SALGOT - Stroke Arm Longitudinal study at the University of Gothenburg, prospective cohort study protocol

Kinematics describes movements of the body through space and time, including linear and angular displacements, velocities and accelerations, but without reference to the forces involved. Kinematic data can be achieved by optoelectronic systems where multiple high-speed cameras send out infra red light signals and detect the reflection from the markers placed on the body. Kinematic variables provide objective, precise and detailed measures of movement performance and quality.

Kinematic movement analysis has become a useful assessment tool within rehabilitation and is employed routinely for gait analyses. Few studies have used kinematic movement analysis to examine the upper extremity in a longitudinal design. In one of these studies the kinematic data was obtained from an isolated fast elbow extension [15, 19] and in the other a targeting fast reaching movement [20]. In order to better understand the situation of a person with impaired upper extremity function, information is needed regarding activities of daily living. It is known that the motor activity of the upper extremity is dependent on the meaning of the task and on the shape and placement of the object [21]. Thus, it is meaningful to study natural purposeful movements with real-life objects. In an earlier study we have developed a test protocol and a program for data analyses of the kinematic variables for the activity of drinking from a glass, which has been applied in a control setting [22] and in stroke subjects [23].

Virtual reality (VR) can be described as the world perceived in a computer. VR systems that include a haptic device can provide tactile feedback to the user through the force feedback. If the system detects a collision between the device and virtual objects, it transmits a reaction to the user's hand, which interacts with perception of the test or training situation [24]. In the real world, objects are usually perceived in the same location whether the sense involved is vision or touch (haptic). In the virtual world, the precise co-location of haptics is technically harder to achieve, but when the co-location is accurate the realism of the manipulation is very high and the user's performance is improved [25]. The knowledge about effects of using VR in assessments and training after stroke is still limited, but sufficiently encouraging to justify additional clinical trials in this population [26‐31].

WHO approved in May 2001 the model on International Classification of Functioning, Disability and Health (ICF) [32] to assess the consequences of a disorder or a disease on the individual person. The ICF model provides a multi-perspective approach to the classification of functioning and disability as an interactive and evolutionary process. In the model an individual's functions in a specific domain is an interaction or complex relationship between the health conditions (physical or mental) and contextual factors (social and physical environment as well as personal factors). The components of ICF can be used to indicate problems (e.g. impairments, activity limitations or participation restrictions summarized under the umbrella term disability) in different areas. This approach forces health professionals to look wider than the usual perspective, which has traditionally lain in the domain of body function and structures. The model boosts the traditional rehabilitation ideology where the focus has not been on the organ but on the person and thereby requiring different treatments depending on that person's goal. In order to assess the consequences of a disease we need to look at different components of the ICF.

Longitudinal studies are difficult to perform. Sweden has a unique situation since people are quite easy to trace through the civic system and moving from one region to another is not so frequent. In addition, the representativeness for the disease is good since all patients within a catchment area are usually referred to the same hospital as private alternatives are scarce and thereby the possibilities to generalize the results are good.

The purpose of this study is to describe the recovery of upper extremity function longitudinally in a non-selected sample with first ever clinical stroke admitted to a stroke unit.

The specific objectives of the present study are to:

This study will evaluate the recovery patterns after first ever stroke without any intervention except standard rehabilitation planning and procedures. All included participants will be assessed eight times during the first year after stroke. Assessments are performed at day 3 and 10, week 3, 4 and 6, month 3, 6 and 12 after onset of stroke. Tests are administrated in block randomized manner in order to minimize the systematic testing bias. The test order and the reason for missed or unsuccessful test results will be recorded in a protocol. All tests are performed by three experienced physical therapists, undergoing a training period together for the assessment battery prior to the study start. ICF classification of functioning, disability and health is used as framework for the selection of assessment measures (Figure 1).

Demographic data will be collected during the first assessment. Stroke subtype will be confirmed by CT and/or MRI scans. Ischemic strokes will be classified for subtype and site for lesion by using TOAST [35] and Bamford classifications [36]. Treatments of thrombolysis or thromboectomy will be registered. Additional data will be extracted from the national quality register for stroke - Swedish Stroke Register [1]. The Self-Administrated Comorbidity Questionnaire (SCQ) will be used to collect additional information on relevant medical conditions and problems [37]. Cognitive function is evaluated at every test occasion using Barrow Neurological Institute Screen for Higher Cerebral Functions (BNIS) [38]. The three prescreen items scoring the level of consciousness/alertness, cooperation and basic communication skills and the item of auditory comprehension will be assessed. The level of physical activity is recorded by a 6-grade scale of Physical Activity Classification [39, 40]. This instrument is valid, short and suitable for longitudinal studies and takes account the activity level both during domestic and fitness activities [40]. Exact time points for all assessments are listed in Table 1.

