OSE works pretty well and it's easy to house rule up and down. Perhaps poison isn't almost always a death sentence? Or perhaps just start at level 3 or 4?
Alternative:
Let 1st level characters have HP equivalent to what
would have been a level 3 or 4 character before, but reduce the amount of HP they gain from levelling up thereafter. That way, the first level is not likely a death sentence, but death remains a reasonable possibility.
This is what 4e did. The precise numbers may not have been where they should be, but the concept is sound. It allows you to have that cushion without forcing a ton of mechanics (e.g. being level 3-4!) The only missing component, then, is "novice level" rules, which would allow folks who really
want that grim-and-gritty. "surviving to level 3
is your backstory" experience to have their cake, without taking away the cake of the folks who don't want to deal with that.
That combination--higher base HP, slower HP scaling, and novice levels--is the mirror image of what actual OSR games have done to address the "it's really pretty annoying/bothersome to
have to spend months on end waiting to get that character who survived to 3rd level." That is, those games (specifically DCC first, others have since copied them) developed the
funnel dungeon, where you run a bazillion low-level characters through, and the ones that survive become your starting stable of characters to draw upon.
I get how "Your character is the one who survived to level 3" is in fact true and also "Living to level 3 is your character's back story" is also true. A rite of passage and all that.
As noted above, it's a fun challenge for those who desire it, and thus should be supported. But "a fun challenge for those who desire it" is a bad state to force EVERY player to go through--and right now, there are no good options for people who don't want that. Either you get the nice, simple, easy-to-make characters
and really difficult survivability, OR you get survivable but much more complicated characters. That's not great. DCC's funnel solution doesn't work for something like 5e (or really any other version of WotC D&D or PF). A different solution is needed, like the one above. That's not saying it's the only possible solution, but it is one that accepts what the game has become while declaring that those who like what the game used to be cannot be left behind.
That's why I advocate so strongly for developing robust, flexible, well-tested "novice level" rules, preferably ones that can spool levelling out almost indefinitely for those folks who really really like taking multiple years to reach even teen levels, let alone max. I have no use for those rules--but I consider them the second-most-important thing for future D&D game design, second only to making actually robust well-made encounter building tools (like what 4e had, specifically).
Though I suppose on the flip side, OSE/BX characters are so simplified and easy to play compared to 5E making, understanding, and playing a whole new character is fairly simple. But yeah I want them to be Big Damn Heroes with some actual danger.
Much to consider....
This, for me, is why having really really well-made, highly accurate encounter-building rules is so important. When you have such rules, you KNOW (up to the reasonable variation of statistics) that what you're doing is going to push the players to their limits, if that's what you want it to do. You can KNOW that, if you
want a fight to be a cakewalk, it should in fact be a cakewalk
unless the players do something to make things worse. Etc.
At that point, the vagaries of the dice are no longer the driving factor in encounter difficulty. Instead, it is player skill and DM skill that are the driving factors. Player skill because good plans lead to much better results, and bad plans lead to much worse results. DM skill because knowing how to push your players hard
but not too hard is something only a living, breathing DM can do. It's something only a human with sensitivity and discernment can do. The rules can, in fact, help you to make the combat be as much what you want it to be as possible, but it's on you to actually DO that, to actually make it come to life and be scary or relaxed as
you see fit.
Well-made tools make it easier to make beautiful finished work. They cannot replace skill--but excellent tools in the hands of an experienced user make the final result much, much better. Shoddy, flawed tools can still be brute forced into producing acceptable results, but they will be working against the maker, not with them. And while good tools in inexperienced hands will still produce rough results, the best tools include guidance for their use--and DMing tools are no exception.
We can, we should, we
must demand better of the tools and teaching we impart to our DMs.