ARMS Uses Sensors and Mobile Apps to Monitor and Analyze Patients' Therapy After Arm TraumaMarch 2014
Topics: Human Computer Interaction, Veterans Affairs, Modeling and Simulation
When working with a patient who is recovering from arm trauma—which could be the result of everything from an injury to a stroke to amputation—physical therapists have a variety of treatments from which to choose. To make sure they're prescribing the right treatment for the patient, therapists use such methods as motor tests and surveys to monitor the patient's response to therapy.
However, current methods for gauging therapeutic response can be expensive, time consuming, and imprecise. To achieve the best results, patients must often be tested in a laboratory environment. But labs don't necessarily offer an accurate appraisal of patients' progress in their normal lives.
"What therapists need is a method for measuring patients' response to therapy that is based on direct measurements of their day-to-day activities outside of the laboratory. This method would allow for a more accurate match of therapy to patient," said Adam McLeod, an artificial intelligence engineer at MITRE.
He is the technical lead on a research project team—led by Elaine Bochniewicz—designed to address this problem. "Our team works closely with the National Rehabilitation Hospital, a key MITRE research partner, which has been involved with the design and implementation of this project from day one," explained Bochniewicz.
The team has come up with a possible solution—a prototype system called Analysis for Rehabilitative Motion Sensing (ARMS), which monitors and reports on a patient's arm activities over weeks or months of time. It continuously analyzes data from a sensor mounted on the impaired arm and streams the results to a remote server for presentation to the caregiver. This system leverages sensing, mobile computing, machine learning, and data visualization technologies.
ARMS Gives Therapists More Data
The ARMS sensor comprises a small inertial measurement unit, a microcontroller, and a Bluetooth modem. Attached to the wrist of the patient's affected arm, the sensor periodically records movement data and transmits it to an app on the patient's smartphone. The app labels each sample of movement data as either "Nonfunctional" or "Functional" arm activity. (These metrics are associated with standardized Activities of Daily Living, or ADLs.) An aggregation server transmits the results to the therapist. The therapist then uses them to monitor the rate of change as the patient gains functional arm movement and immediately see the effects of therapeutic actions.
"Instead of having to rely on imprecise survey answers or on narrowly focused laboratory tests, therapists can use ARMS to acquire precise information about patients' movements over time as they conduct their daily activities," McLeod says. Effective therapy should show accelerated rates of change compared to ineffective therapy—information that allows the therapist to modify treatments in a more responsive and personalized manner than is currently possible.
Testing the Prototype
This team has succeeded in building a working prototype of the full ARMS system, which can act as a persistent collection and reporting system for many types of sensors.
"We have tested ARMS on 27 patients from the National Rehabilitation Hospital, which recruited patients for us and has provided many important components for this project. We are using the data we collected and annotated (based on a standard known as FAABOS, the Functional Arm Activity Behavioral Observation System for measuring hemiparetic arm use) to improve the ARMS system," added McLeod.
The data and its annotations are also being used to support the classification model of the sensors and will be made available to other researchers to use in their work.
The team plans to complete more experiments in the field this year, monitoring patients as they go about their daily activities. They'll use the results to further improve the visualization available to medical practitioners reviewing the data.
"Our goal is to transition this prototype to interested organizations that can supply products to patients and caregivers," McLeod says. "This would give them a valuable tool that could improve upper limb rehabilitation. At the same time, it could lower rehabilitation costs by—for example—allowing remote observation of the patient's progress."
—by Beverly Wood