A Novel Walking Cane With Haptic Biofeedback Reduces Degenerative Loading in the Arthritic Knee

Overview

The most commonly prescribed mobility aid, the walking cane, is often underloaded and therefore fails to reduce knee joint loading and provide symptomatic relief for those with knee osteoarthritis. For this study, a novel walking cane with haptic biofeedback was designed to improve cane loading. The purpose of this study was twofold; 1) to determine the effectiveness of a novel walking haptic biofeedback cane to encourage proper cane loading compared with a conventional cane, and 2) to determine whether scale training or haptic feedback influences short term retention of cane loading. It is hypothesized that haptic biofeedback would increase cane loading (H1) and decrease knee loading (peak knee adduction moment (H2) and knee adduction angular impulse (H3)) when compared to naïve cane use.

* Up to 40 individuals who have experience using a walking cane and who self-report a clinical diagnosis of knee OA can participate in this study. Participants will be excluded if they had a knee replacement in the knee diagnosed with OA, had undergone knee surgery within the past year, could not perform cane walking continuously for 30 minutes, or exhibited other neurological and/or rheumatologic conditions that would impact gait. * Participant biometrics, such as height, weight, and age will be collected in addition to information about OA and cane use history. Participants will complete the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire, from which scores between 0 and 96 can be used to evaluate the impact of knee OA on the individual's quality of life. * A haptic biofeedback cane was designed and manufactured to measure the user applied axial cane load, compare the load to the targeted 20% BW threshold, and then deliver a vibrotactile feedback in the cane handle when the loading was greater than threshold. The cane data acquisition was temporally synchronized to the motion analysis system and recorded axial cane loads at 100 Hz for later analysis. The haptic biofeedback cane consisted of a conventional bariatric walking cane (Patterson Medical Ltd) with a loadcell (Digi-Key100 LBS. Load Cell) placed inside a 3D printed modular cane foot. A microcontroller (Sparkfun™ Pro Micro), SD disk storage (Sparkfun™ OpenLog), eccentric rotating mass vibration motor (model 306-109, Precision Microdrives Limited), USB communications port, and battery completed the system. * Each participant will attend a single laboratory data collection session for the acquisition of kinematic, kinetic, and biometric data. First, the participant will be asked to walk 20 m down a hallway with a conventional cane to establish a self-selected walking speed (SSWS). Participants will then change into motion capture conducive clothing and retro-reflective markers were placed using a modified Plug-In-Gait model with additional markers on the foot, thigh, and shank, via the Kadaba Model, for observation of the Kenn Adduction Moment (KAM) during level ground walking. Markers will also placed on the cane to track its motion. * Participants will then be instructed to walk at their SSWS along an approximately 9-meter walkway containing five force-plates but will remain naïve as to the purpose of the plates throughout data collection, to avoid influencing gait. Data were collected under each of the following five conditions: 1) naïve, 2) scale training, 3) scale recall, 4) haptics training, and 5) haptics-only (Table 2). The order of the scale and haptics conditions was randomized at the beginning of data collection to avoid a learning bias. During all conditions the instrumented cane will be used, and axial loading recoded. However, haptic biofeedback from the cane will only be provided in the haptic conditions. * For scale and haptic conditions, the participants will be instructed to hold the cane in the hand opposite their most affected knee. Before the recall and haptics-only conditions, a five-minute break will be given to test short-term instruction retention. For each condition, between five and eight successful steps will be collected. A successful walk trial is defined as at least one gait cycle with the affected limb's foot cleanly striking the force plate and the cane contacting the floor, not the force plate, during stance phase. Successful trials will also have to be within ± 10% of the subjects SSWS. * Naïve condition * Participants will be instructed to "walk across the lab and use the cane as is typical." During these trials the cane will record axial load but provided no biofeedback (i.e. the cane will act as a conventional cane). * Scale conditions * During scale training, participants will practice applying 20% BW to their canes while standing, using a beam scale set to the correct weight, until they feel comfortable recreating the technique. Participants then practice applying the 20% BW to their canes while walking, no external feedback regarding the load applied to the cane will be provided, however feedback on cane technique will be. Data will be collected during these walking trials. In the scale recall trials participants are to load the cane to 20% during walking without any feedback from a scale or the cane. * Haptic conditions * During haptic training, participants will be instructed to walk around the lab and load the cane during each step until they feel a vibration in the handle. The participants will then familiarize themselves with the biofeedback mechanism and practice loading the haptic cane while walking around the lab space. In the haptic-only trails participants will be instructed to load the cane during walking until they feel a vibration in the handle * Three-dimensional kinematics will be collected using a camera 12 Vicon Nexus system (Vicon Motion Systems, Oxford, United Kingdom) with a sampling frequency of 120 Hz. Five force plates (AMTI) sampling at 1200 Hz were used for collection of ground reaction forces (GRFs) and identification of gait events during walking trials. Two time gates placed 2 meters apart will measure walking speed. * The KAM will be calculated for the arthritic knee during stance phase of each processed step using the inverse kinematics packages in Visual 3D then normalized to body weight and height. KAMs will then be exported to MATLAB for further processing and analysis. Cane load will be normalized by weight and peak cane load per step will be determined in MATLAB. PKAM is calculated as highest KAM and KAAI is calculated as the area under the KAM curve.