Quantitative Analysis of Mobilisation and Functional Status among Older Patients in Hospital Settings

Vivian Hoi Yee Cheung (2011). Quantitative Analysis of Mobilisation and Functional Status among Older Patients in Hospital Settings PhD Thesis, School of Medicine, The University of Queensland.

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
s41305584_phd_finalabstract.pdf Final Abstract application/pdf 68.29KB 0
s41305584_phd_finalthesis.pdf Final Thesis application/pdf 4.04MB 0
Author Vivian Hoi Yee Cheung
Thesis Title Quantitative Analysis of Mobilisation and Functional Status among Older Patients in Hospital Settings
School, Centre or Institute School of Medicine
Institution The University of Queensland
Publication date 2011-03
Thesis type PhD Thesis
Supervisor Len Gray
Mohanraj Karunanithi
Total pages 244
Total colour pages 26
Total black and white pages 218
Language eng
Subjects 11 Medical and Health Sciences
Abstract/Summary Prolonged bed rest and immobilisation are major factors contributing to functional decline among older patients in hospital. As patient activity is rarely documented quantitatively in day-to-day clinical practice, there is no clear guideline of how much physical activity is sufficient to maintain or improve the functional capacity of older patients. The use of accelerometry in activity monitoring may provide an objective and efficient method to measure the amount of daily activity accurately. This information can be used to promote and manage activity in rehabilitation settings. There has been limited validation work in the area of accelerometry monitoring conducted with a large sample of older mobility impaired patients in the hospital settings. A literature review was conducted to identify all studies that validated the use of raw acceleration signals from accelerometer devices together with an activity classification algorithm, to classify human postural movements and mobility into categories of lying, sitting, standing, walking, sit-to-stand and stand-to-sit transitions. Fifty-four studies were identified, and most of these studies involved attaching triaxial accelerometer devices onto the waist of their subjects. The subjects in these studies were healthy individuals (n=37) and patients (n=17). Of the 17 studies with patients, only six of them included older patients in hospital with rehabilitation needs. This review highlighted that activity monitoring with multiple accelerometer devices was shown to be accurate on healthy individuals. However, limited validation work was conducted on older mobility impaired patients with a single waist-mounted triaxial accelerometer device. The aims of this research project were to validate the activity classification algorithm using acceleration signals to classify the physical activities of older patients, and to determine whether the accelerometer devices can be applied in the geriatric rehabilitation settings as activity monitors for everyday clinical use. A framework for a series of studies was adopted to assess the efficacy of transforming the accelerometry activity monitoring method through to application in everyday clinical practice. There are six levels in this framework: technical efficacy, diagnostic accuracy efficacy, diagnostic thinking efficacy, therapeutic efficacy, patient outcome efficacy, and societal efficacy. The context of this thesis covered the first four levels of the framework and future work would be required for the last two levels. This research project consisted of four individual clinical trials outlined as follows. Clinical Trial 1 covered the technical efficacy level. The aims of this trial were to evaluate the quality of the acceleration data collected from biaxial and triaxial accelerometer devices, and to evaluate the performance of a rule-based and a wavelet-based activity classification algorithms. This trial was conducted with five healthy individuals. The triaxial accelerometer devices with the wavelet-based activity classification algorithm were selected based on their better performance over the other options. Clinical Trial 2 addressed the diagnostic accuracy efficacy level. This trial was aimed to determine the accuracy of the wavelet-based algorithm in detecting movements of older patients. Validation data were collected from 10 geriatric rehabilitation patients while they performed a standardised activity routine. The wavelet-based algorithm was found to have a low sensitivity (63%) for stand-to-sit transitions, therefore improvements to the algorithm were required. Clinical Trial 3 addressed three levels of efficacy: technical, diagnostic accuracy and diagnostic thinking efficacy. Twenty patients were attached with the device everyday during their waking hours until discharge. Their activity data were not disclosed to clinicians. The first aim of this trial covered the technical efficacy level, which was to determine the device acceptability by patients and nursing staff. Feedback of the device were mostly positive, where most patients were happy to wear the device in hospital, and most nursing staff reported that the device was easy to put on and practical for patients to wear. The second aim covered the diagnostic accuracy efficacy level, which was to collect more development data to improve the accuracy of the algorithm. An upright angle algorithm using development data collected from patients was implemented to replace the wavelet-based algorithm. This algorithm performed better than the wavelet-based algorithm, and has a sensitivity and specificity of greater than 89% for each of the activity category (lying, sitting, standing, walking, sit-to-stand and stand-to-sit transitions). The third aim covered the diagnostic thinking efficacy level, which was to determine whether clinicians were aware of their patient’s mobility level. Clinicians were asked to estimate the average amount of walking for each patient in the previous week, and these estimates were compared with the device measurements. A majority of the clinicians overestimated the amount of walking by an average of 2.64 times of that measured by the devices, and less than 40% of the clinicians’ estimates were within 15 minutes of the device measurements. Clinical Trial 4 assessed the diagnostic accuracy and therapeutic efficacy of accelerometry in activity monitoring. This trial was aimed to validate the algorithm with a new set of test data that was not used in the development of the algorithm, and to determine whether providing patient’s activity information to clinicians in the form of weekly activity reports can influence clinical decision making to promote more walking activity. Results showed that the algorithm was able to classify movements of older patients accurately with a sensitivity of >78% and a specificity of >93% for all the activity categories. Feedback from clinicians on the activity reports was mostly positive. In 88% of the surveys, clinicians reported that the activity reports were useful in determining patient’s progress, and 71% agreed that the reports were able to motivate them to increase patient’s activity level. This dissertation provides evidence for the efficacy of accelerometry in activity monitoring. In particular, it is acceptable for both patients and nursing staff; it can accurately measure the activities of older patients; it can provide activity information of patients that were previously unknown to clinicians; and the activity information can motivate clinicians to promote patient’s activity levels. More research is required to address the patient outcome efficacy and the societal efficacy. This includes whether providing activity information to clinicians and patients can increase activity levels via explicit goal setting, and constructing a cost benefit analysis on using accelerometry in rehabilitation settings.
Keyword Accelerometer
Human movement
Walking activity
Activity monitoring
Additional Notes Colour: 56, 72, 76, 79, 88-91, 100-106, 108, 111-112, 115-116, 126, 134-135, 152-154 Landscape: 47-48, 188-205, 231-238

Citation counts: Google Scholar Search Google Scholar
Created: Sun, 11 Dec 2011, 18:31:51 EST by Vivian Cheung on behalf of Library - Information Access Service