You finished a course of CBT-I. For a few months, things were better — you fell asleep faster, woke up less often, and the days felt more manageable. Then, slowly, the old patterns crept back. The racing mind at 2 a.m. The heaviness in the morning. The doubt: did the therapy even work, or was I just having a good stretch?
If that sounds familiar, you are not alone. And it is not your fault. A significant number of people who complete Cognitive Behavioral Therapy for Insomnia experience some degree of relapse, especially after the first year. A major study published in Frontiers in Neuroscience followed over a thousand insomnia patients for up to four years and found a clear answer to the question of who is most likely to slide back — and it is not who most people expect. The single strongest predictor of long-term CBT-I outcomes was not anxiety. It was baseline depression.
That finding reshapes how we should approach insomnia treatment for adults in Saskatchewan who are weighing whether CBT-I is the right path — or trying to understand why it didn’t hold.
TL;DR
A four-year study of 1,022 insomnia patients found that CBT-I produces real gains in the first year — but a mild relapse pattern emerges after month 12. The single strongest predictor of who struggled long-term was baseline depression severity (PHQ-9 score), not anxiety, not insomnia severity. When depression is addressed alongside insomnia, outcomes for both conditions improve substantially. Routine depression screening before CBT-I and planned follow-up after it are the two most clinically important takeaways from this research.
What the New Research Found
Huang and colleagues (2026) tracked 1,022 insomnia patients over a follow-up period of up to 48 months — four full years. Participants received mobile-delivered CBT-I alongside standard pharmacotherapy. The researchers measured sleep quality (PSQI), depression (PHQ-9), anxiety (GAD-7), daytime sleepiness, and somatic symptoms at multiple time points from two weeks out to four years (Huang et al., 2026, https://doi.org/10.3389/fnins.2026.1753131).
The results were encouraging and sobering at the same time. In the first 12 months, patients improved significantly across every domain — sleep quality, mood, anxiety, and physical symptoms. Median PSQI scores dropped from 15 at baseline to around 9–10 by month three, and depression scores fell by roughly half. These are real, meaningful gains.
But after the 12-month mark, a mild relapse pattern emerged. Scores on sleep quality, mood, and somatic symptoms showed small but statistically significant upward trends — symptoms creeping back. This was not a full return to square one, but it was enough to matter clinically.
Here is the headline finding: when the researchers looked at what predicted who would do well over the long haul and who would struggle, baseline PHQ-9 score — the measure of depression severity at the start of treatment — was the single strongest predictor of outcomes across every symptom domain. Higher depression scores at intake predicted worse long-term trajectories for sleep quality (β = 0.091, p = 0.008), daytime sleepiness (β = 0.154, p < 0.001), and somatic symptoms (β = 0.122, p < 0.001). Depression at the front door predicted what happened years down the road.
The study’s conclusion was direct: routine depression screening before CBT-I is essential for risk stratification and follow-up planning.
Why Does Depression Undermine CBT-I (Even When Anxiety Feels Like the Problem)?
Most people who come to us for insomnia treatment identify anxiety as the culprit. And that makes sense — lying awake with a racing mind feels like anxiety. But the Huang et al. data revealed something counterintuitive: once depression was accounted for, baseline anxiety (GAD-7) showed limited independent prognostic value. Anxiety on its own did not reliably predict who would struggle with CBT-I over time. Depression did.
Clinical insight: Depression — not anxiety — was the single strongest predictor of long-term CBT-I outcomes across sleep, mood, and somatic symptoms. This held even after controlling for anxiety severity. (Huang et al., 2026)
This does not mean anxiety is irrelevant to your sleep. It means that for many people, what looks like anxiety-driven insomnia may actually be insomnia shaped by an underlying depressive process — one that drains energy, flattens motivation, and narrows the cognitive bandwidth needed to do the work that CBT-I requires.
CBT-I is not passive. It asks you to keep a sleep diary, restrict your time in bed, get out of bed when you cannot sleep, and challenge your own thought patterns about rest. These are active tasks that demand consistent effort over weeks. Research on CBT-I adherence has found that only about half of patients closely follow prescribed sleep restriction schedules in the first four weeks. When depression is layered on top of insomnia, the executive function and motivation required to stick with these strategies are already compromised. The divided focus that comes with managing a mood disorder alongside a sleep disorder can reduce day-to-day engagement with the treatment itself (Muench et al., 2022, https://pmc.ncbi.nlm.nih.gov/articles/PMC8808745/).
That is why CBT for insomnia not working is rarely about willpower. It is more often about an unaddressed condition making the treatment harder to complete as designed.
