Does Wearing Earplugs and an Eye Mask in the ICU Actually Prevent Delirium? What the Sleep Data Shows

Here's a finding that hit me sideways because it's so low-tech it feels almost insulting: a randomized trial found that giving hospital patients earplugs, an eye mask, and a scheduled quiet period at night measurably reduced the incidence of delirium. Van Rompaey and colleagues randomized ICU patients to sleep with earplugs versus standard care and saw fewer patients develop confusion and disorientation, with the effect concentrated in the first 48 hours (see [1]). No drug. No device. Just cutting the ambient noise and light that a modern hospital dumps on a sleeping brain all night.

The mechanism is the same thing I harp on for healthy people optimizing their sleep at home: fragmentation and circadian disruption. ICUs run around 60 decibels average with peaks past 80, and lights stay on for monitoring. That's a recipe for suppressed melatonin and shredded sleep architecture, and the frailer the brain, the worse it tolerates that. A broader review of non-pharmacological sleep interventions in critical care reached the same direction of effect — protecting sleep with simple environmental controls was associated with lower delirium rates, though the authors are honest that trial quality varies and the confidence intervals are wide (see [2]). This is real signal, not a slam dunk, and I want to flag that honestly: effect sizes across these studies bounce around, and a single small trial is noise until it replicates.

Why does this matter to a quantified-self crowd who mostly aren't in the ICU? Because it's the cleanest natural experiment you'll find for the light-and-noise hypothesis. If stripping two inputs out of a chaotic environment changes a hard clinical outcome like delirium, then your bedroom's stray lumens and your partner's snoring are not trivial. The same circadian machinery is running in both cases. I've measured this on myself — capping bedroom light and killing intermittent noise tightens up deep sleep percentage and overnight HRV, which you can watch trend over weeks in /sleep rather than judging off one night.

If you want an N=1 version: run a two-week A/B, eye mask and earplugs on odd nights, off on even nights, and let the wearable log deep sleep, wake events, and morning HRV. One good night proves nothing. A two-week split with a clear separation between conditions actually answers the question.

This is general information, not medical advice — talk to a qualified clinician about your own situation.