Implementation of the HiBalance training program for Parkinson’s disease in clinical settings: A feasibility study

Abstract Background Translating evidence into practice requires adaptation to facilitate the implementation of efficacious interventions. A novel highly challenging balance training program (HiBalance) was found to improve gait, balance, and physical activity in persons with Parkinson’s disease (PD) in an earlier randomized controlled trial. This study aimed to describe the adaptation process and feasibility of implementing the HiBalance program for PD within primary healthcare settings. Method Feasibility was assessed in terms of study processes and scientific evaluation. Nine persons with mild–moderate PD were enrolled in this pre–post feasibility study. The dose of the original program was adapted by reducing therapist‐led training sessions from three to two times weekly. Outcome measures were substituted with ones more clinically feasible. One group (n = 5) received HiBalance training three times weekly for 10 weeks while another (n = 4) trained twice weekly plus a once weekly home exercise program (HEP). Balance performance was the primary outcome, while secondary outcomes (e.g., gait speed, physical activity level, concerns of falling, and health‐related quality of life) were also evaluated. Results Regarding process feasibility, attendance was high (approximately 90%) in both groups, and experiences of the group and home training were positive. Newly selected outcome measures were feasible. The scientific evaluation revealed few adverse events and no serious injuries occurred. Concerning outcomes per group, the average change in balance performance and gait speed was equal to, or exceeded, the minimally worthwhile treatment effect commonly used in PD. Conclusion The findings support the feasibility, in terms of process and scientific evaluation, of the adapted HiBalance program for implementation within clinical settings. A sufficiently powered study is required to ascertain whether the newly proposed program offers similar short and long‐term effects as the original program.


| INTRODUC TI ON
The implementation of efficacious interventions into healthcare settings is necessary for improving the health of larger patient groups (Bradley et al., 2004). However, the rate at which promising interventions are embedded within healthcare remains suboptimal because of challenges related to implementation (Proctor et al., 2011).
The initial demands and costs to society are considerable for implementing new evidence (Krisberg, 2010) therefore, investigating feasibility aspects (e.g., process and scientific evaluation) is important for ensuring optimal uptake of interventions.
Parkinson's disease (PD) is a neurodegenerative disorder typically resulting in deterioration of gait and balance abilities which predispose individuals to more frequent falls and injuries (Bloem, Grimbergen, Cramer, Willemsen, & Zwinderman, 2001). To combatassociated symptoms, the HiBalance training program has been designed, including highly challenging and progressive exercises targeting dysfunctions of subsystems of balance control among persons with mild-to-moderate PD (Conradsson, Lofgren, Stahle, Hagstromer, & Franzen, 2012). To date, the intervention has been found to be feasible (Conradsson, Lofgren, Stahle, & Franzen, 2014) and effective at improving balance and gait performance, activities of daily living, and physical activity levels in more controlled settings, that is, randomized controlled trials (Conradsson et al., 2015).
In order to reach a larger proportion of people with PD, the next step involves testing the clinical applicability and implementation of the program.
Translating research protocols into clinical practice is not a straightforward task because a complex set of decisions and compromises needs to be made in order to be considered by healthcare planners and implementers. However, protocol changes are rarely examined in the literature, especially related to rehabilitation interventions which are inherently complex. The objective of this study was to describe the adaptation process, specifically the necessary changes needed to translate a research protocol into clinical practice, as well as procedural and scientific feasibility of the adapted HiBalance program for implementation within primary healthcare settings.

| Design
This feasibility study had been approved by the regional board of ethics in Stockholm. A typology for feasibility studies, according to Thabane et al., (2010) was used to gather insight into aspects related to translating an intervention from more controlled to clinical settings. This typology provides useful information prior to conducting large-scale studies by providing insight into the following assessments: (1) processes, (2) resources, (3) management, and (4) scientific (effectiveness). For the purpose of this study, the specific process and scientific aspects of feasibility, as summarized in Table 1, were investigated.

| Participants
Participants were consecutively recruited from a convenience sample of referrals at one primary care rehabilitation clinic in central Stockholm. Fourteen participants were invited for initial screening, where nine met the following inclusion criteria: (1) a diagnosed of idiopathic Parkinson's disease; (2) mild-moderate disease severity according to a Hoehn & Yahr, (1967) score of 2 or 3; (3) absence of noteworthy cognitive impairment; (4) age ≥60 years; (5) ability to independently ambulate indoors without the use of a mobility aid; and (6) being on a stable dose of anti-Parkinson's medication for ≥3 weeks. Five participants did not meet the preliminary inclusion criteria: one due to nonidiopathic PD; two had severe impairments (cognitive and balance); and two with comorbidities (extreme back and hip pain) which could have impacted study outcomes. Written informed consent was obtained before first assessments were carried out. As per Table 2, the average age was 71 years and subjects were predominantly female (6/9). The average time since diagnosis was 11 years, and one-third experienced a fall during the past 12 months. Note. a Indicates feasibility aspects investigated in this study.
Due to the feasibility design, with our sole interest in assessing the translation of the HiBalance training program from theory (RCT environment) to practice (clinical settings), no sample size calculation was needed and performed. Results from the process and scientific feasibility will be used to inform the larger implementation trial.

