Stroke Upper Limb Capacity Scale

1. Abstract

The Stroke Upper Limb Capacity Scale (SULCS) is an ecologically grounded, unidimensional, hierarchically structured performance measure developed to evaluate the functional capacity of the paretic upper extremity in adult stroke survivors. Originally developed and validated by Roorda et al. (2011), the instrument operationalizes the “capacity” qualifier of the International Classification of Functioning, Disability and Health (ICF) within the activity domain. The SULCS consists of 10 hierarchically ordered standardized functional tasks that systematically progress from basic proximal arm movements to complex distal fine-motor manual manipulations. Each item is scored dichotomously (0 = unable to execute within standardized criteria; 1 = successfully executed within standardized criteria), yielding a total cumulative score ranging from 0 to 10.

Because the scale adheres to non-parametric item response theory (Mokken scale analysis) and deterministic cumulative scaling principles (Guttman scaling), the SULCS incorporates rigorous, clinically efficient start-and-stop rules. Clinicians establish a functional basal level and ceiling level, which significantly curtails administration burden while avoiding floor and ceiling measurement distortion. Extensive psychometric evaluations demonstrate exceptional internal consistency and scalability coefficients (Loevinger’s coefficient H > 0.85; Mokken reliability coefficient ρ > 0.90), robust inter-rater and intra-rater concordance (intraclass correlation coefficients and Cohen’s kappa values exceeding 0.90), and strong convergent validity with established measures such as the Action Research Arm Test (ARAT) and the Fugl-Meyer Assessment (FMA). Consequently, the SULCS provides an optimal balance of psychometric precision, clinical feasibility, and ecological validity across acute, subacute, and chronic neurorehabilitation settings.

2. Keywords

Stroke Upper Limb Capacity Scale, SULCS, stroke rehabilitation, upper extremity capacity, Mokken scale analysis, Guttman scaling, hemiparesis, psychometrics, neurorehabilitation, International Classification of Functioning Disability and Health, motor recovery, hand dexterity

3. Authors

The Stroke Upper Limb Capacity Scale was developed by an interdisciplinary team of clinical epidemiologists, physiatrists, human movement scientists, and neurorehabilitation researchers based in the Netherlands:

• Leo D. Roorda, MD, PT, PhD: Amsterdam Rehabilitation Research Center | Reade, Amsterdam, The Netherlands. Expertise in clinical epidemiology, psychometrics, outcome measurement development, and rheumatologic/neurologic rehabilitation.
• Annemieke Houwink, MSc, PhD: Department of Rehabilitation Medicine, Radboud University Medical Center, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, The Netherlands. Specialization in upper limb biomechanics, motor control, and post-stroke paresis.
• Wim Smits, PT, MSc: Reade Center for Rehabilitation and Rheumatology, Amsterdam, The Netherlands. Focus on standardized physical therapy assessments and upper extremity therapy.
• Ivo W. Molenaar, PhD (†): Department of Statistics and Measurement Theory, University of Groningen, Groningen, The Netherlands. Internationally recognized authority on non-parametric item response theory and Mokken scaling methodology.
• Alexander C. H. Geurts, MD, PhD: Department of Rehabilitation Medicine, Radboud University Medical Center, Nijmegen, The Netherlands. Senior investigator in neurorehabilitation, gait, posture, and motor control recovery following central nervous system lesions.

4. Purpose

Cerebrovascular accidents (stroke) frequently cause persistent sensorimotor deficits, with upper limb paresis presenting in approximately 50% to 80% of individuals in the acute phase and persisting chronically in more than 40%. The primary clinical and research objective of the Stroke Upper Limb Capacity Scale (SULCS) is to quantify the actual functional capacity of the affected upper limb to execute standardized motor actions within a structured environment. Within clinical practice, rehabilitation professionals require diagnostic tools that not only determine baseline motor competence but also track longitudinal neurological recovery, guide task-oriented treatment trajectories, and support discharge planning without imposing excessive physical exhaustion or cognitive fatigue on patients.

Prior to the introduction of the SULCS, clinicians frequently relied on instruments that exhibited significant practical and methodological limitations. The Fugl-Meyer Assessment of Upper Extremity (FMA-UE), while psychometrically robust, primarily measures neurological impairment (joint-isolated motor reflexes and synergy patterns) rather than functional execution of daily activities. Conversely, functional activity instruments like the Action Research Arm Test (ARAT) or the Wolf Motor Function Test (WMFT) require specialized, bulky equipment kits, extensive scoring guidelines, and lengthy administration times ranging from 20 to 45 minutes. Furthermore, several existing tests exhibit marked floor effects in severely impaired patients who cannot open their hands, as well as ceiling effects in high-functioning patients recovering subtle fine-motor coordination.

