Background noise lowers sleep quality.

What this means

Background noise lowers sleep quality. The effect shows up most in hospital settings, especially intensive care units where equipment, alarms and staff activity run through the night. People in noisier spaces fall asleep less well, wake more often and rate their own sleep as poorer. The same pattern reaches into ordinary bedrooms, where traffic and neighbour noise cut into how much of the night is spent actually asleep. Quieter rooms, earplugs and eye masks push sleep quality back up.

What the research shows

Horsten (2017) put healthy sleepers under recorded ICU noise and counted 9.59 more arousals per night than at baseline (95% CI 2.48 to 16.70). Basner (2023) found sleep efficiency 4.7% lower in the noisiest quintile of exposure than in the quietest (p<0.0001), and Chen (2026) put the drop at 1.81% across the range, widening to 2.73% for the highest quartile against the lowest.

Arık (2020), Warjri (2021), Huang (2026) and Higa (2022) each recorded sleep improving once noise or light dropped. Arık (2020) raised Richards-Campbell sleep scores from 33.50 to 80.61 by cutting noise and light. Warjri (2021) improved sleep quality on three straight nights with a white-noise app (Z = −3.996 to −2.822, p ≤ 0.005). Huang (2026) reported larger falls in PSQI scores in an integrated noise-reduction group (p<0.05), and Higa (2022) found lighting cuts and eye masks lifted subjective sleep quality by 11.89 points (95% CI 8.0 to 15.76). Hu (2010) recorded poorer perceived sleep, more light sleep and less REM under simulated ICU noise and light.

How certain this is

The firmest number here is Basner (2023): a dose-response drop in sleep efficiency at p<0.0001, adjusted for multiple testing, which Chen (2026) reproduces in a separate cohort. Basner (2023) sits within a supporting set of eleven studies spanning cohort measurements, controlled exposure experiments and reviews, several of them recent, running from objective arousal counts to validated sleep scales. The one contrary result, Thomas (2012), came from a ward already down at 35 to 40 dB at night, where quieting it further changed nothing. The finding holds firmly.

In practice

A room running at 35 to 40 dB overnight, the level Thomas (2012) measured, is already past the point where cutting noise further changes sleep quality. Above that floor, the dose matters: Basner (2023) and Chen (2026) both show the loss in sleep efficiency growing with exposure, so a space in the top quintile or quartile of night noise stands to gain the most from bringing levels down, while a quiet bedroom or a low ward has little to gain. ICUs and other settings with alarms, equipment and staff movement through the night sit at the exposure end where Horsten (2017), Hu (2010) and Arık (2020) recorded the largest changes, so noise control there works on measurable sleep loss rather than a marginal effect. Before treating a noisy room as the cause of poor sleep, check where its night-time level actually sits against a baseline, since a ward already down near 35 to 40 dB will not show the same response that Arık (2020) or Warjri (2021) found through earplugs, eye masks or white noise. Where levels are already low, other causes of poor sleep are worth ruling out before noise reduction is tried as the fix.

Dose and thresholds

The harm tracks the exposure level rather than switching on at a single point. Basner (2023) measured the 4.7% efficiency loss in the top quintile of noise against the bottom, and Chen (2026) found the steepest drop, 2.73%, in the highest quartile compared with the lowest. The lower the exposure, the smaller the loss, which is why a ward already at 35 to 40 dB (Thomas 2012) showed little room for further gain.

Where it is contested

Thomas (2012) tried sleep-promoting interventions on a ward where night-time noise was already 35 to 40 dB and saw no improvement from them. That reads as a floor effect: with levels this low there was little noise left to cut, so the null says nothing about louder rooms. The effect concentrates in loud settings, above all ICUs, where Horsten (2017), Hu (2010) and Arık (2020) all found large changes once noise was added or removed.

Why it happens

Noise at night keeps the nervous system on alert and disrupts melatonin, the hormone that sets the circadian clock and lets sleep deepen. Yaşar (2017) argues that regulating light and noise in ICUs may aid recovery after major surgery by letting melatonin production rise, since the hormone also has anti-inflammatory properties. When noise holds melatonin down, sleep stays lighter and arousals climb.