1) Surgery-over-IP would be designed within the limitations of the existing infrastructure, which is more than capable of handling high resolution real-time video.
2) Surgery-over-IP would be useful mostly in areas that are unlikely to have reliable high speed internet access (poor countries, battlefields, etc.).
3) Netflix traffic is streamed to the customer using a strategy that tries to minimize the number of hops it has to take (i.e. content delivery network). This reduces congested links for something that would be a point-to-point application like remote surgery.