Reframing Insomnia Digital Therapeutics as Platforms for Clinical Chronobiology

Heon-Jeong Lee

doi:10.33069/cim.2025.0069

Digital therapeutics for insomnia have moved beyond the experimental stage and are increasingly recognized as evidence-based treatment options. A meta-analysis of randomized controlled trials reported that fully automated digital cognitive behavioral therapy for insomnia (dCBT-I) yields moderate or greater reductions in insomnia severity compared with control conditions, supporting dCBT-I as an accessible and effective strategy for insomnia management [1]. More recently, a multicenter, single-blind randomized clinical trial of a mobile app–based CBT-I program (Somzz), conducted at three university hospitals in Korea, showed that, compared with an active sleep hygiene education app, Somzz produced greater and sustained improvements in insomnia severity, sleep efficiency, and mental health and quality-of-life outcomes, with low attrition [2]. A pilot study of the prototype digital therapeutic LUIT-K, also conducted in Korea and published in Chronobiology in Medicine, demonstrated that delivering CBT-I content via a mobile application over six weeks was feasible and led to significant improvements in insomnia severity and sleep-related indices in adults with insomnia [3]. Together, these findings indicate that digital therapeutics can preserve the core advantages of CBT-I while substantially increasing its applicability and scalability in real-world clinical practice. In this perspective, we argue that such insomnia-focused digital tools should now be explicitly reframed and developed as clinical chronobiology platforms, rather than being treated merely as delivery vehicles for CBT-I.

Digital therapeutics have also expanded into mood disorders, where circadian rhythm disruption is a core pathogenetic feature. A prospective case–control study showed that a system combining a smartphone application with a wearable activity tracker significantly reduced recurrence of mood episodes over one year, compared with usual care alone, while simultaneously promoting healthier daily behaviors in patients with major depressive disorder and bipolar disorder [4]. In this program, activity data from a wearable device and daily self-reports were used to build machine learning models that predicted mood episode risk, and risk information was fed back to patients to promote circadian-related lifestyle adjustments. Subsequent work has demonstrated that wearable-derived sleep and circadian rhythm features can be used by deep learning models to predict mood episodes with high accuracy, and that temporal patterns of sleep and estimated circadian phase exert causal influences on mood symptom trajectories over time [5,6]. These findings illustrate that digital platforms combining continuous monitoring, prediction, and feedback can operationalize circadian concepts as concrete treatment strategies in routine psychiatric care.

Wearable technology has, more broadly, become a central tool in contemporary chronobiological research. A recent review in Chronobiology in Medicine summarized how consumer and research-grade wearable devices can continuously measure movement, heart rate, skin temperature, and light exposure, allowing analyses of circadian phase, amplitude, and stability in real-world environments, and discussed emerging clinical applications of these data [7]. A state-of-the-science report in Sleep systematically described the types and accuracy of data provided by wearables and outlined recommendations for their use in sleep and circadian research, emphasizing both the promise and limitations of digital biomarkers [8]. In psychiatry, a wearable-based monitoring system has been developed that combines physiological and behavioral signals with clinical information to predict and detect self-harm, aggression, and psychiatric instability in inpatients in real time [9]. Collectively, these developments imply that key circadian properties can now be measured in diverse clinical populations outside the laboratory and directly linked to clinical decision-making.

Despite these converging lines of evidence, most insomnia-focused digital therapeutics are still conceptualized and evaluated primarily as “digital CBT-I.” We propose moving beyond this narrow view and instead treating insomnia digital therapeutics as platforms for clinical chronobiology that support real-time assessment and intervention on circadian rhythms in everyday practice.

DIGITAL CBT-I AS IMPLICIT CHRONOTHERAPY

The core elements of CBT-I exert direct effects on circadian and sleep–wake homeostatic mechanisms, even when such mechanisms are not explicitly emphasized in treatment materials. Sleep restriction and fixed wake times stabilize the sleep–wake cycle and strengthen sleep drive, while stimulus control reduces time spent awake in bed and promotes consolidation of nocturnal sleep into a single episode. The combined use of these components consistently yields the greatest clinical effects on insomnia symptoms in both traditional CBT-I trials and digital CBT-I meta-analytic data, and is now supported by mobile app–based randomized clinical trial evidence [1-3]. Many protocols also recommend increasing morning light exposure, reducing bright light and screen use in the evening, and maintaining regular daily routines, thereby manipulating the timing and strength of environmental zeitgebers.

When these strategies are implemented through applications that record bedtimes, wake times, naps, activity levels, and sometimes light exposure with time stamps, they function as structured chronotherapeutic interventions, regardless of whether they are labeled as such. Nevertheless, most current digital therapeutics primarily convert this temporal information into familiar sleep metrics—such as sleep efficiency, total time in bed, and symptom scores—and use it to guide relatively coarse adjustments in sleep restriction or stimulus control. Pragmatic circadian indices, including habitual midsleep time, day-to-day variability in sleep onset and offset, and discrepancies between weekday and weekend sleep–wake patterns, are rarely highlighted to clinicians or patients and are seldom targeted as primary treatment outcomes. In contrast, digital systems developed for mood disorders have placed wearable-derived circadian features and predictive models at the center of their design, using them to drive relapse-prevention feedback over extended follow-up periods [4-6]. These experiences suggest that insomnia digital therapeutics could likewise bring latent circadian information to the foreground and actively use it to personalize treatment and evaluate outcomes.

