I fear, though, that the same players who adopt the 15mAD approach will seamlessly adjust. After all, their logic remains the same: there's no cost to the players in doing so, but a potentially nasty consequence for going on at less than 100%.
I have faced such a group, once, though they weren't quite so pathological - they made camp in the wilderness, having taken some wounds and used most of their spells. During the night, they faced a random encounter, meaning they lost a few more hit points and used up some of their spells (and, it being 3e, the Clerics thus used up some of their spell slots from the next day). As a consequence of which, they insisted on staying put for a full day and night, so they didn't have to proceed while wounded and low on spells.
After that, I always made sure they had easy access to a wand of cure light wounds. It's a horrible piece of equipment, and a bad fix for the issue, and something I'm not at all sorry to see go... except that giving them that easy access made the game flow much more smoothly, because I didn't have to constantly fight their wish not to proceed at less than 100%.
I don't think that these players are pathological. They have noted (i) that D&D has resource-gaming elements to it, but that (ii) ingame time is not, by default, a resource, and hence (iii) it costs them nothing to trade ingame time to maximise their access to the resources, like spells and hit points, that really do matter to action resolution.
If you look at a game like Burning Wheel, that tries to build resting time into the system, you notice a couple of things: first, ingame time
is a resource, because instead of spending it healing you can spend in training; second, players have an incentive to engage challenges at less than full health because, due to the way the BW advancement system works (it cares about the actual degree of mechanical challenge you face, but doesn't care whether you win or lose at that challenge), it can be easier to get the experience you need to advance if you are wounded (and hence challenges are, by default, more difficult).
That's a game at the gritty end of the spectrum. Marvel Heroic RP, at the other end, still draws a distinction between Action Scenes and Transition scenes, and there are things you can do in one sort of scene and things you can do in another sort of scene (including different rules for recovery in each sort of scene).
I think it's bad game design simply to rely on players' aesthetic sensibility to give them a reason to treat ingame time as a meaningful resource if nothing in the mechanics makes it so.
Nobody is realistically going to delve half-way though the "Depths of Destruction" only to say "oh wait, we're too beat up, time to head home" then wait a week or more to fully heal, and then head back, having to basically clear the dungeon all over again, and risking losing any possible hidden treasure if it abandoned by it's occupants.
Returning to town generally has a large minimum time requirement and the PCs risk events occurring in their absence to which they can't respond, but once there hp are gained relatively quickly and reliably. Staying in the field generally has a low minimum time requirement and the PCs can be more responsive to local events, but the hp are gained more slowly and less reliably due to possibility of disruptions. On average neither of those should be the "right" choice
What I see in both these cases is that the GM is making ingame time into a resource in a way which is not at all transparent to the players. Like two identical corridors, one to the left and the other to the right, I don't really see meaningful choice here.
Of course particular scenarios might be different - eg the PCs learn from an oracle that they are on a clock. But in the typical Gygaxian AD&D module there is no data that the player can gather via skillful play that will tell them how much time they have, and what the costs are in terms of loot foregone if they rest. Furthermore, the XP system makes quite a difference here: in Gygaxian AD&D, for instance, missing out on loot is a big deal XP-wise and extra reinforcement adversely affect the fight-to-XP ratio, whereas in 2nd ed AD&D missing out on loot isn't as big a deal, and extra combats due to reinforcements are a boost to XP rather than a drag upon XP.
the conditions should change the rate of recovery, and maybe even your condition should change your rate of recovery. So, as a suggestion for the most basic (and gritty) of rules, the party can rest in a poor, average or good environment; in a poor environment you don't recover any hp but you do get spells and such, in an average environment you recover 1hp/level and in a good environment you recover 2hp/level.
Resource recovery needs to work on a unified timeline. It doesn't matter if it takes a minute between fights or a month, the game will only play in a consistent manner if refilling HPs, refilling spells, and refilling other "daily" abilities works at the same rate.
I very much agree with Kinak here. Hit points are a fighter's main "daily" resource. Putting them on a different clock from the casters is a recipe for balance breakdown. It also increases the pressure on the cleric to be a healer, if spells come back more quickly than hit points and hence speed up the overall recovery process (in my experience this was certainly the default in classic D&D play).