Nearly Half of Medical Students Show Insomnia Symptoms During Exams, Study Finds

Nearly one in two medical students sitting their fourth-year examinations at a German university screened positive for clinically relevant insomnia symptoms, according to a new cross-sectional study published in BMC Medical Education. The research, conducted by a team at University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, offers one of the most detailed snapshots yet of how sleep collapses under the pressure of exam season, and it points to a surprising culprit: not the stress of the exams themselves, but the way students think about sleep.

The study surveyed 124 medical students during a fourth-year examination period, a window of intense psychosocial stress that the researchers deliberately chose to disentangle which factors drive insomnia. Students completed a battery of validated instruments: the Insomnia Severity Index (ISI) as the primary outcome, selected items from the Pittsburgh Sleep Quality Index, the Patient Health Questionnaire-9 for depression, the Generalized Anxiety Disorder-2 scale for anxiety, a visual analogue scale for perceived stress, the Dysfunctional Beliefs and Attitudes about Sleep Scale-16 (DBAS-16), and a reduced Morningness–Eveningness Questionnaire to capture chronotype. Academic performance was also recorded, allowing the team to test whether sleep problems left measurable marks on grades.

The headline number is stark: 46.0 percent of participants, 57 of the 124 students surveyed, scored above the clinical threshold of 10 on the Insomnia Severity Index, indicating symptoms serious enough to warrant attention. That figure is far above the rates typically reported in the general adult population and underscores how vulnerable this group is precisely when their cognitive performance matters most. Insomnia, in the clinical sense, is not simply a few bad nights; it involves difficulty falling asleep, staying asleep, or waking too early, combined with daytime impairment, at least several nights a week.

The objective sleep parameters told a consistent story. Students with clinically relevant insomnia symptoms slept an average of 391.0 minutes per night, about 6.5 hours, with a standard deviation of 76.5 minutes, compared with 438.2 minutes, roughly 7.3 hours, among their peers without significant insomnia, a difference the authors report as statistically significant at P < 0.001. More dramatic still was sleep onset latency, the time it takes to fall asleep: affected students took an average of 51.1 minutes to drift off, versus 21.6 minutes for the rest, again with P < 0.001. Sleep efficiency, the proportion of time in bed actually spent asleep, dropped to 82.2 percent in the insomnia group compared with 90.1 percent in the non-insomnia group.

To understand what was actually driving insomnia severity, the researchers built a multivariable linear regression model that weighed state factors, such as current mood and perceived stress, against trait factors, such as chronotype and sleep-related beliefs. The model performed remarkably well, explaining 57.2 percent of the variance in ISI scores, with an adjusted R² of 0.523. In the adjusted analysis, two variables emerged as independent positive correlates of insomnia severity: depressive symptoms measured by the PHQ-9, with a standardized coefficient of 0.30 and P = 0.01, and dysfunctional beliefs about sleep measured by the DBAS-16, with a coefficient of 0.28 and P < 0.001.

Two variables were inversely associated with insomnia severity. Higher sleep efficiency carried a standardized coefficient of −0.30 with P < 0.001, which is expected, since efficient sleepers by definition have fewer insomnia complaints. More intriguing was the finding that better academic performance was independently linked to lower insomnia severity, with a coefficient of −0.15 and P = 0.037. In other words, students who were performing well academically tended to sleep better, even after accounting for mood, stress, and sleep beliefs, while those whose grades were suffering were more likely to lie awake at night.

Perhaps the most consequential negative result concerns perceived stress, anxiety, and chronotype. Despite the fact that the survey was conducted during an examination period, when stress levels would be expected to peak, none of these three variables remained an independent correlate of insomnia severity once depression and sleep beliefs were taken into account. This challenges the intuitive assumption that exam stress alone explains why students cannot sleep. Instead, the findings suggest that the psychological machinery of insomnia operates through more specific channels: low mood and entrenched, maladaptive cognitions about sleep itself.

The DBAS-16 subscale analysis sharpened this picture further. In a secondary regression model examining the individual subscales of the sleep beliefs questionnaire, which explained 33.2 percent of the variance with an adjusted R² of 0.307, two dimensions stood out: Worry and Helplessness about sleep, and beliefs about the Perceived Consequences of poor sleep. Students who catastrophized about what a sleepless night would do to them, and who felt powerless to control their sleep, scored significantly higher on the insomnia index. This cognitive profile is precisely the target of cognitive behavioral therapy for insomnia, or CBT-I, the first-line treatment for chronic insomnia in adults, which works by restructuring exactly these dysfunctional beliefs and extinguishing the anxiety that keeps sufferers awake.

The researchers frame dysfunctional sleep beliefs as a particularly promising intervention target because they are relatively stable cognitive traits, yet remain amenable to change through structured therapy. Unlike exam schedules or curriculum demands, which are largely fixed features of medical education, sleep-related cognitions can be addressed through screening, psychoeducation, and CBT-I programs delivered by student health services. The authors conclude that sleep health and mental well-being deserve explicit attention within medical education and student health services, a recommendation that carries practical weight given how many students in the sample crossed the clinical threshold.

The study has limitations inherent to its design. As a cross-sectional survey at a single German medical school, it captures associations rather than causal pathways; it cannot determine whether depressive symptoms drive insomnia, whether sleeplessness deepens depression, or whether both feed a common loop. The modest sample of 124 students and the focus on a single examination period also limit generalizability to other institutions, other years of study, and calmer periods of the academic calendar. Nevertheless, the strength of the regression model, the use of validated instruments, and the convergence between subjective insomnia scores and objective sleep parameters lend the findings considerable credibility. For a profession that will spend its career advising patients on healthy sleep, the message that future physicians are themselves at high risk, and that their beliefs about sleep may matter as much as their stress levels, is one that medical educators cannot easily ignore.

Subject of Research: Insomnia prevalence and its psychological and academic correlates among medical students during an examination period

Article Title: Insomnia in medical students during an examination period: associations with depression, dysfunctional sleep beliefs, and academic performance

Article References: Bailer, M., Almazury, K., Sonnauer, F., Sprügel, M. I., Mestermann, S., Utz, J., Kornhuber, J., & Spitzer, P. (2026). Insomnia in medical students during an examination period: associations with depression, dysfunctional sleep beliefs, and academic performance. BMC Medical Education. https://doi.org/10.1186/s12909-026-10565-4

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10565-4

Keywords: insomnia, medical students, sleep disorders, depression, dysfunctional sleep beliefs, academic performance, exam stress, chronotype, Insomnia Severity Index, CBT-I, mental health, medical education