Interdisciplinary Comprehensive Arm Rehabilitation Evaluation (ICARE): a randomized controlled trial protocol

Participants are adults with diagnosis of ischemic stroke or intraparenchymal hemorrhagic stroke. CT or MRI scans are used to confirm stroke diagnosis. If a scan is not available, the diagnosis is confirmed by clinical criteria. Specific Inclusion and exclusion criteria are provided in Table 2.

There are four evaluation points in ICARE: baseline, completion of intervention, and 6 months and 1 year after randomization (Figure 1). The primary time point at 1 year is used to test the primary and secondary hypotheses using the analysis plan described herein.

The log-transformed WMFT time score at the 1 year endpoint is the primary outcome measure for the trial and the basis of a priori sample size and sensitivity estimates. The SIS hand domain and full SIS constitute the secondary outcome measures. The WMFT and SIS are described in detail below. A full list of assessments is included in Table 3, which describes the timing for data acquisition by each instrument including the WMFT and SIS. The full battery of assessments is designed to provide information about muscle strength, cognition, digit sensation-perception, functional ability, depression, self-efficacy, life satisfaction, reintegration, and subjective quality of life. These are listed in the Additional file 1 and arranged roughly into categories using the International Classification of Functioning and Disability Framework [36] (ICF).

Trained licensed occupational or physical therapists blinded to group assignment (Blinded evaluators, BE) and certified for administration, perform all evaluations. The primary outcome assessment is filmed using a digital video camera according to detailed procedures outlined in the Manual of Procedures (MOP). Identical computer, digital video camera, editing, compression and conversion software are standard equipment that were issued to each ICARE site during the start-up phase. A FTPS (File Transfer Protocol Secure) server is maintained to enable digital video review and standardization and data file transfer across sites and to the trial’s administrative team.

The WMFT is a laboratory-based functional assessment of the arm and hand [62]. The instrument includes 15 measures of speed (timed tasks), 2 measures of strength (arm, grip), and 15 measures of movement quality using the Functional Ability Scale (FAS) of the upper extremity [63].The FAS movement quality score is determined from video review by a panel of raters blinded to treatment group and assessment time point. Items include a hierarchically arranged set of functional movements of the shoulder, elbow and hand that progress from proximal movements to distal hand tasks including grasp control and manipulation. The WMFT has been shown to be both reliable and valid in the stroke population [64, 65]. It has been used previously, in numerous studies including the EXCITE and VECTORS trials.

The Stroke Impact Scale (SIS) 3.0 is a well-established 59-item interviewer-administered assessment of health-related quality of life for individuals after stroke [66, 67]. The SIS captures 8 domains of function organized into subscales as well as an overall measure of perceived recovery from stroke. Both the full SIS 3.0 (ICF Participation level) and the hand function subscale (ICF Activity level) are used in ICARE as secondary outcome measures. The hand function subscale requires participants to indicate the difficulty experienced over the past 2 weeks with 5 activities involving the paretic hand, including: carrying heavy objects, turning a doorknob, opening a can or jar, tying a shoelace, and picking up a small coin.

Within 30 days of randomization, the recruiting site team initiates monthly telephone interviews with each participant to ascertain information about health status, healthcare utilization, medications, other therapies, and adverse events.

A multiple question survey interview is administered by a non-treating, unblinded member of the recruiting team upon completion of the intervention phase. At the post-intervention evaluation, each participant is asked a set of questions to assess the extent to which critical components of the investigational intervention (e.g. impairment mitigation, session intensity, participant chosen tasks, therapist-participant collaboration) were incorporated into each assigned therapy group [68].

At the final 1 year evaluation, participants complete another set of questions regarding activity since the end of the intervention phase. Participants are also asked to report the perceived value of the intervention and study participation.

