Approximately 75% of people with Parkinson’s disease (PD) have an impaired balance [1], which constitutes one of the most distressing symptoms [2]. People with PD are particularly unstable when perturbed backwards due to impaired postural reflexes [3–5], which is suggested to be evaluated clinically by using an unexpected shoulder pull [6]. Already early during the disease, turning difficulties are common [7] and an unsteadiness while turning is also associated with having more severe freezing of gait (FOG) [8]. Walking difficulties are also common and mainly characterized by a decreased gait speed and shuffling gait. Gait and balance problems are also related to non-motor features (e.g. cognitive dysfunction) of PD and are exacerbated by dual tasking [9].

People with PD have an increased risk of falling as compared to healthy individuals at the same age, but also in relation to people with other neurological disorders [10–12]. They usually fall while performing activities such as walking, turning, transferring to/from sitting, bending forwards or while reaching [13]. It is also common for people with PD to experience near falls, which can be defined as “a fall initiated but arrested by support from a wall, railing, other person, etc.” [14]. A recent review scrutinized specific factors associated with recurrent falls among people with PD, and fear of falling (FOF) was then highlighted as one of the risk factors [15]. In addition, FOF has been shown to be a predictor for community walking [16] and a major barrier to engaging in exercise [17]. FOF can be defined as a lack of confidence (low self-efficacy [18]) to be able to perform activities without falling, i.e. low fall-related self-efficacy.

Among people with PD, FOF is common and about 70% report activity limitations due to FOF, which also may cause social isolation [10, 19]. Although FOF influences activity and participation negatively among people with PD, there is yet limited knowledge regarding contributing factors. Such knowledge is highly warranted in order to develop means that efficiently tap causal factors. At present, there are four published studies that used multivariate analysis to investigate contributing factors to FOF in PD [1, 20–22]. Two out of these four studies were postal surveys and lacked clinical data [1, 21], and none of them have been replicated [1, 20–22]. More importantly, no study has included independent variables targeting functional balance performance, dual tasking, and gait speed or used an unexpected shoulder pull when assessing postural instability. Since gait speed and functional balance performance have been shown to correlate to FOF in bivariate analyses [23, 24], these aspects may tentatively be of importance when investigating contributing factors to FOF. Dual-tasking might also be of interest since it worsens gait impairments in PD and may lead to wrong prioritization, i.e. the “posture second” strategy [9, 25]. There is thus a need for a more thorough understanding of contributing factors to FOF in PD in order to address this efficiently in clinical practice and research.

This study aimed at determining factors associated with FOF (conceptualized as low fall-related self-efficacy) among people with PD. More specifically, the aim was to determine whether previous postal survey based findings could be replicated in an independent clinical sample and, secondly to investigate whether additional and previously unexplored motor aspects (e.g. gait speed, functional balance performance) as well as cognitive features independently may contribute to FOF.

All people diagnosed with PD receiving care at a south Swedish university hospital during 2007–2011 were considered eligible for inclusion (n = 273). Exclusion criteria were age above 80 years old (n = 106), inability to stand without support (n = 17), inability to understand instructions (n = 8) or being mentally or medically unstable (n = 7). The remaining 135 patients were invited to participate. Twenty-eight (12 women) participants declined to participate, and they did not differ significantly (p ≥ 0.07, the Mann–Whitney U test) from the included ones with respect to age and PD-duration. Three additional participants (2 women) were excluded due to missing data on the dependent variable: the Swedish version of the Falls Efficacy Scale, i.e. FES(S). The final study sample consisted of 104 participants.

