Optimizing alertness in a 24-hour society: the role of personalized digital tools

The invention of artificial light marked a turning point in human history, transforming society by enabling activities that extended beyond natural daylight hours. Coupled with advancements in air travel, these technological breakthroughs ushered in the era of the 24-hour society, altering work, travel, and social patterns. While we all enjoy the benefits of 24-hour services and fast travel, this also introduced a new challenge for our circadian clocks, which control the timing of all physiological processes in our body [1].

Circadian misalignment—where our endogenous circadian clocks become desynchronized from environmental and behavioral time cues—has become increasingly widespread, impacting millions daily [2–4]. Short-term circadian misalignment, such as that during jetlag, results in inadequate sleep, increased sleepiness and fatigue, gastrointestinal and immune disturbances. It affects millions daily, reducing their well-being and productivity, but it disappears after re-entrainment to the destination time zone is achieved. Conversely, shiftwork and, to a lesser degree, social jetlag are examples of chronic circadian misalignment, where the short-term effects are compounded by the long-term consequences of an increased risk of metabolic, cardiovascular, and mental disorders [5]. With ~20 per cent of the workforce in developed countries engaged in shiftwork and an estimated 87 per cent of daytime workers affected by social jetlag [4, 6], there is an urgent need for interventions that can be implemented in day-to-day life.

Laboratory experiments have provided critical insights into the mechanisms of circadian misalignment, demonstrating that light is the key Zeitgeber for the master clock in the Suprachiasmatic Nucleus of the brain [7], while meals and exercise entrain peripheral circadian clocks [8, 9]. Mistiming these time cues relative to our circadian phase results in disruptions of hormones and key autonomic functions, such as temperature regulation and sleep, ultimately reducing alertness and impairing cognitive function. The phase- and dose–response curves for the effects of light on the circadian phase [10–12] have been instrumental in developing and testing interventions for re-aligning the clock, which have proven to be highly effective [13–15].

However, there is a gap between the highly controlled laboratory studies and the real-world implementation of interventions. While the core mechanisms remain unchanged, in the real world, we face numerous factors that make interventions challenging. These include

• inter-individual variability in circadian and sleep physiology, e.g. sensitivity to light [16], chronotypes [6], age [17], and sex [18];

• differences in behavior, including various social and cultural constraints and commitments;

• the ever-changing daily schedules and commitments outside of work; and

• the indefinite number of potential shift combinations and flight itineraries, as well as sleep and light exposure history.

Ideally, this problem requires a personalized solution with real-time feedback about a person’s sleep and circadian phase, so that interventions account for the factors above while also learning and optimizing based on personal history and goals. Twenty years ago, such a solution would have seemed unfeasible. Today, however, advances in circadian and sleep science, mechanism-based modeling, wearable sensors, mobile applications, and machine learning algorithms make this a likely direction for optimizing circadian and sleep health in daily life.

Mechanism-based mathematical models of sleep, circadian rhythms, and alertness provide a critical link between the physiological mechanisms uncovered in the laboratory and the real-world interventions. These models enable the prospective prediction of sleep and circadian dynamics based on physiological mechanisms and have been used to inform the design of interventions. So far, they have been primarily tested for group-level predictions and optimizations [19–21], while individual predictions, especially for shiftwork, are a work in progress. Providing these models with inputs related to a person’s sleep, light, alertness, and behavioral constraints is necessary to enable training and calibration for individuals.

The study by Song et al. [22], in this issue of SLEEP, represents an important step forward in this field. The authors present a mobile app, SleepWake, that utilizes mathematical modeling of sleep–wake cycles and real-time sleep input to improve the alertness of shiftworkers. The key innovation in this study is the continuous updating of sleep recommendations based on an individual’s sleep history as opposed to providing fixed recommendations. In a prospective crossover trial of n = 19 male shiftworkers, they demonstrated that greater adherence to SleepWake recommendations resulted in greater improvements in self-reported alertness.

Interestingly, greater adherence to the recommendations and higher alertness were not associated with a significant increase in total sleep time or sleep quality, whereas lower adherence was associated with a decrease in both. The recommended sleep times were based on the model-predicted sleep, i.e. the times when the alignment between the homeostatic and circadian drives is most conducive to sleep under the shift schedule constraints. This supports the authors’ conclusion that it is the sleep timing, rather than quantity, that enables the improvement in alertness.

The results and the mobile app developed by Song et al. are encouraging, providing a platform for future studies. In particular, what is the effect of recommendations on the objective measures of alertness, such as reaction time on a psychomotor vigilance task (PVT) or working memory? Another question arising from the study is what prevented the low-adherence group from following the recommendations? Understanding this will be crucial for the future adoption of this and similar apps in daily life.

Another mobile app, SleepSync, with a similar aim of improving sleep and alertness of shiftworkers, was earlier reported by Varma et al. [23, 24] In addition to sleep timing recommendations based on the model of arousal dynamics [25], SleepSync provided educational resources on managing lifestyle as a shiftworker; however, unlike SleepWake, it did not account for real-time sleep history. In their crossover trial of n = 13 defence personnel, the team demonstrated that the use of SleepSync sleep recommendations has significantly improved working memory and insomnia scores, but had no effect on reaction time measured using a 3-minute PVT test.

Similarly, an improvement in the insomnia severity index and Sleep Hygiene was reported in a randomized controlled trial in a sample of 58 paramedics using a mobile app, SleepFit [26]. No differences in alertness were reported between the intervention and control groups. Interestingly, the authors reported a high dropout rate (91.4 per cent) for the 3-month follow-up of the study, indicating that the perceived improvements may not have been sufficient to sustain app use.

Overall, peer-reviewed mobile apps for shiftwork so far have focused on sleep recommendations, and it is encouraging to see that they improve multiple, albeit different, cognitive and sleep metrics. Moving forward, it will be essential to validate these findings with objective metrics of alertness and in more diverse populations. A more sensitive, 10-minute PVT test may be required to detect changes, especially with small sample sizes.

Manipulation of light and dark exposure is another important tool that is yet to be utilized in evidence-based shiftwork apps, and it is expected to further improve alertness and reduce circadian misalignment. Personalized light recommendations would highly benefit from knowledge of individuals’ circadian phase, which, so far, cannot be measured with wearable sensors. In the meantime, mathematical models incorporating the dynamic circadian oscillator are a useful tool for predicting circadian phase [19–21].

A future vision for such apps may combine sleep and light interventions with recommendations for meals and exercise times, providing a unified approach to circadian health that targets not just alertness and productivity, but also overall health and well-being. Much more research is needed to understand what these optimal timings are and how the different time cues and circadian clocks interact with each other. Understanding the potential user uptake of the apps is also critical, given the high reported dropout rates.

Funding

None declared.

Disclosure statement

Financial disclosure: The authors have no remaining financial disclosures.

Non-financial disclosure: None.