Digitising clinical assessment and using it for evidence-based prediction of outcomes

Bookmark and Share
Published: 15 Apr 2019
Views: 848
Prof Jorge Nieva - Keck School of Medicine, Los Angeles, USA

Prof Jorge Nieva speaks to ecancer at the 2019 American Association for Cancer Research (AACR) meeting about the digitising of clinical assessment and using it for evidence-based prediction of outcomes.

Prof Nieva explains that they have started using smartphones and video game technologies in order to characterise how the patient is physically coping throughout treatment.

He reports that, as well as using an infrared camera to analyse patients' movements, they are given a fitness monitor which records their physical activity.


Yesterday we talked about digitising cancer and digitising the patient. One of the things we want to do is be able to understand how cancer evolves over time and so have a way of characterising the tumour as it evolves from one treatment to the next because really it’s different, it’s no longer the same cancer after it’s been exposed to therapeutic pressure. The other thing we talked about is digitising information about the patient, so we talked about how we use tools from video gaming technologies to smart watches to really characterise what’s happening with the patients when they’re out of the office as well as when they’re in the office. We use tools such as an Xbox Kinect to measure how the patient moves in the office in order to provide a reproducible, verifiable measure of performance status. We all know that performance status is the most important metric when it comes to determining how a cancer patient will do. So by doing this in the office we can predict which patients are likely to have complications from their treatments.

How do patients interact with the tech?

We try to make it as simple as possible for patients who participate in our clinical trials. We issue them a smart phone, a wearable fitness monitor as well as a connected scale. All that information all uploads seamlessly into the smartphone that they’re provided with and then all that data goes into our servers. Sometimes for the measurements that are done in the office we’ll actually image the patient while they’re walking from the chair on to the examination table and that gives us a measurement of how well they do that movement which tells us a lot about their level of fitness.

What kind of measurements do you take?

By using a 3D infrared camera instead of a regular optical camera we can actually quantify the patient’s movements. We identify where their skeleton would be and then we can see that patient moving in three dimensional space, measuring their mean acceleration as they get on to the exam table which tells us a lot about whether or not they’re fit or frail.

Do these measurements involve apps and games the patient can use?

For the Xbox we don’t actually use an Xbox itself, we only use the Kinect camera part of the Xbox and we connect it actually just to a desktop computer. The patient doesn’t actually get to play the game, it would be fun if they did, but actually all they get to do is just get on to an exam table. We want the assessments to be as simple and streamlined as possible and make it possible for the physician to use those things in the office.

From the standpoint of the home kit that we provide to the patients they get a wearable fitness monitor and they just walk around and do whatever they’re going to do. What we want to know is what’s their time that they’re physically active versus the time that they’re sedentary. We know that the more hours that they spend in physical activity the lower probability that they’ll have a complication from difficult cancer treatment.

How do you analyse such a quantity of data?

We use a lot of big data – experts, mathematicians and other engineers, to try to get at this information. But what we find is that for patients when they’re getting on to the exam table you don’t have to measure everything about their movement. It’s really the movement at their spine base and the non-pivoting leg and that really simplifies the data collection process. Also when we’re looking at patients moving around in the community using the fitness monitor it’s really all about measuring their energy expenditures and really capturing the number of hours that they spend at a level which we define as low physical activity.

How do you take into account confounding variables such as diet or previous fitness?

We absolutely don’t want to get rid of those confounding factors because we think those confounding factors are actually where all the data is. So a patient who is fit, who is exercising, if you’re on your StairMaster for three hours a day we’re not too worried about whether or not you can take difficult treatments. But if you’re somebody who really can’t do much more than get out of bed then that’s a thing that’s important for us to know.

It’s really not about defining who are the uber-fit from the slightly fit, it’s really a much lower level than that that we’re looking at. We’re looking at the people who are pretty much entirely sedentary, being at an activity level for less than 5-10 hours in a 60 day period, from the people who are just modestly active.

What trends have you found so far?

What we’ve found is that unexpected healthcare encounters – those trips to the doctor’s office when you’re not feeling well, emergency room visits, trips to go into the day hospital for extra hydration or anti-nausea medicines – all of those things are higher if the patient is less physically active.

Is there anything you’d like to add?

I’d just like to add that I think these tools are a very useful tool for looking forward at how we do clinical trials and how we do clinical research. We want patients who are enrolling in clinical trials to be able to be evaluated using these tools so that we can make apples to apples comparisons of different patients treated at different centres and know that the information that we’re getting when analysing clinical trial data is accurate, is reproducible, is verifiable, no matter where that patient was treated. The hope is that we’ll ultimately be able to do clinical trials in more places, in larger communities, knowing that the information about patient health is something that is objective.

Could this be implemented in lower income countries worldwide?

The beauty of the way that we’ve set up these programmes is we’re using off the shelf consumer technology. We know that off the shelf consumer technology gets cheaper every year and gets better every year. We’re not trying to develop new technologies that compete with the big technology companies. We want to let those big technology companies make the tools for us, make them as cheaply as possible and as disseminated as possible and then we’ll leverage their capabilities on the back-end. So we know that eventually these technologies will all actually be very inexpensive.