By comprehensibly investigating contributing factors to FOF among people with PD and by using multivariate analyses, this study confirms previous observations suggesting that walking difficulties in daily life is the strongest contributing factor in addition to independence in daily activities and fatigue. Although some previous PD-studies have shown similar results [1, 20], none included independent variables that targeted functional balance performance, dual-task difficulties, and gait speed. A novel finding in this study is that functional balance (that is of importance in daily activities) was identified as an additional significant independent contributor to FOF, whereas a reactive postural response after an external perturbation (and other motor or cognitive aspects) was not. Including functional balance performance in the model increased the explanatory power from 66% to 73%, whereas other motor and cognition aspects do not appear to provide any improvements beyond the first model. The present findings may have important implications for physical therapy and rehabilitation targeting PD.

Several variables that showed highly significant bivariate relationships with FOF (e.g. cognition and falls) were not independently associated with FOF when controlling for other independent variables. This illustrates a major pitfall in relying on bivariate analyses and highlights the importance of using multivariate analyses in this type of studies. Although it may appear surprising that falls did not contribute to FOF, this finding is in line with other PD-studies using multivariate analyses [1, 20, 22].

Our first regression model represents an independent replication of a prior study based on self-reported postal survey data [1]. The replication corroborates walking difficulties as a major contributing factor to low fall-related self-efficacy. This implies that walking difficulties should be a primary target when attempting to reduce FOF.

Although generally confirming previous findings, the present study did not identify turning hesitations as an independent contributor to FOF as shown in the study by Nilsson et al. [1]. This discrepancy is probably not related to differences in the dependent variable (i.e. FOF, operationalized as low fall-related self-efficacy), since the present median FES(S) score was similar to the one obtained in the study by Nilsson et al. (117 and 114, respectively) [1]. However, sample differences may still have contributed, as the present sample seemed to be less affected by their PD than the previous sample, e.g. proportions of fallers (33% versus 45% in the study by Nilsson et al. [1]) and of people needing help in daily activities (10% here versus 27%). An alternative explanation for the discrepancy may be that all independent variables were not identically assessed in the two studies.

Walking difficulties in daily life was identified as a major explanatory variable in both models, accounting for almost two thirds of the variance in FES(S) scores. This suggests that walking ability may be a primary therapeutic target for alleviating FOF. Functional balance performance (BBS scores) was significantly associated with FOF, whereas the NRT was not. The clinical implication of this finding is that balance training probably should focus on challenges induced by self-generated perturbations rather than external perturbations, if aiming at reducing FOF. In other words, it seems like interventions should target functional balance performance and not reactive postural responses if aiming at reducing FOF among people with mild PD.

FOF among people with PD needs specific attention since it has been identified as a risk factor for recurrent falls [15], a barrier for exercise [17], and a predictor for community walking [16]. Furthermore, FOF causes activity restrictions and avoidance as well as social isolation [10, 19, 23]. A recent Cochrane review concluded that physical therapy can yield short-term improvements in walking, mobility and balance as compared with no intervention in people with PD [39]. However, the review did not support reduction of FOF by physical therapy. This may be explained by several factors. For example, few of the reviewed studies included FOF as an outcome; compromised methodological quality of the included studies; or that the key ingredients of the interventions did not address walking difficulties in daily life. Future trials targeting walking ability and including FOF as an outcome are thus needed. Importantly, FOF may be such a complex construct that it best benefits from using an interdisciplinary approach. The latter may be supported by the fact that dependence in daily activities as well as fatigue was independently associated with FOF. Interestingly, it has been suggested that poor walking economy among people with PD may contribute to fatigue [40]. However, the exact role of this enigmatic complaint remains speculative [41–43] and cannot be addressed based on the current study.

This study was able to replicate previous main findings in an independent sample of people with PD by identifying everyday walking difficulties as a primary FOF associated factor, and additional independent contributions by fatigue and the need for help in daily activities. Furthermore, functional balance performance was found to be the only factor among a range of additional clinical motor and cognitive variables that was able to account for additional significant proportions of the variance in FOF. These observations imply that walking difficulties and balance performance in daily life are candidate therapeutic targets in order to reduce FOF in PD. However, longitudinal studies are warranted in order to gain an increased understanding of predictors of FOF in PD and who is at risk of developing a FOF.