Time analysis from video footage gives a simple outcome measure of surgical practice against a measured model of use. The added detail that can be produced, over simply recording the usual surgical process data such as tourniquet times, allows us to identify and time the sequence of surgical procedures as stages, to describe issues, and the identification of idiosyncratic behaviours for review and comparison.

Makoplasty (Mako surgical corp. FL, US) partial knee operation times were compared using this technique with those from the Oxford (Biomet, IN, US) partial knee. Three experienced surgeons were observed over 19 Makoplasty procedures ([Consultant 1] 11, [Consultant 2] 5, [Consultant 3] 3) and 2 experienced surgeons over 11 Oxford partial knee procedures ([Consultant 1] 5, [Consultant 2] 6). Times were refined into separate stages that defined the major operative steps of both the Makoplasty and Oxford processes as used by the surgical team at the Glasgow Royal Infirmary, UK. The videos were reviewed for start and stop times for pre-defined actions that would be expected to be observed during each surgical process and from these stage lengths were calculated. For both the Oxford and Mako system 12 comparable stages were identified for comparison and the timing of the various episodes was tested for statistical significance using a Two-Sample, two tail, t-Test. assuming Equal Variances. [Stages: 1. Setup time, 2. Patient on table, 3. Skin incision, 4. Joint Prep, 5. Robot registration (Not in Oxford), 6. Tibial resection, 7. Femoral resection, 8. Trials, 9. Finishing, 10. Cementing and Washout, 11. Closure and dressing, 12. Off table]

The MAKOplasty procedures were on average longer than Oxfords by 27 minutes. This can largely be accounted for in the additional setup stage 4, where in addition to the usual joint preparation taking a couple of minutes approximately 17 minutes were spent in the MAKO cases undertaking image registration and in stage 5 where nearly five minutes were spent in setting up the robot in the MAKO cases.

In conclusion while operative times fell for the Makoplasties across the learning curve they remained elevated once the plateau was reached. It should be remembered that the surgeons had much less experience with the Makoplasty procedure and were undertaking a randomised clinical trial of outcome and hence were not minded to perform the surgery quickly but to the best of their ability and that this may account for some of the elongated surgical time. Indeed other Makoplasty surgeons report an average surgical time of 30–45 minutes per case and 6 cases per day. What is striking is that the additional steps of registration and robot positioning account for a large proportion of the differences and these are mitigated to some extent by quicker trialling of the implant and finishing of the cuts suggesting more confidence in the suitability of the cut surfaces. There is clearly a need to reduce the registration time to produce more cost effective surgeries.

There is an increasing prevalence of haptic devices in many engineering fields, especially in medicine and specifically in surgery. The stereotactic haptic boundaries used in Computer Aided Orthopaedic Surgery Unicomparmental Knee Arthroplasty (CAOS UKA) systems for assistive milling control can lead to an increase in the force required to manipulate the device; this study presented here has seen a several fold increase in peak forces between haptic and non-haptic conditions of a semi-active preoperative image system.

Orthopaedic Arthroplasty surgeons are required to apply forces ranging from large gripping forces to small forces for delicate manipulation of tools and through a large range of postures. There is also a need for surgeons to move around and position themselves to gain line of sight with the object of interest and to operate while wearing additional clothing such as the protective headwear and double gloves. These factors further complicate comparison with other ergonomic studies of other robotics systems. While robotics has been implemented to reduce fatigue in surgery one area of concern in CAOS is localised user muscle fatigue in high volume use.

In order to create the conditions necessary for the generation of fatigue in a realistic user experience, but in the time available for the participants, an extended period of controlled and prolonged cutting and manipulation of the robotic arm was needed. This pragmatic test requirement makes the test conditions slightly artificial but does indicate areas of high potential for fatigue when interacting with the system in high volume instances.

The surgeon-robotic system interaction was captured using 3 dimensional motion analysis and a force transducer embedded in the end effector of the robotic arm and modelled using an existing upper body model in Anybody software. The kinematic and force information allowed initial calculations of the interaction between the user and the Robotic system. Validation of the model was conducted using Electromyography assessment of activity and fatigue. Optimisation of the model sought to create an efficient cutting regime to reduce cutting time with reduced muscle force in an attempt to reduce users discomfort/fatigue while taking into account anthropometric variations in the users and minimising overall energy requirements, burr path length and maximum muscle force.

From the assessment of a small group of three surgeons with experience of the Robotic system there was little to no experience of above normal localised fatigue during small volume use of the system. Observation of these surgeons operating the robot state otherwise with examples of reactions to discomfort. There is also anecdotal evidence that fatigue becomes more problematic in higher volume work loads.