Consider designing combat with specific goals other than annihilation, design more flexible combats, and prepare fallthrough options
This is a specific example of a very common problem with D&D 5e combats! One of my guiding principles of combat encounter design in 5e is that balancing those encounters is not as mechanically tight as might be preferred-- it's very hard to knowingly hit the "sweet spot" of a challenging, but survivable, encounter, and more PCs introduce more variables that are hard to take into account. Another is that combat is much more brittle for low-level PCs, particularly below level 3.
The usual advice of "include more individual enemies per combat" will bring the action economy imbalance more into line, but it's easy to overshoot and make combat much harder than you'd intended. So my advice is to put your effort into giving yourself more flexibility during the encounter rather than trying to achieve a very precisely balanced encounter in advance. That way you can make it more likely that a combat will have at least the degree of challenge you want while reducing the risk that it will be way too difficult.
1. Goals other than annihilation
It's very common for D&D encounters to be treated as fights to the death for all involved. That's not necessarily a problem, but when the stakes for PCs are death (without convenient access to spells that fix death) it is hard to present a challenging combat without a real risk of PCs dying. If your players are interested in a campaign with that level of challenge and risk then you're all set! But if not, designing combats that can end before one side is totally destroyed can help.
This allows one side of the combat to be overmatched (so you don't have to worry so much about balance) without that imbalance causing them to be in mortal peril. If you overload the enemy group with individual enemies, they can accomplish whatever they need to do and then depart (or the PCs can struggle to survive long enough to accomplish their goal, and then escape). Similarly, even if the enemies are going to fight to the finish, they don't need to kill the PCs. They may want to take prisoners, or send a message, or any number of other things. This approach is basically a way for imperfect balance (or the brittleness of low-level combats) to be kept from unraveling the whole session or game.
2. More flexible combats
This is my most-used option, by far. I design combats with waves of enemies, and if it turns out that one wave was too difficult the later waves don't need to arrive at all. Players, who don't know what you have planned, will never know the difference! If a combat seems too easy, you can send another wave in. This lets you adjust combats in the moment, which again gets around the need for ultra-precise, pre-session planning.
It is worth noting, as Ben Barden pointed out in comments, that this is essentially a gimmick. It works well when used to plan combats that you are especially worried about balancing, but is not a great option to use all of the time. Players will catch on if use waves all the time.
Another spin on this approach is to have conditions for retreats or routs, and to encourage the players to be aware that withdrawing is an option. Enemies don't have to drop to 0 HP before they're out of the fight. Being at 25% of your maximum HP might well convince a creature to flee rather than risk death. PCs, when overmatched, might be well advised to run away and return later with more resources and a better plan.
3. Fallthrough options
It can be very situational, but it may suit your story to have "in case of emergency"-type options ready. A friendly group of soldiers might arrive when the situation starts to look grim for the PCs, evening the odds or even effectively ending the combat. It is important to plan for an intrusion like this in advance, at least as a sketch-- suggesting to your players that you'll get them out of tough combats can introduce awkward elements to the game.
When I use this option (which is rare), I like to add a mini-goal to make it seem more like the PCs are earning their salvation. Something like the PCs catching sight of the banner of the approaching soldiers, and informing them that help will arrive in four turns. Now they don't have to win the combat (however that's defined), but instead they need to survive for four turns. It can be a very exciting situation: can they do it?
Honorable mentions
These aren't quite as on-topic as my main three suggestions, but they can also be helpful:
Put more effort into planning adventuring days rather than individual encounters.
This allows you to include a mix of encounter difficulties, and you can adjust subsequent encounters based on how your players have fared. If a combat you expected to be hard turns out to be deadly, the next one they face before their long rest can be an easy combat drawn from your list. If a combat you expected to be hard turns out to be easy, it can be followed up with another hard encounter.
I usually prepare my combats in D&D Beyond, and it's easy to add or subtract enemies from planned encounters and mark them according to how difficult they are. Then I just pick what seems appropriate from "Highway Bandits (Trivial)", "Highway Bandits (Easy)", "Highway Bandits (Hard)", "Highway Bandits (Unreasonably Deadly)", and so on.
Vary tactical ability of enemies
Most stat blocks offer a variety of things an enemy can choose to do, and some of those choices work particularly well between different individuals. An easy example is a trait like Pack Tactics: enemies with that trait will gain a lot from positioning themselves in ways that allow them to use it. A combat that involves most of the enemies having Advantage on their attack rolls is much more dangerous than one where they rarely do. You can play enemies in ways that make the most of their abilities, to the PCs' peril, or the enemies can be less capable and give the PCs an easier time. This is also something that you can adjust during combat, if you feel the need.
Offer training wheels
Adjusting combat rewards can be an effective way to offset a combat that was harder than you'd intended. If you provide extra healing potions after a brutal fight, the effects of the combat being difficult may be less pronounced.
