Study Design
Sixty people who use LLP, with low balance confidence (< 80) will be randomized to the CBPT intervention condition or control condition (Figure 1). Participants will be assigned to one of two randomization plans based on etymology of amputation: vascular causes (e.g., diabetes, peripheral vascular disease) or non-vascular causes (e.g., trauma, osteomylistis). For each plan, a block randomization will be used with a block size of 4 to ensure balance among the exercise groups throughout the study. The randomization plan will be developed by researcher uninvolved with the project using www.randomization.com and sealed envelopes will be used to conceal the randomization schedule until enrolled participants are assigned to a group. For scheduling purposes, to ensure that the necessary clinicians are on site at the time that group assignment is revealed to the participant (at visit occurring ~ 1week following enrollment), envelopes will be opened after the participant leaves the university after completing the enrollment visit.
Participants in the intervention condition will complete eight weekly training sessions that include a virtual reality active gaming component (i.e., PT component), with CBT strategies that increase balance confidence by addressing subject-specific avoidance behaviors and maladaptive cognitions. Each session will last approximately 1.5 hours. Participants in the control condition will complete a series of at-home upper-body seated exercises during this eight-week period after being taught by a physical therapist how to perform the exercises. At the start of the study and 8- (post-intervention), 16- and 24-weeks following group allocation, participants will complete survey-based outcome measures and activity monitoring. At baseline and post-intervention, participants will complete performance-based assessments of balance and of functional mobility. Outcomes measures are described below.
To minimize loss to dropout between baseline and final follow-up, research coordinators will routinely contact participants by telephone and email to maintain engagement in order to maximize participant adherence. If a participants wishes to drop out of the study they will be queried as to their reasons for doing so and every attempt will be made to address their concerns and to obtain the minimum amount for data necessary to address hypotheses related to the primary outcome the stated hypotheses (i.e. ABC scale and self-reported measures at 0 and 25 weeks).
Setting and Ethical Approval
The study will be performed at the Human Performance Laboratory at Rosalind Franklin University of Medicine and Science, North Chicago, IL, in conjunction with investigators at the Captain James A. Lovell Federal Health Care Center, North Chicago, IL, Edward Hines Jr. Veterans Administration Hospital, Hines, IL and Northwestern University, Chicago, IL. This protocol described herein has already received Institutional Review Board (IRB) approval from all study sites. Any modifications to the approved protocol or reporting of adverse events will be submitted to the IRB by the study PI according to published IRB policies.
Figure 1. Flowchart of study participation
Participants
Participants will be Veterans recruited directly from local VA Medical Centers or civilians/Veterans recruited from the community. Recruitment at local VA hospital centers will take place through a two-step process. In step 1, medical records will be reviewed for persons having a unilateral transtibial amputation and prescribed prosthesis at least 6 months prior. Veterans found through medical record search will be sent a form letter indicating they may qualify for a study and should respond to the letter only if interested in additional information. The letter will ask them to complete and return a copy of the Activities-specific Balance Confidence (ABC) scale. In step 2, participants with ABC scores less than 80% will be called by research coordinators to determine eligibility through the use of an IRB-approved script. Recruitment of civilians/Veterans from the community will primarily take place through flyers posted at local prosthetic clinics, support groups, and physical therapy offices.
All participants must meet the following criteria: at least 18 years of age, unilateral transtibial amputation, at least 6 months experience with using a definitive prosthesis, an ABC score less than 80, and self-reported concern about balance. Participants will be excluded for open wounds on weight-bearing surfaces, inability to stand unassisted and weight-shift for 2.5 minutes (i.e., inability to perform the CBPT intervention), currently receiving physical therapy, ill-fitting or ill-functioning prosthesis as determined during a screening with the study prosthetist, prohibited by primary care physician or research physician to engage in mild exercise or history of neurodegenerative diseases and stroke. Screening and all other study-related activities will occur only after the participant provides written informed consent, which will be taken by a study team member.
