Background: After shoulder surgery for joint stabilization, patients often report that shoulder function improves and positive signs in the ‘apprehension test’ disappear. However, objective validation of the outcomes of shoulder surgery has never been provided. We inquired first about the characteristics of arm movements in healthy individuals and found that in the literature that healthy 2D motion of the arm (e.g., movements performed on table) tend to be smooth and follow shortest amplitude paths with symmetric and unimodal tangential velocity profiles (Minimum Jerk model; Flash and Hogan 1985). In this study, we assumed that such smoothness criterion could be used as an objective indicator of healthy arm movements also in 3D, and thus, we compared the motor outcome before and after different but common surgical procedures for shoulder stabilization (arthroscopy versus open surgery).

Methods: Data were obtained from 3 consecutive point-to-point arm movement trials carried out in each of 3 speed conditions (fast, preferred, slow) and 4 different targets locations towards one central target above the head (Speeds and Movement Directions were repeated measures while Groups were the between-subjects factor). Trials were collected from 14 healthy control subjects (group C), 11 patients before surgery (group B), 3 patients after arthroscopy for stabilization of the shoulder (group A) and 10 patients that underwent open surgery (group D). 3D data were captured by a motion tracking system at a rate of 100 Hz from reflective markers attached to the right arm (acromion and the distal end of the humerus). The kinematic data were pre-processed using MatLab routines. Statistical analyses were based on the following objective measures of smoothness: Time-to-peak speed (TT P), peak-to-mean amplitude ratio (PAR), speed similarity index (SSI) and number of peaks in the tangential velocity of the arm (NO P). Descriptive statistics and multiple 2-way ANO VAs were carried out using these dependent variables (p< 0.05).

Results: Significant effects of the Group factor were observed in the ANOVAs using TT P, PAR and NOP as dependent variables, but not SSI. Post hoc comparisons showed that Group A differed significantly from all others. Patients in group D did not significantly differed from healthy subjects (group C), but patients before surgery (group B) differed from all others. Notably, patients after arthroscopy were also closer to the maximal smoothness scores predicted by the minimum jerk model than even healthy subjects.

Discussion: The results show that kinematics measures may be used to objectively assess the success of one surgical procedure over another. The maximal smoothness criterion seems to be a sensitive measure describing shoulder performance, and thus, parameters derived from this assumption allow for a discrimination of healthy motion from pathological motion. As it stems from the current study, arthroscopy seems to be the preferable intervention since objective measures of smoothness showed that these patients outperform others after surgery. A test of based on slow movement may enhance these differences among procedures because slow movements may rely more on proprioceptive input.