SUMMARY

• Incorrect walker is associated with increased fall risk in older adults, but little is being done to correct the issue.
• Our product, the StrideTech Go, measures the two most common methods of incorrect walker use: 

1. Excessive weight through the handles (measured as left- and right-hand force on the left- and right-hand handles)
2. Excessive distance between the user and the device (measured as the distance between the user’s hip and the frame of the walker) 

• Calibrating and characterizing weight placed on the walker had continual challenges, including placement, calibration, and sensor drift
• StrideTech has successfully produced the first system to measure weight placed on the walker that utilizes commercially viable sensors. 

‍

INTRODUCTION

Stride Tech Medical Inc.’s mission is to prevent falls. Seniors widely use walkers to maintain mobility while reducing the risk of falls. Despite the benefits, habitually poor walker use, marked by excessive weight on the walker handles and/or excessive distance between the user and the walker, can lead to muscle atrophy, poor posture, and falls. A widely publicized investigation in 2009 showed that over 87% of severe falls with an assistive device occurred with a walker. They recommended increased time devoted to fitting and education on proper use. Eleven years later, most seniors still do not receive individualized fitting or training on how to use their walkers.

Our product, the StrideTech Go (STG), is an attachable walker accessory that integrates sensors and biofeedback onto existing walkers to correct common misuses in real-time. Grip covers are embedded with sensors and Velcro over the handles of a walker. An additional sensor is mounted to the frame which measures the distance from the user’s hip to the walker frame. The grip covers vibrate if the sensors detect either of the two primary indicators of walker misuse:

• Excessive weight through the handles
• Excessive distance between the frame of the walker and the use

StrideTech Go is the first commercial product to help fill the urgent need for long-term walker use training. This white paper will outline the technical background and testing done to establish efficacy and briefly outline the next steps and planned product improvements. 

‍

BACKGROUND

When designing and testing the weight bearing system we had three key criteria that were essential to the overall product:

1. Cost
2. Accuracy 
3. Size

These criteria narrowed sensor selection to a type called force sensitive resistors (FSR). FSRs correlate a change in force to a change in electrical resistance, which can be read as a change in voltage. The change in voltage is then calibrated against known forces or loads.  

The purpose of this calibration procedure is to demonstrate the effectiveness of using low profile, consumer electronic grade sensors to measure loading at the handles, as opposed to load cell sensors that cost up to thirty times as much. 

‍

PROTOCOL

StrideTech defines characterization as the ability to distinguish between types of walker misuse using data alone. The data presented was collected in Denver University’s Human Dynamics Laboratory. Calibration was first established by hanging weights of known values off the midpoint of each handle system. To calibrate against user scenario weight bearing, a team member would stand on a laboratory force plate. Then, they would slowly lean onto the walker handles until the handles were completely loaded. Since the weight of the team member could be measured off the force plate, the weight placed on the walker handles could be calculated from subtracting the force plate readout from the team member’s starting weight.  

To characterize the FSR system, the team member used the walker under four different walker misuse tasks.

The four walker misuse tasks were:

1. Upright, correct walker use
2. Incorrect walker use: high hip distance, correct weight-bearing
3. Incorrect walker use: correct hip distance, high weight-bearing
4. Incorrect walker use: high hip distance, high weight-bearing

For each of the incorrect use scenarios, the team member would start from standing, exaggerate the incorrect use to their maximal range of motion, then begin walking while attempting to maintain position. The StrideTech hip-distance and weight-bearing data were graphed and categorized to establish if patterns in the data could be established and to ensure reliable accuracy of our system’s sensors as compared to a research standard instrument. 

‍

RESULTS

Case Study 1: Calibration of the FSR system

Due to the noise and nature of FSR data, raw data cannot be used to calculate pound-force or weight on the walker. The first step in establishing the accuracy and reliability of the system was filtering data. Next, we selected a calibration model that could consistently approximate pound force from the raw ADC output of the FSRs. Below is a graph of actual weight data (both from hanging weights and user loading) in blue, and predicted weight data in red. 

The accuracy and fit of this model were deemed sufficient for use as calibration of raw FSR output in the device to pound-force values and thresholds for feedback.

‍

Case Study 2: Characterization of Walker Misuse measured by the FSR system

Next, the team tested to establish differences in pound force for different walker misuse scenarios: proper walker use, high hip + high force, high hip + proper force, and proper hip + high force scenarios. Raw data was filtered, then calculated to their corresponding voltage output (mV) for ease of visuals and passing through different calibration models.

Even simulated walker use scenarios resulted in much larger ranges of FSR readings than stationary walker loading, which the team theorizes is due to difficulty maintaining a constant handle loading while moving.

For proper posture trials the average left FSR output was 214 mV, and the average right FSR output was 152 mV. For high hip distance, and proper force loading, the average left FSR output was 503 mV and the average right FSR output was 279 mV. For proper hip distance, high force loading, the average left FSR output was 910 mV and the average right FSR output was 592 mV. For the high hip distance and high force loading trial the average left FSR output was 917 mV and the average right FSR output was 622 mV. 

As expected, significantly larger FSR output values can be seen in high force scenarios, usually increasing by about 300 mV. Interestingly, higher average FSR values were also seen for the high hip, proper force loading posture. This may indicate that increasing distance between the walker and the hips naturally leads to an increase in force loading at the handles. 

Combined with hip distance readings, a clear visual difference in walker use can be seen in these four scenarios. Overtime, this data may be useful to users and health professionals to offer specific suggestions for improvement on walker use.  

‍

CONCLUSION

Improvements to sensor and calibration procedures are necessary. Most notably, significant differences between left and right hand FSR outputs could be seen throughout the scenarios. The team theorizes this is due to different “baseline” force loadings on the FSR system before any weight is placed on them. This was likely due to a difference in how tightly each grip cover was secured over the FSR insert. If the left grip cover was secured more tightly, the left FSR assembly would be under higher compression and would therefore display a higher mV readout than the right FSR assembly- even though no physical weight is on either handle. 

Since this testing session, StrideTech has implemented a calibration procedure which minimizes the effects of the grip cover on the FSR before weight is detected. 

To be able to effectively correlate data with real behaviors, more data collection is needed, particularly with older adult walker users as there may be an additional mode of misuse we can not replicate. Additionally, more testing with people of different heights, weights, and mobility will allow us to have a greater repository of weight bearing ranges that may be normal vs. unhealthy.