If by "well-designed scenario" we mean "scenario designed to be challenging to D&D PCs having given suite of game-changing abilities", then it is practically tautologous that a well-designed scenario will take accoount of those abilities (and from memory this is the line that the 3E DMG runs).
If by "well-designed scenario" we mean "scenario that seems fit within the genre and promises to evoke the genre/thematic experience that high fantasy RPGers are looking for", then I am less persuaded that game-changing abilities can't ruin a well-designed scenario. For instance, negotiating with the god of death, or sneaking past him/her, to rescue a loved one who died before his/her time sounds like a reasonable scenario for high level fantasy PCs, but in a standard D&D game there is a risk that it degenerates into the PCs trying to bust through Hades teleport wards, which in play can tend not to be that epic.
I ran a similar scenario in 1e. Hades had tricked Artemis to sit in his chair of forgetfulness. The PCs got a visit from the head cleric of Artemis and asked to make a run to free her. The PCs had no illusions they were a match for the god of death in his demesne. They felt their only hope lay in stealth; they were somewhat incorrect in that I felt they were other non-confrontational ways to succeed, but stealth was certainly workable.
They felt trying to bust through a major god's defences in his home is likely both quixotic and career-limiting (in the sense that being permanently removed from play ends a career). This sentiment I agreed with.
I think some sort of balance is therefore required, between honouring the game's mechanical tradition and making room for genre/thematic appropriate scenarios. Personally I liked where 4e drew the balance, but I can see that D&Dnext is going back a bit more towards tradition. I don't have a strong sense if it goes as far in that direction as 3E did (which I personally find too far, and hence making good scenarios too hard to make work).
I see [MENTION=20323]Quickleaf[/MENTION]'s suggestions about divination as another way to try and strike the balance. My practical concern about that is that it still rests on a mechanical technique that is not that popular among the trad D&D crowd: simple rolls for stuff that is ancilliary to scenario (opposed Streetwise check, opposed Arcana check) and then complex resolution for the core stuff (the divination example, or searching the village house-to-house). This idea of scoping up or down based not on ingame complexity but at-table significance is a fine approach to RPG mechanics, but I think would be hard to successfully introduce into D&D. (Eg 4e's solo-standard-minion-swarm rules can be seen as a version of it for combat design - scope the monster based on at-table significance rather than ingame complexity - but they are rather contentious.)
As one in the "trad" crowd, scoping up works well for me when there are no consequences for failure. I will typically "scope up" random outdoor encounters when the group is traveling through areas which have encounter levels that pose no effective threat and go to narration of encounters unless the players wish to involve themselves. I'd almost certainly zoom out for followers doing a door-to-door search of a friendly town as well -- getting direction fro the PCs as to strategy and pattern and giving them opportunities to respond to hue and cry or other disturbances of note.
The problem with scoping up at other times is the systems used to determine consequence aren't built to change scale.
I adapted the Pathfinder mass-combat rules and it was well-received by the group as a scaling system. It left a lot of PC choice and consequence available which is why I think it worked. I found variants of skill challenges also accepted for general high-level action though I think the maths on the base system need to be more obvious to casual users.