It is expected that 70 participants will be recruited in order to obtain a total of 60 eligible participants (30 intervention and 30 control participants), assuming a 15% drop-out rate. Using G*Power 3.1, a sensitivity analysis (with regard to Hypothesis 1a) was conducted, specifying a power of.80, an alpha of.05, and an average correlation of observations of 0.50. According to the analysis, with 30 participants in each group we will be able to detect a small to medium effect (f =.152) of intervention on ABC score over time with this design, which is consistent with the size of effect observed for older adults. 10,11 We will have similar power for all hypothesesin Aim2, which use the same analysis as Hypothesis 1a; we will have power to detect a medium effect size (d = 0.46) for Hypothesis 1b and a correlation of r≥0.42 for Hypothesis 3.
CBPT Intervention
Below is a summary of the primary tasks undertaken during each of the 8 sessions.
Session 1: The initial goal of session 1 is to introduce the participant to the 8-session CBPT- intervention and develop rapport. The behavioral therapist will deliver psychoeducation on how the use of the video games can improve balance as well as how balance confidence can impact behavior. Participants’ baseline physical skills on active video games will be assessed using two games (Trash Bin and Italian Alps; see Table 1 for full description of games) and the behavioral therapist will introduce the importance of goal setting in behavioral change. The behavioral therapist will introduce the Behavior Recording Form for Prosthesis Use, which is a standard behavior record with modifications to assess for activity avoidance due to prosthesis concerns. Between-session assignment includes completion of the Behavior Recording Form for Prosthesis Use (Appendix A).
Session 2: Session 2 will begin with a brief participant check-in regarding health and any falls during the prior week to ensure that the participant is safe to continue physical therapy. Next, the participant will be introduced to the four new video games, not used during session 1, that will be played during the course of the intervention (Table 1). After completing the video games, the behavioral therapist will review the completed Behavior Recording Form for Prosthesis Use, with the participant to gain an understanding of the participant’s activity restriction and avoidance behavior. Specific behavioral goals will be developed with the participant. The behavioral therapist will review the Breathe to Relax form with the participant, which teaches diaphragmatic breathing for relaxation (Appendix B)
Between-session assignments will include completion of the Behavior Recording Form for Prosthesis Use, reviewing the Breathe to Relax form, as well as practice the breathing skill.
Session 3: After a brief check in to ensure the participant is well, the participant will play all four video games from Session 2. The behavior therapist will then conduct an in-depth review of between-session assignments to continue skill building in behavior recording and goal setting. Additionally, the behavioral therapist will review and practice diaphragmatic breathing with the participant as well as refine initial behavioral goals. The behavioral therapist will then introduce systematic exposure to the participant in the context of reducing fear and changing behavior. This will be followed by an exposure exercise. The exposure exercise will be developed through a collaborative process between the behavioral therapist and the physical therapist based on an emerging understanding of the participant’s fear exposure hierarchy and physical ability. A fear exposure hierarchy is composed of a series of gradated tasks based on level of provoked fearfulness. An example of an exposure exercise is asking a participant that avoids reaching for an object above their head to practice reaching up high. Exposures will be conducted in the lab under the guidance of the behavioral therapist as well as at home in between sessions.
Between-session assignments will include completion of the Behavior Recording Form for Prosthesis Use, practicing Breath to Relax, and an at-home exposure exercise.
Sessions 4 through 7: The physical therapist will continue to work on games with the participant. The behavioral therapist will continue to use and review previously introduced procedures including physical skill acquisition during physical therapy intervention, behavior recording, diaphragmatic breathing, and community-based exposure exercises.
Between-session assignments continue to include completion of the Behavior Recording Form for Prosthesis Use, using the Breath to Relax form, and performing an at-home exposure.
Session 8: The final session will focus on summarizing the participant’s behavior change goals and progress made towards completing goals, as well as physical progress in the component and any learned behavioral strategies. The behavioral therapist will discuss strategies to continue to further gains as well as strategies for preventing relapse.
Key Informant Interview: At the end of the CBPT intervention (or at drop-out), participants who received the intervention will complete a 30–60 minute, recorded, individual interview with investigators to discuss their experience with the intervention. A semi-structured interview will be used to assess the following areas:
Intervention Experience: Participants will be asked to discuss what they liked and disliked about the intervention, their preference for the length of the intervention and the length of the individual sessions, the reasons they attended or missed sessions, their thoughts on combining the CBT and PT strategies, and the role of the CBT and the gaming exercises in addressing their balance confidence and activity restriction. We will also obtain feedback on the CBT forms and assignments including information on comprehension, burden and usefulness for addressing balance confidence and activity restriction.
