What "more durable" actually means in the numbers
The clearest comparison comes from a network meta-analysis, a statistical method that combines many trials and lets treatments be compared even when they were not tested head-to-head in the same study. Pooling 13 randomized trials of adults with chronic insomnia who were not already on hypnotic medication, it found that people who began with CBT-I were more likely to be in long-term remission than those who began with pharmacotherapy, with high certainty in that result [2][3].
The size of the gap is easiest to grasp as remission rates. Against an average long-term remission rate of 28% in the groups that started on medication, starting with CBT-I corresponded to a remission rate of about 41% [2][3]. In plain terms, roughly four in ten CBT-I starters versus roughly three in ten medication starters were doing well at follow-up. That is a meaningful edge, not a cure for everyone.
"Long-term" here has a specific meaning: the longest follow-up available in each trial, ranging from about 3 to 12 months, with a median of roughly 24 weeks [2][3]. So the durability being measured is measured in months, not years.
Why a talking therapy might outlast a pill
CBT-I targets the thoughts and behaviors that keep insomnia going, rather than sedating sleep directly. One case report describes a man whose sleep problems appeared to be worsened by growing tolerance to his sleeping pills; the therapy was aimed at those underlying cognitions and behaviors, and his sleep quality improved over the long term [4]. Another case describes a woman with more than ten years of poor sleep who had tried various insomnia medications with minimal improvement and adverse reactions, and who was referred for a multi-week course of virtual CBT-I; that report emphasizes CBT-I's long-term benefits and its ability to reduce reliance on medication [5].
These are individual cases, so they illustrate a mechanism rather than prove it. The trial evidence is what establishes the pattern: CBT-I produced better long-term remission and fewer dropouts than pharmacotherapy, and short-term outcomes also favored CBT-I except for total sleep time [2][3]. In other words, the therapy's advantage is not that it makes people sleep longer right away, but that more people stay better.
Improvement can fade, and depression predicts who fades most
Durability does not mean a straight line. In a four-year follow-up of 1,022 patients treated with a mobile-based CBT-I program combined with medication, sleep quality, anxiety, depression, and somatic symptom scores improved markedly during the first 12 months, then showed mild relapse afterward, while daytime sleepiness stayed largely stable [1]. That pattern is consistent with the trial finding that many people are not in full remission even after treatment [2].
Who relapses matters too. In that same follow-up, higher baseline depression scores predicted poorer outcomes across sleep, mood, and somatic domains, while baseline anxiety added only modest independent predictive value once depression was accounted for; older age was associated with somewhat better treatment response [1]. The practical implication the authors draw is routine depression screening to guide follow-up for people with chronic insomnia [1]. Because this cohort received CBT-I plus medication together, it speaks to how durable combined treatment is over years, not to a clean head-to-head between therapy and pills.
Is combining therapy and medication the best of both?
Combination treatment looked somewhat better than medication alone, but the evidence there was weaker, and there was no clear difference between CBT-I alone and the combination [2][3]. The trial authors concluded that combination therapy may beat medication alone but is unlikely to be worth the extra burden over CBT-I by itself [2][3].
So the practical answer for someone choosing a first treatment is to favor CBT-I, while recognizing that the advantage is a moderate shift in odds rather than a guarantee, that follow-up in these trials is measured in months, and that depression is a useful signal for who may need closer monitoring [1][2].
Sources used in this answer
Long-term follow-up of insomnia patients: symptom trajectories and predictors of treatment outcomes in a mobile-based CBT-I program.
In a four-year follow-up of 1,022 patients given mobile-based CBT-I plus medication, sleep, mood, and somatic symptoms improved markedly in the first 12 months then mildly relapsed, and higher baseline depression predicted poorer long-term outcomes.
Initial treatment choices for long‐term remission of chronic insomnia disorder in adults: a systematic review and network meta‐analysis
In a network meta-analysis of 13 trials (823 participants), starting with CBT-I produced higher long-term remission than pharmacotherapy (odds ratio 1.82), with estimated remission rates of 41% versus 28%, and fewer dropouts.
Initial treatment choices for long term remission of insomnia disorder in adults: a systematic review and network meta-analysis
The same network meta-analysis reported CBT-I was more beneficial than pharmacotherapy for long-term remission with high certainty, while combination therapy showed only weaker evidence of benefit over pharmacotherapy and no clear difference from CBT-I alone.
Cognitive behavioral therapy for insomnia (CBT-I) for an adult with multiple sclerosis
A case report of a man with multiple sclerosis describes CBT-I targeting cognitions and behaviors, with improved long-term sleep quality, in a context where growing tolerance to sleeping pills was implicated.
Cognitive Behavioral Therapy for Insomnia
A case report of a 62-year-old woman with over ten years of insomnia and minimal benefit from prior medications describes long-term benefits of virtual CBT-I, including reduced reliance on pharmacotherapy.