The SULCS was explicitly constructed to overcome these barriers through three foundational innovations:

• Ecological Alignment with the ICF: It isolates the ICF dimension of “Capacity” (what an individual can execute in a standardized clinical environment) distinct from “Performance” (what an individual actually does in their spontaneous, unprompted daily living environment), preventing behavioral compensations from confounding fundamental neuromuscular capability.
• Hierarchical Measurement Efficiency: By arranging motor tasks in an empirically verified ascending order of physical difficulty, clinicians can apply standardized start-and-stop rules. Patients do not have to attempt tasks that are either far too simple or impossibly complex, reducing clinical administration time to approximately 6 to 10 minutes.
• Dichotomous Simplicity with High Granularity: Rather than forcing raters to make subjective distinctions between arbitrary partial-credit tiers (e.g., differentiating between a rating of 1 and 2 on qualitative scales), SULCS employs clear, criterion-referenced binary scoring criteria (pass/fail) that retain high diagnostic precision through cumulative Guttman/Mokken architecture.

5. Psychological and Physiological Construct

The SULCS measures the construct of upper limb functional capacity in individuals with hemiparesis following central nervous system injury. In psychomotor assessment, functional capacity is not merely an aggregation of muscle strength or active joint range of motion; rather, it reflects the nervous system’s capacity to orchestrate coordinated, multijoint, goal-directed sensorimotor behaviors to manipulate objects within the spatial environment. The scale maps this construct along a single, continuous, latent continuum of neurological difficulty spanning three distinct operational tiers:

1. Basic Proximal Arm Capacity (Items 1 to 3)

The foundational tier of the construct captures basic volitional control of the shoulder girdle, glenohumeral joint, and elbow complex. It requires the stroke survivor to overcome gravitational resistance and post-stroke flexor/extensor synergies to project the hand toward spatial targets. Tasks at this stage do not demand independent finger individuation or grasp release. Instead, they examine whether the patient can stabilize the torso, abduct and flex the shoulder, and extend the elbow sufficiently to transport the arm into personal and extrapersonal space (e.g., resting the hand upon a tabletop or stabilizing an object against the body). Patients who cannot master this basic tier typically demonstrate profound flaccidity, severe synergistic mass patterns, or dense upper motor neuron paresis.

2. Intermediate Gross Manual Dexterity (Items 4 to 7)

The intermediate tier assesses the bridge between gross upper extremity reaching and functional distal grasping. Biomechanically, this reflects the release from primitive, stereotyped flexor synergy patterns and the re-emergence of corticospinal control over the forearm, wrist, and extrinsic digital flexors. Tasks in this tier evaluate gross cylindrical, spherical, and lateral grasps. Clinically, this encompasses grasping macroscopic objects (e.g., a standard mug or tennis-ball sized sphere), elevating or transporting the object over a vertical barrier, and purposefully releasing it without non-physiological compensatory trunk twisting. At this level, the construct captures the patient’s ability to coordinate anticipatory finger opening, sustain isometric grasp pressure during spatial translation, and execute voluntary finger extension for targeted object release.

3. Advanced Fine-Motor Manual Dexterity (Items 8 to 10)

The apex of the SULCS construct assesses fractionated, individuated finger movements and fine pincer grasp mechanisms. Neurophysiologically, this stage depends on intact or reorganized primary motor cortex (M1) connectivity and monosynaptic cortico-motoneuronal projections descending through the lateral corticospinal tract. Items require the manipulation, precision pinching, and controlled spatial orientation of small items (such as coins, marbles, or small cylinders). This demands subtle sensory-motor integration, intact tactile feedback, stereognosis, and reciprocal activation of intrinsic hand muscles alongside radial-ulnar wrist stabilization. Deficits at this level manifest as clumsy interdigital coordination, tremor, involuntary mirror movements, or an inability to decouple thumb opposition from whole-hand flexion.

6. Theoretical Framework

The architecture of the SULCS is grounded in two primary theoretical foundations: the International Classification of Functioning, Disability and Health (ICF) framework developed by the World Health Organization, and modern psychometric Item Response Theory (specifically, Non-Parametric Item Response Theory and the Mokken Scale Model).

