Furthermore, elderly with PD have shown reduced levels of physical activity [9] and a nine times increased risk of injurious falls, compared with the healthy elderly [10]. As PD progresses, balance problems gradually increase and are generally non-responsive to or worsen with levodopa treatment [11-13]. Currently, there is a growing body of research that highlights the role of physical exercise as an essential part of managing the disease, with potential neuroprotective mechanisms [14,15]. The effects of balance training on balance and gait performance are promising for individuals with PD [16-18], although several questions remain unanswered, particularly regarding dose, intensity and duration, as well as regarding specific exercises to improve balance control in the different stages of the disease [17-21]. Additionally, there is a lack of knowledge of the long-term effects of training interventions on participation in everyday living, such as physical activity and QoL [16].

Balance control relies on the interaction of several physiological systems (the musculoskeletal, neuromuscular, cognitive and sensory systems) with environmental factors and the performed task [22]. Balance impairment or postural instability in individuals with PD is often associated with poor or absent reactive responses following external perturbations, such as performing a rapid step following a slip or trip [23]. However, as degeneration of the basal ganglia affects many physiological systems essential for balance control, balance disorders in PD cannot be addressed to one single function but, rather, are the result of impairments of multiple systems [24-26]. First of all, dysfunction of the basal ganglia influences sensory integration, i.e. the ability of the central nervous system to transform different modalities of sensory information (somatosensory, visual and vestibular) into a single reference frame, which is important for estimation of limb and body position in relation to the environment [27]. Particularly, individuals with PD have shown impaired proprioception [13,27], overestimation of the amplitude of movements [28] and over-reliance on vision for balance control [29]. Moreover, the central motor drive has shown to be impaired in PD [30], causing bradykinetic movements and poorly timed and scaled anticipatory postural adjustments (APAs), i.e. feedforward control to stabilise the posture before and during voluntary movements [31,32]. Furthermore, the PD-characteristic “stooped” posture, decreased joint range of motion, narrow foot stance and axial rigidity increase the risk of instability since these musculoskeletal constraints reduce the functional limits of stability, i.e. the area in which the body centre of mass (CoM) can be moved with maintained stability and without changing the base of support (BoS) [33,34]. Taken together, sensory-motor interaction is essential for balance control and consequently, as discussed by Konczak et al. [27], sensory problems may degrade movement responses and cause instability to individuals with PD.

The functions of the basal ganglia incorporate motor program selection and adaptation [35,36], here referred to as motor agility, which involves maintenance of coordination between body parts, task-specific adjustments of movement and quick shifts from one task to the next. Deficient motor regulation in PD, manifested as poor inter-segmental coordination [37], problems adopting postural synergies [35,38] and delayed change of motor commands when shifting from one task to another [39], is believed to contribute to increased instability and freezing of gait during fall-related activities [40,41]. Another critical aspect of balance control in PD is the ability to divide attention and simultaneously process multiple tasks (motor or cognitive), i.e. multi-tasking. While performing a multi-task activity, individuals with PD compared with healthy subjects are more inclined to shift attention away from the balance task, which can lead to falls [24].

To be successful, all types of training require the application of basic training principles of (i) specificity, i.e. being specific to the targeted function; (ii) progressive overload, i.e. providing a challenging overload to the physiological system through a certain level of intensity and regularity; and (iii) varied practice, i.e. promoting variation between exercise conditions [42,43]. To be specific, balance training needs to target functions, or impairments, of balance control associated with PD symptoms. As previously mentioned, there is increasing scientific knowledge concerning the interaction between and impact of different physiological systems for balance disorders in PD [24-26]. Previous intervention studies targeting balance performance in PD have, however, mainly emphasised non-specific training or compensatory training [17,44]. Moreover, to stimulate relearning of physiological systems important for balance control, the challenge level (i.e. intensity of the training, the difficulty level of balance exercises and the total training dose) needs to be considered. Allen et al. [17] recently concluded, based on a meta-analysis of exercise interventions targeting balance performance in PD, that previous studies in general have used low challenge levels, suggesting that future exercise interventions should focus on more challenging aspects of balance exercises and a higher training dose. Furthermore, to enhance generalisation of learned motor skills to a wide variety of situations and promote a multi-faceted repertoire of movement strategies, it is essential to guarantee practice through a wide variety of balance exercise conditions [43].

We hypothesise that a new conceptual framework for balance training, emphasising critical aspects of balance control specifically related to PD symptoms through highly challenging, progressive and varied exercises, will improve, or maintain, balance and gait performance in elderly with PD. Moreover, we predict that increased physical functioning following training will lead to greater confidence in everyday participation, improved QoL and increased levels of physical activity.

Balance impairment in PD comprises a wide spectrum of disability and health [16]. Detailed long-term follow-up, 1year post-intervention, in the proposed study will give a unique opportunity to address important questions concerning potential effects of and limitations for group-based balance training. First of all, the proposed study will be one of the first trials to provide knowledge of potential effects of exercise intervention according to the whole spectra of ICF domains, including laboratory measurements, clinical tests, and free-living and self-perceived assessments (see Table ​Table3).3). Such knowledge will be essential for designing and improving exercise interventions in the future. Secondly, long-term follow-up will provide knowledge of the maintenance of potential training effects. This is particularly important for PD given the progressive deterioration of the disease in which regular periods of supervised balance training may be required to sustain physical performance.

Since this intervention will not be restricted to a fixed program with specific exercises, several steps will need to be taken to standardise the intervention by educating physiotherapists in the framework and implementation of the training approach, as well as supervising the physiotherapists during the course of training and monitoring training performance as well as progression. During the course of training, each training session is planned with regard to the different phases of the intervention (see Table ​Table2)2) and adapted by the physiotherapist to meet individual capacities of the participants. The physiotherapists conducting the training will be responsible for transferring the conceptual framework into practice, using a pragmatic approach rather than restricting the intervention to a standardised protocol. Thus, translation of this research intervention into the clinical setting will be enhanced.