Barriers to Participation: Participants will be asked about barriers to integration that they felt were not addressed during the intervention, as well as what the intervention could include to address those barriers.
Opinions on Recruitment: Participants will be asked about advertising messages and recruitment venues that would increase the appeal of participating in similar future studies.
Training and Oversight of Behavioral Therapist and Physical Therapist
Two advanced, clinical psychology doctoral students will be trained by two licensed clinical psychologists to deliver the CBT portion of the intervention. The behavioral therapists will review the treatment manual, which will be followed by four 2-hour training sessions covering balance confidence, CBT, the integration of CBT into a PT setting, and understanding the psychological and physical health sequelae of a leg amputation. Two training sessions will focus on the delivery of the CBT strategies. This will include a review of CBT theory and specific CBT strategies used to address balance confidence and activity restriction. In addition, behavioral therapists will role-play the delivery of the CBT strategies to ensure intervention fidelity and to provide feedback on adherence of the intervention protocol. Potential problems with the delivery of the CBT content will be discussed (e.g., participant fails to routinely complete assignments).
Once the intervention begins, behavioral therapists will receive supervision by licensed clinical psychologists. Supervisors will listen to audio recordings of the first 8 sessions for each behavioral therapist and provide corrective feedback; thereafter 10% of sessions will be reviewed. The behavioral therapists will also attend weekly supervision meetings to discuss their participant’s progress and ensure fidelity to the intervention. Behavioral therapist and auditor checklists for each session will be used to ensure that treatment objectives for each CBPT session are met. The behavioral therapist will complete the behavioral therapist checklist after each session. A random selection of audio-recorded sessions for each participant will be audited by supervisors to monitor treatment fidelity. They will complete the appropriate auditor checklist. When a session is reviewed with less than 85% of treatment-specific objectives met, supervisors will remediate training as needed. This process will continue through all treatment waves so that drift can be swiftly corrected.
The research teams’ senior licensed physical therapist will train a physical therapist to complete the PT components of the intervention. After reviewing the training manual, the sessions will include training on the operation system used for administering physical therapy exercises, and all safety features in the laboratories. The physical therapist will also learn how to recognize adverse events, including excessive exertion and be taught to monitor and assess vital signs throughout the intervention to avoid excessive exertion. Most importantly, the physical therapist will be instructed on how to regress and progress the level of difficulty for each game following the protocol.
Sham-Control Condition
Participants assigned to the control condition will be introduced to three exercise modules (Modules A—C) that include seated exercises to improve core and arm strength and flexibility. All of the exercises are demonstrated to the participant by a physical therapist at the initial session. Participants will be provided a manual with thorough descriptions and visual aids of the seated exercises. Some of the exercises include use of a resistive (plastic air-filled) ball, which is also provided to the participant. Participants are instructed to start with Module A and continue for weeks 1—3, followed by Module B for weeks 4—6, and to finish with Module C for weeks 6—8. Participants will be told to follow the modules at their own pace. Each module includes approximately 8 exercises that take 15–20 minutes to complete and participants will be told to work on the exercises three days per week. Participants will be provided a worksheet to record exercise days. Participants will be contacted by phone every 2–3 weeks during the 8-week intervention period to maintain engagement and encourage completion of the follow-up assessments.
The control is intended to account for a potential placebo effect and is considered a sham-control in that none of the exercises target lower extremity strength or flexibility which is more likely to induce functional improvements and promote balance confidence during weight-bearing activities.
Outcome Measures
Outcomes will be collected before the start of the intervention (baseline) and again at 8 weeks (post-intervention), 16 weeks, and 24 weeks following group allocation.
Balance Confidence (Primary outcome measure): Balance confidence will be measured through the Activities-specific Balance Confidence (ABC) Scale. The scale asks participants to rate their level of confidence in completing 16 complex functional tasks without losing balance. 21 It has strong reliability (ICC = 0.91), internal consistency (Cronbach α = 0.95), and good construct validity in samples of older adults with fear of falling. 22 The ABC scale has also been validated to assess balance confidence in individuals with lower limb amputations.23
Community Integration: The following measures will be used to quantify community integration (secondary outcomes):
Activity Monitoring: Participants will receive a SmartWatch3 Activity Monitor (modus health, WA), which records stride count data for every minute of use, and a QStarz BT-Q1000XT (Qstarz International; Taipei) GPS travel recorder, which records a participants longitude, latitude, local time and date and position every ten seconds. The SAM and GPS will be worn for one week on their prosthetic pylon and then returned in a pre-paid envelope. Community integration will be assessed as the average number of steps/day, average number of steps/day taken outside the home and total time spent outside the home per day.