The upper extremity motor function will be assessed using the Fugl-Meyer Assessment for Upper Extremity (FMA-UE) [41], and a maximum score of 66 corresponds to normal motor function. The psychometric properties of Fugl-Meyer Assessment have shown excellent reliability and validity [41‐43]. The non-motor domains of FMA-UE, sensation, passive range of motion and pain during passive joint motions will be completed as well.

Action research Arm Test (ARAT) is a performance test for upper extremity function and dexterity [44]. The ARAT uses ordinal scoring on 19 items divided into four hierarchical subtests: grasp, grip, pinch and gross movement. Each upper extremity is evaluated individually and the test can be completed in 5-15 minutes [44, 45]. ARAT has been shown to have good validity, sensitivity to spontaneous and therapy-related gains after stroke both in acute and chronic phase [44, 46]. The ARAT has shown good responsiveness [47] and excellent inter-rater and intra-rater reliability [44, 48].

Spasticity will be assessed with the Modified Ashworth Scale (MAS). The muscle groups of elbow flexors and extensors, wrist flexors and extensors will be evaluated. The MAS is the best alternative for spasticity assessment in clinical setting available and has been shown to have fair reliability for these joints [49, 50].

The grip strength will be measured using the Jamar Hand Dynamometer. Standardized positioning and instructions are followed and the average of three trials is used as test outcome [51]. Reliability for the grip strength measure is very high [52].

Three-dimensional motion analysis of upper extremity during drinking task will be performed with a 5-camera optoelectronic ProReflex Motion Capture System (MCU240 Hz, Qualisys AB, Sweden). The tracing of the three-dimensional coordinate positions of the markers is completed automatically by Qualisys Track Manager, 2.0. The capture data is then transferred to MATLAB (The MathWorks Inc) software for custom-made analysis. A standardized drinking task with stable test-retest reliability will be used [53]. The participant is sitting in front of the table with tested hand resting on the edge of the table (Figure 2). A drinking glass, filled with 100 mL water is placed 30 cm from the table edge in the midline of the body. The drinking task includes reaching, grasping, and lifting the glass from the table and taking a drink (one sip); placing the glass back on the table behind a marked line; and returning to the initial position. Participants are instructed to sit against the chair back during the whole task, but the sitting position is not restrained, and compensatory movements are allowed. All participants perform the drinking task at a comfortable self-paced speed, starting with their non-affected arm, after practicing a few times. The mean of the three middle trials of total five will be used for statistical calculations. A total of 9 spherical 12-mm retroreflective markers are placed on the third metacarpophalangeal joint of hand, styloid process of ulna on wrist, lateral epicondyle of elbow, middle part of acromion on right and left shoulder, upper part of sternum, forehead and on the upper and lower edge of the glass. The procedure has been described in more detail previously [53, 54].

In the VR test [55], the participant reaches into a virtual space and interacts with 3D objects. The VR equipment consists of a semi-immersive workbench with haptic device and stereoscopic glasses. In our set-up, the haptic equipment looks like a stylus shaped instrument attached to a lever system and it is freely movable in all directions (Figure 3). During the test, the position of the stylus is tracked, and resistive force is applied to the stylus when it comes into contact with the virtual object, providing force feedback. In addition to the visual perception, the haptic device creates an illusion of manipulation and sensation of the virtual objects. The participant moves the stylus in a realistic environment, experiencing the sense of moving inside the computer screen. The precise co-location of haptics is achieved by projecting the virtual image onto the same location as the user's hand through the mirror setup. The VR-test, developed by our group, is a precise quantitative kinematic measurement tool for arm and hand movements and has been shown to have a good test retest reliability [31, 56, 57]. During the test the participant has to move the haptic stylus to 32 different targets in the virtual environment (VE) generated by the computer. The targets appear one after the other and disappear when touched. Each target consists of a whole circle (diameter 3.0 cm viewing angle). The 32 target placements in the VE are random to the subject but are actually set according to a pre-set kinematic scheme for evaluation purposes. In each test occasion the participant have one or two training trails before the measurements starts. Both dominant and non-dominant hand is measured, starting with the non-dominant hand. The participant performs the test as fast as possible.

