The way we move is an indication of how healthy we are. Declines in mobility are linked to the progression of several diseases including Parkinson’s disease, multiple sclerosis, chronic obstructive pulmonary disease (COPD) and more, and the Mobilise-D project set up an easy-to-use, clinically-validated system to monitor people’s walking and gait patterns in real-world settings. This can provide an accurate and meaningful measure of disease status.
The method involves attaching a single wearable device to a patient’s lower back over a seven-day period, which gives clinicians and researchers a clearer picture of exactly how well a person is moving in real life. Algorithms developed by the project then translate the device readings into digital mobility outcomes that can indicate a change in a person’s disease status, how it is evolving over time and even predict future health events.
How the algorithms work
There are several layers of algorithms involved in the Mobilise-D offering. The first task that these algorithms do is identify those periods when a person is walking as opposed to sitting down, lying down, standing but not moving, or transitioning from sitting down to standing up. Once those periods of walking are extracted, the algorithms take four different features of a person’s mobility and use them to calculate twenty-four digital mobility outcomes that have been both analytically and clinically validated as part of the project. The digital mobility outcomes measure the amount and pattern of walking activity - as well as discrete measures of gait that capture the pace, rhythm and variability of gait patterns in the real world. Depending on how you compare to the digital mobility outcomes, your state of health can be assessed.
A clinician may decide that the best way to monitor how well your medication is working is to use the Mobilise-D system and do an at-home test. They will send you home with a device either attached to your back with a special waterproof patch made by the project, or encased within a belt. Your level of mobility will then be monitored using Mobilise-D’s 24 digital mobility outcomes over the course of a week. This system gives clinicians and researchers access to more accurate, real-world data about the patient’s mobility and represents a big leap forward for monitoring and measuring a person’s walking pattern and gait in clinical trials.
“A patient sees their clinician for a 15 minute period on one day out of 365 days of the year,” says Brian Caulfield, professor of physiotherapy at UCD Ireland, who was involved in Mobilise-D. “It’s a snapshot in time, and it really only indicates what they are capable of doing as opposed to what they actually do.”
Could I use a smartwatch or a fitness watch instead?
The difference between Mobilise-D’s offering and the fitness watches currently on the consumer market is in the validity of the data collected. While the information collected and processed by a fitness watch is interesting and pretty good for monitoring your health, the standard of data is not technically or clinically validated. The watches are often developed for and tested in young, healthy adults rather than patient populations and don’t hold weight against the stringent requirements that regulatory agencies demand for clinical trials. It’s also not proven whether consumer fitness watches are sophisticated enough to detect the slighter changes in movements that may indicate a health decline. That means that for clinical trials and patient monitoring, a consumer smartwatch won’t cut it.
Who’s using the Mobilise-D algorithms?
EMPATICA – a spin-off company from Massachusetts Institute of Technology (MIT) that specialises in clinical grade digital wearables and the algorithms that interpret their outputs – is one of the companies that’s incorporated Mobilise-D’s algorithms into their products and services.
EMPATICA provides clinical trial researchers with a one-stop-shop monitoring platform that keeps tabs on a suite of real-world data collected from clinical-grade sensors that is deemed suitable for use in clinical trials. Called the EMPATICA Health Monitoring Platform, it works by having the participants in a clinical trial wear a sensor and / or a digital wearable device – including clinical-grade fitness watches and pods that can be placed on the back – and the monitoring platform then reads out the relevant health data and draws insights based on clinically validated algorithms. The key here is that the information must all be robust and reliable enough to be used in a clinical trial – and so, Mobilise-D’s algorithms fit the bill perfectly.
“Gait and mobility measures are so versatile,” says Marisa Cruz, Chief Medical Officer of EMPATICA. “They can obviously be studied for particular gait disorders but they can also be used as quality of life measures. The Mobilise-D framework allowed us to use algorithms that have been validated in specific clinical populations that are of interest to many of our clients.”
Other companies that are using Mobilise-D’s outputs include Clario, Actigraph and McRobert’s.
Real-world data and decentralised clinical trials
The COVID-19 pandemic helped to prove that decentralised and hybrid clinical trials could be done effectively, and since then more and more pharmaceutical companies have been investigating how clinical studies could be carried out at home. Using clinical-grade digital wearable systems seems like an ideal solution.
“We’ve seen a lot of enthusiasm for either entirely decentralised or hybrid clinical study designs,” says Cruz. “The ability to collect data with more granularity and more precision in real-world situations allows for trend analysis and visibility into signs of disease progression, and that is very attractive.”
This is a win for patients too. “The feedback we’ve gotten from patients is that they are quite happy to do things at home. It even improves compliance because people feel more empowered and they find it much easier,” says Mike Jackson, director of iXscient, who was involved in the Mobilise-D project.
Open source leading to take-up
The decision to make Mobilise-D’s algorithms open-source and to translate them into the Python programming language was labour-intensive but vital to ensure the take-up of Mobilise-D’s results, according to Caulfield.
“For us to achieve real impact we needed to develop a version of our pipeline that was open source,” he says.
Their efforts have paid off - to date, there have been more than 500 downloads of the source code and four companies have integrated the Mobilise-D algorithms into their own offerings.
Mobilise-D is supported by the Innovative Medicines Initiative, a partnership between the European Union and the European pharmaceutical industry.