A964. ROUTINE CLINICAL OUTSOME ASSESSMENT OF THE SHOULDER IS VALID USING INERTIA SENSOR BASED MOTION ANALYSIS.

Abstract

Introduction: In clinical orthopaedics questionnaire based outcome scores such as the DASH shoulder score suffer from a ceiling effect, subjectivity and the dominance of pain perception over functional capacity. As a result it has becomes increasingly difficult to clinically validate medical innovations in therapy or implants and to account for rising patient demands. Thus, objective functional information needs to be added to routine clinical assessment. Motion analysis with opto-electronic systems, force plates or EMG is a powerful research tool but lab-based, too expensive and time consuming for routine clinical use. Inertia sensor based motion analysis (IMA) can produce objective motion parameters while being faster, cheaper and easier to operate. In this study a simple IMA shoulder test is defined and

1. its reliability tested,

2. its diagnostic power to distinguish healthy from pathological shoulders is measured and

3. it is validated against gold standard clinical scores.

Methods: An inertia sensor (41x63x24mm3, 39g) comprising a triaxial accelerometer (±5g) and a triaxial gyroscope (±300°/sec) was taped onto the humerus in a standardised position. One-hundred healthy subjects without shoulder complaints (40.6 ±15.7yrs) and 40 patients (55.4 ±12.7yrs) with confirmed unilateral shoulder pathology (29 subacromial impingement, 9 rotator cuff pathology, 2 other) were measured. Two motion tasks (‘hand behind the head’ and ‘hand to the back’) based on the Simple Shoulder Test (SST) were performed on both shoulders (three repetitions at self selected speed). Motion parameters were calculated as the surface area described by combing two angular rate signals of independent axes (ARS) or by combing the angular rate and the acceleration of a single axis (COMP score). The relative asymmetry between two sides was scored.

Results: The test produced high intra-(r2≥0.88) and inter-observer reliability (r2≥0.82). Healthy subjects scored a mean asymmetry of 9.6% (ARS) and 14.6% (Comp). Patients with shoulder complaints showed > 3× higher asymmetry (ARS: 34.1%, Comp: 42.7%) than the healthy controls (p< 0.01). Using thresholds (ARS: 16%, Comp 27%) healthy and pathological subjects could be distinguished with high diagnostic sensitivity (e.g. ARS: 97.5% [CI: 85.3–99.9%]) and specificity (e.g. COMP: 85.5% [CI: 76.1–91.1%]). Both asymmetry scores were strongly intercorrelated (r2=0.76) as were the clinical scores (r2=0.62, DASH-SST). Asymmetry and clinical scores were hardly correlated (r2< 0.14).

Discussion: The IMA shoulder test and asymmetry scores showed high reliability meeting or exceeding common clinical scores. With a fast assessment of a simple ADL tasks (test duration < 60s) it was possible to provide diagnostic power at clinically usable level making routine clinical application feasible even by nonspecialist personnel. Weak correlations with the clinical scores show that the new test adds an objective functional dimension to outcome assessment which may have the potential to differentiate new treatments or implants required to trigger new therapeutic innovation cycles. Similar motion tests and parameters could also serve lower extremity outcome assessment.