A panel of 3 experienced therapists provide FAS ratings for the WMFT. Inter-rater reliability is refined via teleconferences. The Data Management and Analysis Center (DMAC) convert and compress the site submitted edited digital WMFT test administration. Media stream access is provided to the members of the rater panel through the secure ICARE database website, thereby enabling remote viewing and scoring capability. Intra-rater reliability and inter-rater reliability were initially assessed using a sample of 8 videotaped WMFT test administrations procured from a database of previous studies. Raters rated each video at that time and after 1 month. A web/teleconference followed to review and discuss discrepant individual item ratings. This resulted in further refinement of the scoring guidelines. Two additional test-retest iterations were completed, each time with a set of 21 filmed ICARE WMFT test administrations. These were analyzed using weighted Kappa statistics, Spearman correlation coefficients and Bland-Altman plotting to determine points of disagreement. Currently, each of the 3 panel members rate each video independently, to yield 3 independent ratings per video. Specific instances of item score disagreement are identified by the DMAC for review and discussion by the panel before adjudication by majority vote. The WMFT FAS Panel meets via tele-conference monthly to view and rate a new video as a group.

The Accelerated Skill Acquisition Program (ASAP) is a fully defined, principle-based protocol that integrates three fundamental elements including: skill acquisition through task-specific practice, impairment mitigation to increase capacity, and motivational enhancements to build self-confidence (Figure 2). ASAP is grounded in the evidence-based expectation that effective rehabilitation of the paretic upper extremity is achievable and based upon the provision of challenging, intensive, and meaningful task practice for motor skill acquisition, mitigation of associated linchpin impairments and dysfunctions of movement, and the confidence to integrate use of emerging skills into daily life activities [2, 19, 20, 71, 72].

The ASAP structured protocol includes an initial evaluation and orientation session (Session A) and 30 1-hour visits of an individualized practice program focused on function of the arm and hand most affected by the stroke. This integrated and evidence-based intervention for upper extremity recovery arose from diverse but converging, complementary, and interdisciplinary literatures of basic and applied science as well as recent translational and stroke clinical trial research of upper extremity recovery. Unique aspects include the structured framework by which intensity and progression of practice is managed and which fosters participant skills and confidence through therapist-patient collaboration. Participants randomized to ASAP are given a customized package of therapy that includes challenging, intensive, and meaningful practice of activities related to participant chosen real-world tasks (e.g., carrying groceries, handwriting) that engage the arm most affected by the stroke. Participants are offered a mitt to wear on the less affected hand during the time outside of therapy to promote use of the weaker arm and hand; however, the participant is not required to use the mitt if so chosen. Outside of therapy specific assignments (i.e., Action Plans) are given to encourage self-managed, confident, safe, and effective arm use at home and in the community.

Table 4 summarizes the eight evidence-based, non-exclusive (over-lapping) operating principles that are currently used to guide ASAP intervention sessions: 1) ensure challenging and meaningful practice [20, 73‐76], 2) address important mutable impairments through task-breakdown [77‐80], 3) enhance motor capacity through overload and specificity [81, 82], 4) preserve natural goal-directedness in movement organization [83, 84], 5) try to avoid artificial task breakdown when engaging in task-specific practice [85], 6) assure active patient involvement and opportunities for self-direction [86, 87], 7) balance immediate and future needs for efficient motor skill and capacity enhancement with the development of confidence and self-management skills [88, 89] and 8) drive task-specific self-confidence (self-efficacy) high through performance accomplishments [90, 91]. Together, these principles are designed to emphasize and drive the development of skilled movement performance.

The program begins with an orientation and evaluation session to accomplish the following 10 goals: 1) prepare the collaborative ‘real-world’ task list to be used during training; it includes 6 tasks the patient most wants to perform with at least two bimanual activities, two strength-dependent activities including the most affected arm, and two activities requiring dexterity of the most affected hand, 2) designate a priority or benchmark task from the collaborative task list, 3) determine fundamental impairments and the priority-task challenge threshold or movement breakdown point(s), 4) prepare a collaborative schedule for the first day of training, 5) orient the participant to the mitt and its function including safety precautions, 6) identify appropriate conditions for mitt wearing and not wearing, 7) orient the participant to a recurring brief self-efficacy question, 8) orient the participant to out-of-lab action plans, 9) orient the participant to their collaborative role with the trainer and 10) obtain the participant’s agreement to be a collaborative partner. The orientation and evaluation Session A may last up to two full hours.

