview with Mr. Anand N. Venkata.

Assistant Professor in the Department of Internal Medicine Division of Pulmonary and Critical Care Medicine at the University of Arkansas for Medical Sciences (UAMS). 

Through your years of working in this field have you encountered any use cases of Machine learning or AI in healthcare?

It is seen on a daily basis. The amount that we are seeing over the last 10 -15 years has gradually increased. Before we used Machine learning for simple repetitive tasks, such as checking if medicines have been administered. We use Machine learning now on a daily basis to see things such as  WBC counts of what type of blood cells are in picture, or to look at how medicines interfere which may be difficult for a human to see. It can quickly analyze different medications and their impact on the patient. Machine learning quickly adapts and detects the appropriate doses for medicines. These are more recent changes.

When you say recent changes, approximately how long has it been in use for?

More recent changes like looking at complete blood count, differentials, or EKGs and analyzing it have been there for the last 10-15 years. Some of the other applications such as looking at interactions between medications and some of the other applications have been increasing over the past five years. Some softwares used for patient care are capable of  more and some less so it purely depends on the software available in a place.

Are you taught or exposed to Machine learning in medical school?

I dont think Im the right person to ask because when I went to medical school the computing power was very different. So current teaching in medical school may be different. But there are avenues for Machine learning in  medical school and now there is specialized training for Machine learning in medicine. However,  in medical school there is a lot to be taught and since there is a lot that needs to be taught as foundational learning in medical school, I’m not sure where Machine learning fits into it.

Do you think  Machine learning can be used in conjunction with a doctor to improve diagnostic accuracy?

No doubt about it, with the increase in computing power, there has been a steady increase in the amount of Machine learning we use. We currently use it for quite a few things. There are many avenues for which we incorporate Machine learning in healthcare both on a day to day basis and also on a public health basis when you are looking at patient population. In addition it is used with clinical research to see if a project is doable and what the outcome may be before embarking on a long and expensive project. 

What advancements in this field are you most excited for in the future?

So,for my field of pulmonary critical care, the things that are showing a lot of promise are using Machine learning algorithms to look at potential what are called lung spots or lung nodules. To Identify with a higher accuracy of what will be benign and what will be malignant. This has an accuracy of over 80-90 percent however it is not widely used. Expansion of tis application can help guide both clinicians and patients into deciding if we should pursue a diagnostic procedure such as a biopsy versus not taking action.