**Acknowledgements**

*Assistive and Rehabilitation Engineering*

on the decision threshold.

(accuracy percentage > 90%).

evaluations were grouped into two categories that corresponded to incorrect or normal movement execution. The average of the four PTs scores were retained for the analyses. In case of a tie, the trial was discarded. Then, each average of this dichotomy variable (incorrect versus normal) were compared with the evaluation made by the algorithm. It was expected that the result would be different depending

The results showed that the algorithm corresponds in an 88% with the evaluation made by the PTs. There is certain variability in the performance of the assessment from one exercise to another. For example, the best results are obtained for the evaluation of the ROM for hip abduction and the forward and backward sequence

The results of other studies were similar to these showing that it is possible to capture and analyze movement using a low cost system [32]. Consequently, the assessment of different types of movements, both in healthy people and in patients, has a high consistency rate among PTs and the system that is intended to be incorporated into a low-cost rehabilitation platform. A recent study validated the use of the Kinect within a TR system for THA (KiReS) [65]. This system allows PTs to define sets of exercises recorded in front of the Kinect. Successively, the patient executes the movement and his performance is compared to that of the PT. The movements analyzed were flexion, extension and abduction of the left and right hip. The unipodal equilibrium and squat movement executed with the lower left and right limb were also analyzed. The results showed a 91.88% consistency

Our results are consistent with this study although our objective was to compare the simultaneous clinical assessment of several PTs with the system. However, the algorithms proposed in the three preceding studies not only evaluated and analyzed the limbs movement but also the compensatory movements. Patients often use compensatory strategies to facilitate movements [66]. Compensatory movements limit the limb affected functional recovery and can cause pain. This is the reason why a TR system should capture, limit and quantify the compensations. Compensatory movements adopted by patients create altered movement patterns that could limit

The Kinect camera works well especially when the user is facing the device, but the recognition of the skeleton from the top and side is not accurate. Moreover, in one of our studies the patient's clothing made it difficult to recognize the skeleton. The TR could help solve accessibility limitations to rehabilitation services for many patients with THA. In addition, it can help reduce health care costs by allowing patients to empower themselves in their rehabilitation process. However, one of the biggest challenges is to convince the patient of the remote service. One study evaluated the feasibility of introducing a TR program for patients with THA, through a survey. The results indicated that TR in patients with THA is feasible from the perspective of access, feelings and preferences about technology [69]. Another study compared the implementation of a telerehabilitation program at home with a conventional rehabilitation program [70]. The results showed that the patients who got TR reached the same results for the evaluations on functional activities, exercises and patient education. Finally, a study comparing a TR program to a conventional THA program did not find

The systematic review found 19 studies that explained the conventional physical therapy for THA. Commonly, the functional recovery is composed of three stages

between the therapist's performance and that of the patient.

the proper use of the affected limb in ADL [67, 68].

significant differences in the results between the two programs [71].

**22**

**8. Conclusions**

This research has been partially supported by the Consorcio Ecuatoriano para el Desarrollo de Internet Avanzado (CEDIA), Grant CEPRA-XI-2017-2115.