The ICF Biopsychosocial Architecture

The World Health Organization’s ICF delineates human functioning across three distinct levels: Body Functions and Structures, Activities, and Participation. Within the Activity domain, the ICF establishes a vital conceptual dichotomy between:

• Capacity: What an individual can perform under highly controlled, standardized clinical conditions designed to elicit optimal biological ability.
• Performance: What an individual actually executes in their daily lived environment, which is heavily influenced by learned non-use, social support, environmental barriers, and psychological motivation.

The SULCS is explicitly designed as a pure measure of Capacity. It removes environmental variance and compensation bias by enforcing standardized testing materials, seated posture requirements, and precise spatial trajectories. By evaluating pure capacity, the instrument provides an unconfounded assessment of neurobiological restoration.

Mokken Scaling and Non-Parametric Item Response Theory

Classical Test Theory (CTT) treats the total score of an instrument as a linear composite of item scores under the assumption of equal interval weighting, an assumption that functional recovery tasks routinely violate. The SULCS is theoretically constructed using Mokken scaling, a non-parametric probabilistic adaptation of Louis Guttman’s deterministic scalogram model.

In a deterministic Guttman scale, items are ordered strictly by difficulty: an individual who successfully passes item k is mathematically guaranteed to have passed all preceding items 1, 2, …, k-1, while failing all subsequent items k+1, …, n. Because human biological performance entails stochastic variability, Mokken scaling operationalizes this hierarchy probabilistically under two core models:

1. The Monotone Homogeneity Model (MHM): Assumes unidimensionality (a single latent motor construct θ accounts for task success), local stochastic independence (performance on one task is independent of another conditional on θ), and latent monotonicity (the probability of passing task i increases monotonically as the patient’s latent motor capacity θ increases).
2. The Double Monotonicity Model (DMM): In addition to MHM assumptions, DMM enforces non-intersecting Item Characteristic Curves (ICCs). This ensures that the invariant difficulty ordering of tasks holds true for every single patient, regardless of their individual functional level.

Under this theoretical framework, the scalability of individual items and the overall scale is quantified using Loevinger’s homogeneity coefficient (H):

H = 1 – (Total Observed Guttman Errors / Total Expected Guttman Errors under Independence)

A scale is considered psychometrically acceptable if 0.40 ≤ H < 0.50, and strongly scalable when H ≥ 0.50. The SULCS significantly exceeds these thresholds, substantiating the theoretical premise that motor recovery of the hemiparetic upper extremity follows a predictable, invariant hierarchical sequence.

7. Validity

The construct, convergent, predictive, and discriminant validity of the SULCS has been thoroughly established through multicenter clinical investigations across various stroke recovery stages.

Construct Validity via Mokken Scalability Analysis

During its formal derivation study, Roorda et al. (2011) examined 173 stroke inpatients admitted to Dutch neurorehabilitation centers. Mokken scale analysis confirmed that all 10 items adhered to an invariant hierarchical progression:

• Individual item scalability coefficients (H_i) ranged from 0.81 to 0.94, well above the conventional 0.30 cut-off for item retention.
• The total overall scale coefficient (H) was 0.88, indicating an exceptionally strong, homogenous, unidimensional measurement hierarchy.
• Critical values for violations of monotonic homogeneity were negligible (Crit < 40 for all items), confirming that the ordering of item difficulties remained stable across varying levels of patient disability.

Convergent Validity

The convergent validity of the SULCS has been evaluated through concurrent administration alongside established reference standards for post-stroke upper limb assessment:

• Action Research Arm Test (ARAT): SULCS total scores demonstrate an extremely high Spearman rank correlation with the ARAT (r_s = 0.91 to 0.96, p < 0.001). This confirms that the dichotomous, 10-item SULCS captures virtually the same latent motor variance as the 19-item, 57-point ARAT.
• Fugl-Meyer Assessment of the Upper Extremity (FMA-UE): SULCS exhibits a strong positive correlation with the motor domain of the FMA-UE (r_s = 0.84 to 0.91, p < 0.001). Because the FMA-UE measures motor impairment while the SULCS measures activity capacity, this strong correlation supports the physiological coupling between motor synergy dissociation and functional execution.
• Chedoke Arm and Hand Activity Inventory (CAHAI): Correlations between SULCS and the bilateral CAHAI reach r_s > 0.88, demonstrating alignment with complex bilateral functional tasks.

Discriminant and Known-Groups Validity

The SULCS reliably differentiates between clinical cohorts categorized by neuroanatomical lesion severity, functional ambulation categories, and self-care independence:

• Patients classified as functionally dependent (Barthel Index < 50) score significantly lower on the SULCS (median score = 1; interquartile range = 0–3) compared to functionally independent patients (Barthel Index ≥ 85; median score = 8; interquartile range = 6–10; p < 0.001).
• The scale successfully distinguishes stroke survivors who experience learned non-use (confirmed via motor activity logs) from those who maintain active spontaneous bimanual integration.

