Abstract
Introduction
Measured outcomes from knee joint arthroplasty (TKA) have primarily focused on surgeon-directed criteria, such as alignment, range of motion measured in the clinic, and implant durability, rather than on functional outcomes. There is strong evidence that subjective reporting by patients fails to capture objective real-life function.1,2 We believe that the recent emphasis on clinical outcomes desired by the patient, as well as the need to demonstrate value, requires a new approach to patient outcomes that directly monitors ambulatory activity after surgery.
We have developed and tested a system that: 1) autonomously identifies patients who are not progressing well in their recovery from TKA surgery; 2) characterizes patient activity profiles; 3) automatically alerts health care providers of patients who should be seen for additional follow-up. We anticipate that such a system could decrease secondary procedures such as manipulation under anesthesia (MUA) and reduce hospital re-admission rates thereby resulting in significant cost savings to the patient, the care providers, and insurers.
Methods
The components of the system include: 1) A sensor package that is mounted correctly in relation to the knee joint (Figure 1a) and is suitable for long term use; 2) An application that runs under the Android operating system to communicate with the sensor and to gather subjective information (pain, satisfaction, perceived stability etc. together with a photograph of the surgical site (Figure 1b); 3) Software to upload the data from the phone to a remote server; 4) An analysis and reporting package that generates, among other metrics, a profile describing the patient's activity throughout the day, trends in the recovery process, and alerts for abnormal findings (Figure 1c). The system was pilot tested on 12 patients (7 females) who underwent TKA. Complete days of data collection were scheduled for each patient every two weeks until 12 weeks, starting during the second week after surgery.
Results
Patients tolerated the system well and datasets of up to 13 hours long were recorded. There was a considerable variation between patients in the use of the prosthetic knee joint at a given time point after surgery. At 6 weeks post-surgery, for example, some relatively inactive subjects had less than 50 excursions per hour while active subjects exhibited more than 750 excursions per hour. It was notable that, in activities of daily living, subjects rarely used the extremes of the flexion range that had been measured during post-operative clinic visits. Examples of activity recognition during free-living will be presented.
Discussion
A remote knee monitoring system has been designed and successfully tested in an outpatient setting. The system has revealed discrepancies between knee function measured during clinic visits and that measured remotely during free living. Remote monitoring after orthopaedic procedures adds an important new dimension to the assessment of patient outcome.
Acknowledgments
This work was supported by grants from the Washington Research Foundation and The Wallace H. Coulter Translational Partnership at the University of Washington.