Training sessions are to be 1 hour per day, 3 times per week for a total of 30 hours, with rest breaks allowed to maximize engaged time-on-task. Following a debriefing of the previously assigned Action Plan and the acquisition of several physiologic measures (i.e., BP, heart rate, pain, muscle soreness), each regular training session begins with a collaborative ordering of the real-world tasks identified during Session A (orientation and evaluation). The real-world tasks may change as interests and goals evolve over the 30 sessions, however the priority task may not change for benchmarking purposes. Task and movement analysis are done for each real-world task to determine the key movement dysfunctions or impairments. The goal of intervention training is to focus attention and effort directly on the problematic area (i.e. dysfunction, impairment) to facilitate skill acquisition without simply providing a compensatory strategy as an easy and quick fix to the problem. Classic exercise-overload principles (e.g. intensity, periodicity) are used to drive progression and build motor capacity (e.g., muscular strength and endurance, coordination). Practice activities within the three categories of real-world functional tasks (i.e., strength, dexterity, bimanual) are selected based on patient task preferences. Training is collaborative and interactive with the participant providing problem identification as feasible (e.g., “What is the limiting factor when you perform that task?”) and solutions through self-assessment and trainer feedback/suggestions. Confidence building and empowerment are embedded in the training and education during each session. Task-specific (i.e., priority task) self-efficacy assessment is performed 4 times throughout the 30 sessions using a Brief Self-Efficacy Rating Scale. The participant is asked to provide a number between 0 and 10 in response to the question: “How confident are you that you can (fill in specific priority task activity)?” The confidence rating is followed by a question asking, “What can we do this week to increase your capabilities and raise your confidence?” As the participant gains insight and develops self-assessment and management skills, the response to the follow-up question is expected to become more insightful and rich in self-knowledge and ways to drive challenge and improvement. One strategy for building capacity is to perform inter-session ‘Action Plans’ or out of clinic activities. All 30 sessions have associated Action Plans including some with handouts that provide background information related to ASAP principles and evidence. Table 5 lists the knowledge-based handouts provided to participants and the planned distribution order across the 30 sessions. Individual participant-specific Action Plans are also generated and used throughout the 30 sessions. These Action Plans encourage specific practice in the home or community settings. Examples include finding a challenging task involving food preparation or eating or reading an education handout about motor recovery. At the beginning of each session, participants are asked to report on the effectiveness of their Action Plan activities on the next day of training. The therapist also rates the participant’s engagement in the Action Plan on a scale from 1 = no effort/did not attempt, 2 = some effort, and 3 = great effort/engagement, apparently practiced diligently.

Figure 3 illustrates the clinic environment and two examples of tasks selected by participants to practice during ASAP sessions. To prevent cross-contamination between study groups, further details of the investigational intervention as outlined in the ASAP MOP have been embargoed until study completion.

To ensure data quality and consistent administration of the intervention among all ASAP therapists, a staged standardization process (Phase 1 and 2, Phase 3 and Re-certification) was developed. At each stage, an expert reviewer (experienced members of the ICARE Investigative team) assesses the digital video footage accompanying documentation for therapist mastery of each ASAP principle implemented during a 1 hr session. Phase 1 and 2 standardizations must be attained working with a volunteer with stroke before a therapist may initiate therapy with a randomized participant in the ICARE trial. Once Phase 1 and 2 standardizations have been attained, the therapist may begin working with an ASAP-randomized participant to complete Phase 3 standardization. After Phase 3 standardization is acquired, the therapist must re-certify according to the schedule directed by the reviewer. If in the worst case scenario, a therapist fails to demonstrate ASAP treatment fidelity at Phase 3 certification or recertification thereafter, all post-randomization data accrued for a participant seen by that therapist will be designated unusable for outcome analysis.