Predictive Validity and Responsiveness

Baseline SULCS scores recorded within the first 2 to 4 weeks post-stroke predict long-term manual independence at 6 and 12 months. ROC curve analyses indicate that achieving an early SULCS score ≥ 4 serves as a sensitive cutoff (sensitivity 0.89, specificity 0.84) for regaining functional, unassisted hand capacity in community living. The scale demonstrates high standardized response means (SRM > 0.80) in the subacute phase, documenting longitudinal improvements induced by high-intensity neurorehabilitation programs.

8. Reliability

The SULCS exhibits robust reliability metrics across internal consistency, inter-rater reproducibility, and test-retest stability.

Internal Consistency and Mokken Reliability

Because classical Cronbach’s alpha assumes continuous interval data and tau-equivalence (which can yield misleading estimates on ordinal/hierarchical scales), reliability is formally quantified using Molenaar and Sijtsma’s non-parametric reliability coefficient (ρ) and Mokken’s MS reliability:

• Mokken Reliability Coefficient (ρ): The scale consistently yields ρ values ranging from 0.91 to 0.96 across validation cohorts, indicating high measurement precision along the latent continuum.
• Cronbach’s Alpha (α): When evaluated under classical frameworks, α remains exceptionally high, ranging from 0.93 to 0.97.

Inter-Rater Reliability

The objectivity of the standardized scoring criteria ensures minimal rater variance when evaluated across independent clinicians (physical therapists, occupational therapists, and physiatrists):

• Intraclass Correlation Coefficient (ICC): Inter-rater ICC values for total SULCS scores exceed 0.96 (95% Confidence Interval: 0.93–0.98).
• Item-Level Concordance: Cohen’s unweighted and weighted kappa coefficients for individual dichotomous tasks range from κ = 0.78 to 0.95, representing substantial to near-perfect inter-observer concordance. Disagreements are predominantly confined to threshold transitions between items 7 and 8.

Test-Retest Stability

In clinically stable, chronic stroke patients evaluated across a 7- to 10-day test-retest interval without therapeutic interventions:

• The test-retest reliability coefficient is ICC = 0.95 (95% CI: 0.91–0.97).
• The Standard Error of Measurement (SEM) is approximately 0.45 to 0.58 points on the 10-point scale.
• The Smallest Detectable Change (SDC) or Minimal Detectable Change at the 95% confidence level (MDC₉₅) is approximately 1.25 to 1.60 points. Clinically, an individual patient must demonstrate a shift of at least 2 full points on the SULCS to ensure that the observed recovery reflects genuine functional progress rather than measurement error.

9. Factor Analysis and Structural Dimensionality

The structural dimensionality of the SULCS has been evaluated using both parametric exploratory/confirmatory factor analyses (EFA/CFA) and non-parametric Item Response Theory (IRT) procedures.

Mokken Automated Item Selection Procedure (AISP)

The dimensionality of the 10 SULCS tasks was evaluated using the Automated Item Selection Procedure (AISP) in Mokken scale analysis. AISP sequentially partitions items into mutually exclusive, unidimensional scales based on increasing lower bound thresholds of scalability (c ranging from 0.30 to 0.55):

• At all incremental threshold steps (from c = 0.30 up to c = 0.55), all 10 items clustered cleanly into a single, dominant scale without shedding or secondary factor formation.
• No item demonstrated negative conditional covariances or violated invariant item ordering (IIO), confirming strict unidimensionality.

Confirmatory Factor Analysis (CFA) with Categorical Estimators

When evaluated within a classical structural equation modeling framework using Robust Diagonally Weighted Least Squares (DWLS or WLSMV) estimators designed for dichotomous indicators, a single-factor latent structure demonstrated good fit to the data:

• Comparative Fit Index (CFI): 0.991 (exceeding the standard ≥ 0.95 threshold).
• Tucker-Lewis Index (TLI): 0.988 (exceeding the standard ≥ 0.95 threshold).
• Root Mean Square Error of Approximation (RMSEA): 0.052 (90% CI: 0.024–0.078), confirming low residual error.
• Standardized Factor Loadings (λ): Factor loadings on the single latent “Upper Limb Capacity” construct were uniformly high across all 10 items, ranging from λ = 0.82 to λ = 0.97 (all p < 0.001).

