D1 – Sensor-enhanced Activity Management (SEAM)
Project team
Project PI – David Reinkensmeyer
Danny Zondervan, Naveen Khan, Mike Jones, George Collier, Veronica Swanson, Chris Johnson, Vicky Chan, Raeda Anderson, Amanda Rabinowitz, Edgar de Jesus Munoz Ramos, Frank DeRuyter
Purpose/aims
Participating in rehabilitation at home improves function and health. However, current rehabilitation practice struggles to effectively implement home-based rehabilitation and remains dependent on an often-ineffectual strategy – instructing patients to follow printed handouts of exercises. The overall goal of this project is to improve the quality of home-based rehabilitation by incorporating remote activity sensors into a therapy management app. Our working hypothesis is that we can substantially improve effectiveness of home-based rehabilitation by replacing printed handouts with a combination of two innovative digital health technologies: 1) app-based therapy management (Pt Pal) and 2) sensor-based, activity tracking and gamified exercise (FitMi).
We call the proposed resulting platform SEAM – Sensor-Enhanced Activity Management. The combined system will allow development of a first-of-its kind therapy management platform and “Big Data” analytic toolkit (described in D2) that will be used in this study and beyond by the wider rehabilitation community for better understanding and optimizing outpatient rehabilitation outcomes.
Aim 1 is to combine Pt Pal and FitMi to provide persons with a disability with access to an enhanced home rehabilitation environment (Years 1-2).
Aim 2 is to pilot test the combined platform with outpatients and clinicians at UC-Irvine and Shepherd Center in order to obtain detailed user feedback and refine the system based on that feedback (Years 2-3).
Status – On target
We developed an app-based version of the FitMi gamified exercises (about 40 exercises). The “FitMi subapp” runs on Android and iOS. We modified PtPal so that therapists can prescribe the FitMi exercises from PtPal, creating what we call SEAM “game mode”. This included implementing Bluetooth connectivity for the FitMi sensor puck in PTPal.
Therapists also expressed a desire to have the FitMi sensor measure repetitions of exercises from the large library of exercises already contained in PTPal. So, we also established a SEAM “rep mode” that allows this. Essentially, the puck counts reps for any exercise, reporting each rep to PTPal via Bluetooth.
A team of testers provided extensive feedback on the SEAM integration, which led to several iterations of improvements, mainly for ease of use. In addition, therapists expressed a desire to have a smaller sensor available for “rep mode”. So, we implemented connectivity with the Flint “MigoClip”, which is a smaller sensor.
We performed a pilot study with SEAM with three rehabilitation therapists and ten outpatients at UC Irvine. We tracked therapist’s attitudes and experiences, quantified patient adherence, and demonstrated success in using RPM codes for reimbursement. The therapists reported that remote monitoring and the use of a physical movement sensor was motivating to their patients and increased adherence. The paper was published:
Swanson VA, Chan V, Cruz-Coble B, Alcantara CM, Scott D, Jones M, Zondervan DK, Khan N, Ichimura J, Reinkensmeyer DJ (2021) A pilot study of a sensor enhanced activity management system for promoting home rehabilitation exercise during the COVID-19 pandemic: Therapist experience, reimbursement, and recommendations for implementation, Int. J. Environ. Res. Public Health 2021, 18(19), 10186
Based on the results of the SEAM pilot study we improved the usability of the system by improving the Bluetooth workflow, the rep counting algorithm, the visual feedback, and formfactor of the sensor we used (moving to a small clip away from the FitMi puck). We pilot tested the improve system at Shepherd center beginning in April 2022 with 10 patients. The % of exercises that were completed increased to 30% from 25% and we successfully acquired rep data for 63% of the exercises, an increase from 22% in the first pilot study. We published the results in an ACRM abstract:
Swanson VA, Johnson C, Khan N, Dzivak J, Zondervan DK, Chan V, Cowart S, Jones ML, Reinkensmeyer DJ, Uptake Rates of a Sensor for Monitoring Home Exercise Programs for Physical Rehabilitation, American Congress of Rehabilitation Medicine, 2022.
