There seem to be two different definitions of "reverse profile" on this thread. One definition is doing a shallow dive, then an SI, then a deeper dive. The other definition is a dive that gets progressively deeper. Avoiding these profiles does not seem warranted using modern computers. Here's the language from SSI regarding the second definition:
Going deep, then shallow, then deep again will yield unreliable results. Since a dive computer is an actual computer, an old axiom in the computer business is applicable here: "garbage in equals garbage out." In other words, the computer model must be supplied with the correct input to calculate decompression status properly. Since the computer gathers its input using a depth gauge and a watch, if a diver dives improperly the computer will calculate improperly (see Multi-Level Diving). This illustrates the third limitation, which is training.
Multi-level dives, in general, are those where divers spend time at progressively shallower depths during the dive, which allows extended bottom time. Typical multi-level dives are on walls, large coral formations and wrecks where the diver can descend to the deepest depth and work up to the surface.
If divers go to the deepest depth first, and proceed to successively shallower depths, then multi-level diving may be relatively safer. If divers alternate between deep and shallow depths, however, then it is less safe. The reasons why are complex and have to do with the solubility of gases under pressure, but to illustrate in simple terms, imagine a sponge absorbing water. Sponges absorb water in varying amounts, but at some point will be saturated—they will hold no more water. The tissues at increased pressure are much like sponges; they will absorb nitrogen until they are saturated (for that pressure). At reduced pressure, tissues release nitrogen. If they are re-subjected to increased pressure, they will again absorb nitrogen, but the residual nitrogen will affect that process.