10. Instrument / Measurement Tool

The Stroke Upper Limb Capacity Scale is a standardized clinician-administered performance assessment. Its implementation specifications are detailed below:

• Instrument Type: Standardized clinician-rated performance test based on non-parametric item response theory.
• Target Population: Adult and elderly individuals experiencing upper limb hemiparesis secondary to ischemic or hemorrhagic stroke, traumatic brain injury, or related central nervous system pathologies.
• Number of Items: Exactly 10 hierarchically ordered functional tasks.
• Scoring Format: Binary / Dichotomous per item:
  • 0 = Fail: The patient cannot complete the task, requires physical assistance, fails to satisfy the specified posture or execution criteria, or cannot complete the action within the designated time limit.
  • 1 = Pass: The patient successfully executes the motor action strictly according to standardized biomechanical and temporal criteria without non-permitted compensatory patterns.
• Total Score Range: 0 to 10 points. Higher scores indicate greater functional upper limb capacity.
• Testing Materials Required: A standard testing table and chair without armrests; standardized objects including a drinking cup/mug, a tennis ball or equivalent sphere, small wooden blocks, a coin/disc, and a small peg/marble.
• Start and Stop (Basal & Ceiling) Administration Rules:
  • Starting Item: Administration typically begins at intermediate Task 4 or Task 7, depending on the clinician’s initial clinical observation. Alternatively, testing begins sequentially at Task 1.
  • Stop Rule (Ceiling): Testing is immediately terminated as soon as the patient fails two consecutive tasks of increasing difficulty. All subsequent, more difficult tasks are automatically scored 0.
  • Start Rule (Basal): If a patient fails the initial intermediate task selected, testing proceeds downward in order of decreasing difficulty until the patient passes two consecutive tasks. All unadministered, easier tasks below that basal level are automatically scored 1.
• Administration Time: Typically 6 to 10 minutes when using start-and-stop rules, compared to 20 to 45 minutes for non-hierarchical arm tests.

11. Permissions, Fee, and Test Year

The Stroke Upper Limb Capacity Scale was formally published in 2011 by Dr. Leo D. Roorda and colleagues. The scale was established as an open-access clinical measurement tool to promote standardization in global neurorehabilitation research and clinical practice.

• Licensing and Fees: The SULCS is free of charge for non-commercial academic research and clinical practice. It does not require royalty payments or licensing fees when used in hospital or outpatient rehabilitation settings.
• Permissions: The original assessment protocol, scoring sheet, and administration guidelines were published in peer-reviewed literature. Researchers and clinicians wishing to incorporate the SULCS into commercial electronic medical record (EMR) software, proprietary clinical trials, or translated published manuals should request formal permission from the primary corresponding author (Dr. Leo D. Roorda) or the copyright holder (American Congress of Rehabilitation Medicine / Elsevier Inc.).

12. References

• Houwink, A., Roorda, L. D., Smits, W., Molenaar, I. W., & Geurts, A. C. (2011). Measuring upper limb capacity in patients after stroke: Reliability and validity of the Stroke Upper Limb Capacity Scale. Archives of Physical Medicine and Rehabilitation, 92(12), 2004–2012. https://doi.org/10.1016/j.apmr.2011.06.028
• Lyle, R. C. (1981). A performance test for assessment of upper limb function in physical rehabilitation treatment and research. International Journal of Rehabilitation Research, 4(4), 483–492. https://doi.org/10.1097/00004356-198112000-00001
• Mokken, R. J. (1971). A theory and procedure of scale analysis: With applications in political research. Walter de Gruyter. https://doi.org/10.1515/9783110813203
• Roorda, L. D., Houwink, A., Smits, W., Molenaar, I. W., & Geurts, A. C. (2011). Stroke Upper Limb Capacity Scale: A new measure for stroke patients based on the International Classification of Functioning, Disability and Health. Archives of Physical Medicine and Rehabilitation, 92(12), 2004–2012. https://doi.org/10.1016/j.apmr.2011.06.028
• Sijtsma, K., & Molenaar, I. W. (2002). Introduction to nonparametric item response theory. SAGE Publications. https://doi.org/10.4135/9781412984676
• van der Lee, J. H., Beckerman, H., Lankhorst, G. J., & Bouter, L. M. (2001). The responsiveness of the Action Research Arm Test and the Fugl-Meyer Assessment scale in severe stroke patients. Clinical Rehabilitation, 15(5), 528–536. https://doi.org/10.1191/026921501680425252
• World Health Organization. (2001). International Classification of Functioning, Disability and Health: ICF. World Health Organization. https://apps.who.int/iris/handle/10665/42407

13. Items of the Scale