The Phase 1 and 2 standardization processes include demonstration of mastery in the following three task areas: 1) administration of the Brief Self-Efficacy Question; 2) determination of the challenge threshold and fundamental movement problem for two arm and hand activities, and 3) determination of appropriate task practice intensity and activity/task progression. Additional mastered elements for Phase 1 and 2 include: generalized ASAP protocol awareness via a knowledge test and therapist self-assessment of at least 90% mastery. Phase 3 standardization is attained via demonstrated mastery of the ASAP protocol through review of the 1^st and 20^th treatment sessions administered to the new therapist’s first ASAP randomized participant. Recertification is maintained through demonstration of continued mastery, on a target schedule of approximately every 6 months. During the trial start-up period, a 3-day training workshop was held for all ASAP therapists to attain Phase 1 standardization. Intervention therapists in all stages of standardization and the protocol developers facilitate ongoing training via monthly ASAP therapist teleconferences with video and literature review, individual conferencing, an on-line discussion board, and site visits for digital video review and discussion of ASAP principles and clinical experiences.

The UCC group receives Outpatient Occupational Therapy for the upper extremity as determined by each participant’s individual occupational therapist, based upon usual and customary practice standards. In accordance with those practices, the participant attends a standard Occupational Therapy evaluation session prior to initiating intervention sessions (treatment). The number of visits and frequency (dose) for the intervention is determined per the therapy prescription generated at the evaluation, site-specific usual and customary practices, and payer guidelines (e.g. private insurance, HMO, Medicare). The evaluation may have occurred prior to randomization. The dose may be modified per usual and customary practices as the intervention ensues. The treating therapist(s) documents each session per the treating facility’s protocol and documentation methods. The CSC of the clinic that randomized the participant is responsible for procuring the evaluation and treatment records and gleaning the necessary information for ICARE using standardized forms. Vital signs, as well as any symptoms of fatigue or stress are monitored during each session in accordance with usual and customary standard practice.

The DEUCC group initially receives Outpatient Occupational Therapy for the upper extremity as determined by each participant’s individual therapist, based upon usual and customary practice standards in a manner identical to the previously described UCC group with respect to content. Differentiating this group from UCC is its dose-equivalency with the ASAP group; each receives 30 hours of distributed treatment. Prior to initiating treatment (intervention) and in accordance with usual and customary practices, the participant receives a standard Occupational Therapy evaluation, where a recommended number of visits and frequency (dose) are determined, in accordance with site-specific usual care practices. This evaluation may have occurred prior to randomization. Immediately following the initial evaluation, the therapist and patient are notified of group assignment and 30 sessions are scheduled. According to protocol, if more than 30 treatment sessions are scheduled, the CSC will indicate that the treatment window is closed and a post-test will be scheduled. Provision of further treatment is documented via the monthly telephone interviews. Documentation and data are collected in a manner identical to that for the UCC group.

Thirty hours of training was chosen for the ASAP protocol because this dosage fell within the range of previous rehabilitation intervention trials that were shown to be effective in the post-acute outpatient setting. A training schedule of 30 hours of treatment, ideally distributed into 1 hour sessions, 3 days per week over 10 weeks is the anticipated experimental frequency for both the DEUCC and the ASAP groups. Both interventions vary in content and structure only. Pilot data from a multi-site outpatient survey conducted prior to initiating ICARE suggested that 30 hours of training would be higher than that commonly prescribed in the outpatient post-acute setting, and the 1 hour session length was designed to afford practicality and allow patients to participate in other concurrent therapy services (e.g., physical therapy and speech therapy). Those randomized to DEUCC or ASAP are expected to receive at least 90% (≥ 27 hours) of the 30 scheduled hours between 10 to 16 weeks post-randomization to be considered adherent to the dosing protocol. Therapy may extend up to 16 weeks to compensate for delays in initiating treatment post-randomization, missed sessions due to illness or other unavoidable absences. Precise tracking of therapy dose is maintained on all participants including: frequency and duration of individual treatment sessions, duration of total treatment sessions, and number of missed and canceled appointments. Additionally, for the UCC and DEUCC groups, initial OT prescription, prescription change, treatment venue and change in treatment venue are documented.