BEYOND SYMPTOM RELIEF: TOWARD CIRCADIAN HEALTH

A first conceptual shift is to regard circadian alignment and rhythm robustness as explicit therapeutic goals in insomnia treatment, rather than as secondary by-products of symptom improvement. In practice, this would mean that digital therapeutics routinely compute and present practical circadian indices—such as midsleep time, variability in sleep onset and offset, weekday–weekend differences, and the strength of 24-hour rest–activity rhythms—alongside insomnia severity and sleep diary variables [7,8]. In patients with comorbid mood disorders, such indices could help clinicians determine whether improvements in insomnia symptoms are accompanied by stabilization of the underlying circadian system, which is likely to be important for long-term relapse prevention [4-6]. This view is consistent with recent arguments that behavioral medicine principles should be more deliberately integrated into digital chronobiology tools to enhance circadian health [10].

A second shift is from static, session-based protocols to time-sensitive personalization. Because digital therapeutics continuously record behavior, they can adapt not only the content but also the timing of interventions to the patient’s current circadian state. At a simple level, educational modules and behavioral suggestions can be delivered at times of day when the patient is most likely to be awake and receptive, or recommendations regarding light exposure and sleep scheduling can be tailored according to chronotype. More advanced approaches, informed by wearable-based prediction and causal modeling studies, could dynamically adjust the intensity of sleep restriction, daytime activity promotion, social rhythm support, and notification frequency and timing when subtle changes in circadian features signal emerging risk [5,6,8]. When both “what” and “when” are jointly optimized in this way, digital therapeutics become genuinely time-based interventions.

A third shift concerns integration into existing health-care systems across specialties. To function as chronobiology platforms, digital therapeutics must summarize circadian information in formats that clinicians can quickly interpret and apply. Rather than presenting raw actograms or opaque proprietary “sleep scores,” weekly or monthly reports could provide concise overviews of changes in sleep–wake timing, regularity, and rhythm amplitude, together with insomnia severity, mood, and functional status [1-4,7]. In mental health settings, such summaries could guide decisions regarding the timing of pharmacotherapy, psychotherapy sessions, and adjunctive light therapy, and could be combined with crisis-monitoring systems for inpatients at high risk of self-harm or aggression [9]. In primary care and cardiometabolic clinics, circadian reports could be interpreted alongside traditional indicators such as blood pressure, glycemic markers, and weight, thereby elevating daily temporal structure to the status of an additional “vital sign.”

Circadian concepts are also expanding beyond sleep and mood. Recent chrononutrition reviews in Chronobiology in Medicine have synthesized evidence that irregular eating schedules, late-night caloric intake, and disrupted metabolic rhythms contribute to obesity and cardiovascular risk, and that realigning meal timing with endogenous circadian phase can improve metabolic health [11]. These findings suggest that digital platforms that already monitor sleep and activity could eventually be extended to capture meal timing and composition, enabling integrated interventions that simultaneously target sleep, circadian rhythms, and nutrition [7,11].

OUTLOOK

Digital therapeutics for insomnia initially emerged as practical solutions to the limited availability of clinicians trained in CBT-I and the need to reduce reliance on hypnotic medications. Evidence from meta-analytic studies, mobile app–based randomized clinical trials, and early clinical pilots now indicates that these tools can safely improve insomnia symptoms across diverse populations [1-3]. In parallel, research on wearable-integrated mood programs, predictive models, and crisis-monitoring systems suggests that digital platforms can combine continuous monitoring, machine learning, and behaviorally oriented feedback to implement circadian-based management strategies in real-world psychiatric care [4-6,9].

The next challenge is to make this circadian dimension explicit. By incorporating standardized circadian metrics into design, outcome assessment, adaptive algorithms, and clinical interfaces, insomnia-focused digital therapeutics can evolve from disorder-specific tools into general-purpose instruments of clinical chronobiology. Such platforms will not replace sleep laboratories or specialized chronobiology clinics, but they can extend circadian perspectives into primary care, mental health services, and the everyday lives of patients. For a journal devoted to biological rhythms and medicine, this reframing carries both scientific and practical implications. Scientifically, it encourages trial designs that treat circadian indices as key endpoints and allow sufficient follow-up to test whether improvements in alignment mediate long-term benefits in sleep, mood, and physical health [5,6,8]. Clinically, it suggests that smartphones and wearables, when rigorously validated and thoughtfully integrated, can serve not only as channels for delivering CBT-I but also as gateways through which circadian science is translated into scalable, patient-centered care.