This study was exempt from institutional review board oversight and a waiver of authorization was granted prior to initiating the study. Records related to providing a HIS device (the Ermi Shoulder Flexionater—Fig. 1) were reviewed for all patients who were prescribed the device between January 1, 2010 and December 31, 2015. Patients were being treated due to a variety of conditions including frozen shoulder, rotator cuff repair, impingement syndrome, biceps tenodesis, superior labral tear from anterior to posterior (SLAP), chondroplasty, microfracture surgery, lysis of adhesions, acromioclavicular joint sprain, distal clavicle excision, subacromial decompression, humerus fracture, mastectomy, and shoulder replacement surgery.
The data collected included: (1) initial ROM; (2) last recorded ROM; (3) the number of days between those measurements; and (4) the number of days between device delivery and the last recorded ROM. The initial recorded ROM was the measured ROM that was taken closest to the time of delivery of the Shoulder Flexionater. Another set of ROM measurements was required when a re-certification of medical necessity occurred (generally on a monthly basis). The last recorded ROM measurement was the most recent measurement available in the physical therapy notes which was generally from the beginning of the last re-certification period. This measurement is most likely not the patient’s final ROM after completing treatment. The inclusion criteria for patient data were: (1) The patient must have 2 sets of ROM measurements for at least one plane of shoulder motion; (2) the initial ROM measurement must be taken within 30 days of device delivery (could be up to 30 days before delivery or 30 days after delivery); (3) the second ROM measurement must be at least 10 days post-device delivery; and (4) there must be at least 14 days between measurements. As an example, a patient could have their initial measurement 4 days before device delivery and their last measurement at 10 days after delivery.
The HIS device was prescribed when patients reached a plateau in their recovery after at least 4 weeks of physical therapy and were unable to meet their treatment goals [20]. These goals were based on injury type, surgical procedure, contralateral shoulder ROM, age and sex of the patient, and preoperative ROM. The device could be configured to stretch in external rotation, abduction, or internal rotation (Fig. 1). A representative of the company set up the device and trained the patient on its use at the time of delivery. The patient never changed the plane of stretch on their own. Generally, the device was setup to stretch in external rotation for approximately one month of treatment and then switched to stretch in abduction by the representative of the company. These are the two planes in which motion is most commonly lost in adhesive capsulitis patients [21]. In some patients, internal rotation was stretched in a subsequent month. In some cases, this protocol was different at the direction of the treating clinician. Patients used the device to stretch their shoulder during three sessions per day. In each session, the patient stretched the shoulder up to the maximum tolerable stretch (end-range of motion) for 10 min and then released the stretch for 10 min. This was followed by another 10 min period of end-range stretch. The patient can apply a torque in small increments of up to 125 Nm as shown in laboratory testing. Patients were instructed to stretch to the point of discomfort but not the point of pain. Therefore, patients would generally apply a torque well under the maximum level that the device can generate. The hydraulic nature of the device allowed for fine control of the stretch by the patient while allowing the patient to feel feedback as the stretch was advanced. The completely patient-controlled nature of the stretch allowed the patient to push to the end range of motion without the fear and muscle guarding that can occur when the stretch is applied by a physical therapist.
Records from 15,133 patients were reviewed. The ROM measurements taken using a goniometer were recorded from physical therapy progress notes. All measurements in a single patient were from the same clinic. A patient could have measurements from one or more planes of shoulder motion: external rotation, abduction, forward flexion, and internal rotation. Available measurements in the clinical notes were passive ROM in the majority of patients, but in rare cases (< 5%) only a complete set of active ROM measurements or a mix of active and passive measurements was available. In mixed cases, the data were only used if the initial measurement was a passive measurement. Since passive ROM is generally larger than active ROM, the patient’s improvement in ROM would most likely be underestimated, and not overestimated. Two or more ROM measurements that met the inclusion criteria for at least one plane of shoulder motion were available for 1871 patients. There were 1727 patients with data for external rotation, 1443 patients with data for abduction, 851 patients with data for forward flexion, and 457 patients with data for internal rotation. Patient populations were different between motions because available recorded measurements varied from patient to patient. Patients who did not have 2 sets of measurements for at least one plane of motion were excluded from the analysis. Measurements for forward flexion were included in the study even though the device did not stretch in that plane of motion because it is an important indicator of successful recovery.
Pre- and post-treatment ROM for external rotation, abduction, forward flexion, and internal rotation were compared using two-sample, equal variance t tests after F tests ensured equal variances. One-way analysis of variance was performed when comparing ROM measurements between age groups.