Sample size estimates were computed for the primary a priori aim of detecting a clinically relevant difference in log-transformed WMFT time score between ASAP and DEUCC at 1 year follow up with α = 0.05. With the proposed total sample of 360, we would have sufficient power to detect a moderate difference in treatment effect of Δ = 0.40 for an attrition rate of 17%, and Δ = 0.42 for an attrition rate of 25% using a two-group 2-sided t-test with a type I error of 0.05 and 80% power. The EXCITE trial showed treatment effect sizes of 0.50 and 0.63 for the high and low functioning groups, respectively. For the small sample phase II VECTORS study, a similar analysis showed a treatment effect size of 0.20. Given the outpatient timing of ICARE, a less dynamic period of change than VECTORS, we expect to have sufficient power to detect the effect size in log WMFT time score. We also should have sufficient power to detect a difference in the proportion of participants who change ≥ 25 points on the normalized SIS at 1 year using a χ² statistic. The minimal success rate difference that can be detected with 80% power ranges from 13.5-20% (with 17% attrition) to 14.0-21% (with 25% attrition); the variation in the estimates reflects potential variation in improvement in the DEUCC group. To achieve this recruitment goal (N = 360) in 40 months, each of the three centers aims to randomize 3 participants per month.

Descriptive statistics of demographics, baseline characteristics, and distributions will be performed for the total sample and for each group. Group comparisons will be made to examine whether any of these variables need to be accounted for in further analysis due to an imbalance across groups. Continuous variables will be compared using ANOVA for normally distributed variables to test for mean differences, and Wilcoxon rank sum test will be used for non-normally distributed variables to test for median differences. χ² or Fisher’s exact test will be used for categorical variables to test for frequency differences. Participant characteristics that differ at baseline will be included as covariates in the analysis of the primary and secondary outcomes.

This trial addresses two different questions separated by primary aim and secondary aim: 1) Is ASAP superior to DEUCC (primary aim)?, and 2) Is ASAP or DEUCC superior to UCC, (secondary aim) For all analyses, assumptions required for the data distribution (e.g., normal distribution) will be assessed. Any transformations of data or alternative methods necessary to analyze the data will be determined by examining the structure of the data. All analyses will be performed in accord with the intent-to-treat (ITT) principle (e.g., group status is determined by randomization at baseline). A p-value < 0.05 will be used to indicate statistical significance for all analyses.

The analytic approach to the secondary aims will follow that described for the primary aims above, with the exception of comparing differences between groups receiving 30 hours of treatment (ASAP and DEUCC) to UCC. For other secondary outcomes, baseline to 1-yr change will be compared between ASAP and DEUCC groups and ASAP and UCC groups using analysis of covariance or logistic regression models (see Additional file 1). In addition, a composite physical domain, which includes strength, hand function, ADL/IADL, and mobility, will be created and compared in a similar way. Bonferroni adjustments will be made of the critical p-value for these secondary analyses, as there are two comparisons: ASAP to UCC and DEUCC to UCC; a p-value < 0.025 will be used to indicate significance.

