You've got two ways of proceeding: a) You can try to invalidate the patent. Finding prior art--published, used or sold--dated at least a year before the patent application was filed, Dec. 22, 1994--does that.
If you find it, and it really describes *exactly* what's patented here, the patent's invalid. If you find prior art that is something that would make this patent "obvious" to someone "skilled in the art" of electronic design, the patent's also invalid, but those requirements are subjective; the invalidation is not a given.
You could then ask the Patent Office to formally cancel that patent by filing a request for reexamination based on the new evidence. I wouldn't bother if you find identical art. You might consider it if you find prior art that suggests but is not identical to the patented invention. Depends on your comfort level, confidence, etc. It costs $2k to file, and you'd probably need a patent attorney to handle it.
b) You can use another circuit.
It's important to remember the only thing patented is exactly what's in the claims.
To infringe a claim your device has to include EACH and EVERY element described in the claim; (if you didn't use as many elements you'd have done the same thing with a simpler circuit and would have grounds for patenting an improvement; if you use distinctly different elements, or in a different configuration, your device is novel, and not infringing.)
These guys claim 1) a back-to-back zener AND 2) a MOSFET switch in 3) the positive supply line in each of their five claims.
You could make a different circuit not including all those elements and not infringe, IMHO.
Here's some prior art:
The following appears in Bob Pease's "Troubleshooting Analog Circuits", (c) 1991, ISBN 0-7506-9184-0, p164:
Holy crap. I thought I was so tricky designing this into a customer's low voltage gizmo. Took a couple minutes of "original thought" to realize this circuit. Fortunately, I don't violate this patent as I only use one or zero zener diodes, sometimes put a capacitor between gate and source, and sometimes use an N-chan FET on the negative lead.
Pretty soon, someone will find a current patent for a novel way of making a voltage divider using two resistors in series.