What Happens After Month 12? The 12-Month Drift
The trajectory data from Huang et al. paint a clear picture. Gains from CBT-I are real and substantial in the first year. But after month 12, there is a gentle drift back toward symptom return across sleep, mood, and somatic domains — what we might call the 12-month drift.
This is not failure — it is a known pattern, and it is manageable with the right plan.
Think of it like physiotherapy for a knee injury. The initial course of treatment gets you functional again. But if you stop all exercises and never check in with your therapist, there is a reasonable chance the knee will stiffen up. Sleep works the same way. The skills you learn in CBT-I are durable, but they benefit from periodic reinforcement.
That is where booster sessions come in. A booster is not a full repeat of treatment. It is typically one to three sessions, spaced months apart, where you and your therapist review your sleep diary, identify any drift in habits, and recalibrate the strategies that worked the first time. A single session usually runs about 50 minutes and covers three things: what has changed in your sleep since the last check-in, which CBT-I strategies have slipped, and whether your mood warrants closer monitoring. For someone whose insomnia is complicated by depression, boosters may also include a brief PHQ-9 re-screen — because a dip in mood and a dip in sleep quality tend to travel together.
Without proactive follow-up, the mild relapse the study describes often goes unaddressed until the person is back in a full-blown insomnia episode, wondering what went wrong. The answer is usually not that CBT-I failed. The answer is that no one checked in.
What This Means If You’re Starting CBT-I in Saskatchewan
If you have depression alongside insomnia, this research does not mean CBT-I will not help you. In fact, the evidence points in the opposite direction: treating insomnia can significantly improve depression outcomes.
Key finding: Adding CBT-I to antidepressant medication nearly doubled the rate of depression remission — 61.5% in the CBT-I-plus-medication group compared to 33.3% in the medication-only group. Among those who achieved insomnia remission, 83% also achieved depression remission. (Manber et al., 2008)
A pilot study by Manber and colleagues established exactly that (https://pmc.ncbi.nlm.nih.gov/articles/PMC2279754/). A meta-analysis published in npj Digital Medicine confirmed that digital CBT-I produces significant improvements in both insomnia and depressive symptoms, with a moderate effect size for depression (SMD = −0.42) and a large effect for sleep outcomes (SMD = −0.76). Higher treatment adherence strengthened the effect — those who completed more of the program saw greater improvements in depression, anxiety, and sleep (Lee et al., 2023, https://www.nature.com/articles/s41746-023-00800-3).
So the message is not “depression makes CBT-I pointless.” The message is that depression changes what a good treatment plan looks like.
Depression should be assessed and addressed at the same time as insomnia. Walking into CBT-I with unrecognized or untreated depression is the strongest modifiable risk factor for poor long-term outcomes. That does not mean you need to resolve the depression first — it means both conditions should be on the table from the start.
Follow-up beyond the initial treatment block matters. An eight-session course of CBT-I is a starting point, not a finish line. If your baseline depression score is elevated, planned follow-up at three, six, and twelve months can catch early signs of relapse before they become entrenched.
A therapist who can treat both conditions makes a real difference. Sleep hygiene tip sheets and generic sleep apps have their place, but they cannot adapt to the interaction between depression and insomnia in your specific case. Working with a clinician who understands both mood disorders and CBT-I outcomes — and who knows the healthcare landscape in Saskatchewan, including the access challenges that come with rural and northern communities — means the treatment can flex when it needs to. Our CBT-I group therapy program in Saskatchewan (https://sasksleep.com/group-therapy) integrates this kind of assessment from the intake stage.
Good News for Older Adults
One finding from the Huang et al. study that may encourage older readers: older age was associated with better treatment response. Older participants showed lower anxiety and depression scores over time compared to younger participants (β ≈ −0.05 for both GAD-7 and PHQ-9).
Why might this be? The study did not isolate a causal mechanism, but the picture makes intuitive sense. Older adults often have more regular daily routines, which align naturally with the structured scheduling that CBT-I requires. They tend to have fewer competing life demands — no young children, less workplace volatility — which makes it easier to prioritize sleep habits. And there is good evidence that older adults bring strong motivation to health-related behaviour change, especially when the impact of poor sleep on daily functioning is hard to ignore.
If you are an older adult in Saskatchewan considering CBT-I, the data suggest you are well positioned to benefit — and that is genuinely worth knowing.
Should You Screen for Depression Before Starting CBT-I?
Yes — and this is not a formality.
The PHQ-9 is a nine-item questionnaire that takes about two minutes to complete. It is one of the most widely validated screening tools in primary care and mental health settings, asking about the frequency of symptoms like low energy, poor concentration, changes in appetite, and feelings of hopelessness over the past two weeks. Scores range from 0 to 27, with higher scores indicating more severe depression.