| Adaptation of the HiBalance program
The theoretical underpinnings of the HiBalance program have previously been described in detail (Conradsson et al., 2012Leavy, Kwak, Hagströmer, & Franzén, 2017). This program targets subsystems of balance control typically affected in PD: (1) sensory integration, (2) anticipatory postural adjustments, (3) motor agility, and (4) stability limits. To target these progressive symptoms, motor learning principles, that is, specificity, progressive overload, and variation, were used as foundation to challenge individual progression. The program is group-based (4-7 persons) and is facilitated by two physical therapists (PT's) who were trained to develop exercises according to the balance control framework used in this study.
Additionally, the program incorporates gradual integration of dualtasking (DT)-cognitive (e.g., counting or remembering items) and motor task (e.g., carrying or manipulating an object)-to target mildassociated cognitive impairments.
Adaptation of certain aspects of the initial balance training program was necessary to facilitate its implementation within clinical settings. Firstly, the three times weekly training dose during the efficacy trial was reduced to align with policy regarding rehabilitation reimbursement in the Swedish healthcare system. Training dosage was discussed during a workshop by an expert group consisting of PT's from hospitals and primary care facilities, including some with previous experience of the intervention, and the researcher group. The final decision was to reduce therapist-led sessions to twice weekly for 10 weeks. This decision was also informed by data from a previous qualitative study where participants perceived that 30 training sessions were too great a time commitment (Leavy, Roaldsen, Nylund, Hagstromer, & Franzen, 2016).
To compensate for this marked reduction, a home exercise program (HEP) was proposed and developed over several months by to ensure clinical feasibility. Consequently, the test battery was condensed to fewer performance-based and self-reported measures . The primary outcome, that is, balance performance measured with the Mini-BESTest, remained unchanged.

| Process assessment
For the assessment of process feasibility, attendance/adherence rates, eligibility criteria, the appropriateness of data collection tools/outcome measures, time taken to complete all measurements, and participants' experiences of the program were investigated.

| Scientific assessment
Safety of the intervention was assessed by recording the nature and frequency of adverse events which included falls, injuries, fatigue, and pain. To assess the effectiveness (responsiveness) of the newly proposed dose, participants were randomized to (1) receive training three times weekly (Group 3x), (2) receive group training twice weekly and perform the HEP once a week (Group 2x + HEP).

| Data analysis
Data were analyzed descriptively, using the proportion of individuals per group who changed their outcome status at post-testing. This study was not sufficiently powered to test between-group differences and report on the statistical superiority of one training dose over the other. However, we discussed the observed changes in relation to their clinically meaningful important differences.

| Process assessment: attendance, data collection procedure, and experiences
Group 3x

| Scientific assessment: safety and effectiveness
Concerning adverse events during therapist-led sessions, three participants fell without causing injury; all stumbled over artificial hindrances that were created to challenge balance. Another participant reported pain associated with the training which lasted more than 2 days, and another felt dizzy during the session but recovered after a resting period. No participant reported negative events during the HEP.

| D ISCUSS I ON
Results of this study reveal that the adapted HiBalance program was feasible in mild-moderate PD within a primary healthcare setting, in that attendance rates were high, adverse events few, and the effects on balance performance and secondary outcomes were detectable and favorable. The results support F I G U R E 1 (a) Mini-BESTest scores and (b) gait speed pre-and postintervention initial feasibility of the intervention which was adapted to suit the healthcare context. However, a larger and more rigorously designed multisite study is required to establish whether the newly adapted program offers similar short-and long-term effects as the original program.
According to the feasibility typology, assessment of the process revealed good adherence as approximately 90% of the total sessions were attended by participants. The eligibility criteria were similar to that used in the RCT (Conradsson et al., 2015) however, the recruitment process revealed that younger patients (<60 years) are commonly referred to rehabilitation in primary healthcare settings. To promote access, the implementation trial will include persons of all ages. Participants completed the newly selected outcome measures with ease. This could be because some of the selected measures, such as the EQ-5D and 6MWT, are part of existing routine practice.
However, PT's reported that almost 2 hr of data collection per participant was not clinically feasible, and the 6MWT was resultantly removed from the test battery, while the Walk-12, FES-I, and EQ-5D were suggested to be completed by participants at home. Overall, participants' experiences were positive and all felt challenged by the program.
As Training resulted in improved balance performance. Despite the small sample size, both groups improved their balance performance by at least one point on the Mini-BESTest, which is also the minimal detectable change in mild-moderate PD (Löfgren, Lenholm, Conradsson, Ståhle, & Franzén, 2014). Similar positive trends were found for secondary outcomes, that is, gait speed, concerns of falling, and exercise endurance. In fact, a similar (average) difference in gait speed was found following training in the Group 2x + HEP when compared to the efficacy trial (Conradsson et al., 2015). The fact that the findings of the two studies corroborate are an indication of the potential effectiveness of the adapted training dose in the current study.
This study presented with several limitations. Firstly, the sample size was too small to sufficiently power outcomes and to adequately evaluate adverse events. Also, the lack of a control group under- In conclusion, this study demonstrates that the adapted HiBalance program is feasible from both a process and scientific/ effectiveness perspective. A larger multisite study is needed to test the effectiveness of the adapted program on balance and gait outcomes as well as to inform future widespread implementation.

ACK N OWLED G M ENTS
We would like to thank all participants for their time and adherence to the testing and evaluation procedures during the study period.
We are grateful for the financial support from the Swedish Research

CO N FLI C T O F I NTE R E S T
All authors have no conflict of interest to declare.