ABILHAND [58, 59] is a questionnaire aiming to assess manual ability in persons with chronic stroke. It is interview based and focused on perceived difficulties in everyday activities. A Swedish version has been validated [60]. ABILHAND is a Rasch-based assessment; it is unidimensional and can be used as linear measure [58, 59].

Stroke Impact Scale (SIS) [61] is a questionnaire on different aspects of the stroke recovery where the person replies on their perception regarding their life after the stroke. The 59 questions are divided into 8 domains; strength, memory, emotion, communication, activities of daily living, mobility, hand function and social participation. Items within the domain are ordered hierarchically based on clinical perspective and Rasch analysis [62]. Only the first four sections are used for the test occasion at day 10.

Impact on Participation and Autonomy (IPA-E) is a generic outcome measure for adults with chronic conditions where the person estimates perceived limitations in participation and autonomy related to dependency in the current living surrounding [63‐65]. The subscales include autonomy indoors, family role, autonomy outdoors, social life and relationships, work and education. Additionally, IPA-E identifies the extent to which limitations in life are experienced as problematic in areas of mobility, self care, activities, economy issues, social life, work and education. IPA-E is valid, reliable and sensitive to change after stroke [63‐65].

EuroQol Quality of Life Scale (EQ-5D) will be used to measure the health status related to the quality of life. It is a widely used generic measure and includes five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression [66, 67].

The kinematic data in the drinking task is filtered with a 6-Hz second-order Butterworth filter, resulting in zero-phase distortion and fourth-order filtering. The drinking task is broken down into five logical phases: reaching for the glass, forward transport of the glass to the mouth, drinking, back transport of the glass to the table, and returning the hand to the initial position. The selection of kinematic variables and data analysis calculations will be based on our earlier studies [53, 54]. Movement onset is defined as the time when the tangential velocity of the hand marker exceeds 2% of the maximum velocity in the reaching phase. Movement offset is detected when the velocity of the hand is less than 2% of the maximum velocity in the returning phase. Start of forward transport phase is defined as the time when the tangential velocity of the glass exceeds 15 mm/s. The drinking phase is identified by a 15% increase or decrease of the steady-state distance between the face and glass marker. The start of the returning phase is defined as the time when the tangential velocity of the glass is less than 10 mm/s. Movement times are calculated for the whole movement and separately for each phase. Peak tangential velocity and angular velocity of the elbow joint are computed for the reaching phase. Smoothness of movement is quantified by computing the number of movement units during the reaching and forward transport phases [53]. Angular joint motions are computed from the 3D position data for elbow flexion/extension, shoulder flexion/extension in the sagittal plane, and abduction/adduction in the frontal plane [53]. Compensatory trunk movement is computed for the entire drinking task as the maximal displacement of the thorax marker from the initial position [53]. Interjoint coordination between the shoulder and elbow joint angles for reaching phase is computed using cross-correlation analysis of zero time lag [53].

In the VR-test hand position data (haptic stylus end-point) will be gathered. The position of the stylus is tracked and resistive force is applied to it when it comes into contact with the virtual model, providing force feedback. All measurements generate time-stamped motion data (x, y, z) at 1000 Hz. Different parameters such as reaction- and movement time, velocity, acceleration and deceleration times are calculated. To obtain the movement quality of the hand trajectory, a hand path ratio, corresponding to the length of the pathway is calculated. The selection of kinematic variables and data analysis calculations will be based on our earlier study [30].

The raw scores from the ABILHAND questionnaire are analyzed using a Rasch analysis computer program and expressed as logistically transformed probability measures, logits [68]. In the Rasch model the raw scores are used to estimate the linear ability for each subject and linear difficulty for each item of measurement around a unidimensional continuum. Thus, the Rasch model converts the ordinal score of subject's manual ability into an equal interval linear measure.

Prior longitudinal studies stroke cohorts at Sahlgrenska University Hospital have had a dropout rate of 30%. With a power (1-β) at 0.8 and a significance level (α) at 0.05, we need a sample of 88 patients (two-sided test) to determine a medium effect of 6 points change (10%) on ARAT. Therefore, we aim to include 120 persons.