Repair tissue healing after rotator cuff repair remains a significant clinical problem, and excessive shoulder activity after surgical repair is believed to contribute to re-tears. In contrast, small animal studies have demonstrated that complete removal of activity impairs tendon healing and have advocated for an “appropriate” level of activity, but in humans the appropriate amount of shoulder activity to enhance healing is not known. As an initial step toward understanding the relationship between postoperative shoulder activity and repair tissue healing, the objectives of this study were to assess the precision, accuracy, and feasibility of a wrist-worn triaxial accelerometer for measuring shoulder activity. Following assessments of precision (±0.002 g) and accuracy (±0.006 g), feasibility was assessed by measuring 1 week of shoulder activity in 14 rotator cuff repair patients and 8 control subjects. Shoulder activity was reported in terms of volume (mean acceleration, activity count, mean activity index, active time) and intensity (intensity gradient). Patients had significantly less volume (p ≤ .03) and intensity (p = .01) than controls. Time post-surgery was significantly associated with the volume (p ≤ .05 for mean acceleration, activity count, and mean activity index) and intensity (p = .03) of shoulder activity, but not active time (p = .08). These findings indicate this approach has the accuracy and precision necessary to continuously monitor shoulder activity with a wrist-worn sensor. The preliminary data demonstrate the ability to discriminate between healthy control subjects and patients recovering from rotator cuff repair and provide support for using a wearable sensor to monitor changes over time in shoulder activity.