In the Huang et al. study, nearly a quarter of participants (23.39%) entered treatment with severe depression (PHQ-9 of 15 or above), and another 20% had moderate depression. These were the patients most likely to show attenuated gains and earlier relapse — and they are exactly the patients who benefit most from a treatment plan that accounts for depression from the outset.
Research by Sandlund and Norell-Clarke has similarly found that among primary care patients with insomnia, baseline depressive symptoms — not baseline insomnia severity — predicted whether patients achieved remission after CBT-I. Those with more severe depression at the start were less likely to fully remit from insomnia (as cited in Michel et al., 2023, https://www.sciencedirect.com/science/article/abs/pii/S0165178123004778).
At STG Health, the PHQ-9 is part of every intake assessment for insomnia. If your score is elevated, we adjust the treatment plan — not by withholding CBT-I, but by building in concurrent mood monitoring, more frequent check-ins, and, when appropriate, coordination with your prescribing physician. If you complete a PHQ-9 before starting treatment and your score comes back in the moderate-to-severe range, the right next step is to bring that to your therapist’s attention directly — it should shape how your CBT-I is structured from the outset, not be set aside until things feel worse.
The CBT-I depression predictor finding is too strong to ignore. Offering insomnia treatment without knowing where someone’s mood stands would be a disservice.
If CBT-I Has Not Held for You, the Answer Is Not “Try Harder”
If you have been through CBT-I and the gains did not last, or if you are wondering whether it will work given what you know about your own mood, the Huang et al. study offers a clear and — we think — hopeful reframing. The problem was probably not that you did not try hard enough. The problem may be that depression was part of the picture and was not adequately addressed alongside the insomnia.
That is fixable. Depression is treatable. Insomnia is treatable. And when both are treated together, the outcomes for each improve. The research is consistent on this point — from the Manber study showing doubled depression remission rates, to the Huang et al. data showing that screening and stratification can guide better long-term care.
You do not need to figure this out alone.
Frequently Asked Questions
Does depression affect how well CBT-I works?
Yes, significantly. A four-year longitudinal study found that baseline depression severity (PHQ-9 score) was the single strongest predictor of long-term CBT-I outcomes — stronger than anxiety severity or insomnia severity. Patients with higher depression at intake showed worse trajectories for sleep quality, daytime sleepiness, and somatic symptoms over time (Huang et al., 2026).
Can CBT-I help with depression, not just sleep?
Yes. Research shows that treating insomnia with CBT-I can significantly improve depression outcomes, particularly when combined with antidepressant medication. One study found that adding CBT-I to medication nearly doubled the rate of depression remission compared to medication alone (Manber et al., 2008).
Why do CBT-I gains sometimes fade after a year?
This is a documented pattern — the 12-month drift. CBT-I produces substantial improvements in the first year, but a mild return of symptoms is common after month 12, particularly for patients who had elevated depression at baseline. Booster sessions and planned follow-up are the most effective way to sustain gains (Huang et al., 2026).
Should I address my depression before starting CBT-I?
You do not need to resolve depression before starting CBT-I — but both should be on the table at the same time. Entering CBT-I with unrecognized or untreated depression is the strongest risk factor for poor long-term outcomes. A good clinician will assess and monitor both from the outset.
References
Huang, X., et al. (2026). Long-term outcomes of mobile-delivered CBT-I in insomnia patients: A four-year follow-up study. Frontiers in Neuroscience. https://doi.org/10.3389/fnins.2026.1753131
Lee, E., et al. (2023). Digital cognitive behavioral therapy for insomnia and comorbid depression and anxiety: A meta-analysis. npj Digital Medicine. https://www.nature.com/articles/s41746-023-00800-3
Manber, R., et al. (2008). Cognitive behavioral therapy for insomnia enhances depression outcome in patients with comorbid major depressive disorder and insomnia. Sleep, 31(4), 489–495. https://pmc.ncbi.nlm.nih.gov/articles/PMC2279754/
Michel, F., et al. (2023). Predictors of CBT-I remission in primary care insomnia: The role of baseline depressive symptoms. Psychiatry Research. https://www.sciencedirect.com/science/article/abs/pii/S0165178123004778
Muench, A., et al. (2022). Treatment adherence in cognitive behavioral therapy for insomnia. Journal of Sleep Research. https://pmc.ncbi.nlm.nih.gov/articles/PMC8808745/
If you are in Saskatchewan and struggling with insomnia — especially if low mood has been part of the picture — get in touch with STG Health (https://sasksleep.com) to discuss whether CBT-I is right for you and what a personalised plan might look like.
