The problem you have here is actually not that uncommon. I'll assume that you're currently using stand-alone thick access points? There are a few ways you can attack this problem.
1) Most of the newer enterprise systems are "light" or "thin" APs with a wireless controller. Most of the controllers will automatically adapt the RF power/channel and use load balancing to try to adjust the coverage area to provide optimal performance.
2) Capacity planning. There are software packages out there, that allow you to model your environment, identify the ceiling height, number of floors, material types of the walls, doors, windows, staircases, etc., and define coverage zones with specific bandwidth and user counts. The software understands the material properties of the various building materials, understands the attenuation values of those materials, and allows you to visualize the actual RF propagation from all of the access points in your actual environment. It will also tell you the exact number of APs required for your needs, and show you where to place them on your floor plan for optimal coverage. You can also manually manipulate the AP placement, should you want to do so.
3) Most enterprise systems have multiple SSIDs. So, you can assign the guest traffic to that SSID, limit the bandwidth amount and give it "low priority" Quality of Service (QoS). Then on your hospital SSID you can give unlimited bandwidth and the "high priority" QoS. This will assure that your clinical/internal data always gets the majority of bandwidth, and only a very specific amount is available for guests.
This was first published in August 2006