Physical activity in Parkinson’s disease : measurement, correlates and effects of balance training

Abstract

Aim: The overall aim of this thesis was to investigate correlates of physical activity in older adults with Parkinson’s disease and to evaluate the effects, both short- and long term, of the HiBalance program on physical activity and sedentary. Further, to investigate the associated factors of a training effect on physical activity. To this end disease-specific physical activity estimates are needed.

Methods: In Paper 1 accelerometer cut points for different walking speeds were defined. Thirty older adults with mild-to-moderate Parkinson’s disease walked at self-defined speeds of brisk, normal and slow speeds for three minutes in an indoor corridor. Walking speed was used as a reference measure, and cut points were generated using ROC curves. The cut points were cross-validated and Cohen´s quadratic weighted Kappa was calculated. In Paper II, correlates of both total physical activity and cut point-defined brisk walking were investigated by applying correlation analysis followed by multiple linear regression to accelerometer data. In Paper III, short- and long-term effects of the HiBalance program on physical activity and sedentary behavior were evaluated, using mixed analysis of variance and a multilevel model. Further, associated factors to a training effect on physical activity were investigated using a multiple linear regression.

Results: Optimal cut points for the vertical axis were ≤328 and ≥730 counts / 15 seconds for walking speeds at ≤1.0 m/s and >1.3 m/s, respectively. Sensitivity ranged between 68-100 %, with specificity between 75-82 %, whilst validation and Kappa analysis showed 74% absolute agreement and a substantial agreement of κ = 0.79 (95% CI 0.70–0.89), respectively. Exploration of correlates of total physical activity and amount of brisk walking led to two linear regression models. Motor impairment, physical function, body mass index and dyskinesia were significantly associated with total physical activity, explaining 34% of the variance, whilst physical function and balance control were significant factors related to brisk walking, explaining 22% of variance. Short- and long-term effect analysis revealed that brisk walking was the only factor showing a significant interaction effect of group and time. Moreover, the effect on brisk walking dissipated 6 months after intervention. Analysis of the training effect on physical activity revealed that intervention group affiliation and spring season were significantly associated to an increase in brisk walking, while increased balance after training was not.

Conclusion: This thesis provides cut points for physical activity measurement in older adults with Parkinson’s disease. Results also suggest that correlates of total physical activity and brisk walking differ, and evidence of factors not previously shown to be associated with PA in this population, is provided. Moreover, the HiBalance program leads to an increased amount of brisk walking in daily living, yet this increase is not linked to improved balance control. Also, clinicians should be aware of the seasonal effect on ambulatory activity in this population, and that the intervention effect dissipates after half a year, thereby warranting recurrent training.