Community Reintegration of Service members (CRIS): CRIS measures activity and participation with three subscales, Extent of Participation, Perceived Limitation, and Participation Satisfaction. The first two of three scales are included in this study. The extent of participation scale evaluates the frequency of engaging in specific activities using 7-point scales (e.g. “In the past two weeks on average, how often did you engage in recreational activities, not including watching TV?”). The perceived limitation scale uses a 7-point scale to evaluate self-perceived limitations in participation (e.g., “I avoided going to crowded places such as the mall of community gatherings”). The CRIS has been validated in a sample of injured veterans.24
Short-Form Health Survey (SF36): The SF36 consists of 36 questions that measure eight dimensions of health-related quality of life.25 Three subscales, Role Limitation due to Physical Health Problems, Social Functioning, and Role Limitations due to Emotional Problems, will be separately examined as secondary outcomes to assess community integration.
Frenchay Activity Index (FAI): FAI is a 15-item behavioral scale that primarily measures social participation and daily activities (i.e., domestic/chores, work/leisure, and outdoor activities).26 Frequency of participation in activities over the last 3–6 months is rated on a four-point scale resulting in a single summary score, ranging from 0 (inactive) to 45 (highly active). The FAI is a valid and reliable measure among users of a LLP (ICC = 0.79).27 We include the measure based on its association with the ABC scale in users with a LLP.4
Additional secondary outcomes include quality of life and functional ability as described below:
Quality of Life (QoL): The Well-being scale of the Prosthetic Evaluation Questionnaire (PEQ) will be used to assess QoL.28 The PEQ is a survey that was developed to quantify function and prosthesis-related QoL for persons with lower limb amputation. The user marks a visual analog scale with two anchor points to respond to the following: “Over the past four weeks rate”…1) “How satisfied have you been with how things have worked out since your amputation” (anchors: “extremely dissatisfied” and “extremely satisfied”)? and 2) “Your quality of life” (anchors: “worst possible life” and “best possible life”). The simplified assessment allows participants to easily provide a summary of experiences by essentially considering QoL as unidimensional. This particular scale has good consistency (Cronbach α = 0.83) and reliability (ICC = 0.89). 29
Functional abilities: the Berg Balance Scale (BBS) and L-test of walking will be used to measure functional abilities. Blinded assessors with prior training in administering these measures will assess performance.
Berg Balance Scale: This test requires participants to complete 14 tasks of increasing difficulty beginning with sit-to-stand and progressing to unipedal stance. It is a well-established outcome measure initially designed to assess balance in older adults30 and has been validated for persons with lower limb amputation.8
L-test: Participants rise from a chair without armrests, walk 3m, turn right, walk 7m, turn around and trace their path back to the start. The time to complete the task is recorded. The test has strong reliability (ICC = 0.98) and convergent validity (r = 0.93 with the Timed-up-and-Go test (TUG). 31 The L-test is better than TUG at distinguishing changes in walking ability of fit, younger LLP users and it is as strong a predictor of social participation, as is balance confidence.4
Data Management and Sharing
Data collection methods include the use of: 1) a commercially available step count monitors and GPS trackers; 2) completion of pen-and-paper versions of validated clinical surveys; 3) scores on validated performance-based functional tests of balance The methods for managing each type of data are nearly identical. Specifically, when a potential participant arrives at a study site to obtain informed consent, they will be assigned a study ID number that contains no identifiable information. Should that participant enroll in the study, then all data will be labeled and referenced only with the study ID number. For example, data from the monitors will be downloaded onto password protected computers, assigned the appropriate ID, and then analyzed offline. All pen-and-paper data will only have ID numbers listed on them and no identifiable information (e.g., participant name) to ensure confidentiality. A single list linking names and codes will be kept on a password protected computer locked in the PIs laboratory. All performance based data and pen-and-paper based outcomes data will be entered into an IRB-approved secure on-line data collection tool, Qualtrics, using the appropriate subject ID. Summative scores from Qualtrics will then entered into electronic files on password protected computers, to further ensure confidentiality. De-identified data on Qualtircs can be shared among all research team members. The study-PI will periodically check the data values to ensure they are within range and will verify entries based on the original pen-and-paper versions which will stored in locked cabinets in the PIs laboratory to further ensure. All data will be stored for no less than three-years following the submission the final financial report. A separate data monitoring committee is not required by the IRB or funding source for this study. However, the Human Research Protection Office (HRPO) of the U.S. Army Medical Research and Development Command can perform an on-site audit at any time. To ensure confidentiality, information about enrolled participant will only be shared with team members, unless instructed otherwise by HRPO.
