Counts describe this library, not the strength of evidence.
Cómo interpretar la evidencia
Sleep health has several dimensions. Regularity cohorts and sleep-extension trials answer different questions; a prediction from observational data is not proof that changing a score prevents disease. Exercise effects also differ between sleep quality and daytime sleepiness.
In two population-based German samples of 855 children aged 4-9 and 1,047 adolescents aged 10-17, 22.6% of children and 20.0% of adolescents had a problematic amount of sleep difficulties. The data are parent- or self-reported questionnaires from one cohort, so they describe patterns and cannot show why they arise.
In wearable data from 1,824 city-dwelling working adults aged 21-40, time in bed rose by 20 min and resting heart rate fell by ~2 bpm during lockdown, while physical activity dropped an average of 42%. The least active group (~51% of the sample) lost the most activity, but the data cover only young working adults in the early lockdown.
In 320 Hong Kong adults aged 60 or older with chronic insomnia, 12 weeks of tai chi or conventional exercise raised wrist-actigraphy sleep efficiency by about 3.5 percentage points versus no intervention (tai chi +3.4%, 95% CI 1.6-5.1), and gains lasted to 24 months. The improvements were modest in absolute terms and came from a single research site.
Among 1,111 young adults aged 18-25 in New Zealand and the United States, sleep quality was the strongest predictor of depressive symptoms and well-being, ahead of sleep quantity, physical activity and diet. The data come from a single online self-report survey, so the patterns are correlations and cannot show that improving sleep changes mood.
Of 494 articles screened, 23 studies of time-restricted eating (TRE) were reviewed; adherence averaged 80% and weight loss averaged 3%, with fat mass also falling. The abstract gives no per-study sizes, confidence intervals or control comparisons, and the authors themselves call for more rigorous studies.
Among 7538 physicians with error data, moderate, high and very high sleep-related impairment were associated with 53%, 96% and 97% higher odds of a self-reported clinically significant medical error (ORs 1.53, 1.96, 1.97) than low impairment, after adjusting for burnout. Only 40% of attending and 51% of house staff physicians responded, and everything was self-reported.
In an internet survey of 6,519 Italian adults during the COVID-19 lockdown, 3,562 (55.32%) reported poor sleep quality, and poor sleep went with female gender, changed sleep-wake timing and extremely severe stress, anxiety and depression. The sample was self-selected online and measured at one time, so it cannot show that lockdown or distress caused the poor sleep.
Among 667 members of the Netherlands Sleep Registry, a quarter of people with pre-pandemic clinical insomnia had a meaningful improvement in sleep quality during lockdown, while 20% of pre-pandemic good sleepers slept worse. The sample was a self-selected registry assessed online, so the percentages describe these volunteers, not the Dutch population.
Pooling 7 cross-sectional studies, the estimated prevalence of sleep disturbances was 34.8% (95% CI 24.8-46.4%) among 3745 nurses and 41.6% (95% CI 27.7-57%) among physicians caring for COVID-19 patients. Heterogeneity was very high (I2 of 97.3% or higher), so these figures are broad estimates, not precise rates.
In 20,065 older adults from England and China, sleeping 4 hours or less (β -0.022 SD per year, P = .001) or 10 hours or more (β -0.033 SD per year, P = .003) per night was linked to faster cognitive decline than the 7-hour reference. Sleep was self-reported once, so reverse causation and recall error remain possible.
In an online survey of 1,078 Iranian young adults, problematic social media use was associated with psychological distress and insomnia both directly and through fear of COVID-19, and with distress also through COVID-19 misunderstanding. Because all measures came from one self-report survey, the mediation pathways are statistical patterns, not proof of cause.
In an online survey of 1,212 people in Sweden (mean age 36.1 years, 73% women), rates of depression, anxiety and insomnia were 30%, 24.2% and 38%, with poor self-rated health and past mental health problems the strongest predictors. The sample was a self-selected online group with many women, and there was no pre-pandemic comparison.
European psychiatry : the journal of the Association of European Psychiatrists
Among 52,599 Chinese adults, those who reported consistently low sleep (4.2 to 4.9 hours) had the highest hazard of cardiovascular events (HR 1.47) and death (HR 1.50) over about 6.7 years. Sleep was self-reported, and the lowest group had only 879 people, so estimates are wide.
In an online survey of 1,563 medical staff in China, 564 (36.1%) had insomnia symptoms, and worry about infection or uncertainty about controlling the outbreak went with higher odds. It is a self-selected, one-time questionnaire, so it cannot show what caused the poor sleep.
In 92,457 US adults wearing a wrist tracker, mean daily resting heart rate was 65 bpm, with individual averages ranging from 40 to 109 bpm. Within a person it was much steadier, though 20% had at least 1 week in which it shifted by 10 bpm or more. The users chose to buy and consistently wear a tracker, so they may not represent the general population.
Of 14 moderate- or strong-quality studies of exercise programmes in generally healthy adults aged 60+, all but one found at least one significant improvement in sleep, and none reported harmful effects. Only 6 were randomised controlled trials, and 11 of 14 relied on subjective sleep measures.
The international journal of behavioral nutrition and physical activity
Among 1726 community-dwelling Chinese adults aged 70-87, poor sleep quality (PSQI > 5) was associated with higher odds of frailty (OR 1.78) and pre-frailty (OR 1.51) after adjustment, and sleeping 9 h a night or more was also linked to both. This is a cross-sectional snapshot, so poor sleep may be a result of frailty as much as a cause.
In 58 papers from the last five years, sleep disruption predicted the onset and persistence of paranoia and hallucinations, and psychological sleep treatments reported large improvements in sleep with possibly modest gains in psychotic experiences. Most of the studies were correlational (38 of 58) and only 7 tested treatment, so a causal role is suggested rather than proven.