Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring
<p>Current orthopaedic practice lacks low-burden tools for objectively quantifying patient recovery in the first weeks following knee replacement surgery. Postoperative functional tests are burdensome to the clinical care team and unsuited for assessment of at-home recovery, while most patient...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: |
| _version_ | 1797107598786297856 |
|---|---|
| author | Small, SR |
| author2 | Khalid, S |
| author_facet | Khalid, S Small, SR |
| author_sort | Small, SR |
| collection | OXFORD |
| description | <p>Current orthopaedic practice lacks low-burden tools for objectively quantifying patient recovery in the first weeks following knee replacement surgery. Postoperative functional tests are burdensome to the clinical care team and unsuited for assessment of at-home recovery, while most patient reported outcome measures are not validated for perioperative use and may not reflect objective improvement in postoperative mobility. The aim of this thesis is to establish a methodology for objectively assessing the trajectory of early postoperative recovery following knee arthroplasty.</p>
<p>A scoping review was conducted to evaluate the use of wearable sensors in knee arthroplasty clinical care, finding wearable-based gait analysis and physical activity monitoring as the most common applications of wearables in this area. This review highlights a significant missed opportunity to leverage physical activity monitoring for longitudinal assessment of patient recovery in the early postoperative period.</p>
<p>Initially, fifty-four healthy volunteers were recruited for 24 hours of free-living activity monitoring. Reduced sampling rate was found to underestimate overall physical activity at both the wrist and hip body placements using the Axivity AX3 accelerometer. Transformations were subsequently generated to enable the direct comparison of accelerometer data collected at 25 and 100 Hz. Ground truth steps of thirty-nine healthy volunteers were recorded during 1 hour of free-living activity. A hybrid machine learning model was developed to count steps from a wrist-based accelerometer. In external validation, this model predicted steps with a 0.8% mean bias and 5.1% mean absolute percent error, compared to >20% error common to most current wrist and hip-based sensors. When step counts were derived for UK Biobank participants, daily steps and peak cadence were significant factors differentiating participants with lower limb arthritis and non arthritic matched controls.</p>
<p>In a clinical study, patients undergoing primary knee arthroplasty were recruited from the Nuffield Orthopaedic Centre. Baseline preoperative physical activity data was collected from 141 patients, while 105 patients contributed up to 42 days of postoperative activity data, with high wear compliance during both preoperative and postoperative monitoring periods. Increase in week-to-week postoperative physical activity was documented, culminating in a median step count and acceleration of 3,983 steps/day and 19.5 mg, respectively, during the sixth postoperative week. Significantly increased recovery of physical activity was observed in patients undergoing unicompartmental versus total knee arthroplasty.</p>
<p>Postoperative physical activity monitoring was highly acceptable in the clinical population and demonstrated utility for tracking recovery at the individual level and differentiating outcomes between surgical cohorts. Further implementation of activity monitoring in the clinical orthopaedic setting can be a useful means for assessing early postoperative patient recovery while additionally serving as an objective clinical trial outcome measure.</p> |
| first_indexed | 2024-03-07T07:18:17Z |
| format | Thesis |
| id | oxford-uuid:5a9e41bf-9da1-4e5c-951b-aa2fc9a52cca |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:18:17Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | oxford-uuid:5a9e41bf-9da1-4e5c-951b-aa2fc9a52cca2022-09-07T10:32:46ZIdentifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoringThesishttp://purl.org/coar/resource_type/c_db06uuid:5a9e41bf-9da1-4e5c-951b-aa2fc9a52ccaTotal knee replacementActivity trackers (Wearable technology)OrthopaedicsEnglishHyrax Deposit2022Small, SRKhalid, SDoherty, APrice, A<p>Current orthopaedic practice lacks low-burden tools for objectively quantifying patient recovery in the first weeks following knee replacement surgery. Postoperative functional tests are burdensome to the clinical care team and unsuited for assessment of at-home recovery, while most patient reported outcome measures are not validated for perioperative use and may not reflect objective improvement in postoperative mobility. The aim of this thesis is to establish a methodology for objectively assessing the trajectory of early postoperative recovery following knee arthroplasty.</p> <p>A scoping review was conducted to evaluate the use of wearable sensors in knee arthroplasty clinical care, finding wearable-based gait analysis and physical activity monitoring as the most common applications of wearables in this area. This review highlights a significant missed opportunity to leverage physical activity monitoring for longitudinal assessment of patient recovery in the early postoperative period.</p> <p>Initially, fifty-four healthy volunteers were recruited for 24 hours of free-living activity monitoring. Reduced sampling rate was found to underestimate overall physical activity at both the wrist and hip body placements using the Axivity AX3 accelerometer. Transformations were subsequently generated to enable the direct comparison of accelerometer data collected at 25 and 100 Hz. Ground truth steps of thirty-nine healthy volunteers were recorded during 1 hour of free-living activity. A hybrid machine learning model was developed to count steps from a wrist-based accelerometer. In external validation, this model predicted steps with a 0.8% mean bias and 5.1% mean absolute percent error, compared to >20% error common to most current wrist and hip-based sensors. When step counts were derived for UK Biobank participants, daily steps and peak cadence were significant factors differentiating participants with lower limb arthritis and non arthritic matched controls.</p> <p>In a clinical study, patients undergoing primary knee arthroplasty were recruited from the Nuffield Orthopaedic Centre. Baseline preoperative physical activity data was collected from 141 patients, while 105 patients contributed up to 42 days of postoperative activity data, with high wear compliance during both preoperative and postoperative monitoring periods. Increase in week-to-week postoperative physical activity was documented, culminating in a median step count and acceleration of 3,983 steps/day and 19.5 mg, respectively, during the sixth postoperative week. Significantly increased recovery of physical activity was observed in patients undergoing unicompartmental versus total knee arthroplasty.</p> <p>Postoperative physical activity monitoring was highly acceptable in the clinical population and demonstrated utility for tracking recovery at the individual level and differentiating outcomes between surgical cohorts. Further implementation of activity monitoring in the clinical orthopaedic setting can be a useful means for assessing early postoperative patient recovery while additionally serving as an objective clinical trial outcome measure.</p> |
| spellingShingle | Total knee replacement Activity trackers (Wearable technology) Orthopaedics Small, SR Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title | Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title_full | Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title_fullStr | Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title_full_unstemmed | Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title_short | Identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| title_sort | identifying the trajectory of normal recovery following knee arthroplasty via physical activity monitoring |
| topic | Total knee replacement Activity trackers (Wearable technology) Orthopaedics |
| work_keys_str_mv | AT smallsr identifyingthetrajectoryofnormalrecoveryfollowingkneearthroplastyviaphysicalactivitymonitoring |