While the final dataset will be stripped of identifiers prior to release for sharing, there remains the possibility of deductive disclosure of subjects given the somewhat small number of persons with transtibial amputation in the Chicago-land VHA system meeting the inclusion criteria of the study. Thus, we will make the final research data (both the raw and processed data) and any relevant documentation available to users only under a data-sharing agreement requiring investigators to commit to: 1) using the data only for research purposes and not for purposes of identifying participants; 2) properly securing the data; 3) destroying or returning the data after analyses are completed; 4) an agreement to share any new data sets resulting from the initial sharing-agreement. The research and clinical community will be made aware of the data when the study results are presented at scientific meetings and in final publication form.
Analysis Plan
To test Hypothesis 1awe plan to use a 2 (group) x 4 (time point) mixed ANOVA to examine if there are significant differences in changes in ABC scores between the groups (i.e., an interaction effect), accounting for any missing data. In the event of a significant interaction, a post-hoc paired t-test will be used to compare ABC scale in the intervention group at 0 and 25 weeks and independent samples t-test to compare ABC scale between groups at 25 weeks.
To test Hypotheses 1b and 1c we will use a paired t-test to compare BBS scores and L-test times pre- and post-intervention. H1b and c do not compare time-related changes in performance between groups as they are only intended to assess the immediate effects of intervention.
The same analyses from Hypothesis 1a will be used to test all hypotheses in Aim 2. For Hypothesis 2a, a group x time interaction on self-reported community participation measures (CRIS, FAI, SF36 scales) and objective measures from the activity monitors will be tested. A false-discovery rate correction method will be used to minimize the likelihood of a Type I errors. For Hypothesis 2b, an interaction on the QoL score will be tested. In the event that H2a is not supported, alternative measures to assess integration, if
To test Hypothesis 3changes in ABC scores (0 weeks to 25 weeks) will be correlated with changes in self-reported and objective measures of community participation. A false-discovery rate correction method will be administered to minimize the likelihood of a Type I errors. In the unlikely event that Hypotheses 1a and2a were not supported, this would not exclude the possibility of an association between subject-specific changes in confidence and participation. In such a case, a significant correlation would suggest that improving balance confidence canincrease participation, but for optimum effectiveness of the intervention, one should only train individuals who are most likely to respond to the intervention (show large increase in ABC scores). As an initial step to identify commonality among “responders” we would employ a suite of supervised learning techniques (e.g., cluster and canonical correlation analysis) to explore multidimensional relationships among responses to individual items on the ABC, BBS, CRIS and SF36.
To analyze intervention efficacy, the interviews will first be transcribed verbatim. After 15 interviews, all investigators will review the transcripts to generate an initial list of codes. After 5 additional interviews, they will again re-view transcripts to assess whether new codes emerged. If saturation is not reached after 20 interviews, another 5 participants will be interviewed and saturation will be reassessed. Once saturation is reached, lead investigators will review all transcripts to generate an initial list of codes using a content coding framework. The creation of the coding framework will begin by the identification of anything that participant’s mention related to the CBT strategies, the gaming exercises, activity restriction or fear of falling/balance confidence. From this broad coding framework, responses will be sorted according to the themes. Brief descriptions of final codes will be created. A lead investigator will then meet with research assistants to discuss the coding structure. Research assistants will code 3 transcripts, meet to review discrepancies and revise the codes, if necessary. Remaining transcripts will be coded individually; research assistants will then meet to resolve discrepancies. Final codes will be entered into NVivo, version 9 for analysis.