VALIDATING INERTIA SENSOR BASED MOTION ANALYSIS AS A CLINICAL OUTCOME MEASURE FOR SHOULDER FUNCTION

Abstract

Introduction: Clinical outcome scores such as the DASH shoulder score suffer from subjectivity, a ceiling effect and pain dominance masking functional changes which shall be assessable to address rising patient demands and improve the clinical validation of modern therapeutic improvements. Lab based motion analysis may provide such data but it is too costly, time consuming and complex for routine clinical follow-up. Inertia sensor based motion analysis (IMA) can produce objective movement parameters while being fast, cheap and easy to operate.

In this study, a simple and clinical feasible inertia sensor based motion analysis (IMA) shoulder test is defined and tested for its

1. reliability,

2. diagnostic power to recognize pathological movement and

3. validity against gold standard clinical scores.

Methods: An inertia sensor (41x63x24mm, 39g) comprising 3D accelerometers (±5g) and 3D gyroscopes (±300°/sec) was taped onto the humerus in a standardised position. Healthy subjects (n=100, 40.6 ±15.7yrs) and 50 patients (55.6 ±12.7yrs, m/f 17/33) with confirmed unilateral shoulder pathology (39 subacromial impingement, 11 other) were measured. Two motion tasks (hand-behind-head, hand-to-back) were performed on both shoulders. Using automated algorithms, a simple motion parameter was calculated by adding the peak-to-peak angular rates per axis. The relative asymmetry between both shoulder sides was scored (healthy control within subject). Patients were also assessed using the DASH score and the Simple Shoulder Test (SST).

Results: The test produced high intra- (r2=0.90) and inter-observer reliability (r2=0.83). Asymmetry was > 3 times higher in patients (36.3%) than healthy controls (9.6%, p< 0.001). Using a threshold (> 16% asymmetry) healthy and pathological subjects could be distinguished with high diagnostic sensitivity (98.0%) and specificity (81.0%). The non-affected shoulders of the patient group did not differ from the shoulder of the healthy group (p=0.18). Sub-group analysis comparing the 30 best performing healthy to the 10 highest asymmetry pathological shoulders also revealed sign. lower range of motion, shorter motion path and longer cycle times (p< 0.01). Visual signal analysis exposed specific motion patterns (e.g. healthy: overshooting at point of task achievement, pathological: drift or tremble at rest position). IMA asymmetry was only weakly correlated with DASH or SST (r2< 0.25).

Discussion: The IMA shoulder test and asymmetry score showed high reliability and diagnostic power meeting or exceeding common clinical scores. The fast assessment (t< 60s) of a simple motion tasks makes it suitable for routine clinical follow-up to supplement classic scores. Weak correlations with DASH and SST show that the test adds an objective functional dimension to outcome assessment.