Anxiety disrupts sleep, and disrupted sleep makes anxiety worse. That much is obvious. What’s less obvious is that once insomnia has been running for a few months it develops its own maintaining mechanisms, and treating the anxiety alone frequently doesn’t fix it.
The two patterns
Difficulty falling asleep is the classic anxiety picture. You lie down, the day’s interruptions stop, and the mind starts. Common in generalized anxiety, and worry postponement helps here more than anywhere else.
Waking at three or four and not getting back is more characteristic of depression, but occurs in anxiety too — often driven by alcohol, by a cortisol rise, or by nocturnal panic.
Why sleep hygiene mostly doesn’t work
The standard advice — dark room, no screens, consistent bedtime, no caffeine after two — is reasonable and isn’t a treatment. Trials consistently find sleep hygiene alone performs poorly for established insomnia.
Worse, for anxious people it can become another performance to fail at. Optimising your sleep environment and then lying there monitoring whether it worked is a recipe for exactly the arousal you’re trying to avoid.
What does work
CBT-I — cognitive behavioural therapy for insomnia — is first-line, outperforms medication over the long term, and is genuinely different from sleep hygiene. It has two components people find counterintuitive and slightly brutal.
Stimulus control. The bed is for sleep only. If you’re awake more than about fifteen or twenty minutes, get up, go elsewhere, do something quiet and dull, and return only when sleepy. Repeat as many times as necessary. This breaks the association between bed and lying awake — which, after months of insomnia, is a genuinely conditioned response.
Sleep restriction. The hard one. You temporarily reduce time in bed to roughly the amount you’re actually sleeping, which builds sleep pressure and consolidates fragmented sleep. The first week is unpleasant and it’s the component that does the most work.
Plus the cognitive part, which for anxious people is often the crux: the catastrophic beliefs about not sleeping. I will be useless tomorrow. I can’t function on five hours. This will make me ill. Those beliefs create the arousal that prevents sleep, and testing them directly matters.
The most useful reframe
Trying to sleep is the one thing guaranteed to prevent it. Sleep isn’t a behaviour you can perform; it arrives when you stop doing anything, including trying.
So the target of CBT-I isn’t sleeping more. It’s being all right about being awake — at which point the monitoring stops, arousal drops, and sleep tends to sort itself out.
On sleeping tablets
Effective short-term, with tolerance and dependence risks, and evidence for CBT-I outperforming them over time. Benzodiazepines and z-drugs are best kept brief. If you’re on them long-term, coming off is worth doing alongside CBT-I and with medical supervision rather than alone.
Also worth checking alcohol, which people use for sleep and which reliably wrecks the second half of the night.