Does treating insomnia improve depression and anxiety? Yes

The patient who did everything and still says nothing changed

You know the visit. Third follow-up, labs better than March, every supplement taken, everything asked of them done. Then: "I don't feel any different." Most of us reach downstream. Another adaptogen, a different amino acid, eventually a referral.

One arm is usually still untouched, and it sits on the intake form you read past. Sleep is almost never sequenced first, because most of us were trained to read insomnia as a symptom that clears once the real work lands. It is a cousin of the pattern in why patients don't follow treatment plans.

So the question worth settling is whether treating insomnia improves depression and anxiety on its own, or whether sleep merely tracks the mood picture.

One boundary before the evidence, because it governs everything below. What follows describes what moves across populations in trials. It is not a method for diagnosing or treating an individual patient's mood disorder. A patient with significant mood symptoms stays under your medical management, and behavioral sleep support sits alongside mental health care rather than in place of it.

Why fragmented sleep keeps yesterday's emotional load in play

During REM sleep, when most vivid dreaming happens, the brain strips the emotional charge off the previous day's memories while keeping the memories themselves. Van der Helm and colleagues showed amygdala reactivity to emotional images fell overnight in a group that slept and rose in a group kept awake, tracking REM physiology (1).

Wassing and colleagues found the boundary condition: that adaptation ran only when REM was consolidated, in long unbroken stretches. Restless, fragmented REM impeded it (2).

The clinical point sits in that word "fragmented," because fragmentation, not short duration, is the signature of insomnia. A patient can report seven hours in bed and still carry yesterday's emotional load into today, because the nightly reset is not running. That is often why a mood picture stays anchored despite a good plan.

What the trials show when you treat insomnia directly

Three lines of evidence matter. Magnitude first. A 2023 systematic review pooled 22 randomized trials and 10,486 participants and found digital CBT-I improved depression by a standardized mean difference of -0.42 (95% CI -0.56 to -0.28, 21 trials) and anxiety by -0.29 (95% CI -0.40 to -0.19, 18 trials) (3). Across pooled trials, treating insomnia improves depression and anxiety. Small to moderate, real, not a cure.

Direction second. The OASIS trial randomized 3,755 people to ten weeks of digital CBT-I or usual care. Insomnia improved substantially (d = 1.11), paranoia (d = 0.19) and hallucinations (d = 0.24) modestly. A mediation analysis, which tests whether change in one variable accounts for change in another, found the insomnia improvement mediated those gains (4). Be precise about scope: OASIS measured psychotic experiences, not depression. It establishes direction of travel.

Mechanism third. A pathway review published in April 2026 mapped four routes by which CBT-I appears to produce antidepressant effects: HPA-axis and stress reactivity, inflammation, behavioral activation, and cognitive-affective change including emotion regulation and repetitive negative thinking (5).

The review also cites a single pilot study in patients with mild depression, in which CBT-I alone produced a 57% reduction in depression severity and more than 85% no longer depressed at follow-up (5). One study inside a review, not a pooled estimate, and not something to promise a patient with significant symptoms.

"Bidirectional" is true, and it tells you nothing at the visit

Search this question and nearly every result lands in the same place. Sleep and mood are bidirectional, each worsens the other. Accurate, and clinically inert: a true fact with no instruction about what to move first.

Mediation evidence is different in kind. When change in sleep accounts for change in the mental health outcome, you are not looking at two symptoms that travel together. You are looking at an arm that moves and carries something with it, which is the sequencing rule you actually need in the room.

So does treating insomnia improve depression and anxiety, or only sleep?

The honest answer holds two things at once. Yes: treating insomnia improves depression and anxiety, not only sleep scores. And no, not dramatically for everyone. An SMD near -0.4 is a real population shift, not a promise of remission for the patient in front of you.

It is still worth sequencing first: the sleep effect is large (d = 1.11 in OASIS), the treatment is drug-free, and it stacks with the plan you built. Most first moves ask you to add something. This one asks you to change the order.

What treating sleep first changes in your next follow-up

Knowing that treating insomnia improves depression and anxiety changes one thing in the room: the order. When a patient's mood has not moved on a good plan, treat sleep as the next intervention, not the next thing you hope resolves. Say it plainly: "Let's move your sleep first and look at your mood again in six weeks."

Then hold everything else steady, so in six weeks you can read what happened using the mood measure you use.

The same sequencing applies in populations you already treat. In perimenopausal patients the sleep arm is often the one nobody has worked, the argument in behavioral sleep support for menopausal patients.

The uncomfortable move most providers avoid

At the visit where a committed patient says nothing is working, this rule asks you to add no new agent. To them, and to you, it can feel like doing less. It is not. You are moving the arm with the best evidence of carrying the others, then holding the plan still long enough to read the result.

Why holding steady is harder than adding

Providers avoid it because a prescription is a visible act of care and "we are changing your sleep for six weeks" is not, unless something real happens between visits. Sleep behavior change happens across the nights between appointments, not in the room, the problem in how to improve patient adherence between visits.

What not to do when you treat insomnia first

Do not tell a patient with significant mood symptoms that their depression or anxiety is "really just a sleep problem." That is not what the evidence says, it is not yours to conclude from a sleep history, and it can push them away from care they need.

Do not stop, taper or adjust psychiatric medication on the strength of a sleep intervention. That decision belongs with the prescriber, inside the patient's mental health care.

Do not hand over a sleep hygiene sheet and call it treating insomnia. The evidence above is for structured behavioral treatment delivered over weeks, with support.

Where behavioral sleep support stops and mental health care begins

Treating insomnia improves depression and anxiety at a population level. That does not make it a substitute for mental health care, and the line needs to be explicit with the patient and in your notes.

A patient with significant depressive or anxiety symptoms stays under medical management: yours, a psychiatrist's, a therapist's, or all three. Nothing in the sleep evidence changes screening obligations, safety assessment, or referral thresholds. Where risk is present, that pathway runs first and independently.

What behavioral sleep support does is address a modifiable driver upstream of mood, inside your scope, without the patient leaving your care. At Clementine the embedded coaching service line is built that way: a between-visit layer patients work with under your plan, not a referral out. Staffing it in-house or delivering it remotely is covered in what integrating telehealth behavioral support looks like.

A checklist for your next three follow-ups

Treating insomnia improves depression and anxiety by a modest, real margin. Sleep is not the whole answer to a stuck mood picture, but it is the arm with the best evidence of moving, and sequencing it first is a judgment you can defend.

1. 01

   Ask about continuity, not hours

   When mood has not moved on a good plan, ask about sleep continuity, not hours in bed.

2. 02

   Name the sequence out loud

   Tell the patient you are moving sleep first, and set the six-week re-check at that same visit.

3. 03

   Hold the rest of the plan steady

   Keep medications, supplements and referrals unchanged through the window so the result is readable.

4. 04

   Support the nights between visits

   Make sure something carries the behavior change across the nights between appointments, or the plan stays theoretical.

5. 05

   Keep mental health care running

   Behavioral sleep support runs in parallel with the patient's mental health care, never in place of it.