Electronic Thesis and Dissertation Repository

Degree

Doctor of Philosophy

Program

Epidemiology and Biostatistics

Supervisor

Dr. Mark Speechley

Abstract

Community mobility (CM) is an important instrumental activity of daily living associated with quality of life and independence. Measuring the CM of older adults, particularly those with gait disorders such as Parkinson’s disease (PD), is an important way to understand how to help people maintain mobility in the real life setting. CM is measured using self-report measures and emergent technologies, such as wearable Global Positioning System (GPS) sensors. However, the measurement properties of most available assessments have not been compared within a PD population to determine their appropriateness and identify any feasibility issues.

The primary objective of this project was to compare a novel instrumented measure (Wireless Isoinertial Measurement unit with GPS; WIMU-GPS) with a self-report diary and the Life Space Assessment (LSA). To accomplish this aim, a review of literature was first conducted to show that the validity and reliability between mobility measures were seldom assessed in existing comparison studies. Then, seventy people with early to mid-stage PD (67.4 ± 6.5 years, 67.1% men) wore the WIMU-GPS and completed the self-report diaries and LSA for a 14 day period. Moderate agreements were observed between WIMuGPS and diary for trip frequency and duration (Intraclass correlation coefficient [ICC] = 0.71, 95% CI = 0.51 to 0.82; 0.67, 95% CI = 0.42 to 0.82, respectively). Disagreement between these two measures was higher for duration, particularly among individuals who regularly worked or volunteered. Convergent validity and good reliability was attained for trip frequency (Spearman correlation coefficient [rs] = 0.69, 95% CI = 0.52 to 0.81; ICC = 0.714, 95% CI = 0.51 to 0.82) and duration outside (rs = 0.43, 95% CI = 0.18 to 0.62; ICC = 0.674, 95% CI = 0.42 to 0.82) measured by the WIMU-GPS and diary. However, convergent validity was not observed between WIMU-GPS recordings and LSA reported life space size (rs = 0.39, 95% CI = 0.14 to 0.60). The LSA exhibited ceiling effects and discrimination issues. It should be avoided as a CM measure when it is feasible to use the WIMU-GPS and diary instead.

The secondary objective was to determine the utility and feasibly of using WIMU-GPS to quantify different dimensions of CM in people with PD (PwP). Using a subset of participants, it was first determined that sampling error was minimized in non-discrete continuous outcomes, such as “time outside” and “area size”, when daily WIMU-GPS recordings lasted at least 600 minutes. A shorter recording minimum of at least 500 minutes per day was also suitable for discrete outcomes, such as “trip count” and “hotspot count”. The sample size precluded the determination of the optimal number of days of recording. However, data from at least seven distinct days of recording is required to capture the natural fluctuations in CM between days of the week. From a practical standpoint, a minimum of seven distinct recording days were best attained if the WIMU-GPS was worn for at least eight days. Next, the new minimum GPS recording length was adopted in a larger subset of the sample to show that PwP were regularly in the community, and they preferred vehicular travel over walking when travelling to a destination. Distances walked by PwP increased when they perceived higher levels of PD-related impact on emotional wellbeing (Pearson correlation [r] = 0.40, p < 0.01) and bodily discomfort (r = 0.30, p = 0.03). Complementary diary data also showed PwP were making regular weekly visits to medical facilities.

Finally, the body of work described in this Dissertation culminated in a series of practical recommendations for those interested in the CM of an older PD population or wishing to use GPS sensors for assessing real-life CM. The results of this Dissertation also are useful resources for the development of needed standards on how mobility measurements should be compared, and on the study design, data collection, and reporting of health data using GPS sensors.

Download

Included in

Epidemiology Commons

Share

COinS