We analyzed a snapshot of 2500 users of FitMi who used the system for at least eight weeks to test whether three factors – impairment level, initial success, and steadiness of use – were associated with persevering at home rehabilitation exercise. For each factor there was an optimal range associated with higher perseverance: mild impairment (but not too mild), high levels of success (but not perfect success), and steady, regular use, respectively. We also observed that usage decelerated over time. This deceleration was attributable to an exponentially decreasing probability of initiating sessions, rather than a declining amount of exercise within a session. We published a journal article on these results:
De Jesus Ramos Muñoz E, Swanson VA, Johnson C, Anderson RK, Rabinowitz AR, Zondervan DK, Collier GH, Reinkensmeyer DJ (2022), Using large-scale sensor data to test factors predictive of perseverance in home movement rehabilitation: optimal challenge and steady engagement, Frontiers in Neurology, vol. 13, DOI=10.3389/fneur.2022.896298
The realization that FitMi progressively challenges users in a too aggressive way led to a round of modifications to the FitMi software, which were released in 2020. We have accumulated hundreds of new user’s data for the new software, and are now analyzing whether the software modifications had the desired effect of improving perseverance.
We published the results of a randomized controlled trial that evaluated the efficacy of FitMi compared to a conventional home program using a booklet of exercises following subacute stroke. Participants who used FitMi improved by an average of 8.0 ± 4.6 points on the UEFM scale compared to 3.0 ± 6.1 points for the conventional participants, a significant difference (t-test, P = .029). FitMi participants exhibited no significant changes in UE MAS or VAP scores. Thus, the sensor-based exercise system incorporating a suite of recommended design features significantly and safely reduced UE impairment compared to a paper-based, home exercise program.
Swanson VA, Johnson C, Zondervan DK, Bayus N, Mccoy P, Ng J, Schindele J, Reinkensmeyer DJ, Shaw S (2022) Optimized home rehabilitation technology reduces upper extremity impairment compared to a conventional home exercise program: A randomized, controlled, single-blind trial in subacute stroke, Neurorehabilitation and Neural Repair, 37(1):53-65
It would be valuable if home-based rehabilitation training technologies could automatically assess arm impairment after stroke. We tested whether a simple measure – the repetition rate (or “rep rate”) when performing specific exercises as measured with FitMi – can be used to estimate Upper Extremity Fugl Meyer (UEFM) score. Using linear regression, UEFM score was well estimated using the rep rate of one forward-reaching exercise from the set of 12 exercises (r2 = 0.75); this exercise required participants to alternately tap pucks spaced about 20 cm apart (one proximal, one distal) on a table in front of them. UEFM score was even better predicted using an exponential model and forward-reaching rep rate (LOOCV, r2 = 0.83). We also tested the ability of a nonlinear, multivariate model (a regression tree) to predict UEFM, but such a model did not improve prediction (LOOCV r2 = 0.72). However, the optimal decision tree also used the forward-reaching task along with a pinch grip task to subdivide more and less impaired patients in a way consistent with clinical intuition. These results show how a simple measure – exercise rep rate measured with simple sensors – can be used to automatically assess arm impairment.
Swanson VA, Johnson C, Zondervan DK, Shaw S, Reinkensmeyer DJ (2023) Exercise repetition rate measured with simple sensors at home can be used to estimate upper extremity Fugl-Meyer score after stroke, under review.
Key accomplishments
Developing the SEAM system to the point where rehabilitation therapists and patients used it in daily practice and obtained reimbursement for using it.
Improving the usability of SEAM and testing it in a second pilot study.
Determining how key factors, such as initial success at the exercise game, relate to long-term perseverance with home exercise, using a unique, large data set.
Improving the FitMi software to allow users to better optimize challenge, and obtaining a dataset to test if the changes improved adherence (analysis in progress).
Demonstrating that an optimized, home-based technology reduces upper extremity impairment compared to a conventional home exercise program in a randomized, controlled, single-blind trial in subacute stroke.
Showing how a simple measure – exercise rep rate measured with simple sensors – can be used to automatically assess arm impairment.
Challenges and course corrections
Therapists preferred prescribing their own exercises rather than selecting from the gamified exercises in SEAM game mode. We developed “rep mode” to allow this.
Accurately counting reps for a very large number of exercises is challenging. We developed an innovative solution that assigns rep credit for any movement of the sensor during the rep window.
Therapists expressed a preference for a smaller sensor than the FitMi puck. We integrated the MiGo clip to address this preference.
Usability of the system, including seamless Bluetooth connectivity, is critical for therapists and patients and improving it can improve uptake.