A single interim data analysis of Specific Aim 1 (ASAP vs DEUCC) for only the primary outcome, WMFT time, will be performed when approximately 33% of the participants have completed the 1-year evaluation. The O’Brien-Fleming group sequential method [92] will be used with the significance level defined as 0.005 at the interim analysis to maintain the overall type I error of 0.05. The advantage of the O’Brien Fleming method is: 1) it is conservative for significance at the interim analysis point, and 2) it maintains the conventional p-value in the final analysis. Group assignment will remain blinded with codes X and Y representing either ASAP or DEUCC groups to avoid revealing treatment assignment. Only the primary outcome for the primary aim (log WMFT mean time) will be analyzed as described above. If the test of success rate in WMFT time at the interim analysis point reaches a p-value less than these critical p-values, the DSMB will decide whether analysis of secondary outcomes is necessary to provide additional support of the results. If deemed necessary by the DSMB, conditional power will be computed to make a reassessment of the power at the end of the trial.

The DMAC conducts weekly checks for all expected measurements in the database. Missing data must be identified by the cause of missing data, (e.g. participant not able to complete the measurement or evaluator error, forgot to record the data). Detailed reasons for missing data are recorded in a comment box under each measurement. This missing data form/point control procedure will enable future analysis for missing data patterns (e.g., systematic or random), and thus, support decision making for final data analysis.

The adverse event reporting system was developed with NINDS guidance, and has a pre-planned, automatic work-flow that triggers instant electronic communication. A site team member completes the online “Adverse Event Initial Report Form” within 24 hours of knowledge of an event and in compliance with all local IRB reporting procedures. This initiates the report process. The initial report is instantly added to the ICARE database and triggers an immediate, automatic notification to the study PIs, project manager, originating site’s physician investigator and clinical site coordinator, and DMAC members. The site physician investigator is responsible to obtain all necessary facts to verify and adequately describe the event, determine seriousness, determine the appropriate course of action and report these items via the ICARE “Adverse Event Confirmation Form” to the database within 72 hours after the initial report submission. If the AE is confirmed as serious, an immediate, automatically triggered email notification with a narrative of the SAE is sent to the study PI’s, project manager, originating site’s physician investigator and clinical site coordinator, MSM and the DSMB liaison. The MSM will review the SAE narrative and adjudicate study-relatedness and expectedness within 5 days (120 hours) of the automated email notification of an SAE. If data are not entered within the first 96 hours, an automated courtesy reminder is sent to the MSM. A detailed AE report is generated weekly and available to all study personnel. An AE summary report is generated and reviewed quarterly by the Data Safety and Monitoring Board (DSMB) and the USC IRB. Each site reports to its respective IRB in accordance with its policies and procedures. All AE reports, and amendments to them, are recorded and monitored by the DMAC.

The DMAC consists of the director, who is the unblinded statistician and who is responsible for managing the DMAC and acting as liaison to the DSMB for data-related matters, including preparation of reports and analyses (interim and primary outcomes). The other DMAC members are staff from USC’s Statistical Consultation and Research Center and include a co-director in charge of database development and management, a programmer, data manager and two research assistants. Additionally, a blinded statistician serves on the executive committee and is able to answer statistical questions that arise without danger of breaking the blind of the study.

Microsoft® SQL Server 2000 Edition is utilized for the secure database. This database is configured to utilize transaction logging for auditing and recovery purposes. All data entry/retrieval travels over the Hypertext Transfer Protocol Secure (HTTPS) layer. Backups are conducted daily, weekly, monthly and quarterly. The backup files are stored in a backup server at a secure separate location. The WMFT video files are stored in a multimedia server and streamed to the data entry system with the corresponding accession number, which provides blinding to site and evaluation time point.

Users are issued individual usernames and passwords to gain access to the secure web-based data entry system. Since ICARE is a multisite study, each site has access only to its site- specific data records. Within each site, only designated personnel can enter and modify data. The blinded evaluators are prohibited from accessing any treatment related information in the database, including randomization assignment.

All clinical report forms (CRF) are barcoded to prevent ID error. Sites are required to scan and submit the original CRFs for all key outcome measurements to the DMAC for data entry quality checking. Quarterly, the DMAC performs data quality checking (QC) on 100% of primary outcome data (WMFT and SIS).